University of Wollongong Research Online

Faculty of Science - Papers (Archive) Faculty of Science, Medicine and Health

1-1-2000

Thyroid hormones and their effects: a new perspective

A. J. Hulbert University of Wollongong, [email protected]

Follow this and additional works at: https://ro.uow.edu.au/scipapers

Part of the Life Sciences Commons, Physical Sciences and Mathematics Commons, and the Social and Behavioral Sciences Commons

Recommended Citation Hulbert, A. J.: Thyroid hormones and their effects: a new perspective 2000. https://ro.uow.edu.au/scipapers/29

Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected] Thyroid hormones and their effects: a new perspective

Abstract The thyroid hormones are very hydrophobic and those that exhibit biological activity are 3',5',3,5-ltetraiodothyronine (T4), 3',5,3-l-triiodothyronine (T3), 3',5',3-l-triiodothyronine (rT3) and 3,5,- ldiiodothyronine (3,5-T2). At physiological pH, dissociation of the phenolic -OH group of these iodothyronines is an important determinant of their physical chemistry that impacts on their biological effects. When non-ionized these iodothyronines are strongly amphipathic. It is proposed that iodothyronines are normal constituents of biological membranes in vertebrates. In plasma of adult vertebrates, unbound T4 and T3 are regulated in the picomolar range whilst protein-bound T4 and T3 are maintained in the nanomolar range. The function of thyroid-hormone-binding plasma proteins is to ensure an even distribution throughout the body. Various iodothyronines are produced by three types of membrane-bound cellular deiodinase enzyme systems in vertebrates. The distribution of deiodinases varies between tissues and each has a distinct developmental profile. Thyroid hormones have many effects in vertebrates. It is proposed that there are several modes of action of these hormones. (1) The nuclear receptor mode is especially important in the thyroid hormone axis that controls plasma and cellular levels of these hormones. (2) These hormones are strongly associated with membranes in tissues and normally rigidify these membranes. (3) They also affect the acyl composition of membrane bilayers and it is suggested that this is due to the cells responding to thyroid-hormone-induced membrane rigidification. Both their immediate effects on the physical state of membranes and the consequent changes in membrane composition result in several other thyroid hormone effects. Effects on metabolism may be due primarily to membrane acyl changes. There are other actions of thyroid hormones involving membrane receptors and influences on cellular interactions with the extracellular matrix. The effects of thyroid hormones are reviewed and appear to be combinations of these various modes of action. During development, vertebrates show a surge in T4 and other thyroid hormones, as well as distinctive profiles in the appearance of the deiodinase enzymes and nuclear receptors. Evidence from the use of analogues supports multiple modes of action. Re-examination of data from the early 1960s supports a membrane action. Findings from receptor 'knockout' mice supports an important role for receptors in the development of the thyroid axis. These iodothyronines may be better thought of as 'vitamone'-like molecules than traditional hormonal messengers.

Disciplines Life Sciences | Physical Sciences and Mathematics | Social and Behavioral Sciences

Publication Details This article was originally published as Hulbert, AJ, Thyroid hormones and their effects: a new perspective, Biological Review, 75, 2000, 519-631. Copyright Cambridge Philosophical Society. Original journal available here.

This journal article is available at Research Online: https://ro.uow.edu.au/scipapers/29 Biol. Rev. (2000), 75, pp. 519–631 Printed in the United Kingdom # Cambridge Philosophical Society 519

Thyroid hormones and their effects: a new perspective

A. J. HULBERT Metabolic Research Centre, and Department of Biological Sciences, University of Wollongong, Wollongong, N.S.W. 2522 Australia (e-mail: hulbert!uow.edu.au)

(Received 21 May 1999; revised 30 May 2000; accepted 7 June 2000)

ABSTRACT

The thyroid hormones are very hydrophobic and those that exhibit biological activity are 3h,5h,3,5-- tetraiodothyronine (T4), 3h,5,3--triiodothyronine (T3), 3h,5h,3--triiodothyronine (rT3) and 3,5,-- diiodothyronine (3,5-T2). At physiological pH, dissociation of the phenolic kOH group of these iodothyronines is an important determinant of their physical chemistry that impacts on their biological effects. When non-ionized these iodothyronines are strongly amphipathic. It is proposed that iodothyronines are normal constituents of biological membranes in vertebrates. In plasma of adult vertebrates, unbound T4 and T3 are regulated in the picomolar range whilst protein-bound T4 and T3 are maintained in the nanomolar range. The function of thyroid-hormone-binding plasma proteins is to ensure an even distribution throughout the body. Various iodothyronines are produced by three types of membrane-bound cellular deiodinase enzyme systems in vertebrates. The distribution of deiodinases varies between tissues and each has a distinct developmental profile. Thyroid hormones have many effects in vertebrates. It is proposed that there are several modes of action of these hormones. (1) The nuclear receptor mode is especially important in the thyroid hormone axis that controls plasma and cellular levels of these hormones. (2) These hormones are strongly associated with membranes in tissues and normally rigidify these membranes. (3) They also affect the acyl composition of membrane bilayers and it is suggested that this is due to the cells responding to thyroid-hormone-induced membrane rigidification. Both their immediate effects on the physical state of membranes and the consequent changes in membrane composition result in several other thyroid hormone effects. Effects on metabolism may be due primarily to membrane acyl changes. There are other actions of thyroid hormones involving membrane receptors and influences on cellular interactions with the extracellular matrix. The effects of thyroid hormones are reviewed and appear to be combinations of these various modes of action. During development, vertebrates show a surge in T4 and other thyroid hormones, as well as distinctive profiles in the appearance of the deiodinase enzymes and nuclear receptors. Evidence from the use of analogues supports multiple modes of action. Re-examination of data from the early 1960s supports a membrane action. Findings from receptor ‘knockout’ mice supports an important role for receptors in the development of the thyroid axis. These iodothyronines may be better thought of as ‘vitamone’-like molecules than traditional hormonal messengers.

Key words: thyroxine, triiodothyronine, diiodothyronine, nuclear receptors, membranes, deiodinases, metabolism, growth, development, antioxidants.

CONTENTS

I. Introduction ...... 520 (1) Some prejudices ...... 521 II. Thyroid hormones ...... 522 (1) Structure and physical chemistry...... 522 (2) Plasma concentrations and distribution to the tissues...... 523 (3) Cellular uptake, cellular location and hormone metabolism ...... 534 520 A. J. Hulbert

III. Effects of thyroid hormones...... 542 (1) When studying the effects of thyroid hormones...... 542 (2) Hypotheses regarding thyroid hormone action ...... 544 (a) Nuclear receptors and thyroid response elements...... 545 (b) Thyroid hormones and the membrane bilayer ...... 549 (3) Effects on the thyroid hormone axis ...... 558 (4) Effects on metabolism and thermogenesis...... 560 (5) Effects on excitable tissues ...... 571 (6) Effects on growth ...... 574 (7) Other effects...... 576 IV. Thyroid hormones and vertebrate development...... 580 (1) Thyroid axis during vertebrate development...... 580 (2) Effects of thyroid hormones on vertebrate development...... 586 V. Some perspectives ...... 593 (1) Analogues as a source of knowledge ...... 593 (2) The early sixties revisited...... 596 (3) Knockouts from the nineties ...... 597 (4) Resistance to thyroid hormones ...... 600 (5) Future insights from the human genome ...... 602 VI. Conclusions...... 602 VII. Acknowledgements ...... 606 VIII. References...... 606

I. INTRODUCTION that there are several other iodinated thyronines and their metabolites in the vertebrate body and that On Christmas Day, 1914, thyroxine was first there exists a complex metabolic system involved crystallised by Kendall. This was approximately with their interconversions. 4500 years after the earliest recorded use of seaweed Even before the discovery of T3 many effects of as an effective remedy for goitre, approximately a either an inactive or overactive thyroid gland were hundred years after the first isolation of elemental known and there was intense interest in the means iodine, approximately forty years after the initial whereby the thyroid gland exerted its hormonal association of cretinism with the thyroid gland, effects on the body. Special attention was paid to its twenty years after the demonstration by Magnus- effects on metabolism as possibly being central to Levy of the stimulatory effect of the thyroid on the many of the other observed effects. metabolic rate of humans, and at approximately the In the early 1960s, seminal work by Tata and same time as the discovery by Gudersnatch that colleagues shifted emphasis to protein synthesis. The tadpoles fed thyroid tissue precociously turn into finding of an early stimulation of ribonucleic acid frogs (Pitt-Rivers & Tata, 1959; Hetzel, 1989). synthesis (Tata & Widnell, 1966), followed by the Thyroxine is 3h,5h,3,5--tetra-iodothyronine and is discovery of specific nuclear T3 binding sites (Oppen- nowadays commonly called T4. heimer et al., 1972; Samuels & Tsai, 1973) led to the The use of radioisotopes and paper chromato- nuclear receptor hypothesis to explain thyroid graphy in the 1940s led to the realisation that there hormone effects. were other important iodinated compounds apart As an offshoot of the search for the genetic basis of from thyroxine. In Melbourne in 1948, an unknown cancer, in 1986 two groups reported that cellular iodothyronine was isolated using paper chromato- equivalents of the viral erb-A gene coded for part of graphy and suggested to be triiodothyronine (T3) the thyroid hormone nuclear receptor. The genes for (Hird & Trikojus, 1948). This was two years before the two main types of thyroid nuclear receptor had the generally accepted first publication of the been independently isolated from two different unknown iodinated thyronine (Gross et al., 1950) sources (Sap et al., 1986; Weinberger et al., 1986). that was later isolated and identified as 3h,5,3-- These thyroid nuclear receptors are now recognised triiodothyronine and demonstrated to be more to be members of a superfamily of nuclear hormone potent than T4 in preventing goitre (see Gross, receptors that bind to specific sections of the genome 1993). Since the 1950s with the development of (see Jensen, 1991). sophisticated analytical techniques we have realised There are myriad effects of thyroid hormones that Thyroid hormones and their effects 521 have been recorded this century. A review by Hoch revisit this hypothesis, in order to examine if it still in 1962 on the biochemical actions of these hormones has some validity. alone cited 611 references. Although much of the Another prejudice stems from my early physio- literature about the effects of thyroid hormones logical training. It is that the scientific use of the assumes that such effects are mediated by nuclear word ‘regulation’ involves homeostasis and therefore receptors, a more detailed search of the literature for requires negative feedback to be involved some- the scientific basis of such statements turns up little where. This meaning of ‘regulation’ is also ap- direct evidence. plicable to biochemistry, with respect to the control What is obvious from the literature is that a of metabolism (see Fell, 1997). It is common to read number of thyroid hormone effects are indubitably statements like ‘…thyroid hormones regulate…(a initiated by thyroid hormone receptors within the particular enzyme, activity etc.)h. However, in nucleus (e.g. thyroid-stimulating hormone (TSH) general, negative feedback from thyroid hormone secretion in anterior pituitary, synthesis of malic effects to the thyroid are not evident. For example, enzyme in liver, growth hormone secretion), how- using this rather strict physiological definition, the ever, it is equally obvious that there are effects thyroid hormones do not ‘regulate’ metabolism. If # undoubtably non-genomic in origin (e.g. Ca + fluxes they did, a decreased metabolic rate should result in mammalian erythrocytes, sugar uptake by thymo- either in increased hormonal secretion or increased cytes, action on heart membranes). What a search of hormonal effect (or both) in order to ‘regulate’ the literature does suggest is that the sites of origin of metabolic rate. An opposite response is expected for the majority of thyroid hormone effects are still an increased metabolic rate. The objection to the currently not known. improper use of the word ‘regulate’ is, I believe, Following the discovery of T3, and the fact that more than a pedantic quibble. It is important in that nuclear receptors have a greater affinity for T3 than it influences how we think about what these for T4, the belief developed that T3 is the active hormones are doing. Indeed, often a more ap- thyroid hormone and that T4 is only a ‘pro- propriate phrasing is that thyroid hormones ‘in- hormone’. The finding during the last decade that fluence’ (or stimulate, inhibit, initiate) a certain diiodothyronines (T2s) can also initiate thyroid enzyme, activity, developmental event etc. hormone effects, and in some cases have a potency One exception to the general lack of negative similar to that of T3, yet thyroid nuclear receptors feedback between thyroid hormone effects and the have negligible affinity for T2, has also raised thyroid is, of course, TSH release from the anterior questions regarding the mode of action for thyroid pituitary. With respect to the thyroid, what appears hormone effects. to be ‘regulated’ are the levels of thyroid hormones It is the purpose of this review to attempt a themselves. This is also true when we examine the synthesis of what we currently know of thyroid cellular level of organization. The thyroid hormones hormones and their effects. The literature on the are metabolised within cells via various deiodination thyroid is vast and I am indebted to many excellent mechanisms and in some tissues (most notably the reviews that will be referenced in the appropriate brain), there is evidence of negative feedback sections. While much information is available for between thyroid status and the deiodination systems. humans and the laboratory rat, I have not restricted The consequence of this feedback is that both the myself to these two mammals but have taken a whole organism and some individual cells attempt to broader phylogenetic survey because of the power of maintain a relatively constant concentration of an evolutionary perspective in understanding what thyroid hormone despite varying conditions. Viewed is biologically important. in this light, the thyroid hormones are more akin to an important body constituent, a ‘vitamin-like’ molecule rather than a classical ‘hormonal mess- (1) Some prejudices enger’. I come to this task with a background in the Such a perspective has been raised, in the context physiology of mammalian metabolism and its evol- of the wide phylogenetic distribution of these ution. Many years ago, in youthful exuberance, I molecules, in an interesting recent review by Eales published a hypothesis that several thyroid hormone (1997) where he suggested the term ‘vitamone’ to effects may be caused by changes in the fatty acid describe such a role. In this light, it is also of interest composition of membranes (Hulbert, 1978). Whilst that the superfamily to which thyroid nuclear my ideas have inevitably evolved since then, I will receptors belong includes those for vitamin D and 522 A. J. Hulbert

(vitamin-A-related) retinoic acid (Jensen, 1991). arrangement of thyroglobulin favours the coupling This perspective suggests that due to the relative of some of these iodotyrosines to form T4 which is constancy of their concentration, thyroid hormones secreted from the thyroid follicle. In cases of iodine are probably best regarded not as ‘regulatory’ deficiency, relatively more T3 and less T4 is hormones in the normal sense but as ‘permissive’ synthesised and secreted. This review will not hormones, possibly ‘vitamonesh. concern itself with events in the thyroid follicle but will concentrate on the thyroid hormones themselves and how they exert their multitudinous effects. In II. THYROID HORMONES view of the relative ease of synthesis of these molecules (e.g. both T4 and T3 can be synthesised (1) Structure and physical chemistry by non-enzymatic means), their wide phylogenetic Thyroid tissue consisting of follicles of epithelial cells distribution is not surprising (see Eales, 1997). surrounding a lumen filled with colloid has been The chemical structure of the thyroid hormones identified in all vertebrates examined to date. In T4 and T3 is illustrated in Fig. 1 together with a most vertebrates, these follicles are grouped together molecular model showing the outer phenolic ring into a discrete gland, the thyroid, whilst in others oriented in a plane perpendicular to the inner ring. they are diffusely distributed generally in the Also included in Fig. 1 are another two iodothyro- anterior region of the body (Etkin & Gona, 1974). nines, reverse T3 (rT3) and 3,5,-diiodothyronine, This review will restrict itself to vertebrates. For two that exert some thyroid hormonal effects and will be fascinating recent accounts of the function of thyroid discussed in later sections. Because they are syn- hormones in invertebrates the reader is referred to thesised from natural tyrosine residues, the amino Eales (1997) and Johnson (1997). acid part of the molecule is of the -isomer The thyroid epithelial cells take up iodide and configuration. The molecular mass of T4 and T3 is secrete thyroglobulin protein into the central lumen 777 and 651 Da, respectively. The thyroid hormones of the follicle. Of the more than 100 tyrosine residues (and their metabolites) are the only known molecules per molecule of thyroglobulin, a number are iodin- in the body that contain iodine and most of the ated ortho to the phenolic hydroxyl group to form 3- molecular mass of both thyroid hormones is con- iodotyrosine and 3,5-diiodotyrosine. The spatial tributed by their iodine atoms which are, by far, the

+NH3 COO – +NH3 COO – +NH3 COO – +NH3 COO – CH CH CH CH

CH2 CH2 CH2 CH2

I I I I I I I O O O O

I I I I I OH OH OH OH

T4 T3rT3 T2 Fig. 1. The chemical structures and molecular models of four biologically active thyroid hormones; thyroxine (T4), 3h,3,5-triiodothyronine (T3), 3h,5h,3-triiodothyronine (rT3) and 3,5-diiodothyronine (T2). Thyroid hormones and their effects 523 heaviest elements found in the body (atomic mass the deiodination of T4, both in the thyroid itself and 127). The next heaviest element found in the body is in peripheral tissues. This deiodination will be selenium (atomic mass 79) which is an important discussed in Section II.3. Other iodothyronines are part of the enzymes that deiodinate the thyroid also formed by the deiodination mechanisms and hormones. The iodine atoms also have strong these are present in small concentrations in the electron binding energies and are very electron plasma. Both T4 and T3 are found in the plasma attractive. either bound to plasma proteins or in the free The amino group, carboxyl group and phenolic (unbound) state. Information concerning both the kOH group are all ionizable parts of these total and free concentration of T4 and T3 in plasma molecules. The relative ionization of the phenolic from a diverse range of vertebrates under a number group in different thyronines will be a major of conditions is collated in Table 1. Much of this influence on their relative lipophilicity. Simple information has been obtained from the values for phenolic groups are 50% dissociated at pH of controls in particular experiments, some have been approximately 10 (i.e they have a pK of approx- interpolated from graphs, and many values have imately 10). The presence of a single iodine (with its been recalculated in SI units. Selected values for electron attraction) on the outer ring of T3 reduces earlier stages in the life cycle of some species are the pK of its phenolic group to 8.45, whilst the two also included. Fig. 2 presents plasma T4 and iodine atoms on the outer ring of T4 further reduces T3 concentrations for normal (l euthyroid) adult the pK of its phenolic group to 6.73 (Gemmill, vertebrates relative to their body mass. Many of the 1955). At physiological pH approximately 80% of reports cited in Table 1 did not report body mass of T4 molecules have the phenolic hydroxyl group in the animals examined. For these I have estimated the ionized form whilst only approximately 10% of the adult body mass of the particular species. Table T3 molecules have an ionized phenolic hydroxyl 1 and Fig. 2 also include the ranges regarded as group (Korcek & Tabachnick, 1976). This difference typically ‘euthyroid’ for humans (from Stockigt, will result in different relative hydrophobicities of T4 1996). and T3 at physiological pH. In normal adult vertebrates, the plasma total Both molecules are hydrophobic molecules, with concentration of T4 is in the nanomolar range with T3 being more so than T4 because of its relative lack an average of approximately 35 n and ranges from of an ionized phenolic group. They both have a low below measurable levels in some amphibians to 148 solubility at neutral and acid pHs but dramatically n in the hedgehog, Erinaceus europaeus. The plasma increased solubility in alkaline solutions, because of total concentration of T3 is also in the nanomolar the increased ionization of the phenolic end of the range for vertebrates and averages approximately molecule. For example, the solubility limit of T4 in 3.5 n, ranging from unmeasurable to a maximum aqueous solution rises from 2.3 µ at pH 7, to 4.5 µ of 76 n reported for the parrotfish, Sparisoma spp. at pH 9, and to 260 µ at pH 11 (values taken from The vast majority of both T4 and T3 is bound to Schreiber & Richardson, 1997). This is the basis for plasma proteins with the free hormone being in the the common practice of using alkaline solutions for picomolar range for all vertebrates, approximating injection of thyroid hormones. Whilst the hormones 0.01–0.1% of the total hormone concentration. For are soluble in the injectate, once inside the body they adult vertebrates, the average free T4 and free T3 will return to a more neutral (or even acidic) plasma concentrations are approximately 40–50 p, environment and thus will be more hydrophobic in with respective ranges of 2–600 p and 1–492 p.It character. Hillier (1970) reported that at physio- is the level of ‘free’ thyroid hormones that is the logical pH the ‘partition coefficient’ between phos- more physiologically important concentration (Rob- pholipid (phosphatidylcholine) and the aqueous bins & Rall, 1957; Ekins, 1986; Mendel, 1989). environment was 12000 for T4 and 22000 for T3. In many cases reported in Table 1, the concen- Dickson et al. (1987) report values of 17500 and trations in the early phase of the life cycle are 23500 for T4 and T3 respectively. considerably higher than those observed in the adult phase. This is especially true for periods of ‘meta- morphic’ change in some of the vertebrates. This (2) Plasma concentrations and distribution developmental surge of thyroid hormones will be to the tissues discussed later. Although the hormone concen- Whilst circulating T4 originates from the thyroid trations are presented as a single value for each follicles, circulating T3 comes predominantly from species, of course the normal euthyroid hormone 524 A. J. Hulbert . . (1994) . (1996) . (1987) . (1987) . (1987) . (1987) . (1985) et al . (1994) . (1997) . (1985) et al . (1981) et al et al . (1994) et al et al et al et al . (1998) et al et al et al et al et al et al (1987) (1987) (1987) (1987) (1995) (1997) (1993) (1985) Free [T3] (pM) Reference Free [T4] (pM) Total [T3] (nM) Total [T4] (nM) Body mass (g) 206 2.7 2.4 5.0 3.2 Eales & Shostak in vertebrate plasma ) T3 ( Ammocoete larvaeJuvenileSpawning adultSpawning femaleAdult malesAdultsEggs 50 65.0Peak smolt 12.0 104.0 24.0 200 4.6 282.0 12.0 1.6 116.0 0.7 Weirich 0.4 3.5 19 25.2 1.7 1.1 3.0 19.0 Weirich 8.4 Hornsey (1977) Youson 1.2 Hornsey (1977) Sower AdultsAdultsAdults Ura Tagawa & Hirano 1150 7.1 10.0 11.0 3.3 3.4 9.9 3.2 Gomez Weirich Larsson Condition Adults 0.3 0.0 Fredriksson Pre-smoltTransitionSmoltWild smolt 10Wild parrImmatureImmature 15.0ImmatureAdult 30 31.0 110 9.2Young adults 55 15.0 7.6Male 100 adults 4.5 4.0 2.3 4.6 9.5 400 12.6 4.6 14.0 Swanson 3.5 750 & Dickhoff 7.4 10.2 1–6 1.9 Swanson & Dickhoff 3.6 2–4 Swanson 0.9 & Dickhoff Whitesel (1992) Gelineau 19.3 Whitesel (1992) Reddy & Leatherland Holloway Waring & Eales Brown Cyr & triiodothyronine ) T4 ( Petromyzon marinus Petromyzon marinus Petromyzon marinus Petromyzon marinus Petromyzon marinus Petromyzon marinus Oncorhynchus keta Oncorhynchus masou Oncorhynchus mykiss Salvelinus namaycush Salmo salar Species Phallusia mammillata Oncorhynchus kisutch Oncorhynchus kisutch Oncorhynchus kisutch Oncorhynchus kisutch Oncorhynchus kisutch Oncorhynchus mykiss Oncorhynchus mykiss Oncorhynchus mykiss Oncorhynchus mykiss Salvelinus alpinus Salmo trutta Gadus morhua Concentrations of thyroxine Sea lamprey Sea lamprey Sea lamprey Sea lamprey Sea lamprey Sea lamprey Salmon SalmonMasu salmon Various species Embryos 9–19Rainbow trout Lake trout Baltic salmon 1.5–7.7 Sullivan Ascidean Salmon Salmon Salmon Salmon Salmon Rainbow trout Rainbow trout Rainbow trout Rainbow trout Arctic charr Brown trout Atlantic cod Agnathan Fish Vertebrate Protochordate Table 1. Thyroid hormones and their effects 525 . (1987) . (1987) . (1993) . (1986) . (1986) . (1985) et al . (1986) et al et al et al et al et al et al (1981) (1981) John-Alder (1992) (1991) (1991) comm. Adult malesAdult malesHibernating adultAdult (spring) 58 56 14.5 1.3 3.2 13.0 0.5 John-Alder (1984a) John-Alder (1984b) John-Alder (1983) John-Alder (1984b) AdultPeak-metamorphicPeak-metamorphicAdultsAdultsAdultsFemale adults 9.0 6.0YearlingsCaptive 0.0 adults 2.2Field adults 0.5Captive 0.0 0.05–1 10.7 7.7 3.0 0.1–2 11.1 0–0.2 5.5 0.3 8.3 Suzuki Weil & (1986) Suzuki 13.1 23 3.1 Alberch 8.2 3.2 Tasaki Suzuki & Suzuki Larsson Vandorpe John-Alder & Joos Gerwein & John-Alder & Joos Steinberg Adults 20–35 1.0 3.4 MacKenzie AdultsAdultsAdultsAdultsAdultsAdultsAdults 91 237Adults 65 34.1Adults 72 4.4Adults 13.0 98 14.0 14.4 104Peak-metamorphic 5.4 4.0 74 10.4 599.7 69 5.0 5.7 491.6 30.9 137 Eales 4.8 70.8 & Shostak (1987) 1.9 24.6 6.3 64.4 46.1 17.6 Eales & Eales Shostak 10.0 4.6 & (1987) Shostak 33.8 (1987) 23.2 Eales 2.3 & Shostak 6.5 (1987) 0.3 23.2 24.6 5.4 Eales & 48.9 Shostak 1.6 39.9 (1987) Eales 2.1 & 100.4 Shostak 21.5 (1987) 0.4 Eales & Shostak (1987) 24.6 0.6 Eales Eales & & Shostak Shostak (1987) (1987) Alberch J. M. P. Joss personal spp. Adults 120 54.3 76.2 303.7 419.4 Eales & Shostak (1987) Dipsosaurus dorsalis Dipsosaurus dorsalis Dipsosaurus dorsalis Dipsosaurus dorsalis Eurycea bislineata Rana clamitans Rana catesbiana Rana catesbiana Bufo regularis Rana ridibinda Bufo japonicus Sceloporus undulatus Sceloporus undulatus Sceloporus undulatus Ameiva undulata Carassius auratus Holocentrus rufus Cephalopholis fulva Haemulon flavolineatum Chaetodon striatus Holacanthus tricolor Sparisoma Acanthurus coeruleus Acanthurus bahianus Cantherines pullus Pseudupeneus maculatus Neoceratodus forsteri Eurycea bislineata butterflyfish filefish Lizard Lizard Lizard Lizard Salamander Frog Bullfrog Bullfrog Frog Frog Toad Lizard Lizard Lizard Lizard Goldfish Squirrelfish Coney Grunt Banded Rock beauty Parrotfish Blue tang Ocean surgeonfish Orange-spotted Spotted goatfish Australian lungfish Salamander Reptiles Amphibians 526 A. J. Hulbert . (1987) . (1987) . (1984a) . (1984a) . (1996) . (1985) et al et al et al et al et al et al (1996) (1985) (1985) (1996) (1996) (1996) (1989) (1988) (1987) (1988) (1988) (1988) Free [T3] (pM) Reference Free [T4] (pM) Total [T3] (nM) Total [T4] (nM) Body mass (g) HatchlingAdultAdultHatchlingAdult10 days post hatching10 days post hatching8 38.6 days post hatchingAdult 5.1 21.0Egg 5.0 18.0 23.0Immature 2.5 15.4 0.8 5.1 6.8 22.0 18.5 23.2 2.6 8.7 6.9 23000 12.5 McNabb & McNabb Olson 3.9 4.6 McNabb 15.9 McNabb McNabb & & Cheng Olson 1.6 McNabb & 12.0 Olson 10.0 McNabb & Cheng McNabb & Olson 9.2 3.1 Groscolas & Leloup Williamson & Sechman Davison & Bobek Adult springAdultAdultAdultAdult 85Adult 12.7 520 800 12.5 3.0 1150 1.0 0.3 9 3.0 0.3 0.5 Naulleau 0.15 Hulbert & Williams Licht Hulbert & Williams Hulbert & Williams 1.7 Venditti Condition Growing (expt 1)Growing (expt 2)AdultsAdult hibernating 80 80 100 145.0 83.7 3.4 49.5 1.3 0.3 Denver & Licht (1991) Denver & Licht (1991) Naulleau Etheridge (1993) Coturnix japonica Coturnix japonica Streptopelia risoria Streptopelia risoria Streptopelia risoria Streptopelia risoria Sturnus vulgaris Agelaius phoenicus Aptenodytes forsteri Gallus domesticus Gallus domesticus Viper aspis Chelonia mydas Trachydosaurus rugosus Chelodina longicollis Crocodylus johnstonii Podarcis sicula Species Trachemys scripta Trachemys scripta Thamnophis sirtalis Viper aspis .) .) cont cont Japanese quail Japanese quail Ring dove Ring dove Ring dove Ring dove Starling Redwing blackbird Emperor penguin Chicken Chicken Viper Green sea turtle Lizard Tortoise Crocodile Lizard Turtle Turtle Garter snake Viper Birds Vertebrate Reptiles ( Table 1 ( Thyroid hormones and their effects 527 . (1997) . (1990) . (1990) . (1996) . (1996) . (1993) . (1996) . (1985) et al . (1997) . (1997) . (1998) . (1998) . (1994) . (1994) . (1986) . (1986) . (2000) . (1995) et al et al et al et al et al et al et al et al et al . (1997) et al et al et al et al et al et al et al et al et al (1983) (1983) (1982) (1982) (1987) Adult (30 day)Adult femalesHatchlingAdultAdult 731Juvenile 400Adult 870Adult 34.2 120000Adult 7.6 2.5 1.5 1070Adult 1.8Pouch-young (peak) 760Adults 60.3Adults 3000 107.5Adults 3000Adults 0.9Adults 9.5 1.5 Di 15.2 1500Adults Meo John 15.7Adults 6000 81.2 1.6 22.0 0.7 9000 Dawson 12.2 20.3 3.1 Dawson 1.5 5.2 18.4 Hulbert 3.2 & 4.5 Grant 12.3 0.8 44.9 Hulbert 17.4 Nicol & 29 Grant 45.7 166.0 33 0.4 30.9 Janssens 8.1 43.6 Hulbert & 41.9 45.0 Augee 3.3 Janssens Hulbert & 61.0 Augee 1.4 1.7 Lawson 1.1 Tomasi (1984) Tomasi (1984) Tomasi (1984) Liu Tomasi (1984) Burgi AdultAdult 4300 27.0 22.0 1.3 1.6 Bruggeman Rudas &Severe Pethes infection (1986) AdultTTR null adultAdult malesActive maleActive femaleAdult maleMSG obeseAdultExercised 21.0 adult 146 700 18.9 900 0.4 46.3 200–300 283 39.7 148.0 1.1 103.0 224 0.7 200–300 45.0 1.3 65.6 15.7 51.5 43.8 14.1 0.9 7.6 0.8 Palha 0.9 Burgi 7.6 0.9 15.4 Palha Tomasi & 2.3 16.7 Horwitz Woody Fowler 1.7 (1988) Fowler (1988) Woody Ribeiro Ribeiro Adult 22.0 Larsson Gallus domesticus Columbia livia Struthio camelus Struthio camelus Ornithorhynchus anatinus Ornithorhynchus anatinus Tachyglossus aculeatus Tachyglossus aculeatus Isoodon macrourus Macropus eugenii Macropus eugenii Phascolarctus cinereus Sorex vagrans Peromyscus maniculatus Reithrodontomys megalotis Microtus montanus Microtus brandtii Mus musculus Gallus domesticus Gallus domesticus Mus musculus Mus musculus Mus musculus Mesocricetus auratus Erinaceus europaeus Erinaceus europaeus Rattus norvegicus Rattus norvegicus Rattus norvegicus Rattus norvegicus Gallus domesticus Chicken Pigeon Ostrich Ostrich Platypus Platypus Echidna Echidna Bandicoot Wallaby Wallaby Koala Shrew Rodent Rodent Rodent Brandt’s vole Mouse Chicken Chicken Mouse Mouse Mouse Hamster Hedgehog Hedgehog Rat Rat Rat Rat Chicken Mammals 528 A. J. Hulbert . . . . (1995) . (1985) . (1995) . (1995) et al et al et al . (1983) . (1997) . (1996) et al . (1998) . (1998) . (1987) . (1985) . (1985) . (1985) . (1985) . (1985) . (1985) . (1985) . (1996) . (1998) . (1985) . (1986) . (1986) . (1992) et al et al et al . (1994) et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al et al (1988) (1993) (1993) Free [T3] (pM) Reference Free [T4] (pM) Total [T3] (nM) Total [T4] (nM) Body mass (g) Healthy adultsAdultAdultAdultAdult (fall)Adult (spring)AdultAdult 3600 3900 36.0 32600 34.6 22.0 28.0 15.0 29.0 1.5 22.9 8.4 2.7 Kyle 22.0 Paradis & Rawson Rawson Larsson Page (1996) Peterson Larsson Larsson Condition Adult (young)Adult (old)Adult 63.0 30.0 1.3 1.2 80.0 36.0 15.0 6.0 1.4 Chen & Walfish (1978) 5.0 67.0 Chen & Walfish (1978) 16.0 Weirich Fetal (22 days)Adult (PUFA diet)Adult (SFA diet)Fetal (62 days)AdultHealthy adults 400 400 68.0 6.0 65.0 0.5 0.2 0.4 49.9AdultAdult 26.0 0.3 39.0 66.0 Takeuchi 0.6 Ruiz 1.8 de Takeuchi Ona 24.8 Castro 17.0 4.0 Castro Mooney 40.0 25.0 1.1 19.6 Paradis Miller AdultAdultFetal 13800 48.8 0.4 56.0 100.0 3.0 30.0 Minten 2.1 Wrutniak Larsson AdultAdultAdultAdultAdultMixedAdult 83.0 91000 90000 94.0 25.7 83.7 46.0 29.9 1.0 3.7 0.6 5.1 Larsson 14.0 2.6 Tomasi Colavita 9.8 Larsson Boily (1996) Larsson 2.9 Schumacher Felis domesticus Felis domesticus Felis domesticus Oryctolagus cuniculus Marmota monax Marmota monax Canis familiaris Canis familiaris Species Rattus norvegicus Rattus norvegicus Rattus norvegicus Rattus norvegicus Rattus norvegicus Rattus norvegicus Cavia porcellus Cavia porcellus Felis domesticus Canis familiaris Canis familiaris Canis familiaris Macaca mulatta Ovis familiaris Ovis familiaris Capra hircus Capra hircus Sus scrofa Ursus americanus Phoca vitulina Halichoerus grypus .) .) cont cont Cat Cat Cat Rabbit Woodchuck Woodchuck Dog Dog Rat Rat Rat Rat Rat Rat Guinea pig Guinea pig Cat Dog Dog Dog Monkey Lamb Sheep Nubian goatNubian goat *Goat Goat *Pig Black bear Harbour seal Gray seal Newborn 21 days old 4200 5600 64.4 86.2 235.5 103.0 De la Colina De la Colina Vertebrate Mammals ( Table 1 ( Thyroid hormones and their effects 529 . (1992) . (1985) . (1998) . (1998) . (1985) et al . (1985) . (1994) . (1994) . (1995) et al et al et al . (1997) et al . (1986) . (1986) et al et al et al et al et al et al et al AdultAdult (non-lactating)Adult (lactating)At restAfter exerciseAdultsAdultsAdultsAdultsAdults 64.3Adults 45.1 1.7 525000 49.0 1.7 9.0 19.9 8.6 1.0 1.0 34.8 0.7 Tiirats 38.0 (1997) 11.6 71.0 1.2 91.6 Tiirats 115.0 (1997) 2.6 2.1 2.8 Messer 33.2 1.9 31.0 2.2 Larsson Gonzalez 34.0 8.8 16.4 Irvine Gonzalez & Evans (1975) Arbelle 6.4 Franklyn Arbelle Wrutniak Adult 9.3 1.5 Schumacher AdultsAdult (pre antarctic)Adult (post antarctic)Euthyroid range 78000 79000 94.0 88.8 2.2 2.0 60–140 32.2 29.6 1–3 5.0 4.1 Reed Reed 10–25 1.8 3–8 Stockigt (1996) 14.3 Maes Bos taurus Bos taurus Bos taurus Equus domesticus Equus domesticus Equus domesticus Equus domesticus Mixed species Mixed species Homo sapiens Homo sapiens Leptonychotes weddelli Homo sapiens Homo sapiens Homo sapiens Homo sapiens Cattle Cattle Cattle Horse Horse Horse Horse New world primates Old world primates Human Human Weddell seal Human Human Human Human TTR, transthyretin; MSG, monosodiumnot glutamate; given PUFA, in polyunsaturated the fatty source acid; reference. SFA, saturated fatty acids; *specific name of Nubian goat was 530 A. J. Hulbert

lampreys amphibians birds lamprey TT3 amphibians TT3 bird TT3 teleost fishreptiles mammals fish TT3 reptile TT3 mammal TT3 1000 1000 Total T4 Total T3

100 100 ) M

10 10 Concentration (n 1 1

0.1 0.1 1 10 100 1000 104 105 106 1 10 100 1000 104 105 106 1000 1000 Free T4 Free T3

100 100 ) M

10 10 Concentration (p 1 1

0.1 0.1 1 10 100 1000 104 105 106 1 10 100 1000 104 105 106 Fig. 2. The concentrations of total and free thyroid hormones (T4 and T3) in the plasma of adult vertebrates. The ranges regarded as euthyroid in humans are represented by a vertical line in each graph. All data are from Table 1. concentrations are regarded as a range. For humans in different vertebrate classes. This is not the case for the approximate normal euthyroid range is typically the total thyroid hormone levels where there seem to regarded as 60–140 n for total T4, 1–3 n for total be patterns for different classes of vertebrates. For T3, 10–25 p for free T4 and 3–8 p for free T3 example, apart from fish, there seems to be a trend (taken from Stockigt, 1996). The free hormone of increasing total T4 going from amphibians to values especially are influenced by measurement reptiles, to birds, to mammals. For total T3, the techniques. As can be seen from Fig. 2, the values for levels tend to increase from amphibians to reptiles, to fish span the entire range observed in vertebrates. mammals, to birds. The total levels of hormone will This is the case for total and free levels of both T4 be greatly influenced by a combination of the and T3. The fact that most of these values come from plasma concentrations of thyroid-hormone-binding a single study (Eales & Shostak, 1987) shows that proteins together with their relative affinity for T4 this is a real situation and not an artifact of and T3. In this light, it is of interest that one of the interlaboratory variation. main plasma binding proteins in higher vertebrates, Although there are only a few values for some transthyretin, has recently been shown to have a vertebrate classes, it appears from Fig. 2 that there is higher affinity for T3 than T4 in birds, but higher no consistent pattern of different free hormone levels affinity for T4 than T3 in mammals (Chang et al., Thyroid hormones and their effects 531

1998). Thus, total thyroid hormone levels in plasma 5×10 –3 5×10 –5 5×10 –7 5×10 –9 5×10 –11 5×10 –13 5×10 –15 100 0 are likely to be largely a reflection of thyroid- hormone-binding plasma proteins. They do not 80 Solubility 20 necessarily indicate that one species of vertebrate is limit of T4 i.e. 2 µ M [T4] free more hyperthyroid than another. 60 40 Another factor of note from Fig. 2 is that there is TBG blood i.e. 30 pM no relationship between body mass and the plasma TTR 40 ALB 60

concentration of thyroid hormones. The lack of a [Prot], Mol % [T4·Prot], Mol % correlation between body mass of mammals and 20 80 either free T4 or free T3 levels is of interest in view of the well-known influence of body mass on basal 0 100 metabolism. On a body-mass basis, mice have a 5×10 –3 5×10 –5 5×10 –7 5×10 –9 5×10 –11 5×10 –13 5×10 –15 metabolic rate approximately 20-fold greater than [T4], M large mammals such as horses and cattle (Kleiber, 1961) yet the free thyroid hormone levels in their Fig. 3. The relationship between plasma free T4 con- centration and T4 binding to the human plasma proteins, plasma are similar. In the same manner, free thyroid albumin (ALB), transthyretin (TTR) and thyroxine- hormone concentrations are not greatly different binding globulin (TBG). [T4]blood l blood concentration between ectothermic and endothermic vertebrates of free T4, Prot l protein. Reprinted with permission although the resting metabolism of an endotherm is from Schreiber & Richardson (1997). 5–10 times that of a similar-sized ectothermic vertebrate at the same body temperature (Hulbert, 1980b). for T4 means they can form a multicomponent The binding of T4 and T3 to plasma proteins has, ‘buffer’ system for T4, analogous to a multi- for obvious reasons, been most studied in humans component pH buffer system with different pK where three main thyroid-hormone-binding proteins values for each component (Schreiber & Richardson, have been identified. These are thyroxine-binding 1997). This is illustrated in Fig. 3. Following their globulin (TBG), transthyretin (TTR, previously discovery, these proteins were called ‘transport’ known as thyroxine-binding prealbumin) and albu- proteins. However, the description of their role as min (ALB). Some T4 is also bound to plasma transporters probably does not fully describe their lipoproteins. The affinity for T4 is greatest for TBG function. Their role has been illustrated elegantly by −"! (approximately 1i10 ), intermediate for TTR the classic experiment of Mendel et al. (1987). These −( (approximately 7i10 ) and lowest for ALB authors showed that when liver lobules were perfused −& (approximately 7i10 ). Their affinity for T3 is with buffer, electrolytes and T4 (but no protein) in the same order but in each case is lower than that that the T4 partitioned into the first parenchymal −) for T4, being approximately 5i10  for TBG, cells with which the perfusate came into contact, −( approximately 1i10 M for TTR and approxi- with the result that the tissue uptake of T4 was very −& mately 1i10  for ALB (Robbins, 1991). How- uneven. However, when binding proteins were ever the plasma concentration of each of these added to the perfusate, there was a more uniform thyroid-hormone-binding proteins is in the opposite distribution and uptake of T4 by the cells of the order, being greatest for ALB (640 µ), intermediate perfused liver. It has been suggested that the for TTR (4.6 µ) and lowest for TBG (0.27 µ). descriptor ‘distributor proteins’ is more apt than The combination of abundance and affinity results ‘transport proteins’ when applied to the plasma in the following hormone distribution in human thyroid-hormone-binding proteins (Schreiber & plasma; TBG is estimated to carry 65% of bound T4 Richardson, 1997). The rapid uptake of T4 is and 80% of bound T3, whilst TTR carries 15% of probably related to its high hydrophobicity. This T4 and 9% of T3, and ALB has 20% of T4 and hydrophobicity would likely be exacerbated by the 11% of T3 (Robbins, 1991). It is estimated that in fact that many phosholipid head groups are highly humans approximately one in 30000 molecules of acidic. Thus, when a T4 molecule comes close to a plasma ALB has one T4 molecule bound to it, biological membrane it enters an environment that compared to one in 300 molecules of TTR and one is approximately 1.5 pH units more acid than the in three molecules of TBG (Schreiber & Richardson, general medium, thus reducing the degree of 1997). ionization of the phenolic kOH group, which in turn The difference in the affinity of these three proteins increases the hydrophobicity of the T4 molecule and 532 A. J. Hulbert hastens its partitioning into the membrane lipid the even distribution of thyroid hormones through- bilayer (Hillier, 1970). The same argument would out the body required a more powerful distributor also apply to T3 which is more hydrophobic than T4 protein in their plasma. to begin with, having a smaller proportion of In amphibians, TTR is a plasma protein for only molecules with an ionized phenolic kOH at a short period during metamorphosis, but not in the physiological pH. adult amphibian (Yamauchi et al., 1998). In adult Although humans have three thyroid-hormone- birds, diprotodont marsupials and eutherian mam- binding proteins in their plasma this is not true for mals, TTR is a plasma protein. It is also secreted by all vertebrates. An extensive survey of thyroid- the choroid plexus cells into the cerebrospinal fluid hormone-binding proteins in the plasma of 93 by adult reptiles, birds and mammals. Indeed, it is vertebrate species (Richardson et al., 1994) together the predominant protein secreted by the choroid with the excellent review by Schreiber & Richardson plexus in all such vertebrates examined and has been (1997) shows that ALB is the oldest thyroid- postulated to be a T4 transport system into the brain hormone-binding plasma protein, being found in all (Dickson et al., 1987; Chanoine et al., 1992; vertebrates examined. TTR is a cerebrospinal fluid Southwell et al., 1993). It is probably responsible for protein (secreted by the choroid plexus) in adult the even distribution of T4 throughout brain tissue reptiles, birds and mammals but only in adult birds, (Schreiber & Richardson, 1997). Whilst amino acid polyprotodont marsupials and eutherian mammals residues at the surface of TTR show a rate of is it a plasma protein. This represents at least three evolutionary change similar to proteins such as separate occasions during the evolution of higher albumin, the central T4-binding channel of TTR is vertebrates where TTR became a plasma protein as highly conserved as the histone H4, one of the secreted by the adult liver. Recently, a TTR has most strongly conserved proteins found in nature been reported in a fish species, the sea bream Sparus (Schreiber & Richardson, 1997). Although the aurata (Santos & Power, 1999). amphibian choroid plexus does not secrete TTR into A third T4-binding protein, migrating slower the cerebrospinal fluid bathing the brain, it does than albumin, occurs more sporadically among the secrete a lipocalin which may have a similar function vertebrates. It has been characterized in humans, (Achen et al., 1992). rats and sheep as TBG and in the turtle as vitamin- In view of the findings of Mendel et al. (1987) and D-binding protein. It is present in most large the discussion of Schreiber & Richardson (1997), it eutherian mammals and is probably also TBG in may be that in large vertebrates (such as humans) these species. In other vertebrate species, its precise there is a requirement, because of the distances identity is not known and it is suggested that it may involved, for high-affinity plasma proteins to be one of the apolipoproteins (Schreiber & counteract the tendency for thyroid hormones to Richardson, 1997). TBG is predominantly respon- partition into membranes, and thus to ensure a sible for the high total T4 concentration in the uniform distribution of thyroid hormones through- plasma of adult humans compared to most other out the whole body rather than a hormone gradient vertebrates (Larsson, Pettersson & Carlstrom, 1985). radiating out from the sites of T4 synthesis (the TBG is not present in all primate species. It is found thyroid gland) and T3 manufacture (such as the in adults from the Catarrini and Prosimiae but not thyroid, liver & kidney). the Platyrrini (Seo et al., 1989). Whilst TBG is not A recent observation, that in the tammar wallaby found in significant amounts in the plasma of adult Macropus eugenii, there are plasma proteins which rats or mice, it is present during the early postnatal bind significant amounts of thyroid hormones during period (Vranckx, Savu & Nunez, 1989). development that are not present in the adult (S. R. In present-day adult fish, amphibians and reptiles Richardson, unpublished data), together with the (and presumably early vertebrates), albumin is findings that TTR is secreted by the amphibian liver adequate as the sole plasma thyroid-hormone- only during metamorphosis (Yamauchi et al., 1998) binding protein but in adult birds and most and that TBG appears in the plasma for only a short mammals there is at least one other thyroid- period during the early postnatal life of the rat hormone-binding plasma protein with a higher (Vranckx et al., 1990), suggests it is likely that full affinity than albumin (and in the case of some understanding of the evolution of plasma thyroid- eutherian mammals two such proteins). It may be hormone-binding proteins in vertebrates will have to that at the high metabolic activity and constant wait until more is known about their biology during body temperatures of the endothermic vertebrates, vertebrate development. Thyroid hormones and their effects 533

The presence of multiple thyroid-hormone-bind- deprivation results in reduced serum levels of thyroid ing proteins illustrates a degree of redundancy and hormones in horses (Messer et al., 1995) and of T4 in thus a potential safety factor. For example, in penguins (Groscolas & Leloup, 1989). documented cases where one of these plasma proteins As well as food deprivation, non-thyroidal illness is absent, the individuals concerned show no ill can also reduce plasma levels of thyroid hormones, effects and appear to be healthy and exhibit normal especially T3. In humans, T4 concentrations change development. This is illustrated for mice in Table 1, very little, whilst those of T3 decrease and rT3 where although total plasma T4 concentration is increase during starvation (Danforth, 1986). Similar reduced in TTR-null mice, they have normal free changes occur during non-thyroidal illness (Kaptein, T4 and T3 plasma concentrations (Palha et al., 1986). In cats, non-thyroidal illness results in a 1994). Similar deficiencies in TBG and albumin in significantly lower total T4 concentration but no humans, and albumin in rats, as well as other change in free T4 levels (Mooney, Little & Macrae, inherited or acquired variations in these proteins, 1996). still result in phenotypically euthyroid individuals Blood levels of T4 represent the balance between (see Mendel, 1989; Bartalena & Robbins, 1992). the rate of hormone secretion and disposal. Changes When studied in sheep, thyroid hormones ex- in plasma levels may reflect changes in the regulatory changed between the plasma pool and interstitial system or may simply reflect changes in the steady fluids easily and appeared in the lymph faster than state. Changes in total levels need not reflect changes did plasma proteins. Free T4 levels in the plasma in thyroid status. For example, during pregnancy in and lymph samples were similar (Simpson-Morgan humans, total T4 concentration increases due to & Sutherland, 1976) showing that free T4 con- elevated serum TBG levels, however free T4 con- centration experienced by cells is the same as that centration remains in the normal non-pregnant measured as plasma free [T4]. range (O’Leary et al., 1992). Most reported changes The plasma concentration of T4 is controlled in a are for total hormone levels and whether there are homeostatic manner by the secretion of TSH from similar changes in free levels is often not known. the anterior pituitary. There is a diurnal cycle of Indeed, with such a large protein-bound fraction in TSH secretion from the anterior pituitary. In the plasma, substantial changes in free T4 levels are humans, peak TSH concentration occurs in the probably not easily achieved over short periods of night and the nadir in the late afternoon (Fisher, time. 1996). Diurnal cycles have been reported for thyroid Since T4 disposal appears related to metabolic hormone levels in rats (Cokelaere et al., 1996), rate (see Section II.3), some reported changes in amphibians (Gancedo et al., 1997) and fish (Cerda- thyroid hormone levels may best be interpreted as Reverter et al., 1996; Pavlidis et al., 1997). changes in the steady state following changes in Thyroid hormone levels are also influenced by metabolic rate. For example, in the rat, cold feeding. In young pigs, levels of both T4 and T3 exposure results in increased T4 degradation and a increased after a meal, peaking approximately 60 decrease in plasma total T4 levels (Gregerman, min after feeding; the rise was greater for T3 than 1963). That thyroid secretion rate also increases T4 and was dependent on both the energy content during cold exposure (Lachiver & Petrovic, 1960) is and nutrient composition of the meal (Dauncey et probably best explained as a homeostatic response to al., 1983). This post-meal increase involves both maintain relatively constant plasma T4 levels in the total and free hormone levels, is relatively immediate cold, rather than a hormonal response to increase and although it depends on the energy content of the heat production in the cold. If the latter were the food, is not directly related to changes in blood case we would expect to measure an increase in glucose levels. It is also observed in thyroidectomized plasma T4 levels rather than a decrease. animals and likely related to changes in peripheral Chronic exercise (swimming) in rats results in no deiodination and redistribution of hormonal pools change in either total or free levels of T4 or T3 (Dauncey & Morovat, 1993). The circadian rhythms (Woody et al., 1998). Exercise in thoroughbred in plasma thyroid hormone levels in rats are horses resulted in no change in total T4 levels but an influenced by their feeding pattern and are most increase in total T3 concentration (Gonzalez et al., pronounced when the rats are fed only once a day 1998). A 48 km flight in pigeons resulted in a 35% (Cokelaere et al., 1996). The diurnal cycle in thyroid decrease in total T4, a 60% decrease in total T3 and hormone concentrations is influenced by nutritional a 130% increase in rT3 levels (George & John, state in fish (Cerda-Reverter et al., 1996) and food 1992). 534 A. J. Hulbert

30 Chelonia mydas & Mexican tortoise Gopherus flavo- marginatus) had low plasma T4 levels with no

) Liver significant seasonality, whilst the painted turtle –1

g Chrysemys picta showed significant seasonal variation µ in plasma total T4 levels with the lowest values Kidney 20 associated with the hibernation period (Licht et al,. 1991). In general, thyroid hormone plasma concen- trations appear much less labile than the levels for many other hormones. Often the changes measured in total hormone concentration reflect changes in 10 Heart Spleen thyroid-hormone-binding proteins rather than Testis changes in thyroid status. The protein-bound thy- roid hormones in the plasma represent a significant Tissue thyroxine concentration ( Tissue Muscle hormonal store that will act to dampen changes in the hormone content of some tissues. For example, 0 0 10 20 30 when rats are thyroidectomized, within two weeks Phospholipid content (mg/g) plasma concentrations of T4 and T3 fall dramati- Fig. 4. The relationship between phospholipid content cally by 96-98% whilst muscle, heart and brain T4 and thyroxine (T4) content of various tissues in the rat. and T3 contents only fall by approximately 50% Reprinted with permission from Hillier (1970). over the same period (Obregon et al., 1981). Thyroid hormone concentrations appear most variable be- tween fish species. An examination of seasonal variations in hormone concentrations of humans showed a small but (3) Cellular uptake, cellular location and significant variation in total T3 but none in free T4 hormone metabolism levels (Maes et al., 1997). Prolonged residence in Antarctica resulted in no significant change in either In 1968, Hillier examined the uptake and release of total or free T4 levels in humans but did result in a T4 and T3 by the isolated perfused rat heart. Initial significant decrease in both total and free plasma rates of uptake were similar for both T4 and T3, concentration of T3 (Reed et al., 1986). whilst a steady state was reached after 30 min Animals that do show substantial seasonal vari- perfusion for T4 and after 120 min for T3. T3 ation are often those that have a seasonal inactive equilibrated to a ‘space’ five times that of T4 (which " (hibernation) phase. For example, woodchucks was 15 ml g− heart). This ‘space’ was independent (Marmota monax) show a seasonal variation in free T4 of perfusate hormone concentration (13 p to 1.3 levels with low values at the autumnal equinox µ), and the rates of equilibration were relatively (Rawson et al., 1998) and hedgehogs Erinaceus unaffected by temperature. Washout curves sug- europaeus show a low level of total T4 during winter gested two compartments, a fast-release compart- (Fowler, 1988). In the black bear Ursus americanus, ment which was the same size for T4 and T3 and a which has an inactive period known as hibernation slow-release compartment which was much greater (without the dramatically depressed body temper- for T3 than T4 (Hillier, 1968b). These results atures of other mammalian hibernators) there is a suggested a purely physical process of partitioning small decrease in free T4 and free T3 levels which between the perfusate and a compartment within seems to be related to food restriction (Tomasi, the heart in which the hormones are much more Hellgren & Tucker, 1998). In reptiles, both the viper soluble (T3 more so than T4). He also showed, that Vipera aspis (Naulleau, Fleury & Boissin, 1987) and with serum (containing hormone-binding proteins) the desert iguanid Dipsosaurus dorsalis (John-Alder, in the perfusate, the uptake of hormones was reduced 1984b) have a low total T4 concentration during (more so with T4 than T3), that hormone uptake their annual hibernation period, whereas in the non- became more dependent on perfusate concentration hibernating green sea turtle Chelonia mydas, plasma and that the release of accumulated hormone was total T4 concentration is constant throughout the also accelerated (Hillier, 1968a). Two years later, in year (Licht, Wood & Wood, 1985). In a study of a follow-up study examining the binding of thyroid three diverse turtle species, two species (green turtle hormones to phospholipid membranes he showed Thyroid hormones and their effects 535

Plasma Erythrocytes

Salivary gland Stomach Duodenum Jejunum Ileum Appendix Colon

Liver Kidney Brown fat

Lung

Heart Diaphragm Skeletal muscle

Thymus Spleen Lymph node

Spinal cord Medulla Total brain Cerebral cortex (5%) Cerebellum (14%) Pons (9%) Hypothalamus (1%) Remainder (71%)

Anterior pituitary Thyroid Adrenal Pancreas

Testis Epididymis Prostate

0 1020304050600 2 4 6 8 100 20 40 60 80 [T4] (pmol g–1) [T3] (pmol g–1) % T3 from local deiodination Fig. 5. T4 and T3 content of tissues from the euthyroid rat. The % T3 formed from local deiodination of T4 is also shown. All data are taken from van Doorn et al. (1985). that there was a strong correlation between the from local deiodination of T4 (van Doorn, Roelfsema phospholipid content of different tissues and the & van der Heide, 1985). The T4 and T3 contents of " tissue T4 concentration of the same tissues. This most tissues are in the 1–10 pmol g− range, with the relationship is shown in Fig. 4. liver and kidney having higher tissue T4 contents of " " Since the studies of Hillier (1968a, b, 1970) the 31 pmol g− and 16 pmol g− , respectively. Liver and tissue concentrations of T4 and T3 have been kidney T3 contents were more typical of other tissues " " measured both by isotopic equilibrium analysis and being, respectively, 7 pmol g− and 10 pmol g− .Itis radioimmunoassay of extracted hormones with the of interest that the liver has more T4 per gram than values obtained being similar for the two techniques does the thyroid gland where most thyroid hormones (Obregon, Morreale de Escobar & Escobar del Rey, are found in the form of thyroglobulin. The plasma 1978). In Fig. 5 the T4 and T3 tissue concentrations has the highest T4 content but most of this is protein- for the rat are shown, including the % T3 formed bound with the free concentration being only 536 A. J. Hulbert approximately 20 p (equivalent to approximately from the liver to the plasma which appears to be " 0.02 pmol g− ). It is not possible directly to translate passive (Krenning & Docter, 1986). Carrier-me- " tissue contents in pmol g− to equivalent concen- diated, saturable uptake of thyroid hormones has trations because cells are not homogenous but also been described for isolated adipocytes (Parl et contain compartments such as membranes, in which al., 1977; Landeta, Gonzalez-Padrones & Rod- iodothyronines have a much greater solubility, as riguez-Fernandez, 1987), cultured fibroblasts well as cellular proteins to which they can bind. (Cheng et al., 1980), cultured pituitary tumour Traditionally, intestinal contents have been ex- (GH3) cells (Horiuchi et al., 1982), lymphocytes cluded in studies analyzing the tissue distribution of (Holm et al., 1980) and more recently in cultured thyroid hormones (including Fig. 5) but recent cells from rat anterior pituitary where T3 and T4 studies show that there are significant quantities of may enter by the same carrier (Everts et al., 1993, both T4 and T3 in intestinal contents and that these 1994). It has recently been demonstrated in neonatal are readily exchangeable with other body compart- rat cardiac myocytes (Everts et al., 1996) but not in ments (Nguyen et al., 1993). cultured human muscle cells (Bolhuis et al., 1983), in The thyroid hormones with non-ionized phenolic cultured glial C6 cells (Yusta et al., 1988) nor in groups are amphipathic molecules; although they choroid plexus (Dickson et al., 1987). It seems that are highly soluble in phospholipids, spin-label studies thyroid hormone uptake is controlled differently in suggest that T3 does not flip-flop across membranes hepatocytes and anterior pituitary cells (Everts et al., to any appreciable extent and will tend to remain in 1995). Many studies suggest an interaction between the outer half of the bilayer (Lai et al., 1985; amino acid transporters and the iodothyronines and although an effect of the spin label itself influencing some suggest the iodothyronine transporter is either the property measured cannot be ruled out in this a previously described amino acid or neuro- particular study). Interestingly, the spin-labelled T3 transmitter transporter. It may be that iodo- was measured to have a lateral diffusion constant in thyronine transporters and these other transporters −) # −" the membrane of 3i10 cm s at 31 mC (Lai & are related. In some situations, it appears that Cheng, 1984) which is similar to the estimated rate plasma membrane transport is an important limiting of diffusion through cytoplasm for T3 (Luxon & factor for further metabolism of thyroid hormones Weisinger, 1992). The thyroid hormones are likely within cells (for review see Hennemann et al., 1998) to enter cells by a number of possible means, Thyroxine uptake by hepatocytes from juvenile including in some situations by endocytosis of rainbow trout Oncorhynchus mykiss suggests that the protein-bound thyroid hormones, as well as by basic system in fish liver cells closely resembles that specific carriers. They can be taken up into many observed for mammalian hepatocytes, in that it is a cells by an energy-dependent transport system stereospecific, saturable, energy-dependent carrier- located in the plasma membrane (Rao et al., 1976; mediated process. In these fish hepatocytes, thy- Krenning et al., 1978). The membrane iodothyronine roxine uptake does not depend on extracellullar Na+ transporters have been reviewed by Kragie (1994, but may involve an endocytotic component (Riley 1996). and Eales, 1993). Many uptake studies do not take into account the Of all body tissues of the rat, the brain shows the partitioning of thyroid hormones into lipid mem- greatest contribution of tissue T3 derived from local branes in their measurement of uptake. Studies that deiodination (see Fig. 5). It has a specialized system show saturation suggests a carrier-mediated process of T4 uptake via the unidirectional secretion of newly and the use of various inhibitors suggest that uptake synthesised TTR by the choroid plexus, that enables is often energy-dependent and sometimes also de- T4 to cross the blood-brain barrier (Dickson et al., pendent on the transmembrane Na+ gradient. 1987). Labelled T4 is rapidly accumulated by the Uptake is also often shown to be stereospecific. In the choroid plexus both in vivo and in vitro. This process rat liver cell, there appears to be at least two carriers, appears to be simple partitioning between the blood one which preferentially transports T4 and the other plasma and the choroid plexus membranes. Thus, transporting T3 (Krenning & Docter, 1986). Al- the choroid plexus acts as a transport system for T4 though less than half the T3 found in liver comes to pass from the blood to the brain via the directly from local deiodination (Larsen, Silva & cerebrospinal fluid (Dickson et al., 1987). Kaplan, 1981; see Fig. 5), the tissue is an important Once taken up into cells, thyroid hormones can be source of systemic T3 (from the deiodination of T4). found either (i) ‘free’ in aqueous solution, (ii) There is thus also a substantial efflux of T3 (and T4) ‘bound’ to cellular proteins or (iii) located within Thyroid hormones and their effects 537 the lipid matrix of the various cellular membranes. the interface between lipid droplets and surrounding They will be in some form of steady state between all aqueous environment rather than evenly distributed three classes of sites. I know of no work that has in the lipid body. Kriz, Fong & Goldfine (1981) also attempted, in any cell type, to determine the relative reported evidence for the localization of T3 at amounts of either thyroid hormone in each of these multiple cellular sites. three locations. Hillier (1970) found that in rat liver homogenate As determined by differential centrifugation of 12–15% of T4 was present in the ‘free’ state (in the tissue homogenate, the subcellular distribution of aqueous compartment). Simple alcoholic extracts "#& I-labelled T4 in normal rat liver is 54, 18, 17 and indicated that at least half of the T4-binding capacity 11% in microsomes, supernatant, nuclei and mito- of the homogenates was due to lipid material and chondria respectively, whilst the corresponding boiling the homogenate (which will denature bind- protein distribution is 25, 37, 23 and 15% re- ing proteins but not significantly affect the relatively spectively (Schwartz, Bernstein & Oppenheimer, thermostable phospholipids) did not impair its 1969). The subcellular location of thyroid hormones ability to bind T4. The influence of this hydrophobic has also been assessed in some tissues by the use of compartment of the cell (i.e. its membranes) on the autoradiography. In cultures of developing nervous distribution of thyroid hormones is shown by the "#& tissue, Manuelidis (1972) showed that I-labelled correlation between tissue phospholipid content and T4 accumulated at the cell membrane, mitochon- tissue T4 content (see Fig. 4). dria, endoplasmic reticulum, nucleus and the sy- Measurements of T4 and T3 concentration in napse in order of decreasing intensity. Following the human erythrocytes show that 60% of T3 and 34% "#& injection of 0.3–1.0 nmol of I-labelled T3 into the of T4 were associated with the cell membrane even rat, Bar-Sella, Stein & Gross (1973) examined its after seven washes of erythrocyte ‘ghosts’ (Bregen- sub-cellular localization in the posterior pituitary gaard et al, 1989). The membrane can be calculated and median eminence. The ‘subcellular sites of to account for approximately 1% of the total cell hormone concentration were found to correspond to volume using values for the volume (94 fl) and # intracellular membrane structures, such as the membrane surface area (135 µm ) of an average mitochondria, Golgi apparatus and nuclear en- human erythrocyte (Evans & Fung, 1972). These velope’ and ‘many of the grains counted with the amphipathic hormone molecules are largely asso- nucleus are actually associated with the nuclear ciated with the membrane bilayer because of their envelope and some of the grains counted with the physical chemistry (and thus not ‘bound’ to it in the cytoplasm may be due to labelled hormone localized normal sense). These measurements are likely to be on the plasma membrane’ (Bar-Sella et al., 1973, underestimates in light of the extensive washing of p. 1412). In isolated liver cells, autoradiography of the membrane preparations prior to the measure- labelled T3 showed that the hormone was sub- ment process. The membrane component will in- cellularly associated with mitochondria, endoplas- clude T4 and T3 both in the lipid matrix of the mic reticulum, cytoplasm, nuclei, lysosomes and bilayer as well as that bound to membrane proteins. lipid in decreasing intensity (Sterling et al., 1984). The fact that the membrane to cytosol ratio for the Because of the physics and geometry involved in various iodothyronines was the same in euthy- transmission electron microscopy, it is only possible roidism, hyperthyroidism and hypothyroidism sug- to view membranes along the plane of the bilayer gests a purely physical relationship with thyroid and not perpendicular to the plane of the bilayer. hormones possibly concentrated mostly in the lipid Thus, it is possible that some of the T3 molecules bilayer. High-affinity binding sites have also been assigned to the cytoplasm may actually be mem- measured in both human and rat erythrocyte brane located (e.g. in endoplasmic reticulum that is membranes (Angel, Botta & Farias, 1989) in the plane of the section). Thus, electron micro- Although the amount of thyroid hormones in scopic autoradiography will underestimate the as- membranes may not be great, because of their sociation of thyroid hormones (and other amphi- relatively small volume compared to the aqueous pathic molecules) with membranes because we are compartment of cells, the concentration of thyroid not able to ‘see’ all cellular and subcellular hormones in membranes will be many times greater membranes. Their relative absence in lipid droplets than that in the aqueous compartment of the cell. in cells may reflect the fact that the hormones are Both T4 and T3 bind to a number of cellular amphipathic rather than completely hydrophobic proteins. In the euthyroid rat, it is estimated that less and thus they are more likely to be associated with than 15% of tissue T3 is bound to the specific 538 A. J. Hulbert nuclear receptors, except in the anterior pituitary protein-bound molecules. The measured rate for a where it is approximately 50% (Oppenheimer, saturated fatty acid analogue is approximately −) # −" Schwartz & Surks, 1974). How much of the 0.4i10 cm s (Weisinger, 1996). extranuclear T4 and T3 is bound to extranuclear Intracellular T4 and T3 undergo various meta- proteins is unknown. Extranuclear binding sites are bolic transformations. Knowledge of these metabolic located on plasma membranes, in the mitochondria transformations and their relative importance in and in the cytosol and many have a high affinity for different tissues has exploded over the last two T3. For a list of extranuclear binding sites and their decades and is reviewed in detail elsewhere (see respective affinities see Kragie (1996) and for plasma Hennemann, 1986; Kohrle, Brabant & Hesch, 1987; membrane binding sites see Segal (1989b). Binding Berry & Larsen, 1992; Leonard & Koehrle, 1996; to the nuclear envelope has also been reported St. Germain & Galton, 1997). Most is known about (Lefebvre & Venkatraman, 1984) and a protein the various deiodinative pathways which are re- located on the luminal face of the endoplasmic sponsible for approximately 80% of the daily T4 reticulum and nuclear envelope also binds thyroid breakdown in humans with non-deiodinative path- hormones (Kato, Velu & Cheng, 1989b). One ways (conjugation, deamination and decarboxyl- cytoplasmic binding protein is the monomer of ation as well as oxidative degradation) responsible pyruvate kinase subtype M#, that can bind either T4 for the remaining 20% (Burger, 1986). The non- or T3; binding, in turn, inhibits the enzymatic deiodinative pathways may be responsible for up to activity of the monomer. This enzyme also exists in approximately 56% of T3 metabolism in humans a tetrameric form which has greater enzymatic (Burger, 1986). The pathways of thyroid hormone activity than the monomeric form but the tetrameric deiodination are shown in Fig. 6. form does not bind thyroid hormones (Kato et al., Monodeiodination of T4 results in the removal of 1989a). iodine from either the outer (phenolic) ring or inner Very little is known about either the specific (tyrosyl) ring and produces T3 or rT3, respectively. identity of most of these proteins or the physiological These two molecules can be further deiodinated to importance of the binding. Some of them may be three types of diiodothyronines (3,5-T2, 3,3h-T2 or involved in the transport and the metabolism of the 3h,5h-T2), which in turn can be deiodinated to form thyroid hormones whilst others may have their two monoiodothyronines (3-T1 or 3h-T1), which can functions affected by such binding and thus mediate be converted to the non-iodinated thyronine (see some thyroid hormone effects. In rat liver, one of the Fig. 6). There are three separate types of deiodinases, cytosolic binding sites has been identified as the initially differentiated by kinetics and patterns of glutathione-S-tranferase enzymes and although the inhibition etc., they can now be identified by their thyroid hormones are potent inhibitors of the cDNAs. They have been called type I, type II and activities of these enzymes this may not be physio- type III, and more recently D1, D2 and D3 (St logically important in the euthyroid state (Ishigaki, Germain & Galton, 1997). All three appear to have Abramovitz & Listowsky, 1989). The glutathione-S- a selenocysteine residue at the active site of the transferase enzymes are known to bind many enzyme, with selenium being very important in different lipophilic compounds (Mannervik, 1985). determining catalytic efficiency (Berry & Larsen, That much of thyroid hormone binding to proteins 1992). All require cytosolic thiols for their activity involves hydrophobic interactions is illustrated by and they possess both different affinities for substrates the fact that fatty acids are capable of competitively and have different relative affinities for the range of inhibiting binding both to plasma proteins (e.g. Lim iodothyronines. All three deiodinase enzymes are et al., 1988) and to the nucleus (Wiersinga, Chopra membrane-bound enzymes located mainly in the & Teco, 1988; Inoue et al., 1989; van der Klis, microsomal fraction of tissue homogenates suggest- Wiersinga & de Vijlder, 1989). ing an endoplasmic reticulum and\or plasma mem- Amphipathic molecules diffuse through hepatic brane location. cytoplasm relatively slowly. Estimation of the in- D1 is probably found in all tissues but has tracellular transport rate of T3 suggests a rate of especially high activity in liver, kidney, thyroid −) # −" approximately 3i10 cm s (Luxon & Weisinger, tissue and the central nervous system. In kidney, D1 1992). This is approximately the same rate as that is found on the basolateral plasma membrane of measured for the lateral diffusion of T3 in membrane proximal convoluted tubule cells, whilst in the liver bilayers (Lai & Cheng, 1982, 1984) and is also it is located on the endoplasmic reticulum. It is typical of aqueous diffusion of either amphipathic or capable of both outer- and inner-ring deiodination Thyroid hormones and their effects 539

! 5 ! 3 4 ! 3 5 T4 ! Outer ring 5 -deiodination

rT3 T3

Inner ring ! ! ! ! 3,3 -T 3 ,5 -T2 5 -deiodination 3,5-T2 2

! 3 -T1 3-T1 Iodine Nitrogen T0 Oxygen Fig. 6. Pathways of the deiodination cascade of T4 resulting in the various iodothyronines. Solid arrows represent outer ring monodeiodination and broken arrows represent inner ring monodeiodination. Reprinted with permission from Kohrle et al. (1986). and in many animals is inhibited by the thyroid human placenta has been shown to have an essential inhibitor propylthiouracil (PTU). A PTU-insen- requirement of phospholipids for activity (Santini et sitive D1 deiodinase has recently been described for al., 1992). a teleost fish species (Sanders et al., 1997). All three deiodinases are evolutionarily quite old: D2 has been found in the central nervous system, cDNAs have been isolated for all three types from brown adipose tissue, anterior pituitary and placenta various mammalian, avian, amphibian and fish and is capable of outer-ring deiodination. It has a species and an evolutionary tree determined (Fig. 7). higher affinity for T4 than does D1, and is not The separation of the three deiodinase enzymes inhibited by PTU. D2 is located on the plasma predates the separation of the ancestors of extant membrane in the cerebral cortex. In situ hybridiz- vertebrates. The deiodinases are related to each ation histochemistry in the rat hypothalamus and other and their relative tissue distribution varies pituitary shows that D2 mRNA is not evenly between vertebrate species. It has become obvious distributed but is heavily concentrated in tanycytes, that these deiodinases are an integral part of the which are glial cells that reside in the floor and vertebrate thyroid hormone axis. They are them- lateral walls of the third ventricle having long selves influenced by thyroid status and also appear to cytoplasmic processes that extend into the adjacent be important during development. The tissue dis- neuropil of the medial hypothalamus and median tribution of the various deiodinases suggests that eminence where they are intimately associated with they are responsible for the intracellular concen- axon terminals and blood vessels (Tu et al., 1997). In trations of thyroid hormones in different tissues. the neonatal rat brain, the D2 deiodinase is primarily Their importance in determining tissue-specific expressed in glial cells, rather than neurones intracellular concentrations of T4 and T3 is illus- (Guadano-Ferraz et al., 1997). In the adult rat trated by the recent finding that in hypothyroid rats, brain, the choroid plexus lacks deiodinase activity euthyroidism in all tissues can only be achieved by (Southwell et al., 1993). the combined infusion of both T4 and T3 at D3 is found in the central nervous system and the approximately the relative rates that both T4 and placenta and carries out inner-ring deiodination. It T3 are secreted by the rat thyroid gland: 1.2 nmol is also not inhibited by PTU. The D3 isolated from T4 and 0.23 nmol T3\100 g\day (Escobar-Morreale 540 A. J. Hulbert

Human ultrafiltration) are 30 p T4, 4.8 p T3, 0.6 p rT3, Dog 0.4 p 3,3h-T2 and 0.8 p 3h,5h-T2 (Faber et al., Mouse D1 Rat 1984). We know very little of the concentration of Chick these iodothyronines in other vertebrates. Tilapia The iodothyronines produced from T4 and T3 are Tilapia less lipophilic than either T4 or T3. Information Chick regarding their relative hydrophobicity can be D2 Rana Xenopus garnered from their partitioning between the eryth- Rat Human rocyte cytoplasm and membrane. In human eryth- rocyte ghosts, the following percentages of iodothyro- Human nines partition into the membrane; 60 % for T3, D3 Rat Rana 34% for T4, 23% for 3,3h-T2, 16% for rT3 and 3% Fundulus for 3h,5h-T2, with these values being the same in Fig. 7. Evolutionary tree diagram of the deiodinase hypo- and hyperthyroidism (Bregengaard et al., enzymes (D1, D2, D3) from various vertebrates. Re- 1989). The order of these iodothyronines is similar to printed with permission from St. Germain & Galton their expected hydrophobicity at physiological pH, (1997). when both the lipophilicity of the iodine (the more iodine atoms the more hydrophobic the molecule) et al., 1996). The deiodinases are also involved in the together with the degree of ionization of the phenolic autoregulation of T3 at the cellular level during OH group (the more iodines on the phenolic ring the periods of altered plasma concentrations of thyroid less hydrophobic the molecule) are taken into hormones, especially in the brain (Dratman et al., account. The calculated phospholipid:cytoplasm 1983; St. Germain & Galton, 1997). partition ratios (assuming the cell membrane is 1% The presence of intracellular deiodinases results in of total cell volume) for T3, T4, 3,3h-T2, rT3 and the fact that most of the body’s T3 is produced by 3h,5h-T2 are approximately 150, 50, 30, 20 and 3, deiodination of T4. It is estimated that, in the rat, respectively. These values are likely underestimates approximately 55% of the circulating T3 comes because of the seven washes of the ‘ghosts’ carried from the thyroid gland, with a significant amount of out before measurement (Bregengaard et al., 1989). this being produced by the intrathyroidal deiodin- In normal euthyroid humans (70 kg body mass), ation of T4 by the thyroid’s own D1 deiodinase the production rates of the iodothyronines are " (Chanoine et al., 1993a). The remainder of the estimated to be 130 nmol day− for T4, 48 nmol " " " circulating T3 comes predominantly from the liver day− for T3, 60 nmol day− for rT3, 50 nmol day− −" −" and kidney. Whereas dietary selenium deficiency in for 3,3h-T2, 21 nmol day for 3h,5h-T2, 7 nmol day −" the rat results in a dramatic decline in both liver for 3,5-T2 and 30 nmol day for 3h-T1 (Chopra, selenium content and liver D1 deiodinase activity, 1996). The deamination products of T4 and T3 are " the decline in thyroid selenium content is modest produced at only 1.8 and 5.8 nmol day− respect- " and thyroid D1 deiodinase actually increases in ively, whilst 10 nmol day− is the production rate of activity (Chanoine et al., 1993a). the T3 sulphate conjugate (Chopra, 1996). The relative importance of T3 derived from local In a study of the turnover of six iodothyronines cellular deiodination for the T3 content of tissues (T4, T3, rT3, 3,3h-T2, 3h5h-T2 and 3hT1) in the varies from 2–4% in lung and skeletal muscle to euthyroid rat, DiStefano & Feng (1988) found the 65% in cerebral cortex (see Fig. 5). distribution of all six iodothyronines could best be Deiodination is also responsible for the other characterized by three pools (in increasing size); a iodothyronines found in vertebrates. In human plasma pool, a rapidly exchanging pool and a slowly serum, total rT3 is approximately 600 p, the three exchanging pool. The respective pool sizes for T4 T2s are 55–85 p, and T1 is approximately 60 p (as% of total pool size) were approximately 25, 18, (Chopra, 1996). The total serum concentrations of and 57, whilst for T3 they were 2, 24 and 74. They Tetrac and Triac (the deaminated products of T4 were similar for the other four iodothyronines, being and T3) are 7.2 n and 2.8 n respectively whilst approximately 3–9, 8–22 and 60–90% of total pool the sulphated conjugates of T4 and T3 are, re- size for the plasma, rapidly exchanging and slowly spectively, approximately 20 p and approximately exchanging pools, respectively. The precise identity 75 p (Chopra, 1996). The free concentrations of of these two non-plasma pools is not known but there iodothyronines in adult human serum (measured by appear to be at least two possibilities; they could Thyroid hormones and their effects 541 reflect the tissue distribution of the thyroid hormone variation in BMR is more likely the ‘cause’ of the uptake mechanisms into two groups or they could observed variation in T4 secretion rates in mammals represent two physical compartments, respectively, rather than vice versa. This does not preclude the fact an aqueous non-membrane compartment and a that thyroid status (which is related to the con- lipophilic membrane compartment. The fact that centration of the thyroid hormones rather than their the slowly exchanging pools are the major sites of T4 turnover) influences the metabolic rate of mammals monodeiodination suggests the second scenario. specifically, and vertebrates in general. In the rat, faecal excretion accounts for 24% of It was originally thought that the metabolism T4, 30% of T3, and 6% of T1 turnover but only (especially deiodination) of thyroid hormones was 1–3% of disposal of rT3 and the T2s. The remaining involved in the exertion of thyroid hormone effects; portion of each iodothyronine was completely and however, because of the dominance of the nuclear irreversibly metabolised. The kinetic variables for all receptor paradigm as the only accepted mode of the iodothyronines except T4 and T3 were similar, thyroid hormone action, this idea has not received suggesting that similar mechanisms are responsible much attention. Early work (Oppenheimer et al., for their transport, metabolism and distribution. 1971) that is cited as evidence for a dissociation The production rates (per 100 g body mass) are between hormonal deiodination and hormonal ac- " approximately 1 nmol T4 day− , approximately 0.25 tion is not valid if these hormones act at more than " " nmol T3 day− and 0.13–0.84 nmol rT3 day− (Di- one site in the cell. The phenobarbital-induced Stefano & Feng, 1988). increase in T4 deiodination used by the above study Expressed per 100 g body mass, the human is due to an increased amount of microsomal " production rate is 0.19 nmol T4 day− and 0.07 membrane per cell and would likely result in a " nmol T3 day− . Both these rates are 20–25% of the decreased T4 availability for other sites in the cell. respective production rates measured for the rat and Thus, if T4 acts at multiple sites in the cell, as I will illustrate the body size versus metabolism relationship argue later, we would expect a significant decreased in mammals (see Kleiber, 1961). The basal meta- effect of T4 at these other sites also. This is exactly bolic rate (BMR) of mammals is proportional to what was observed by Oppenheimer et al. (1971); body mass to the 0.73 power. A recent compilation the two effects measured were resting oxygen of T4 and T3 utilization rates by euthyroid mammals consumption and the activity of the mitochondrial (which for T4 equals the hormonal production rate) enzyme, glycerolphosphate dehydrogenase. If there has shown that these rates scale to body mass to the are multiple sites of action in the cell, the possibility 0.74 and 0.81 power, respectively (Tomasi, 1991). that deiodination is involved in some thyroid Thus in mammals there is a correlation between the hormone effects should remain open. BMR and the turnover of T4 (and T4 secretion from In humans, low-energy diets lead to changes in the thyroid). But what is cause and what is effect? Is both deiodinative and non-deiodinative pathways the difference in BMR due to the difference in T4 with consequent changes in plasma iodothyronine secretion rate or is the T4 secretion rate due to the levels. Serum levels of free T4, total 3,3h-T2 and difference in BMR? Because the hypothalamic\ 3h,5h-T2 are unaffected, whilst serum free T3 and anterior pituitary\thyroid axis is organized to main- total rT3 concentrations are reduced and serum tain a constant plasma T4 concentration it is possible total 3,5-T2 concentrations are increased with a low- that the thyroidal T4 secretion rate varies with the energy diet. There is a decrease in the overall plasma 0.74 power of body mass because the metabolic disposal rate, but no change in the percentage of T3 disposal of T4 is related to the overall metabolic rate metabolized by inner and outer ring monodeiodin- of the mammal (which is proportional to the 0.73 ation. For rT3 there is a large decrease in outer ring power of body mass). Insight can be gained from deiodination, a small increase in inner ring deiodin- Fig. 2. If thyroidal T4 secretion is the ‘cause’ of the ation and a large increase in non-deiodinative body-mass-related change in BMR then we might metabolism with a low-energy diet (Burger et al., expect that plasma concentrations of free thyroid 1987). That dietary-induced alterations in deiodin- hormones would be greater in smaller mammals ases can lead to complex alterations in the thyroid because of their greater mass-specific metabolic rate, axis is illustrated by the finding that fasting leads to since presumably the regulatory information ‘seen’ an increase in D2 deiodinase activity in the rat by the cells is the free hormone concentration. There hypothalamus (Diano et al., 1998). In young pigs, is no such allometric trend in free hormone levels in low energy intake results in lowered serum T4 and mammals. Thus, a reasonable conclusion is that the T3 concentrations but low ambient temperature has 542 A. J. Hulbert no effect on serum hormone levels. However, whilst are reported to be secreted from the thyroid gland low temperature results in a decreased fractional (Escobar-Morreale et al., 1996). turnover of T4 and T3, energy intake has no effect In a similar vein, other iodothyronines have also (Macari et al., 1983). been shown to have significant effects. For example, Variations in the metabolism of thyroid hormones rT3 often regarded as solely an ‘inactivated’ thyroid have also been observed during non-thyroidal illness hormone, is sometimes more potent than T3 (e.g. in humans with the degree of thyroid hormone Leonard et al., 1990; Farwell et al., 1995). It has also disturbance being correlated with disease severity. been shown that 3,5-T2 is an ‘active’ thyroid Commonly during such illnesses, serum T3 levels are hormone with respect to the calorigenic action of the lowered and sometimes T4 levels and TSH levels are thyroid hormones (Horst, Rokos & Seitz, 1989). reduced as well (McIver & Gorman, 1997) and However, because of the common practice of serum 3,5-T2 is elevated (Pinna et al., 1997). In some using only T3 when studying thyroid hormone illnesses, there is also interference with plasma- effects both in vitro and in vivo much of the remainder protein binding as well as cellular uptake of thyroid of this review will consider the effects of thyroid hormones (Lim, Stockigt & Hennemann, 1995). hormones that have predominantly been measured as T3 effects. It is the opinion of this reviewer that there are four iodothyronines that have significant III. EFFECTS OF THYROID HORMONES but not identical biological activities and these are T4, T3, rT3 and 3,5-T2 (see Fig. 1). (1) When studying the effects of thyroid The other problem that plagues consideration of hormones literature on thyroid hormone action is that of The first question is which thyroid hormone to use. ‘dose’. As cited above, the T3 production rate for It is a common assumption in the literature that T3 the euthyroid rat is approximately 0.25 nmol\100 g\ is the ‘active’ thyroid hormone and that T4 is only day (approximately 0.16 µgT3\100 g\day). It is a prohormone. It is not uncommon to see T4 common practice for considerably larger doses to be referred to as the ‘inactive’ hormone. Whilst T3 is used (I also plead guilty in this respect). For more active on a ‘per mole’ basis in many situations, example, 200 µgT3\100 g is sometimes used and this does not mean T4 is ‘inactive’. Indeed, T4 is the justified as being ‘receptor saturating’. This is more predominant secretion by the thyroid gland and a than 1000 times greater than the replacement dose of case can be made that T4 is the more physiologically T3. Six hours after the intraperitoneal injection of relevant hormone for in vivo studies, as it allows the half this amount in the rat, the plasma T3 body’s cells to perform any appropriate deiodina- concentration is 3 µ (Hartong et al., 1987). It is not tions. T4 is an active hormone in its own right. T4 unreasonable to expect that for a significant period has been reported to affect directly the expression of of time after the injection of such doses, the plasma thyroid-hormone-sensitive genes (Bogazzi et al., T3 concentration is three orders of magnitude above 1997). normal euthyroid levels. Such high doses (sometimes T4 also can have effects separate from T3. For in the µ range) are also used in in vitro studies. Since some effects, T4 has been shown to be two orders of free hormone concentrations for a wide variety of magnitude more potent than T3 (Leonard, Siegrist- vertebrate species are in the p range (Fig. 2), I have Kaiser & Zuckerman, 1990; Farwell, Tranter & had to assume the results of many studies to be Leonard, 1995). Similarly, studies on birds show relevant but they may instead turn out to be that T3 cannot substitute for T4 regarding re- physiologically unimportant when reexamined at productive timing (Pant & Chandola-Saklani, 1995; more realistic concentrations. Reinert & Wilson, 1997). The use of T3 rather than It is also common practice to give daily injections T4 may be misguided and based upon an assumption of thyroid hormones to induce a hyperthyroid (T4 is only a prohormone) that is not valid in many condition. Because a single daily bolus will result in situations. a transient peak in plasma concentration it is The importance of taking into account both assumed a large dose is necessary to ensure a high cellular uptake mechanisms and cellular deiodinase average concentration between the daily injections. systems is illustrated by the recent finding that, in In other words, it is assumed that the same amount the thyroidectomized rat, normal euthyroid intra- given as a single daily injection will be less effective cellular contents of T4 and T3 were only achieved than if it was given as smaller multiple injections or when both T4 and T3 were infused at the rates they by constant infusion. At least in one case, the Thyroid hormones and their effects 543 opposite is true. The inhibition of TSH secretion by In a similar vein, many cell culture media whilst the rat anterior pituitary was substantially more being adequate for maintaining a living cell, are not sensitive to the same daily dose of T3 when it was necessarily normal and can thus result in non- given as twice daily injections than when it was given physiological conditions. This is especially the case by constant infusion (Connors & Hedge, 1980). The with respect to fatty acid composition of the media pulsatile nature of daily injections of T4 or T3 may and its consequent influence on the membrane lipid give misleading information when correlating hor- composition of some cell cultures. mone effects with plasma hormone concentrations as There is a large reservoir of thyroid hormones in standard practice is often to measure plasma the body, both in the thyroid gland itself and also hormone concentration 24 h after the last injection, bound to the plasma proteins. It appears to be very that is at the trough of the induced large daily cycle difficult to completely deplete the body of thyroid in hormone concentration. This point is illustrated hormones. Although the plasma T4 and T3 levels by one of the few studies to determine the average decrease relatively rapidly, measurable amounts of daily plasma hormone level following thyroid hor- the hormones are still present in the tissues of the rat mone injection (John-Alder, 1983). In this study, for some time following thyroidectomy (Obregon et lizards were given daily T4 injections (20 µg\100 g al., 1981). In addition, there is evidence that there body mass). Whilst control lizards had a total may be extrathyroidal synthesis of thyroid hormones plasma T4 concentration of 3.2 p, the injected in the rat following thyroidectomy (Evans et al., animals had a 24 h post-injection concentration of 1966). 12.9 p, while the 24 h average T4 level was 122 p. There are some hormonal effects that can be For this reason, it is more desirable to use constant observed in vitro but not in vivo. There are two infusion of hormone if at all possible (for example by possible reasons for this. The first is that it may be osmotic minipumps, see Escobar del Rey et al., possible to measure particular parameters in vitro 1989), unless one is examining the time course of a that are impossible to measure in vivo. The second is particular effect following a pulse of hormone. that the observed effect may be due to some Another problem of interpretation of results condition that is found in vitro but not in vivo. Such following thyroid hormone injections is consideration effects, whilst they may give important insight, of how the treatment alters plasma thyroid hormone might have little relevance to understanding what is levels. Generally, total concentrations of the hor- happening in the body. However, we should also mone are measured and it is often not known how remind ourselves that much of our knowledge of the physiologically more relevant free hormone levels biochemistry and cell function comes from in vitro are changed by the treatment. studies. Because of these problems, the hypothyroid- Finally, it is worthwhile to remember that both euthyroid comparison is probably more reliable in hypothyroidism and hyperthyroidism can be re- examining the physiological effects of the thyroid garded as pathological conditions. Our treatments hormones. However, there are also considerations are artifices we create in order to understand the with this comparison. One is that some of the normal euthyroid state (or of course to understand observed effects may be a secondary response rather the pathological state). In the normal (i.e. euthy- than a primary effect (of course, this problem is also roid) adult vertebrate, thyroid hormones, and present in the euthyroid-hyperthyroid comparison). especially their free concentrations, appear generally One of the best cases of this problem is the secondary to remain relatively constant. In the normal state, effects of low growth hormone secretion in hypothy- the thyroid hormone system has both a large roidism and the fact that observed effects of reservoir of hormone stored in the gland (it is hypothyroidism may be due to growth hormone relatively unusual in this respect) and a large deficiency rather than thyroid hormone deficiency. reservoir of easily accessible protein-bound hormone Classical hormonal replacement following induced in the blood. Thus, the concept that thyroid hypothyroidism can resolve some of these problems hormones regulate a process that normally varies, but once again dose is important. whilst thyroid hormone levels themselves are rela- Cell culture studies can avoid many of these tively constant, suggests that the hormones are not pitfalls but the results of such studies should be the most important regulator in such situations. verified in vivo, especially since many cell lines are My purpose in the remainder of this review is to originally derived from tumours and thus may not attempt to understand the myriad of reported behave normally (e.g. Sorimachi & Robbins, 1977). thyroid hormone effects within a few underlying 544 A. J. Hulbert

# basic concepts. I have grouped the various effects 1n T3 induces an influx of Ca + into rat into categories for the sake of convenience and will thymocytes that is evident within 15 s of hormone first discuss reported effects in adult vertebrates administration (Segal & Ingbar 1984). Similarly, before considering the role of the thyroid in application of T3 to GH3 cells in culture (2–5 nl of vertebrate development. 0.1 n T3) resulted in an increase in membrane resistance and a hyperpolarization of the membrane potential. These effects were evident within 1 min of (2) Hypotheses regarding thyroid hormone application, and within a few minutes spontaneously action firing cells became silent (du Pont & Israel, 1987). Because of the dogmatic nature of the oft-repeated These are a few examples of effects which cannot be statement that thyroid hormones act via nuclear mediated via nuclear receptors. receptors, it is difficult to convince many people that It is the thesis of the present review that many they may not do so. Generally, only some very thyroid hormone effects can be explained by an specific effects have been associated with a non- initial action on cell membranes, not just the plasma genomic mode of action. Others have previously membrane but also the other membranes within the questioned the nuclear receptor pathway as the sole cell. I will examine many thyroid hormone effects to mode of action of thyroid hormones (e.g. Segal, see if they can be explained by such an initial site of 1989b, 1990b). A recent review of non-genomic action, rather than the commonly proposed nuclear- actions of thyroid hormones is that of Davis & Davis receptor-mediated mode. First, I should state the (1996). Such questioning of a single genomic mode hypothesis explicitly and say something about of action for a hormone is not unique for thyroid membranes. Several years ago I proposed a similar hormones, as non-genomic actions recently have but more rudimentary thesis for such a site of thyroid been proposed for reproductive hormones (Revelli, hormone action (Hulbert, 1978). At that time, I was Massobrio & Tesarik, 1998), and other steroid unaware of an earlier and different proposal for a hormones including aldosterone (Wehling, 1994) as membrane site of action emphasizing the importance well as for some glucocorticoid effects (Buttgereit, of the iodine atoms of the thyroid hormones Wehling & Burmester, 1998). (Gruenstein & Wynn, 1970). This earlier proposal As stated above, there exists evidence that not all was based around many effects of thyroid hormones thyroid hormone effects are initiated via binding of that are present at supraphysiological levels of thyroid hormones to nuclear receptors. Some extra- hormone and thus may not be relevant to the normal nuclear effects that have been reported have been situation. Dratman (1974) proposed that some observed at supra-physiological concentrations of effects of T4 stemmed from its similarity to (and hormone (as indeed are many nuclear-mediated origin from) tyrosine residues and this proposal has effects) but many have been observed with physio- been recently reviewed (Dratman & Gordon, 1996). logical levels of hormone. Some of the most con- Several other authors have also argued for a vincing evidence is that which demonstrates in vitro perspective of thyroid hormone action that is broader thyroid hormone effects on cells that have no than the nuclear-receptor hypothesis (e.g. Sterling, nucleus, that is, on mammalian red blood cells. In 1979; Muller & Seitz, 1984). Hoch (1988) has # human erythrocytes, membrane Ca +-ATPase ac- presented an extensive review of thyroid hormones tivity is stimulated in vitro by physiological concen- and lipids which has also argued that many thyroid trations of thyroid hormone (Davis, Davis & Blas, hormone effects can be explained as a consequence 1983) and this has in vivo relevance in that of an effect on membrane fatty acids, although he erythrocytes from hyperthyroid and hypothyroid suggests such membrane changes are due to a humans have, respectively, increased and decreased nuclear-receptor-mediated action. enzyme activities (Dube et al., 1986). Lawrence, The proposal is that the thyroid hormones act in Schoenl & Davis (1989) have shown that both T4 more than one way. One of the ways in which they and T3 (0.1 n) stimulated phospholipid-dependent operate is that they associate with the hydrophobic protein kinase activity in rabbit erythrocytes in vitro. interior of membrane systems and change the There is also evidence that in nucleated cells some physical behaviour of the lipid component of the thyroid hormone effects cannot be initiated at the membrane. As a consequence of this, the phos- level of the nucleus. A very convincing aspect of this pholipid and fatty acyl composition of the membrane evidence is the time course of such effects at is altered by normal cellular mechanisms and these physiological hormone concentrations. For example, changes further alter the physical behaviour of the Thyroid hormones and their effects 545 lipid bilayer and consequently that of various twice the average found for anterior pituitary in vivo membrane proteins. For example, due to both the (Oppenheimer et al., 1974) and presumably reflected initial direct effects on the bilayer and also the effects the specific type of anterior pituitary cell from which of the subsequent fatty acyl changes, there are the cell line was derived. In the rat, receptor density changes in membrane permeability and the en- varied between tissues, ranging from approximately zymatic activity of membrane-associated systems. 20 per nucleus in the testes to approximately 8000 These changes may be increases or decreases, and per average anterior pituitary cell. In GH1 cell will, in some cases, alter both the distribution and cultures, the number of receptors was inversely rate of transfer of information between various related to T3 levels (Samuels, Stanley & Shapiro, membrane-limited cellular compartments. Several 1977) whereas in the intact animal similar changes thyroid hormone effects can be explained by such an were not observed (Oppenheimer, Schwartz & initial site of action, especially (but not exclusively) Surks, 1975; Spindler et al., 1975). the effects that are related to the stimulation of In 1986, partly by accident, the genes for these metabolism. This proposal is unusual in that nuclear proteins were discovered separately by two although these effects may be mediated via the groups (Sap et al., 1986; Weinberger et al., 1986). binding to some form of ‘specific’ receptor molecule, The similarity of the viral erb-A oncogene product they are more likely to come about via a less- to the glucocorticoid nuclear receptor suggested that conventional non-specific mode of action, i.e. not the cellular equivalent of this viral erb-A gene may requiring a ‘specific’ receptor in the normal sense of code for a receptor that bound one of the steroid the word. hormones. However, this proved unsuccessful and it Before analyzing the plethora of thyroid hormone was shown instead to bind thyroid hormones with a effects, I will briefly review both the nuclear receptor similar affinity to the previously described nuclear mode of action and following that, some aspects of binding sites, showed the same analog specificities membrane lipids and their function, as well as what and had a similar molecular weight. Sap et al. (1986) we know of thyroid hormone effects on membrane had isolated the gene from a library of DNA cloned lipid behaviour and composition. from the chicken embryo whilst the other group (Weinberger et al., 1986) isolated it from a human placental cDNA library. Since then it has become (a) Nuclear receptors and thyroid response elements apparent that there is a superfamily of related In 1972, specific nuclear binding sites, with limited transcription control factors that all have both a capacity and high affinity for T3, were found in rat DNA binding region and another region that binds liver and kidney (Oppenheimer et al., 1972). Fol- either a steroid hormone (e.g. glucocorticoid, min- lowing this, similar nuclear binding sites were eralocorticoid, progesterone, estrogen), or a vitamin described for a variety of other tissues as well as in (e.g. vitamin D3 and retinoic acid) or the thyroid various cell cultures. Although they were not isolated hormones. Soon after the isolation of the original and purified, the physico-chemical properties and genes, other cDNA sequences coding for thyroid binding characteristics of these nuclear receptors nuclear receptors were also isolated from rat brain were defined (for early reviews see DeGroot et al., (Thompson et al., 1987), human testis (Benbrook & 1978; Oppenheimer & Samuels, 1983; Oppen- Pfahl, 1987), human kidney (Nakai et al., 1988) and heimer et al., 1987; Samuels et al., 1988). two from both GH$ cells (Lazar & Chin, 1988) and The thyroid nuclear receptor was similar in all rat liver (Murray et al., 1988). The thyroid hormone tissues and species examined (Oppenheimer et al., nuclear receptors are thus members of this super- 1987). It was a non-histone protein and was most family of receptors involved in the control of studied in rat liver and in pituitary tumor (GH and transcription (for an early review see Evans, 1988). GC) cell lines. In rat liver nuclei, it was 50.5 kDa in There are two main types of thyroid hormone mass, with approximately 6000 per nucleus and at receptors, an α and a β form. Both forms are endogenous T3 levels it was estimated that approx- sometimes expressed in the same tissue, and both imately 50% of nuclear receptors were occupied forms from rat liver bound thyroid hormones with (Oppenheimer et al., 1974). In GH1 and GC cells, it the same affinity and spectrum of analogue binding was 54 kDa in mass, with approximately 15000 per (Murray et al., 1988). Since the mid 1980s there has nucleus having an average half-life of approximately been intense research interest in this area, much 4.5 h (Samuels et al., 1988). The receptor densities in knowledge has been gained yet the precise mech- these pituitary tumor cell lines were approximately anism of how the system operates, although slowly 546 A. J. Hulbert

Zinc fingers

P box

TRβ-2 514

TRβ-1 100 100 100 461

TRα-1 86 72 82 100 410

86 72 82 492 c-erbAα-2 86 72 82 453

DNA- Hinge Hormone- binding region binding domain domain Fig. 8. Domain organization of thyroid nuclear receptors. The three thyroid-hormone-binding receptors (TRα-1, TRβ-1, TRβ-2) and two isoforms of the non-hormone-binding receptor (c-erbAα-2) are shown. The numbers inside the boxes refer to the amino acid sequence % homology of each TR isoform relative to TRβ-2, and the numbers at the end of each box refer to the number of amino acids in each TR isoform. The letters identify individual amino acids. Adapted from figures in Yen & Chin (1994) and Chatterjee et al. (1997). being unravelled, is still elusive. There are several binds to appropriate parts of the genome. All have very good reviews to which the reader is referred for the same basic structure as other members of the more detailed information (Chin, 1991; Martinez & receptor superfamily which as a group are described Wahli, 1991; Lazar, 1993; Yen & Chin, 1994; as ligand-dependent transcription factors. This basic Ribeiro et al., 1995; Oppenheimer, Schwartz & structure is illustrated in Fig. 8. The amino terminal Strait, 1996; Chatterjee, Clifton-Bligh & Matthews, region of the protein is followed by the DNA-binding 1997; Zhang & Lazar, 2000). I have generally domain, then a hinge region, followed by the ligand- cited these reviews for information rather than the (i.e. hormone\vitamin) binding domain. The β voluminous primary literature. receptors differ only in their amino terminal region. In humans, the α gene resides on chromosome 17 The α gene products have a shorter amino terminal and the β gene on chromosome 3. Alternate region than the β receptors and have a high degree processing of the initial transcript from each gene of amino acid homology in the DNA-binding region, yields two isoforms of each receptor. Both isoforms of hinge region and the ligand-binding region (see Fig. the β thyroid receptor (TRβ-1 and TRβ-2) but only 8). The carboxyl terminal region is highly conserved one of the α receptor isoforms (TRα-1) bind thyroid in all the thyroid-hormone-binding forms and is hormones. The other non-hormone-binding α iso- essential for hormone binding. The ligand-binding form (c-erbAα-2 although sometimes also called region of the receptor is composed of many α helices, TRα-2) is thus not a true thyroid hormone receptor and X-ray diffraction studies of crystals of the region but may be involved in thyroid hormone effects as it in TRα-1, both with and without T3 (and ana- Thyroid hormones and their effects 547 logues) have revealed changes in its three-dimen- consistent with the loss of receptors that has been sional structure with hormone binding (McGrath et observed when isolated liver cells are cultured al., 1994; Wagner et al., 1995). The hormone is without protection of sulfhydryl groups (Yamamoto located in a deep pocket in this protein with the et al., 1992). amino group of the hormone facing the outside. Using antibodies specific to the particular TR Analysis of the 3-D structure suggests that the isoforms (which thus interfere with their T3 bind- receptor adopts different structures to fit the different ing), Schwartz, Lazar & Oppenheimer (1994) were ligands (i.e. T3 analogues). The conformational able to show that all three receptor types are change with hormone binding is probably connected widespread in rat tissues. In liver, the receptor " to transcriptional activation. concentrations (in pmol mg− DNA) were 0.71, 0.17, The DNA-binding domain consists of an area and 0.13 for TRβ-1, TRβ-2 and TRα-1, respectively. relatively conserved between different receptors and The respective values for kidney TRs were 0.10, 0.04 " it is this region that is responsible for recognizing and 0.10 pmol mg− DNA; for heart they were 0.24, " specific parts of the genome. In thyroid receptors, it 0.10, and 0.26 pmol mg− DNA, and for brain cortex " contains 68 residues with basic amino acids or- they were 0.19, 0.07 and 0.40 pmol mg− DNA ganized into two ‘finger-like’ structures that each (Schwartz et al., 1994). Not all cells within a tissue, contain four cysteine residues tetrahedrally co- however, have TRs. For example, within the brain, ordinated with a zinc atom. It is believed these ‘ zinc immunofluorescence studies have shown that oligo- fingers’ interact with a half turn of DNA (see Evans, dendrocytes possess all three receptor isoforms in 1988). Also, an amino acid sequence present at the their nuclei but that astrocytes have none (Carlson et base of the first zinc finger, known as the P box, is al., 1994). However, when astrocytes are examined important for DNA-binding specificity and serves to in cell culture, they do possess TRs (Carlson et al., divide the receptor superfamily into two groups. The 1996). P box is identical in the thyroid hormone receptors The mRNA for the different TR isoforms varies (TRs), retinoic acid receptors (RARs), retinoid X between tissues but not in relation to their respective receptors (RXRs), vitamin D receptor (VDR) and TR content. Between tissues the receptor\mRNA the peroxisomal-proliferator-activated receptors ratio can vary by as much as 15-fold (Oppenheimer, (PPARs). These receptors preferentially bind to the Schwartz & Strait, 1996). nucleotide sequence: AGGTCA (Lazar, 1993). In 1986, the year that the TR genes were Between the DNA-binding and hormone-binding identified, it was found that thyroid hormone nuclear domains, in the hinge region, there is a conserved receptors bound to a particular part of the promoter motif of basic residues that acts as a nuclear regions of the human growth hormone gene and the localization signal for nuclear pore recognition. human placental lactogen gene (Barlow et al., 1986). Within the nuclei of isolated rat hepatocytes, TRs For TRs to activate transcription they must bind to are separable into three compartments that are sections of DNA known as ‘thyroid response ele- distinguished by the conditions needed for their ments’ (TREs). The DNA-binding region of the isolation and thus possibly related to differences in TRs (and some other receptors) has a P box which their DNA binding. Nucleoplasmic TRs (45%) can recognizes the hexanucleotide AGGTCA (Umesono be isolated with isotonic saline, some other TRs & Evans, 1989). (30%) are extracted with high salt (0.4  KCl) In the early 1990s it was recognised that TRs concentrations, whilst the remaining TRs (25%) are could bind to DNA as either monomers or dimers, salt resistant, requiring high salt and 5 m dithio- and that dimers could consist of either the same or threitol for extraction. Pulse experiments with different TR isoforms (i.e. either homodimers or hepatocytes exposed to extracellular labelled T3 heterodimers). It was also discovered that TRs could showed that all were rapidly labelled with T3 form heterodimers with an accessory protein, that (within 1 min) with no preference for any particular augmented TR binding to the genome and was later compartment, that the average occupation time for identified as the retinoid X receptor (RXR). It is T3 on a receptor was approximately 3 mins, and now appreciated that the RXR–TR dimer shows the that T3 did not influence any shift in receptors strongest binding to the genome. Dimerization is between these compartments (Pullen et al., 1994). mediated by at least two protein-protein interactions The sum of these receptors represented 0.2 pmol TR between the adjacent monomers. One interaction " mg− DNA, which is below the normal 0.5–1.0 pmol region is in the DNA-binding region whilst another " mg− DNA reported for liver homogenates but is is a heptad towards the carboxyl end of the hormone- 548 A. J. Hulbert

TREs Nucleotide sequence co-activators replace the co-repressors. In both cases, transcription of the target gene is commenced by allowing the transcription initiation complex to Palindrome (TREpal) commence transcription (Chatterjee et al., 1997). The response elements cited above are all positive TREs, in that they are involved in hormone- Direct repeat (DR) stimulated transcription activation. Thyroid nuclear receptors can also exhibit ligand-dependant tran- scription inhibition but how this occurs is poorly understood. Such inhibition may be mediated by Inverted palindrome (IP) TRs binding as monomers to negative TREs which appear to be half-sites not organized in any particular way, although TR-RXR heterodimers may also be involved (Takeda et al., 1997). consensus TRE Some mutant TRs (either natural or laboratory half-site: created) have been shown to have a dominant negative activity, i.e. the ability to inhibit T3- Fig. 9. The nucleotide sequences for thyroid response dependent TR-mediated transcriptional activation. elements (TREs) involved in the binding of thyroid nuclear receptors to the genome. Reprinted with per- In the syndrome of thyroid hormone resistance mission from Yen & Chin (1994). mutant TRβ have this ability, as does a mutant TRα-1 isolated from patients with hepatocellular binding domain of the receptor. In addition, in carcinoma (K. H. Lin et al., 1997), and c-erbAα-2 dimers both receptors bind to the DNA. This has led (Lazar, 1993). Several studies have also shown that to the recognition that AGGTCA is thus a half-site phosphorylation of thyroid nuclear receptors can for binding of dimers and that TREs for dimers exists increase binding, dimerization and transcriptional as either a direct repeat separated by four bases activity (e.g Lin, Ashizawa & Cheng, 1992; Bhat, (called DRj4), an inverted palindrome (TREpal), Ashizawa & Cheng, 1994; Sugawara et al., 1994). or an inverted palindrome separated by six bases The majority of experiments over the last several (IP). These TREs are shown in Fig. 9. When bound years that have led to the uncovering of the as a RXR-TR heterodimer, the TR is bound to the mechanism of thyroid nuclear receptor operation 3h recognition half-site and it appears that the RXR outlined above, have been carried out using artif- is not ligand sensitive, indeed the TR appears to icially constructed systems. Often the TRs have been block the binding of ligands to RXR upon dimeriz- expressed in bacteria, the TRE has been taken from ation (Glass, 1996). another source and been transiently transfected, Thyroid receptors are known to bind to TREs together with a distinct reporter gene, into a cell both with and without bound thyroid hormone. culture system or combined in an in vitro system. Dimers generally repress transcription of target genes Sometimes the cell culture used doesn’t normally in the absence of bound hormone and activate express TRs and a particular TR gene has been transcription when hormone is bound. The precise transfected into the cells. Indeed, this is sometimes mechanism of this transcriptional activation is not regarded as necessary in that it allows more precise known but it appears that in the N terminal region control of the experimental system. Whilst these there is a weak transcription activation function sophisticated techniques have been powerful in (AF-1) whilst in the C terminus of the ligand- elucidating the complex mechanisms whereby the binding domain there is a more powerful ligand- thyroid nuclear receptor system may work, they do dependent activation function (AF-2). Various ex- not necessarily give insight into the physiological periments suggest that in the absence of T3, TR-TR importance of the nuclear mode of action in homodimers or RXR-TR heterodimers are bound to explaining normal thyroid hormone effects in the TREs and that basal transcription is inhibited by vertebrate body. Indeed, the finding that expression intervening co-repressors between the receptors and of nuclear receptors in cell culture systems sometimes the transcription initiation complex. Upon T3 differs from the in vivo situation means that care binding, TR-TR homodimers dissociate and co- should be taken with interpretation and extra- repressors are released whilst in the case of hetero- polation of results obtained from such systems. dimers both receptors stay attached to the DNA but The finding of increased levels of mRNA for a Thyroid hormones and their effects 549 particular protein after changes in thyroid status is fatty acid will be called 18:0, oleic acid which is an often presented as evidence that the thyroid hor- 18 carbon monounsaturate will be referred to as mones are acting via thyroid nuclear receptors 18:1, whilst docosahexanoic acid a 22 carbon poly- stimulating transcription. However, this may not unsaturate with six double bonds will be called 22:6. necessarily be so, as the level of mRNA is dependent In higher animals, newly synthesised fatty acids are on its rate of production (i.e. transcription) and its saturated (predominantly 16:0 and 18:0) and some rate of disappearance. Even the finding that tran- are converted to unsaturated fatty acids by de- scription is stimulated by changes in thyroid status is saturases, which are membrane-bound multi-en- not necessarily evidence that the thyroid hormones zyme complexes that consume oxygen, insert double are acting via nuclear receptors. In some cases, an bonds at specific places in the fatty acyl chain and equally valid interpretation may be that these are named accordingly. The cells of vertebrates have hormones are acting pre-transcriptionally, by in- ∆5, ∆6, and ∆9 desaturases but lack ∆12 or ∆15 fluencing the rate of information flow to the nucleus desaturases. The products of the ∆12 desaturase are and it is this changed information flow that is called the n-6 (or omega-6) polyunsaturates, whilst responsible for the change in the rate of transcription. the products of the ∆12 desaturase that have been An evidential requirement for a nuclear mode of further processed by the ∆15 desaturase are known action for a thyroid hormone effect is the dem- as the n-3 (or omega-3) polyunsaturates. The onstration of a TRE for the particular gene involved. average membrane bilayer consists of two 18 carbon The finding of a nuclear receptor for any thick monolayers. Membranes without poly- particular molecule does not exclude another mode unsaturates will have no C%C units in the middle of action. For example, retinoic acid as well as acting half of membrane bilayer whilst membranes with through a nuclear receptor can also have effects on only n-6 polyunsaturates will have no C%C units in human erythrocytes, cells which lack nuclei (Smith, the middle third of the bilayer. It is only the n-3 acyl Davis & Davis, 1989) and is also known to inhibit chains that can contribute C%C units to the middle binding of T4 to human TTR (Smith et al., 1994). portion of membranes. The unsaturated fatty acids that vertebrates generally synthesize de novo are the n-9 or n-7 monounsaturates which will contribute (b) Thyroid hormones and the membrane bilayer C%C units to bilayers in two regions both approx- Just as DNA has been described as an ‘eternal imately one-quarter of the way into the bilayer from molecule’, membranes can be thought of as an each side. ‘eternal structure’ and are assemblies of proteins In 1972, Singer & Nicholson proposed the ‘fluid and lipid molecules held together by non-covalent mosaic’ model of membrane structure and although bonds. The lipids in membranes include phos- it has since been modified and extended, it is still the pholipids and glycolipids as well as cholesterol and basis of modern thinking concerning cellular mem- sterols, and at least in vertebrates, it is my assertion branes. Over the last three decades, it has become that cellular membranes normally also contain apparent that the hydrophobic compartments of the thyroid hormones. All membrane lipids are amphi- cell (i.e. its membranes) are a series of complex pathic molecules. microenvironments that influence many aspects of Fatty acyl chains form the hydrophobic section of cellular function. It has also become apparent in the the phospholipid molecule and the core of the last quarter of a century that these aspects of membrane bilayer. Saturated acyl chains are flexible membrane structure are regulated, in that there are because of the possible rotation around the C$C homeostatic systems within the cell to maintain an bonds that form the backbone of the acyl chain The appropriate ‘fluidity’ (or viscosity) of the membrane double bonds of unsaturated acyl chains, on the bilayer when conditions act to change membrane other hand, do not allow rotation around the C%C ‘fluidity’. One of the most obvious of these en- bond. Monounsaturates have a single ClC unit, vironmental factors is temperature but there are whilst polyunsaturates have from two to six such several others. Following its discovery in bacteria, units. The majority of fatty acids in the cellular this response was termed ‘homeoviscous adaptation’ membranes of vertebrates are between 16 and 22 (Sinensky, 1974). It is possibly one of the earliest carbons long and the average chain length is homeostatic responses evolved by cells and has been generally around 18 carbons long. In this review, I shown to also be present in both plants (Thompson will refer to specific acyl chains using a numbering & Nozawa, 1984) and animals (Cossins, Bowler & system. Thus, stearic acid, a saturated 18 carbon Prosser, 1981). Although it involves a number of 550 A. J. Hulbert responses, this self regulation of membrane ’fluidity’ these four initial phospholipid species by deacyl- is mainly brought about by manipulation of mem- ation\reacylation processes (Schmid, Deli & brane fatty acyl composition, particularly the degree Schmid, 1995). of unsaturation of membrane fatty acyl chains. It I have spent some time describing membrane thus involves the lipid desaturases. In a wide variety lipids and have concentrated on the role of acyl of eukaryotic cells, it has been shown that the chains because it is important in the consideration of activity of the lipid desaturases is inversely affected my argument concerning the effects of thyroid by the fluidity of the membrane with which they are hormones. Our understanding of the roles of mem- associated and it has been suggested that this brane acyl composition in influencing many different relationship is one of the mechanisms of the membrane-associated processes is increasing and has regulation of membrane fluidity (Brenner, 1981; been reviewed elsewhere by others (Brenner, 1984; Kates, Pugh & Ferrante, 1984; Thompson & Stubbs & Smith, 1984; Spector & Yorek, 1985; Nozawa, 1984). Accordingly, a decrease in mem- Bernardier, 1988; Murphy, 1990). Because of the brane fluidity (or an increase in membrane rigidity), need to concentrate on the role of the fatty acyl whether it be caused by a decrease in temperature, residues in membrane function in the present review, or the increased membrane incorporation of satu- discussion of membrane bilayer function has by rated fatty acids or of sterols (e.g. Leikin & Brenner, necessity been simplified, omitting for example, 1989) or by other means (such as the experimental consideration of such things as the asymmetry of incorporation of unnatural ‘rigidizers’ etc.) will exofacial and cytofacial leaflets of membranes, as result in stimulation of the desaturase complexes and well as important considerations of phospholipid a consequent increase in the degree of unsaturation head groups. The potential importance of membrane of membrane fatty acyl chains. As in any homeostatic acyl composition and especially the degree of system the converse is also true. This is presumably polyunsaturation of membrane bilayers on the why although changes in the fatty acyl composition cellular metabolic activity of vertebrates has been of the diet may influence the relative occurrence of recently discussed by Hulbert & Else (1999, 2000). individual acyl chains in membrane lipids it gen- Because of the amphipathic nature of the thyroid erally does not significantly affect the overall degree hormone molecules that have a non-ionized phenolic of unsaturation of the membranes (e.g. Withers & kOH, they will be found in the membrane bilayer Hulbert, 1987). Diet can, however, have very in high concentrations and oriented with the significant effects by influencing the balance between phenolic end of the molecule towards the centre of the different types of polyunsaturates in membranes the bilayer. Would their presence affect the physical (e.g. Pan, Hulbert & Storlien, 1994). behaviour of membrane lipids? Dickson et al. (1987) This process is not a perfect homeostatic response have shown that both T4 and T3 quench the light and degrees of homeostasis have been observed. For emitted by fluorescent groups inserted in artificial example the sarcoplasmic reticulum exhibits a lack phosphatidylcholine bilayers. The pH charac- of ‘homeoviscous adaptation’ (Cossins, Christiansen teristics of this quenching suggest that T4 and T3 & Prosser, 1978). Although such responses consist molecules with the non-ionized phenolic kOH are predominantly of changes in acyl chain composition responsible for this quenching. Farias et al. (1995) they also include other alterations such as changes in have also measured the effects of T3 and T4 on the ratio of different phospholipid classes. The liposome bilayers and have shown that both thyroid mechanisms involved include de novo synthesis of hormones ‘rigidify’ membranes in the liquid-crys- phospholipids and direct desaturation of membrane talline phase, that - and -isomers have the same acyl chains but predominantly the membrane effect but that T3 differs from T4 in that when remodelling involves a cycle of deacylation\ membrane bilayers are in a gel state (i.e. below their reacylation of membrane phospholipids which is phase transition temperature) T3 is also capable of accomplished by the combined action of two types of ‘fluidizing’ the liposomal membrane. In this respect, enzymes: phospholipases and lysophospholipid acyl- T3 is similar to cholesterol. The presence of transferases (Hazel & Williams, 1990). Indeed, cholesterol in liposome bilayers also influences the although there are hundreds of different molecular effects of both T4 and T3. In the absence of species of phospholipids in membranes, in rat cholesterol, T3 and T4 were incorporated into the hepatocytes only four molecular phospholipid species liposome at respectively 9 and 6 nmol of hormone are synthesised de novo and the diversity of membrane per 100 nmol of phospholipid. As cholesterol content phospholipid species is due to the remodelling of of liposomes increases a decreased amount of both Thyroid hormones and their effects 551

T4 and T3 is incorporated into the liposome Brasitus & Dudeja (1988) have shown that the membrane. The influence of both T3 and T4 on apical plasma membranes of rat colon are more fluid membrane fluidity and membrane permeability (less rigid) in hypothyroid rats than in euthyroid (both assessed by fluorescent probes) is also de- controls. T4 injections into rabbits resulted in muscle pendent on the cholesterol content of the liposomal plasma membranes that were more rigid than those membrane (Chehin et al., 1995). Although in this from control rabbits (Pilarski et al., 1991). study the T4 and T3 concentrations used for Both T4 and T3 injections into euthyroid rats are measurement of both leakage and fluorescence reported to have an opposite effect on brain polarization were in the µM range, and are thus mitochondria in that they resulted in an increase in supraphysiological, the incorporation measurements fluidity (a decrease in rigidity); T4 was slightly more were performed at physiological thyroid hormone potent at the concentrations used than was T3 concentrations. Incorporation of 50% mol chol- (Bangur, Howland & Katyare, 1995). When the esterol into liposome membranes decreased T3 fluidity of different membrane fractions from the incorporation by 76% and T4 incorporation by 99 brain of hypothyroid rats was compared with % compared to cholesterol-free liposomes (Chehin et euthyroid controls a significant increase in micro- al., 1995). viscosity was measured in the mitochondrial fraction An early and significant study is that by Schroeder but not in the microsomal, synaptosomal or myelin (1982) who showed that T3 in vitro resulted in fraction (Tacconi et al., 1991). rigidification of the rat liver plasma membrane In these in vivo studies cited on hyperthyroid or bilayer. These results, because they occurred im- hypothyroid mammals, there were also significant mediately cannot be due to other thyroid-induced thyroid-status-related changes in membrane lipid changes in membrane lipid composition. They composition (including phospholipid headgroup occurred at physiological concentrations and also composition, cholesterol content, and fatty acyl occurred in a natural membrane. They were composition). Many of these changes themselves measured using two separate membrane fluoro- influence membrane fluidity and thus it is not always phores and the results show the same response as possible to relate the reported in vivo changes in measured by others on artificial membranes (unfor- membrane fluidity\rigidity solely to direct effects of tunately often at higher T3 concentrations). It is the thyroid hormones. tempting to speculate that thyroid hormones ‘rigid- Can thyroid-hormone-induced changes in mem- ify’ the membrane bilayer because the large elec- brane fluidity\rigidity affect the behaviour of pro- tronegative iodine atoms in the thyroid hormones teins located in the lipid bilayer? The Hill coefficient induce increased van der Waals forces in the of a reaction can be used to measure small changes surrounding acyl chains and thereby diminish their in enzyme-membrane interactions. Physiological mobilities. Liver plasma membranes isolated from concentrations of T3 in vitro have been shown to hyperthyroid rats (T3 injections), euthyroid controls change significantly the Hill coefficient for the and hypothyroid rats had a similar fluidity in the allosteric inhibition of two different enzyme systems # absence of Ca + ions (Schroeder, 1982). in erythrocyte membranes from rats (De Mendoza et In an in vivo context, Beleznai et al. (1989) have al., 1977). This study provides evidence that T3 shown that both liver mitochondrial membranes and decreases membrane fluidity and also illustrates mitoplasts from hypothyroid rats are less rigid, three important points. Firstly, thyroid-hormone- whilst those from hyperthyroid rats are more rigid induced changes in membrane fluidity can affect than those from euthyroid rats. Their use of both different membrane-bound enzyme systems within DPH (diphenylhexa-1,3,5-triene) and TMA-DPH the same membrane differently. For example, T3 (1-[4-trimethyl-aminophenyl-phenyl]-6-phenyl- increased the Hill coefficient for erythrocyte Na+,K+- hexa-1,3,5-triene) which, respectively, associate ATPase activity whilst it decreased it for the with the inner hydrophobic core and the outer erythrocyte acetylcholinesterase. Secondly, such area of membranes showed that both parts of the changes in membrane fluidity are general effects in bilayer were similarly affected. Similarly, Parmar et that they can be demonstrated in non-vertebrates: al. (1995) have shown that liver mitochondria from T3 also decreased the Hill coefficient for the allosteric # thyroidectomized rats have less rigid membranes inhibition of Ca +-ATPase in bacterial membranes. than euthyroid controls and that replacement injec- Thirdly, the influence of thyroid hormones on such tions of T4, at physiological levels, resulted in a membrane systems depends on the original fatty acyl rigidification of these mitochondrial membranes. composition of the membrane. 552 A. J. Hulbert . (1986) et al . (1991) . (1991) et al et al . (1991) . (1991) . (1981) . (1981) . (1981) . (1981) . (1981) . (1981) et al et al et al et al et al et al et al et al Patton & Platner (1970) Chen & Hoch (1977) Hoch Hoch Hoch Hoch Ismailkhodzhaeva Raederstorff Raederstorff Paradies Hoch Faas & Carter (1982) Faas & Carter (1982) Faas & Carter (1982) Faas & Carter (1982) Tacconi Shaw & Hoch (1976) Steffen & Platner (1976) Steffen & Platner (1976) Shaw & Hoch (1977) Hoch (1982) diet Pehowich (1995) n-3 diet Pehowich (1995) n-3 dietn-6 dietn-3 Pehowich diet (1995) Pehowich (1995) Pehowich (1995) n-6 diet Pehowich (1995) k k k k k k k k 20:4 0 0 0 0 0 Withers & Hulbert (1987) 0 00 0 0000n-6 j jk k jk k jk k jkkk jkjkjk k k k jk k j jkjk k jkjk jk jk k 0 0 0 kjk k jjk k jjkkk Fatty acyl chain 0 00 kkjk k kjjj kjj jk kjjk k kkj k k 0000000000 0 Hoch 00000 00000 j j kkkjk PI000 \ class 16:0 18:0 18:1 18:2 20:4 22:6 18:2 Comments Reference The influence of hypothyroidism on membrane fatty acyl composition in mammals Species TissueRat Membrane Liver Mitochondria Lipid TL 0 Rat Liver Mitochondria TL 0 0 0 RatRat LiverRatRat LiverRat Liver MitochondriaRat Liver Liver Mitochondria Liver Mitochondria Mitochondria PL Mitochondria Mitochondria PE CL PS 0 PL TL 0 0 0 0 0 Rat Liver Mitochondria PC 0 0 0 Rat Liver Mitochondria PC Rat Liver Mitochondria PE 0 0 RatRatRat Liver Liver Liver Mitochondria Microsomes Microsomes TL PL 0 TL 0 0 0 0 0 0 0 0 Rat Liver Microsomes PC 0 0 0 RatRatRat Liver Liver Liver Microsomes Microsomes Microsomes PE PS PL 0 0 0 0 0 0 Rat Heart Mitochondria CL RatRatRat Heart Heart Heart Mitochondria Mitochondria Mitochondria PE PE CL 0 0 0 Rat Liver Nuclei PL 0 RatRatRat HeartRat HeartRat HeartRat Mitochondria Heart Microsomes Heart Mitochondria Heart Mitochondria Mitochondria TL Mitochondria TL PL PL PC 0 PC 0 0 0 0 0 Table 2. Thyroid hormones and their effects 553 . (1986) . (1986) . (1986) et al et al et al ) . (1991) b . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) . (1991) et al et al et al et al et al et al et al et al et al et al et al et al (1987) (1987) Simonides & van Hardeveld Simonides & van Hardeveld Tacconi Paradies & Ruggiero (1989) Tacconi Tacconi Tacconi Tacconi Tacconi Tacconi Brasitus & Dudeja (1988) Tacconi Syzmanska Valdemarsson & Gustafson (1988) Valdemarsson & Gustafson (1988) Van Doormaal Kirkeby (1972 n-3 diet Pehowich & Awumey (1995) n-6 dietn-6 dietn-3 Pehowich diet & Awumey (1995) Pehowich & Awumey (1995) Pehowich & Awumey (1995) j k k k k k 0 0 0 0 000 Tacconi Tacconi Tacconi 00 00 Van Doormaal Van Doormaal k k jj k 000 0 0 kj jk jk jk jk j j jk jk 0 0 kkj kj j jjk 0 000000 000000 000000 000000 000000 000000 000000 0000 0000 jk kjkkj k jkjjj jkj jjk kk , statistically significant decrease; 0, no significant effect. k , statistically significant increase; Rat Heart Mitochondria TL RatRat MuscleRat Muscle Sarcoplasmic reticulum PC Brain Sarcoplasmic reticulum PE Microsomes PL Rat Brain Mitochondria PC RatRatRat HeartRat Heart Heart Sarcolemma Heart Sarcolemma Sarcolemma Sarcolemma PC PE PC 0 PE 0 0 0 Rat Brain Mitochondria PE Rat Brain Synaptosomes PC Rat Brain Synaptosomes PE Rat Brain Myelin PC Rat Brain Myelin PE Rat Colon Plasma membrane TL 0 0 0 Rat Erythrocyte Membrane PL RatRatRat Erythrocyte MembraneRabbit Erythrocyte MembraneHuman Heart Plasma LeucocyteHuman ErythrocyteHuman Plasma membrane Plasma Platelet membrane Sarcolemma PC PL Plasma PE membrane PL 0 0 PL PL 0 0 0 0 0 0 0 0 0 PL 0 0 0 0 0 0 0 Human PlasmaHuman Plasma PL PL 0 0 Human Plasma PL In some experiments, diets with different fatty acyl composition were examined. TL, total lipid; PL,CL, phospholipid; cardiolipin. PC, phosphatidylcholine;j PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; 554 A. J. Hulbert

From the previous discussion of homeoviscous not statistically significant, but were in the same adaptation of membranes, we would predict that a direction as found for other tissues, namely hypo- thyroid-hormone-induced ‘rigidifying’ effect on thyroidism resulted in an increase in 18:2 and a membranes should stimulate a change in membrane decrease in 20:4 (Tacconi et al., 1991). fatty acyl composition to compensate for the change One of the exceptions to this finding is the results in membrane physical state. Over the last three for sarcoplasmic reticulum (Simonides & van Harde- decades, many studies have shown that membrane veld, 1987). It is of interest that the sarcoplasmic fatty acyl composition is influenced by thyroid status. reticulum is one membrane that is reported not to I have collated these studies into two groups and exhibit homeoviscous adaptation (Cossins et al., summarized their findings. Table 2 summarizes the 1978). Two other exceptions either do not agree with findings showing membrane fatty acyl composition other studies for the same membrane and are for changes induced by hypothyroidism in mammals, total lipids rather than phospholipids (Paradies & whilst Table 3 shows the effects of thyroid hormone Ruggiero, 1989) or are for thyrotoxic rabbits injections on the membrane fatty acyl composition of (Pilarski et al., 1991). mammals. Most of these studies have not been Membrane fatty acid content is not the only published with a view to addressing the hypothesis variable that influences membrane fluidity. For stated above, but rather have been concerned with example, cholesterol also acts to decrease membrane thyroid-induced changes in lipid metabolism. The fluidity. Cholesterol is normally not a large compo- two Tables have been restricted to studies examining nent of intracellular membranes but is an important the fatty acyl composition of phospholipids or, in constituent of plasma membranes. In mammalian some cases, the composition of total lipids where the tissues, the cholesterol\phospholipid ratio is approx- total lipids are predominantly membrane lipids. imately 0.1–0.2 for mitochondria and endoplasmic Only statistically significant effects are shown. The reticulum but is approximately 0.8 for the plasma species include humans, rabbits and rats. The tissues membrane (Schroeder, Wood & Kier, 1998). Al- include the liver, heart, intestine, skeletal muscle, though thyroid status has dramatic effects on plasma brain, erythrocytes and leucocytes. The subcellular cholesterol levels (hypothyroidism is associated with membranes include those from mitochondria, micro- serum hypercholesterolemia and vice versa, see Heim- somes, sarcoplasmic reticulum, nuclei, and plasma berg, Olubadewo & Wilcox, 1985), it appears not to membranes. In some studies, individual phospho- have the same effects on membrane cholesterol lipid classes were also analyzed for changes in their levels. In rats, hyperthyroidism resulted in an acyl composition. increase in both cholesterol and phospholipid con- The diversity of animal treatments, techniques tent of erythrocyte membranes but no change in the and original purposes for the varied studies precludes cholesterol:phospholipid ratio, whilst in hypo- a detailed analysis. For example, the range of thyroid thyroidism the decrease in phospholipid content was hormone injection treatments is huge, ranging from greater than the decrease in cholesterol content, with some that are mildly hyperthyroid to some that are a consequent small increase in the cholesterol: thyrotoxic. However, we can conclude that almost phospholipid ratio (Digiorgio, Cafagna & Ruggiero, every membrane system examined showed a change 1988). Hypothyroidism had no significant effect on in its fatty acyl composition with the most consistent the cholesterol content of rat liver microsomes (Faas change (the few exceptions are discussed below) & Carter, 1982), sarcoplasmic reticulum from rat being an increase in the 18:2 content of membranes fast muscle (Simonides & van Hardeveld, 1987), or with hypothyroidism and a decrease with hyper- human erythrocytes (van Doormal et al., 1986) thyroidism. In addition, there was generally a although it did result in a significant decrease in the reciprocal change in membrane 20:4 content, cholesterol content of rat colonic apical plasma namely a decrease in hypothyroidism and an membranes (Brasitus & Dudeja, 1988). Hyper- increase following thyroid hormone injection. Many thyroidism had no significant effect on the chol- of these studies have only reported the main fatty esterol content of rat liver microsomes (Faas & acyl chains, and in many the influence of thyroid Carter, 1981; Ruggiero et al., 1984) but resulted in a hormones on n-3 polyunsaturates were not mea- significant decrease in cholesterol content in rat fast sured. muscle sarcoplasmic reticulum (Simonides & van The only study to examine thyroid-hormone- Hardeveld, 1987) and liver mitochondria (Ruggiero induced changes in acyl composition of brain et al., 1984). The results concerning the cholesterol phospholipids found only small changes which were content (relative to fatty acid content) of heart Thyroid hormones and their effects 555 mitochondria are contradictory, one study suggest- has shown that it is significantly decreased in ing that thyroid hormones significantly decrease the hypothyroidism (Faas & Carter, 1982). However, cholesterol content (Clejan et al., 1981) whilst the same authors in the previous year (Faas & another suggests there is no effect (Hoch, 1982). The Carter, 1981) reported a significant decrease in ∆6 influence of thyroid status on membrane cholesterol desaturase activity following T3 injections in normal content is not as consistent as that on membrane rats. These doses (two orders of magnitude greater fatty acid profile but should be taken into account than the euthyroid rate of T3 production) are as the presence of cholesterol (and other sterols) in possibly too high to give any physiological relevance membranes influences membrane desaturase activi- to this result. Curiously, the same study found that ties both in vivo and in vitro (e.g. Leikin & Brenner, although measured ∆6 desaturase activity was 1989). reduced, the 18:2 content of the liver microsomes As well as thyroid status affecting the level of was significantly decreased and the 20:4 content plasma cholesterol it also influences the composition significantly increased, which is the opposite of what of plasma fatty acids. In hyperthyroid humans (and one would expect from the reported changes in in a euthyroid individual after a single intake of T4), enzyme activity. In addition, they found the same the 18:2 content of plasma cholesterol esters, plasma dose to result in a very significant increase in ∆6 phospholipids and plasma triglycerides is signif- desaturase activity in food-restricted rats (Faas & icantly reduced (Kirkeby, 1972a,b). Conversely, Carter, 1981). hypothyroidism results in a statistically non-signif- Hoch (1988) also reported that hypothyroidism icant increase in the 18:2 content and a significant results in a significant decrease in ∆6 desaturase decrease in the 20:4 content of plasma phospholipids activity and T3 injection into hypothyroids results in (Kirkeby, 1972b; see Table 2). These changes in a significant increase in enzyme activity but thyro- plasma lipids are the same as those observed for toxic levels of T3 result in a decreased enzyme membranes and it is possible that these plasma activity. Similarly de Gomez Dumm, de Alaniz & changes are the result of changes at the level of the Brenner (1977) also showed a decrease in ∆6 membrane ∆6 desaturase. desaturase activity at grossly thyrotoxic levels of T4 In view of the major mechanisms regulating injections. This illustrates once again the need to use membrane acyl composition, namely the de- physiologically relevant doses of hormones in ex- acylation\reacylation cycle, it is of interest that T4 perimental situations. has recently been reported to stimulate the acylation Are these thyroid-hormone-induced changes in of lysophosphatidylethanolamine in rat heart membrane fatty acid composition mediated by a (Dolinsky & Hatch, 1998). Similarly, in rats, nuclear or an extra-nuclear site of action? Some hypothyroidism results in a decrease in the activities insight into this could be gathered from the time of liver mitochondrial phospholipase A# and cyto- course of such changes. Hoch (1988, p. 204) reports solic lysophospholipase and an increased activity of that within 1–2.5 h after T3 injection into hypo- two liver microsomal acyltransferase enzymes thyroid rats there were significant changes in (Dang, Faas & Carter, 1985). The means whereby membrane fatty acid composition of rat liver the activities of these enzymes are altered due to mitochondria (specifically, decreases in 18:2, 20:4, thyroid status are not known. That they are and 22:6, with increases in 16:0 and 18:0). These responsive to changes in membrane properties and changes had begun by 0.5 h after the T3 injection, are part of the system regulating membrane acyl but were not statistically significant at that time; composition is illustrated by the finding that changes however, there were only two animals in the 0.5 h in its surrounding lipid environment can influence sample. He also reports (Hoch, 1988, p. 221) that the affinity of a rat liver microsomal acyltransferase within 1 h after the T3 injection there was a for its acyl-CoA substrates (Fyrst et al., 1996). significant increase in ∆6 desaturase activity in the Many of the studies listed in Tables 2 and 3 liver microsomes in these rats. Once again the demonstrated reciprocal changes in 18:2 and 20:4 changes were manifest within 0.5 h after hormone content. The ∆6 desaturase is the important step for injection but probably because only two animals the conversion of 18:2 to 20:4 and thus there is were examined there was no statistical significance indirect evidence that the activity of this enzyme is to the difference. These results suggest that the decreased in the hypothyroid state and increased by membrane fatty acid changes occur too quickly to be the presence of thyroid hormones. Direct measure- a nuclear-receptor-mediated effect and may occur as ment of ∆6 desaturase activity in rat liver microsomes a direct effect of the thyroid hormones on the 556 A. J. Hulbert ) a . (1986) et al . (1991) . (1991) . (1984) . (1984) . (1984) . (1984) . (1984) . (1984) . (1984) . (1984) . (1984) et al et al . (1972) . (1980) . (1981) et al et al et al et al et al et al et al et al et al et al et al et al 108 Patton & Platner (1970) Platner 1(1.0–2.5) Hoch (1988) 5 Ismailkhodzhaeva 5 Faas & Carter10 (1981) Steffen & Platner (1976) 3 Hoch (1982) 7 Clejan 167 Gueraud & Paris (1997) Clejan 2 (oral)2 (oral) Raederstorff Raederstorff 1(4) Chen & Hoch (1977) 1(3)1(1) Shaw & Hoch (1977) Shaw & Hoch (1977) 1(2)1(3) Shaw & Hoch (1977) Shaw & Hoch (1977) i i 555 Ruggiero Ruggiero Ruggiero 55 Ruggiero 5 Ruggiero 555 Paradies &5 Ruggiero (1990 Ruggiero 21 Ruggiero 21 Ruggiero Ruggiero Faas & Carter (1981) Faas & Carter (1981) i i i i i i i i i i i i i i i i i i i i i i i i i i i T4; 5 T4; 5 T4; 5 \ \ \ T4; 50 T3; 100 T4; 0.5 T4; 0.5 T4; 50 T3; 25 T3; 25 T3; 25 T4; 2000 T4; 2000 T3; 30 T3; 30 T3; 30 T3; 30 T3; 30 T3; 30 T3; 30 T3; 30 T3; 500 T4; 100 T3; 30 T4; 100 T3; 30 T3; 25 T3; 25 T3; 100 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ E E E E H E H E H H E E E HX HX H E E E E 0HX 0E j j k j 20:4 0E 0 0E 0E 00H 00H 00H jj kk jj k 00 0 0 k kk kj kjj kjj kjj k kjj kjj kjj kjj k k k kjj kjj kjj 0 0 0 0 0 0 0 0 000 0 0 00 0 0 j jk jk j 00 00 j j j jkkjj j jk j jkkjj j Fatty acyl chain 00000 0000 000000E kj kj k jj jj jj kj kj kj j Lipid class 16:0 18:0 18:1 18:2 20:4 18:2 Treatment Reference The influence of thyroid hormone injections on membrane fatty acyl composition in mammals Species TissueRat Membrane Liver Mitochondria TL RatRatRat LiverRat Liver MitochondriaRat Liver MitochondriaRat Liver MitochondriaRat Liver Mitochondria Liver Mitochondria TL Liver Mitochondria PL Mitochondria PL 0 PL 0 PCRat PERat PI Liver 0 Liver 0 Mitochondria Microsomes TL PL Rat Heart Mitochondria PL RatRatRat LiverRat Liver MitochondriaRat Liver Mitochondria Liver Mitochondria Liver MitochondriaRat Mitochondria CL Rat PL Rat Liver PCRat Liver Microsomes PERat Liver Microsomes TLRat Liver 0 MicrosomesRat Liver 0 Microsomes Liver 0 0 MicrosomesRat Liver 0 Microsomes PCRat 0 Microsomes PERat Heart 0 PI 0 Rat Heart Mitochondria TL 0 Heart Mitochondria TL Heart Mitochondria TL 0 Mitochondria PL TL 0 PL 0 PL 0 0 PL 0 0 0 0 RatRat Heart Heart Mitochondria Mitochondria PL PL 0 0 0 0 Table 3. Thyroid hormones and their effects 557 ; . (1989) . (1989) . (1989) . (1989) . (1989) . (1989) . (1989) . (1975) . (1994) . (1991) . (1991) . (1991) et al et al et al et al et al et al et al et al et al et al et al et al (1987) (1987) number of days injected (h after injection). 10 Steffen & Platner (1976) 77 Simonides & van Hardeveld Simonides & van Hardeveld 5 Paradies & Ruggiero (1988) 52077 Paradies 77 Ricquier 7 Ruggiero 7 Ruggiero 7 Ruggiero Ruggiero Ruggiero Ruggiero Ruggiero 77 Pilarska Pilarska 7 Pilarska i i i i i i i i i i i i i i i i i T3; 10 T3; 10 T3; 30 T3; 30 T4; 50 T3; 30 T3; 30 T3; 30 T3; 30 T3; 30 T4; 50 T4; 100 T4; 50 T3; 30 T3; 30 T4; 50 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ E E E E E E E E H H E j 0 00E kk jj j 000E 0 0 kjj kjj kk g hormone(per 100 g body mass) µ 0000E 0 0 kjj jjkk k 0 0 jk j jkj j 000000E 00000 E 0000 kjjk kjkjj jj jkj jkjkjj j j thyroid hormone injected; \ ’, statistically significant decrease; 0, no significant effect. k ’, statistically significant increase; ‘ Rat Heart Mitochondria PL Rat Heart Mitochondria CL Rabbit Muscle Sarcolemma PE RatRatRat HeartRat BAT MicrosomesRat BAT MitochondriaRat BAT MitochondriaRat BAT MitochondriaRat BAT MitochondriaRat BAT TL MitochondriaRat PL BAT Microsomes PL BAT Microsomes PC Muscle Microsomes Sarcoplasmic PE reticulum 0 PC CL 0 Rabbit 0 PL Muscle 0 Sarcolemma PC 0 PE 0 0 PC RatRabbit Muscle Sarcoplasmic Muscle reticulum Sarcolemma PE PL 0 0 j BAT, brown adipose tissue; TL, total lipid; PL, phospholipid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol CL, cardiolipin. ‘ Treatments are given as:E, initial euthyroid; condition H, hypothyroid; HX, hypophysectomized; T4, thyroxine; T3, triiodothyronine. 558 A. J. Hulbert membrane lipid environment and consequently on Although TRs bind T3 with greater affinity than ∆6 desaturase activity. It is of interest that within 0.5 T4, and exogenous T3 can inhibit TSH secretion, h of T3 injection as well as a change in the enzyme serum TSH levels are more negatively correlated activity there was a large change in the Michaelis– with serum T4 than serum T3 levels (see Chopra, Menten constant (Km) of the ∆6 desaturase enzyme 1996). The relative ability of thyroid hormones to which was present for at least 2.5 h following inhibit TSH release will be a combination of their hormone injection (Hoch, 1988). It is possible that uptake, intracellular metabolism and relative bind- the change in ∆6 desaturase activity is primarily ing to the nuclear TRs. Everts et al. (1995) compared responsible for the membrane fatty acid changes the relative uptake, metabolism and inhibition of although more work is necessary to ascertain TSH release (by cultured rat anterior pituitary cells) definitely this as the mode of action. of thyroid hormones and their deaminated products, To my knowledge, nothing is known of the effects Tetrac and Triac. Relative to the free concentration, of other iodothyronines and thyroid hormone ana- cellular uptake rates were in the following order; logues on membrane fatty acid composition. Tetrac " T3 " T4 " rT3. Whereas T3 was not metabolised, Tetrac, T4 and rT3 were all partly metabolised by the cultured anterior pituitary cells. (3) Effects on the thyroid hormone axis The order of potency for reducing TSH release was The concentration of thyroid hormones in plasma Triac " T3 " Tetrac " T4 but it seems that for and tissues is under the control of the thyroid both Tetrac and T4 it is likely that their effect is hormone axis which consists of the following: (1) indirect and possibly due to their respective metab- thyrotropin-releasing hormone (TRH) secretion by olites following intracellular deiodination. Although cells in the hypothalamus, (2) TSH secretion by cells nuclear receptors show negligible affinity for rT3 of the anterior pituitary, (3) the plasma binding and T2, these two iodothyronines exhibited in vitro proteins, (4) cellular uptake mechanisms, (5) in- inhibition of TSH release. These effects were tracellular deiodinases, and (6) nuclear thyroid observed at concentrations above normal and would hormone receptors. thus seem to have little in vivo relevance. They do, In conscious undisturbed sheep, thyroidectomy however, raise mechanistic questions about possible results in increased TRH secretion and T4 re- additional non-nuclear receptor influences on TSH placement reverses this effect (Dahl et al., 1994). secretion. This negative influence of T4 on TRH also occurs in Once secreted into the circulation most thyroid the rat (Rondeel et al., 1988). Transfection of the hormone molecules are bound to plasma proteins. promoter region of the human TRH gene, together TBG has two possible positive TREs, one in intron 0 with thyroid receptor genes into cell cultures has and the other in the 3h region of the gene (Akbari et shown that T3 can repress transcription (Hollenburg al., 1993). In monkey hepatocarcinoma cell culture, et al., 1995). The latter study identified two increasing T4 concentration in the medium from structurally different negative TREs (both single 0.01 p to 10 pM increased TBG appearance whilst half-sites that bind TR monomers) which cooperate further increases of T4 concentration to 10 n to allow negative regulation of the human TRH resulted in a decrease in TBG appearance in the promoter, and suggested that this may be restricted medium (Gershengorn et al., 1976). In humans, to TRβ. More recently, it has been shown that the treatment of thyrotoxicosis resulted in a 24% TRβ-2 isoform, which is most expressed in the increase in TBG concentrations although there was hypothalamus and anterior pituitary, has a greater considerable overlap in TBG levels measured in effect than either TRβ-1 or TRα-1 and that this thyrotoxic and hypothyroid patients (Burr et al., effect is mediated by a novel amino terminal domain 1979). The role of these putative TBG TREs is not of the receptor (Langlois et al., 1997). clear. Another T4 binding protein, TTR, is un- Thyroid hormones are the major regulators of responsive to thyroid status in that the levels of TTR TSH production and inhibit its secretion by anterior mRNA in liver and choroid plexus were unchanged pituitary thyrotropes (Shupnik, Ridgway & Chin, during both hypothyroidism and hyperthyroidism in 1989). TSH is a glycoprotein consisting of an α the rat (Blay et al., 1993). Similarly, plasma albumin (common to some other hormones) and a β subunit concentrations are not known to be changed by (specific to TSH). Negative TREs are found in both alterations in thyroid status in the adult. During the α subunit gene and the TSHβ subunit gene for hypothyroidism in humans there are changes in the both rats and humans (Williams & Brent, 1995). binding of both T4 and T3 to plasma lipoproteins, Thyroid hormones and their effects 559 but no changes in the distribution of rT3 binding to and involves specific interactions between the D2 the various lipoprotein classes (Benvenga & Rob- enzyme, F-actin and T4 molecules (Farwell & bins, 1996). Leonard, 1992). Investigation of the degradation Changes in the thyroid hormone pool in intestinal and recycling of a D2 subunit in astrocytes suggests contents will diminish the decrease in plasma thyroid that T4 influences fundamental processes involved hormone levels following thyroidectomy (DiStefano with the turnover of integral membrane proteins et al., 1993). In euthyroid rats, the amount of T3 in (Farwell et al., 1996). Decreases in D2 mRNA levels the intestinal contents is 20 times the plasma pool in the cerebral cortex of hypothyroid rats have been size and this T3 is exchangeable with the circulation. recorded following injections of both T4 and T3 For example, T4 and T3 levels in the hepatic portal (Burmeister, Pachucki & St. Germain, 1997). The vein plasma are, respectively, 15% and 30% greater inhibitory effects of T4 on the D2 enzyme are rapid, than that in systemic plasma in euthyroid rats. In being evident within 15 min of addition of hormone euthyroid rats, 34% of infused T3 is excreted in the to cultured GH3 cells (St Germain, 1985). T4 is faeces; however, following thyroidectomy this de- more potent than T3 in influencing D2 mRNA creases to 20% of infused T3, and the hepatic portal levels. This thyroid hormone effect is not mediated plasma concentrations of T4 and T3 are 95% and by nuclear TRs. 63% greater than the respective concentrations in Thyroid hormones influence the D3 deiodinase in the systemic plasma. Whereas the plasma T3 a similar manner to their effect on the D1 deiodinase. concentration decreases by 64% in thyroid- Hypothyroidism results in a decrease in skin D3 ectomized rats, T3 concentration in the intestinal activity in the rat (Huang et al., 1985) whilst thyroid contents decreases by 85%. (DiStefano et al., 1993). hormones induce D3 activity in cultured astroglial To my knowledge, it is not known if any of the cells (Esfandiari et al., 1992). However, D3 activity cellular uptake mechanisms for thyroid hormones in the placenta is unaffected by thyroid status in the are influenced by thyroid status. rat (Emerson et al., 1988). Expression of the mRNA Intracellular deiodinases are also influenced by for D3 in cultured cell lines from the amphibian thyroid status. D1 deiodinase activity in rat liver and Xenopus laevis is increased in the presence of T3 (St. kidney is increased in hyperthyroidism and de- Germain et al., 1994). Although to my knowledge no creased in hypothyroidism (Kaplan & Utiger, 1978) TRE has been described associated with the D3 and this appears to be a TR-mediated effect in that gene, this effect of thyroid hormones is likely to be two functional positive TREs have been identified mediated by thyroid nuclear receptors. upstream of the human D1 gene (Toyoda et al., Although levels of mRNA for TRβ-1 in rat heart, 1995). However although mouse D1 gene expression kidney, liver and brain are unaffected, in the is increased by thyroid hormone the upstream region pituitary supraphysiological doses of T3 increase of the mouse gene does not seem to contain a TRE TRβ-1 mRNA levels (Hodin, Lazar & Chin, 1990). and thus its control may be different to that of the Similarly, in the rat around birth there is a dramatic human D1 gene (Maia et al., 1995). increase in TRβ-1 mRNA levels associated with an The activity of the D2 deiodinase is inversely increase in T3 concentration (Strait et al., 1990) and related to thyroid status (Kaplan, 1986). This is the in Xenopus laevis during metamorphosis there is an opposite of the influence of thyroid status on D1 and, up-regulation of TRβ-1 concentration mediated by unlike D1, this has been shown to be a non-nuclear- increasing T3 levels (Kanamori & Brown, 1992). mediated action of the thyroid hormones. In the That these tissue-specific and developmental-time- absence of thyroid hormones, the D2 activity of specific increases in TRβ-1 mRNA levels are TR- astrocytes in cell culture was 10-15-fold greater than mediated effects is supported by the finding of two when they were grown in the presence of serum and TREs in the promoter region of the human TRβ it was demonstrated that this increase in D2 activity gene (Suzuki et al., 1994). The fact that a supra- was due to decreased removal of the enzyme with no physiological T3 dose resulted in an increase in change in its rate of synthesis (Leonard et al., 1990). TRβ-1 mRNA levels in the pituitary in the rat is, T4 and rT3 had similar potency whilst T3 was 100- however, complicated by the observation that it also fold less effective, which is similar to in vivo findings resulted in a simultaneous decrease in levels of TRβ- (Silva & Leonard, 1985). The thyroid-hormone- 2 mRNA (Hodin et al., 1990) which is transcribed sensitive degradation\inactivation of the membrane- from the same gene. In the anterior pituitary of the bound D2 is an energy-requiring process requiring rat, the total nuclear TR binding capacity and an intact actin cytoskeleton (Leonard et al., 1990) isoform distribution is relatively unaffected by 560 A. J. Hulbert changes in thyroid status, either hypothyroidism or It appears that, extracellularly, the ‘regulated’ hyperthyroidism (Ercan-Fang, Schwartz & Oppen- variable is free T4, whilst intracellularly it is T3. heimer, 1996). There is a hierarchy of tissues with respect to the Thyroid hormones have effects at almost every maintenance of intracellular T3 levels. This hi- level of the thyroid hormone axis and together these erarchy is manifest in the results of Escobar-Morreale effects appear to function as a multi-level, hier- et al. (1996) where some tissues (notably cerebellum archical system that, in the adult, acts homeo- and cerebral cortex) maintained a relatively con- statically to maintain a relatively constant cellular stant tissue T3 concentration whilst in others it was level of T3 in some tissues (notably the brain). All variable (e.g. muscle, adrenal, ovary and heart). but two of these effects (namely the effect on D2 and Currently, there is increasing interest in 3,5-T2, a the enterohepatic changes) appear to be mediated deiodination product of T3. The ‘homeostatic-like’ by nuclear thyroid receptors. The effects on TRH influence of thyroid status on the deiodinase enzymes and TSH secretion are part of the classic negative means that plasma 3,5-T2 concentration will also feedback system that acts to maintain a relatively vary less than plasma T3 levels during changes in constant plasma T4 concentration. The entero- thyroid status. For example, Faber et al. (1982) hepatic effects of hypothyroidism will also act to found that although serum T3 levels changed by dampen fluctuations in plasma T4 concentration. j82% and k71% (relative to euthyroid values) in The high-affinity-binding plasma proteins in differ- hyperthyroidism and hypothyroidism, respectively, ent vertebrates can be regarded as mechanisms serum 3,5-T2 concentration changed by only j54% maintaining a relatively constant level of plasma free and k27 %, respectively (this last change was not T4 which in turn will facilitate an even distribution statistically significant). Recently, it has been shown of T4 to the tissues both with respect to time and that 3,5-T2 can inhibit TSH secretion from the body location. The secretion of TTR by the choroid anterior pituitary cells of rats both in vivo and in vitro plexus into the cerebrospinal fluid of higher verte- (Horst et al., 1995) and that it also stimulates brates probably has a similar function. The effect of pituitary D1 activity in rats (Baur et al., 1997). thyroid status on the three deiodinases, and their Although nuclear receptors show negligible affinity respective tissue distribution can be interpreted as a for T2, its in vivo potency is not greatly dissimilar system to maintain a constant level of cellular T3 in from that of T3 and this finding raises intriguing select tissues, especially the brain. During hypo- possibilities concerning the mode of action of thyroid thyroidism, T4 consumption will be minimized in hormones. Measurement of normal intracellular those tissues which have D1 (e.g. liver and kidney) levels of 3,5-T2 are necessary before we can be sure whilst intracellular T3 content will be maintained in of the physiological relevance of these findings. those tissues that have the D2 enzyme, while the decreased D3 activity will minimize T3 removal. (4) Effects on metabolism and This means that especially within the brain there is thermogenesis strong homeostasis of cellular T3 levels. For example, brain T3 levels are at euthyroid levels when The stimulatory effects of thyroid hormones on hypothyroid rats are given T4 at only 10% of the metabolic activity (and thus heat production) have normal replacement dose (Silva & Leonard, 1985). been known for over a century (Magnus-Levy, In hypothyroidism, the rat brain makes almost all of 1895). Indeed, before the advent of modern hormone its own T3 from intracerebral T4, whereas during assays, the measurement of basal metabolic rate hyperthyroidism, it derives most of its T3 from the (BMR) was a diagnostic test for the thyroid status of circulation (Dratman et al., 1983). A similar response individuals (e.g. Dubois, 1936). has been recorded in the rat fetus during hypo- The finding that the allometry of T4 and T3 thyroidism (Ruiz de Ona et al., 1988; Morreale de utilization by mammals has the same exponents as Escobar et al., 1989). Whether such homeostasis is their BMR relationship (Tomasi, 1991) means that present in other tissues is not known but one study there is a stoichiometric relationship in all mammals, suggests that it may also occur in brown adipose whether mice or mammoths, whereby consumption tissue (van Doorn, Roelfsema & van der Heide, of one mole of oxygen is associated with the 1986). This mechanism, unknown at the time, is consumption of approximately 6.3 pmol of T4 and likely to be responsible for the early reports of the approximately 0.63 pmol of T3. brain being unresponsive to changes in systemic Thyroidectomy is capable of decreasing BMR by thyroid status. up to 40% in humans (Dubois, 1936) and has a Thyroid hormones and their effects 561 similar effect in other mammals and in birds. matched (never obese) controls demonstrated that Although it was originally believed that the thyroid the previously obese subjects had a low BMR and was not capable of stimulating metabolism in low plasma free T3 levels (Astrup et al., 1996). In ectothermic vertebrates (and this statement is re- that study, the lower plasma free T3 levels could peated in some recent reviews, e.g. Freake & statistically explain the low BMR of the obesity- Oppenheimer, 1995; Oppenheimer, Schwartz & prone individuals but whether there was a causal Strait, 1995), it was shown some time ago that connection could not be established. These corre- whether thyroid hormones stimulate the metabolism lations between metabolism and T3 concentration of reptiles and amphibians depended on their body may be due to a low metabolic activity being temperature (Maher, 1961, 1964, 1967). It has since associated with low deiodination activity and thus a been confirmed that thyroid status affects metab- low production rate of T3. olism in ectotherms (e.g. John-Alder, 1983) and that Thyroid hormones can also affect other measures this effect is body-temperature-sensitive (e.g. Hul- of metabolic activity. For example, thyroidectomy bert & Williams, 1988). Joos & John-Alder (1990) decreases work rate and work efficiency in goats have shown that T4 supplementation stimulated (Kaciuba-Uscilko et al., 1987) and decreases aerobic standard metabolic rate (SMR, analogous to BMR) capacity (V#max) and locomotor endurance in by 20% in laboratory-captive lizards and by 59% in lizards (John-Alder, 1984a). the field-active individuals. Although night-time That the stimulation of metabolic activity by ambient temperatures were not different between thyroid hormones is largely a cellular event is evident the laboratory and field, day-time ambient temper- from early studies which showed that tissue slices atures were greater in the field than the laboratory. isolated from rats of different thyroid status had An interesting finding from this study is that T4 different in vitro respiration rates (Barker & supplementation resulted in a decrease in plasma T3 Klitgaard, 1952). Most tissues appear to be thyroid levels and SMR was correlated with plasma total T4 hormone sensitive. The conclusion from early studies concentration but not T3 levels. For a review of that brain was refractory to thyroid stimulation of thyroid influence on ectotherm metabolic activity metabolism should be re-examined in view of our the reader is referred to Gupta & Thapliyal (1991). current understanding of the relative constancy of Metabolic activity is quite sensitive to plasma brain T3 levels in different thyroid states. Much thyroid hormone concentration. In a recent study of biochemical work has been carried out on isolated patients receiving chronic thyroid hormone replace- enzymes and sub-cellular preparations (e.g. mito- ment, small changes in their T4 dose significantly chondria) from animals of varied thyroid status in affected their resting energy expenditure. Whilst order to identify the cellular processes important in plasma TSH concentration was most sensitive to this thyroid-stimulated metabolic activity. However, small changes in T4 dose, resting energy expenditure as with many in vitro measurements, the physiological was the next most sensitive parameter measured and relevance of many of these findings has been difficult thyroid-sensitive blood variables such as levels of to determine because of the problem of extrapolation plasma sex hormone binding globulin, triglycerides, to the in vivo situation. The development of metabolic lipoprotein cholesterol levels and angiotensin con- control analysis and its associated techniques over verting enzyme showed no overall change. The low the last quarter of a century has allowed a more T4 daily dose used in this study averaged 98 µgT4 quantitative analysis of the importance of various while the highest dose used was an average 141 µg cellular processes during thyroid stimulation of T4 per day (respectively, equivalent to approxi- metabolism (see Harper & Brand, 1995). For a mately 0.25 µg and 0.35 µg 100 g body mass per review of the use of top-down regulation analysis in day). This small change in daily T4 dose raised understanding energy metabolism the reader is plasma free T4 levels from 18 p to 23 p and referred to Brand (1997). resulted in an average 6% increase in resting energy Basal metabolic rate varies dramatically between expenditure (Al-Adsani, Hoffer & Silva, 1997). vertebrates. Endotherms (mammals and birds) have In euthyroid humans, BMR was found to be BMRs that are 5–10 times those of ectothermic positively correlated with total T3 concentration vertebrates of the same body size (Hulbert, 1980b) (Stenlof et al., 1993). Another study found that daily and BMR varies allometrically with body mass in a energy expenditure of healthy humans was corre- very predictable manner such that per gram body lated with plasma free T3 levels (Toubro et al., mass, a mouse has a BMR that is approximately 20 1996). A comparison of post-obese woman with times that of a horse (Kleiber, 1961). Part of these 562 A. J. Hulbert differences are due to variations in the relative size of the least sensitive and protein synthesis being the metabolically active organs but a substantial part of most sensitive to changes in energy supply (Butt- this BMR variation is related to basic cellular gereit & Brand, 1995). The maintenance of ion differences. A surprising finding has been that, gradients represents a major energy cost of the non- although there is considerable quantitative differ- equilibrium condition we know as life. ence in cellular metabolic activity, the relative In hepatocytes isolated from rats, hypothyroidism composition of cellular metabolic rate is relatively and hyperthyroidism resulted in, respectively, a constant for a particular tissue between different decreased and increased oxygen consumption rela- vertebrates. This has led to the suggestion of a tive to littermate euthyroid controls (Harper & cellular ‘pacemaker’ for metabolic rate. The finding Brand, 1993). Top-down elasticity analysis showed that membrane polyunsaturation is positively corre- that in the hypothyroid-euthyroid comparison, half lated with metabolic activity both in phylogenetic the increase in hepatocyte oxygen consumption was comparisons (Brand et al., 1991; Brookes et al., 1998) due to an increased mitochondrial proton leak and and body size comparisons (Couture & Hulbert, half was due to an increase in non-mitochondrial 1995a, b; Porter, Hulbert & Brand, 1996) has led to oxygen consumption with no net change in ATP the suggestion that membranes are such ‘pace- turnover. In the hyperthyroid-euthyroid compari- makers’ of metabolism (Hulbert & Else, 1999, 2000). son, just under half the increase was due to a greater Although the precise quantitative contribution mitochondrial proton leak, whilst just over half was varies between cell types, the major energy-con- due to an increase in ATP turnover and there was no suming processes of resting cells are: (i) maintenance significant change in the non-mitochondrial oxygen of the mitochondrial membrane H+ gradient, (ii) consumption (Harper & Brand, 1993). Even when maintenance of the plasma membrane Na+ gradient, rat hepatocytes are performing considerable urea- # (iii) maintenance of low cytosolic [Ca +], (iv) genesis and gluconeogenesis, thyroid hormones stim- synthesis of macromolecules (DNA, RNA, and ulate both coupled respiration (that associated with protein), and in some cells (e.g. liver) processes such ATP production) and non-coupled respiration (i.e. as ureagenesis and gluconeogenesis (for a recent either mitochondrial proton-leak-associated respir- review, see Rolfe & Brown, 1997). In active muscle ation or non-mitochondrial oxygen consumption, or cells, considerable energy consumption is associated both) (Gregory & Berry, 1992). with contractile activity. All but the first of these Thyroid status influences proton leak in isolated activities are mediated by ATP consumption. rat liver mitochondria with the hypothyroid-hy- In active cells, the mitochondrial proton gradient perthyroid transition resulting in a sevenfold increase is used to produce ATP for the ATP-consuming in the proton permeability of mitochondria (Hafner processes. In resting cells, substrate oxidation con- et al., 1988). Related to this effect, hyperthyroidism tinues but in this case, when ATP consumption is results in a mitochondrial membrane potential that minimal, proton pumping counteracts the leak of is approximately 30 mV lower than euthyroid protons across the mitochondrial inner membrane. controls when measured in intact resting rat hepato- In this context, the mitochondrial proton leak is cytes (Gregory & Berry, 1991). Other workers have analogous to the ‘governor’ of a steam engine, in shown that in intact rat hepatocytes, the mito- that, just as the governor of a steam engine prevents chondrial membrane potential is greatest in hepato- explosion of its boiler, proton leak (which is non- cytes from hypothyroid rats and least in those from linearly related to the mitochondrial membrane hyperthyroid rats; however, respiratory rate showed potential) acts to counteract the continual build-up the opposite trend being least in hypothyroid and of the trans-membrane proton gradient and thus the greatest in hepatocytes prepared from hyperthyroid mitochondrial membrane potential. For a discussion rats (Bobyleva et al., 1998). As these authors suggest of mitochondrial proton leak, the reader is referred this indicates a physiological effect (rather than a to Brand et al. (1994) and Rolfe & Brand (1997). pathological consequence) of thyroid hormones The maintenance of transmembrane ion gradients decreasing mitochondrial energy coupling. is important in that many other cellular processes Mitochondrial proton leak is found in all the are linked to such gradients. They can be thought of major oxygen-consuming tissues of the rat (Rolfe, as a form of short-term energy storage. Studies on rat Hulbert & Brand, 1994), and in the liver mito- thymocytes show there is a hierarchy of cellular chondrial proton leakiness has been shown to be ATP-consuming processes to changes in energy correlated with the relative polyunsaturation of the supply, with the maintenance of ion gradients being mitochondrial membrane (Porter et al., 1996; Thyroid hormones and their effects 563

Brookes et al., 1998). An examination of the larly, in rat heart mitochondria thyroid hormones mechanistic basis for the increased mitochondrial stimulate pyruvate translocation and cytochrome proton permeability induced by thyroid hormones oxidase activity, without equivalent changes in the suggested that two factors were responsible. These amount of the specific mitochondrial enzyme were: (i) changes in the relative amount of mitochon- (Paradies & Ruggiero, 1989; Paradies et al., 1993, drial inner membrane, and (ii) changes in the 1994). These studies also report thyroid-induced intrinsic permeability of the phospholipid bilayer changes in mitochondrial membrane lipids and from the mitochondrial inner membrane (Brand et suggest that the membrane lipid changes influenced al., 1992). Such permeability changes are not the molecular activity of these mitochondrial mem- restricted to proton permeability in that hypo- brane proteins. In a similar vein, Beleznai et al. thyroidism has also been shown to result in a (1989) demonstrated that T4 dramatically in- decreased inner membrane permeability to other fluenced the physical properties of the lipid micro- cations in rat liver mitochondria (Hafner, Leake & environment of the mitochondrial enzyme, -gly- Brand, 1989). cerol-3-phosphate dehydrogenase. Metabolic control analysis of liver mitochondria It seems likely that thyroid effects on mito- has shown that control of mitochondrial respiration chondrial function can best be described as the is distributed among a number of steps and that consequence of thyroid-hormone-induced changes effective stimulation by thyroid hormones can only in membrane lipid composition, especially changes occur if several enzymes are activated simultaneously in fatty acyl composition. A thyroid-hormone- (Verhoeven et al., 1985). In that study, thyroid induced increase in the relative polyunsaturation of hormone treatment resulted in an increase in the flux mitochondrial membranes (e.g. decreased 18:2 control coefficient of the adenine nucleotide trans- combined with increased 20:4) is proposed to locator, the dicarboxylate carrier and cytochrome stimulate the molecular activity of proteins asso- oxidase. Sterling & Brenner (1995) have convinc- ciated with the membranes (see Hulbert & Else, ingly shown that low doses of thyroid hormones 1999). No mitochondrial enzymes are among those stimulate adenine nucleotide translocator activity in genes reported to have a TRE (see Williams & rat liver mitochondria. Brent, 1995). Many studies have reported an effect of thyroid Thyroid hormone stimulation of cellular ATP status on the activity of a wide variety of mito- turnover requires an increase in ATP consumption chondrial enzymes from a variety of tissues, with an as well as ATP production. As noted above, the increase in thyroid status generally associated with maintenance of the Na+ gradient across the plasma greater enzyme activity. For a review on thyroid membrane of cells is a major cellular energy hormone action on mitochondria, the reader is consumer, the importance of which varies between referred to Soboll (1993). Thyroid hormones both tissues (for review see Clausen, van Hardeveld & result in more mitochondrial membrane and an Everts, 1991). This gradient is maintained by the altered membrane lipid composition (e.g. see Tables ubiquitous Na+,K+-ATPase, also known as the Na+ 2 and 3). Changes in mitochondrial membrane lipid pump. Almost 30 years ago, Ismail-Beigi & Edelman composition are associated with changes in the (1970, 1971) proposed that a substantial portion of Arrhenius kinetics of mitochondrial enzyme activity the enhanced oxygen consumption associated with that is also associated with changes in temperature- increased thyroid status is due to thyroid hormone induced phase transitions of the mitochondrial stimulation of the Na+ pump. Although there was membrane (Hulbert, Augee & Raison, 1976; Par- considerable discussion about the quantitative im- mar et al., 1995). portance of this effect, there was no disagreement The role of thyroid-induced changes in membrane that Na+ pumping was stimulated by thyroid lipid composition is also of interest in relation to hormones. The disagreement about the quantitative changes in the activity of some mitochondrial importance is largely based on methodological enzymes. The transport of pyruvate and tricarboxyl- considerations. For example, in rat liver, ouabain ates is stimulated by thyroid hormones in rat liver inhibition of the Na+ pump results in a 30–40% mitochondria in that the maximal rate (Vmax) of the decrease in respiration (e.g. Hulbert & Else, 1981) process is increased, yet inhibitor studies suggest that whilst measurement of pump activity suggests that there is no increase in the number of any of these its ATP consumption is equivalent to approximately membrane transporters (Paradies & Ruggiero, 10% of liver cell oxygen consumption (e.g. Couture 1990a, b; Paradies, Ruggiero & Dinoi, 1991). Simi- & Hulbert, 1995b). Measurement of Na+ pump 564 A. J. Hulbert

# activity represents the cost of ‘maintaining’ the Na+ (e.g. possibly Ca +, glucose, amino acids etc.). In this gradient, whilst measurement of ouabain inhibition light, it is of interest that the uptake of deoxyglucose of respiration measures the cost of ‘not maintaining’ was significantly increased after 4 h of hormone this gradient and will thus include inhibitory effects exposure, as was lactate efflux (Haber et al., 1988). of elevated intracellular Na+ concentrations as well Similarly, Everts & Clausen (1988) reported that T3 as the energy consumption by the pump itself. induced an increase in passive K+ efflux (measured The initial report that thyroid hormones stimu- as Rb+ efflux) from rat soleus muscle that preceded lated the ouabain-suppressible oxygen consumption the increase in muscle Na+,K+-ATPase activity by of tissue slices (Ismail-Beigi & Edelman, 1970) was 1–2 days. followed by the demonstration that the in vitro This perspective is also supported by the finding enzymatic activity of the Na+ pump was also that short-term administration of T3 to humans increased (Ismail-Beigi & Edelman, 1971). This resulted in no change in Na+ pump numbers in effect has been reported for a wide variety of tissues leucocytes but a significant increase in uptake rate of (including liver, kidney, heart, skeletal muscles and Rb+ (a commonly used experimental analogue for adipose tissue) and several studies have shown that K+) and rate of Na+ efflux. This indicates an thyroid hormones increase pump numbers and sub- increased molecular activity of the Na+ pumps but unit mRNAs (for review see Ismail-Beigi, 1993). A no change in their number in these cells (Turaihi et comparison of sub-unit mRNAs and Na+,K+- al., 1987). Similarly, it has been demonstrated that ATPase protein levels in a range of tissues has shown T3 administration to thyroidectomized rats stimu- there are differential effects of the thyroid hormones lates fluid reabsorption in their kidney proximal both between tissues and between subunits and that tubules (which is related to Na+ pump activity) but these include both pre- and post-translational in- did not change the in vitro Na+,K+-ATPase activity, fluences (Horowitz et al., 1990). suggesting no change in Na+ pump concentration. An obvious question concerns the mode of action Other experiments indicated that this thyroid hor- for this thyroid hormone effect. It is thus of interest mone effect on in vivo Na+ pump activity resulted that Haber & Loeb (1984) reported a thyroid- from an increase in the K+ permeability of the hormone-induced increase in K+ efflux from rat liver proximal tubular cell membranes (Capasso et al., slices that preceded changes in Na+,K+-ATPase 1985). activity. Similarly, Everts & Clausen (1988) reported One of the main pathways of Na+ entry into cells increases in passive Na+ and K+ leaks that preceded is via the (amiloride-sensitive) Na+\H+ exchanger in changes in ouabain binding (a measure of Na+ pump the plasma membrane. Thyroid status has been numbers). The finding that T3-induced increases in shown to stimulate Na+\H+ exchange in renal brush Na+,K+-ATPase activity were not preceded by border membranes isolated from the rat although it changes in intracellular ion concentrations (Philip- is not known if this is due to an increase in the son & Edelman, 1977) may be related to the fineness number of exchangers or an increase in their of the homeostatic control of these variables by the molecular activity (Kinsella & Sacktor, 1985). Na+ pump and the difficulty of measuring in- Whether such changes are related to changes in tracellular ion concentrations. the acyl composition of the plasma membrane is Measurement of thyroid-hormone-induced unknown but it is well documented that the activity changes in a rat liver cell line showed a significant of Na\K ATPase as well as other aspects of 40% increase in Na+ influx 12 h, and a significant transmembrane ion movements are influenced by 10% increase in K+ efflux 8 h after T3 exposure membrane lipids (see Hoch, 1988; Hulbert & Else, (Haber, Ismail-Beigi & Loeb, 1988). The increase in 2000). A survey of the upstream region of the human in vivo pump activity became statistically significant α sub-unit gene failed to find a TREsequence (Shull, 12 h after hormone exposure and in vitro Na+,K+- Pugh & Lingrel, 1989) and none of the Na\K ATPase activity only became statistically significant ATPase sub-units are among the genes reported to after 16 h. In addition, the increase in in vivo pump possess a TRE (Williams & Brent, 1995). activity was consistently greater than the increase in The most parsimonious explanation is that the in vitro enzyme activity (Haber et al., 1988). An increased nuclear expression of Na+,K+-ATPase intriguing possibility is that the increase in passive units is a response to an increase in passive ion leaks. Na+ flux is secondary to a thyroid-hormone-induced This increased leak results in an increased activity of increase in the fluxes of other molecules whose the Na+ pumps already present in the membrane transport is sometimes linked to the Na+ gradient and later an increase in the number of pumps. Such Thyroid hormones and their effects 565 a mechanism, whereby the demand for ion transport membrane and its low concentration in the cytosol # is the primary driving force for Na+ pump synthesis, makes Ca + entry an ideal information transfer has been suggested by Wolitzky & Fambrough process across the plasma membrane. To my # (1986) and is supported by the finding that exposure knowledge, stimulation of Ca + entry represents the of chicken muscle cells to veratridine (which in- fastest known response to physiological concen- creases Na+ leak by opening Na+ channels) results in trations of thyroid hormones, being evident within an increase in both the activity of and the mRNA for 15 s of in vitro T3 exposure in thymocytes (Segal & Na+,K+-ATPase (Taormina & Fambrough, 1990). Ingbar, 1984) and heart tissue where the effect was Erythrocytes from hyperthyroid humans have evident at 10 p (Segal, 1990a). It has been proposed # both an altered membrane phospholipid profile and that this Ca + influx is the first messenger in several # show an increased Na+ and Ca + influx compared to thyroid hormone effects at the plasma membrane euthyroid controls (De Riva, Virgili & Frigato, (for review see Segal, 1989b, 1990b). This thyroid- # 1996). Whilst there were no significant changes in hormone-induced increase in intracellular Ca + either phospholipid profile or ion fluxes 30 days after concentration has also been measured in individual commencement of treatment (plasma thyroid hor- rat myocytes (Lomax et al., 1991). Such a rapid mone levels were reduced but still in the hyper- effect is obviously non-genomic in origin but whether thyroid range), 300 days after treatment commenced it is associated with thyroid-hormone-induced both phospholipid profile and ion fluxes had re- changes in the physical properties of the bilayer is turned to euthyroid levels. Because mammalian not known. It is a transient response with in- # erythrocytes have no intracellular organelles it is tracellular [Ca +] being rapidly returned to normal likely that individual cells are unable to modify their levels. # membrane composition and it is necessary for The plasma membrane Ca +-ATPase has also population turnover to occur for changes to be been shown to be stimulated by in vitro thyroid manifest (De Riva et al., 1996). hormone exposure in both rat erythrocytes (Galo, # The Ca + gradient across the plasma membrane of Unates & Farias, 1981), human erythrocytes (Davis resting cells is the largest known for any ion (being et al., 1983) and rabbit myocardial vesicles (Rud- % #+ approximately i10 ) and is maintained by a Ca - inger et al., 1984). Erythrocytes isolated from # ATPase and a Ca +\Na+ exchanger, both in the hypothyroid and hyperthyroid humans have similar # # plasma membrane, as well as another Ca +-ATPase changes in Ca +-ATPase activities (Dube et al., # in the endoplasmic reticulum and a Ca + carrier in 1986) demonstrating in vivo relevance. # the mitochondrial membrane. T3 rapidly stimulates Plasma membrane Ca +-ATPase activity is known # Ca + uptake by the perfused rat liver (Hummerich & to be affected by membrane lipids and the in- # Soboll, 1989). Ca + cycling has been most studied in teraction between the regulatory calmodulin and the muscle (for review see Clausen et al., 1991). Thyroid pump is influenced by its lipid environment (for hormones have a greater effect on slow muscle than review see Carafoli, 1991). For example, acidic fast muscle and have been shown to result in an phospholipids and unsaturated fatty acids are re- increased amount of sarcoplasmic reticulum, a ported to mimic the stimulatory effect of calmodulin # greater density of Na+ pumps and a decreased on purified erythrocyte Ca +-ATPase (Niggli, # energetic efficiency of Ca + pumping. The relative Adunyah & Carafoli, 1981). Similarly, the in vitro contribution to the BMR depends on the degree of modification of membrane acyl composition of # futile Ca + cycling that occurs in the resting state. It intestinal brush border vesicles has been shown to # is estimated that approximately 30% of the change influence Ca + transport (Kreutter, Lafreniere & in muscle metabolic rate from the hypothyroid to Rasmussen, 1984). # euthyroid state is due to Ca + cycling changes in the In human erythrocyte membranes, long chain # sarcoplasmic reticulum (Clausen et al., 1991) and fatty acids have been shown to modulate Ca +- # that Ca +-cycling is responsible for approximately ATPase activity (Davis et al., 1987) and retinoic acid 6% of the resting metabolism of skeletal muscle, has been shown to be a modulator of the thyroid 18–36% in contracting cardiac muscle and approx- hormone activation of this enzyme (Smith et al., # # imately 2% in liver. It is estimated that this Ca +- 1989). The plasma membrane Ca +-ATPase has cycling in skeletal muscle, heart and liver accounts been most studied in mammalian erythrocytes and it for 3–4% of the BMR of a rat and possibly 4–6% in is obvious that these thyroid hormone effects can not humans (Rolfe & Brown, 1997). be mediated via the nucleus in such non-nucleated # The very high Ca + gradient across the plasma cells. They appear likely to be due to a thyroid 566 A. J. Hulbert influence on the membrane and there is no apparent an effect may be mediated by T3 binding to a reason why such an effect should not also be manifest plasma membrane receptor. in nucleated cells. This influence of thyroid hormones Thyroid-hormone-stimulation of glucose uptake has been reviewed by Davis, Davis & Lawrence by a rat liver cell line (ARL 15 cells) appears to (1989) and these authors have suggested that the involve an increase in both transporter numbers and mechanism of this thyroid hormone effect may the molecular activity of individual GLUT 1 glucose involve membrane lipids. The relative potency of transporters. After 6 hours of hormone exposure various iodothyronines is also different to that for deoxyglucose uptake increased by 40% but GLUT nuclear-receptor-mediated effects in that -T4 is the 1 transporter protein levels did not change. After 48 most potent iodothyronine, with -T3 and 3,5--T2 h of T3 exposure, deoxyglucose uptake increased by having approximately 75% the potency of -T4, 116 % whilst transporter numbers were only 58% and rT3, D-T4 and D-T3 having no stimulatory greater (Weinstein & Haber, 1993). In the rat heart, # effect on the activity of human erythrocyte Ca +- hyperthyroidism has no effect on either GLUT 1 ATPase. protein or its mRNA levels, whilst hypothyroidism The study on rat erythrocytes is interesting in that results in an increase in both GLUT 1 mRNA and these researchers have demonstrated that diet- protein levels (Weinstein & Haber, 1992). The same induced changes in membrane fatty acid com- study showed that, in the rat heart, GLUT 4 mRNA # position influence both basal Ca +-ATPase activity levels are very sensitive to, but that GLUT 4 protein and the response to both T3 and T4 (Galo et al., levels do not change with thyroid status. In rat 1981). adipocytes, hyperthyroidism resulted in an increase Thyroid hormone stimulation of metabolism must in levels of GLUT 4 transporters as well as the also result in increased substrate utilization. In the functional activity of these transporters. The change 1960s, in vitro thyroid hormone exposure was shown in functional activity of the plasma membrane to stimulate amino acid uptake in chick embryo glucose transporters may be related to an increase in bone (Adamson & Ingbar, 1967) and similar effects insulin receptor affinity and insulin receptor kinase were observed in rat thymocytes (Goldfine et al., activity during hypothyroidism (Matthaei et al., 1976). T3 was also shown to stimulate deoxyglucose 1995). Whilst astrocytes cultured in T3-depleted uptake in chick embryo heart cells (Segal, Schwartz media have a decreased number of glucose trans- & Gordon, 1977) and in rat thymocytes, both in vitro porters relative to those cultured in T3-enriched (Segal & Ingbar, 1979) and in vivo (Segal & Ingbar, media, short exposure (10 min) of the hypothyroid 1984). The speed of the response together with the cells to T3 increased the access of the transporters to lack of a requirement for protein synthesis demon- the binding molecules without a change in trans- strated it to be a direct effect on the plasma porter number (Roeder et al., 1988). Thyroxine membrane and not mediated by nuclear receptors. treatment has been shown to have only a minor Since then a number of studies have shown that influence on insulin effects on glucose kinetics in thyroid hormones facilitate substrate uptake across humans (Muller et al., 1995). the plasma membrane of a number of cell types. Thyroid hormones also stimulate the uptake of For example, the uptake of deoxyglucose by cardiac amino acids and the rapid and direct stimulation of and skeletal muscle and adipose tissue is stimulated gluconeogenesis by T3 in isolated-perfused rat liver in vivo by physiological doses of thyroid hormones has been suggested to be due in part to this thyroid (Segal, 1989a). It has been proposed that this hormone effect (Muller & Seitz, 1980). effect is mediated by intracellular increases in When rat hepatocytes are incubated with glucose # cyclic AMP and a transient Ca + influx (see Segal, there is an increased respiration compared to that 1990). observed in a glucose-free medium and this glucose- The stimulation of glucose uptake by chick enhancement of respiration is related to the prior embryo heart cells is biphasic with the initial phase thyroid status of the rat (Gregory et al., 1996). (up to 6 h) involving stimulation of transporters in Measured rates of glucose cycling were greater in the plasma membrane and the second phase (6–24 hepatocytes from euthyroid rats compared to those h) involving the nucleus and requiring protein from hypothyroid rats, however the contribution of synthesis. The production of a monoclonal antibody this enhanced hepatic glucose cycling to the calor- with the ‘configuration’ of T3 has been shown to igenic action of thyroid hormone is minimal stimulate the first phase of this uptake (Gordon et al., (Gregory et al., 1996). 1994). This finding supports the proposal that such Thyroid hormones are also known to stimulate the Thyroid hormones and their effects 567 synthesis of fatty acids (lipogenesis) as well as the euthyroid and hyperthyroid state (Deshpande & hydrolysis of triglycerides (lipolysis). The possibility Hulbert, 1995). That these diets had no effect on that a thyroid-hormone-stimulated lipogenesis\ metabolic rate but significantly influenced body lipolysis ‘futile’ cycle is a significant contributor to mass suggest that lipogenesis was similarly affected. the calorigenic effect has been investigated by Hyperthyroidism diminished the effect of dietary fat Oppenheimer et al. (1991). The early loss of body fat composition on liver malic enzyme activity (Desh- during induced hyperthyroidism in rats suggested pande & Hulbert, 1995). Interaction between that this was the primary substrate for the T3- dietary fat and T3 effects on malic enzyme induction induced increase in metabolic rate as food intake has also been observed in the rat (Clarke & only increased some days after the commencement of Hembree, 1990). The influence of dietary polyun- hyperthyroidism. The food the rats ate (normal saturate content on malic enzyme induction suggests laboratory chow) contained only 4.5% fat and the that energy storage may not be the only factor to be increase in lipogenesis reached a plateau 4–5 days considered but that another aspect of fatty acids, after the commencement of hyperthyroidism. How- their key role in the membrane bilayer, may also be ever, the metabolic cost of this increased lipogenesis important. Possibly, stimulation of lipogenesis is part accounted for 3–4% of the total T3-induced increase of a ‘make more membrane’ message. in metabolic rate. Thus, a futile cycle of lipogenesis\ This possibility is strengthened by the intriguing lipolysis is an insignificant contributor to the calor- findings of Castellani, Wilcox & Heimberg (1991). igenic effects of thyroid hormones. It was suggested These authors measured lipogenesis in perfused liver that the function of the thyroid-induced increase in from both euthyroid and hyperthyroid rats. Whilst lipogenesis was simply to maintain fat stores (Oppen- hyperthyroidism increased both fatty acid synthesis heimer et al., 1991). (lipogenesis) and oxidation, the de novo synthesised The effect of thyroid hormones on lipogenesis is fatty acids were preferentially incorporated into due to the increased expression of genes coding for phospholipids rather than into triglycerides (by lipogenic enzymes, especially the cytosolic malic severalfold) and were poorer substrates for oxidation enzyme and an enigmatic protein called S"% that than were exogenous fatty acids which were preferen- seemed to be associated with lipogenesis in an tially incorporated into triglycerides. These findings unknown way (Goodridge, 1978; Towle, Mariash & suggest that whilst increased lipogenesis will in- Oppenheimer, 1980; Miksicek & Towle, 1982). fluence triglyceride deposition the primary function Recent work suggests that S"% protein is located of enhanced lipogenesis may to provide fatty acids primarily in the nucleus and acts in the induction of for membrane synthesis. In this light, it is interesting mRNAs coding for key lipogenic, glycolytic and that one of the effects of an enhanced thyroid status gluconeogenic enzymes as well as for the D1 is an increase in the amount of cellular membranes, deiodinase (Brown, Maloney & Kinlaw, 1997). That for example increased mitochondrial (e.g. Jakovcic the thyroid hormone effect on lipogenesis is mediated et al., 1978) and sarcolemmal membranes (Szy- by nuclear receptors is demonstrated by the fact that manska, Pikula & Zborowski, 1991). the genes for these lipogenic enzymes (malic enzyme In some tissues, a substantial proportion of oxygen "% and S ) have TREs (Petty et al., 1990; Zilz, Murray is consumed by non-mitochondrial processes. For & Towle, 1990; Liu & Towle, 1994). Indeed, the example, in the rat this is estimated to be 20, 14 and malic enzyme TRE is a common TRE in the 3%, respectively, of the in vivo oxygen consumption construction of transfected systems used to examine of liver, skeletal muscle and heart and represents the mechanism of action of thyroid nuclear receptors. approximately 8% of BMR (see Rolfe & Brown, The induction of malic enzyme in liver is also 1997). Thyroid hormones stimulate non-mito- influenced by diet fat composition. In rats, fat-free chondrial oxygen consumption of rat hepatocytes diets result in elevated activities of malic enzyme and (Harper & Brand, 1993). One of the contributors to other lipogenic enzymes and as dietary fat content non-mitochondrial oxygen consumption will be the (polyunsaturated soya oil) is increased lipogenic desaturase enzyme systems located in the endo- enzyme activities decrease (Carrozza et al., 1979). plasmic reticulum (Pugh & Kates, 1979). It is An absence of polyunsaturates in the diet is a possible that the thyroid-induced changes in mem- powerful stimulus to malic enzyme induction. In brane fatty acyl composition (see Tables 2 and 3) are mice fed diets that differed only in fatty acid responsible for part (or all) of the thyroid-induced composition, polyunsaturates resulted in a significant changes in non-mitochondrial oxygen consumption decrease in liver malic enzyme activity both in the via altered desaturase activity. 568 A. J. Hulbert

Although when considered as whole organisms, substrate oxidation reactions and 3,5-T2 had no vertebrates are aerobic organisms, some individual effect on the proton leak nor on the phosphorylating cell types can adopt anaerobic strategies. Approx- system (Lombardi et al., 1998). Both 3,5-T2 and imately half the heat production of lymphocytes 3,3h-T2 injected into hypothyroid rats stimulated measured in vitro is associated with anaerobic energy resting metabolism (Lanni et al., 1996). The latter metabolism. Lymphocytes from hyperthyroid hu- study also showed that the relative influence of these mans have an elevated heat production due to iodothyronines on cytochrome oxidase activity enhanced anaerobic metabolism (Valdermarsson & varied between tissues. Hypothyroid rats responded Monti, 1994) that in turn appears to be due to to a single injection of either 3,5-T2 or T3 with an thyroid hormone stimulation of adrenoreceptor increase in resting metabolic rate. Although meta- sensitivity (Valdermarsson & Monti, 1995). bolic rate increased to the same degree the response As mentioned above it is often assumed that T3 is to 3,5-T2 was more rapid and resistant to inhibition the ‘active’ thyroid hormone and a large amount of of protein synthesis, whilst the response to T3 was the research cited in this section has involved the use slower and completely abolished by the inhibition of of T3. A decade ago, a surprising report appeared protein synthesis (Moreno et al., 1997). Supra- that fundamentally questioned this assumption. physiological levels of 3,5-T2 in vitro stimulated the Horst, Rokos & Seitz (1989) reported that 3,5-T2, oxygen consumption of human mononuclear blood at a concentration as low as 1 p, resulted in a rapid cells but glucose uptake was unaffected (Kvetny, stimulation (within 30 min) of oxygen consumption 1992). of the perfused liver from hypothyroid rats. T2 was The study giving the most insight into the effects as potent as T3 and exerted its effect more rapidly. of this iodothyronine on mitochondrial activity has Whereas the stimulation by T3 was largely abolished recently shown that in vitro 3,5-T2 stimulates in the presence of an inhibitor of the D1 deiodinase, cytochrome oxidase activity. Part of the normal the 3,5-T2 stimulation of oxygen consumption was control of mitochondrial respiratory chain activity is not affected. Only T3 increased liver malic enzyme due to the fact that at high ATP\ADP ratios, ATP activity with this enzyme being unaffected by 3,5-T2 replaces ADP bound to subunit IV of the cytochrome (Horst et al., 1989). The interpretation of this finding oxidase enzyme complex and allosterically inhibits was that the calorigenic effect of the thyroid its activity. 3,5-T2 was shown to bind to subunit Va hormones was mediated by a direct and rapid effect of the cytochrome oxidase complex and abolish the of 3,5-T2 (derived from the intracellular deiodin- allosteric inhibition due to ATP binding and thus ation of T3) on the mitochondria whilst other longer stimulating enzyme activity (Arnold, Goglia & term effects (such as increased malic enzyme Kadenbach, 1998). T3 has a similar but smaller activity) were mediated via the nucleus. effect. This effect was first apparent at 10 n 3,5-T2 Since this initial report, both 3,5-T2 and 3,3h-T2 and was maximal at 1 µ, although because of the have been shown to be equipotent in stimulating constructed in vitro nature of this system these mitochondrial cytochrome oxidase activity in vitro; concentrations may not have in vivo relevance. the effects are rapid (being evident within 5 min) These studies obviously raise a number of im- and are restricted to these two iodothyronines (Lanni portant questions. They show that: (i) T3 is not the et al., 1994a). When large doses of 3,5-T2 or T3 were only active iodothyronine and that at least two T2s injected into anaesthetized rats that had been treated are also capable of being active thyroid hormones, with the D1 deiodinase inhibitor, propylthiouracil (ii) the T2s can exert effects separate from T3 and on (which will prevent the conversion of T3 to T2), and a molar basis are approximately as equipotent as T3 mitochondria were isolated 1 h later, only 3,5-T2 in stimulating some activities, and (iii) that this injections increased in mitochondrial activity stimulation does not involve the nucleus. The later (O’Reilly & Murphy, 1992). In unanaesthetized studies, however, suffer from the same general rats, 3,5-T2 was also shown to stimulate lipid β- problem that plagues much of the thyroid literature. oxidation (Cimmino et al., 1996). Rat liver mito- The concentrations used are generally too high to chondria have been shown to bind both 3,5-T2 conclude physiological relevance. We currently have (Goglia et al., 1994) and 3,3h-T2 (Lanni et al., 1994b) limited knowledge of the concentrations of these in vitro. Recently, it has been demonstrated that 3,5- diiodothyronines in humans but very little knowl- T2 given as a single large in vivo dose resulted in a edge of the levels in the blood and tissues of rats. stimulation of both state 4 and state 3 respiration of Whilst I know of no reports for free 3,5-T2 rat liver mitochondria; this was due to stimulation of concentrations, the free levels for 3,3h-T2, 3h,5h-T2 Thyroid hormones and their effects 569 and T3 are, respectively, 0.44 p, 0.77 p and 4.8 vast majority of these processes are membrane- p (Faber et al., 1984). For a 70 kg euthyroid associated. Many involve an increase in the passage human, the daily production of 3,5-T2 and T3 is 6.4 (both active and passive) of ions or substrates across nmol and 48 nmol, respectively (Chopra, 1996). membranes, and many involve an increase in the Tissue concentrations of 3,5-T2 and T3 in euthyroid molecular activity of particular membrane-asso- human brains are reported to be approximately 0.1 ciated processes. Some are rapidly induced in vitro " " pmol g− and approximately 1.5 pmol g− respect- whilst some take time to become manifest. Some ively (Pinna et al., 1997). Whilst the levels in adult involve an increase in protein synthesis whilst others euthyroid rats are not known, tissue 3,5-T2 and T3 do not. content of liver from 16-day-old rat fetuses is The evidence that these effects are initiated by " reported to be approximately 0.4 pmol g− and thyroid nuclear receptors is scarce. The only process " approximately 3.5 pmol g− , respectively (Porter- that has definitively been shown to be thyroid field & Hendrich, 1992). Thus, all reports suggest nuclear receptor initiated is the stimulus of lipo- that the normal in vivo levels of 3,5-T2 are less than genesis and this effect is not quantitatively im- those of T3 and suggest that the doses used to date portant. Indeed, a new interpretation of this thyroid are an order of magnitude greater than those hormone stimulation of lipogenesis is that it may be required to demonstrate physiological relevance. part of a ‘make more membrane’ message. The fact Future studies ideally need to include the use of and that membrane lipid composition is altered by confirmatory measurement of the relevant in vivo thyroid status, and that these changes have been concentrations. observed for virtually all sub-cellular membranes These studies may, however, provide an expla- and in a wide variety of tissues, when coupled with nation for the conflicting findings of a direct effect of the finding that membrane composition may be a T3 on isolated mitochondria, in that if mitochondrial ‘pacemaker’ in determining the BMR of different preparations have undetected microsomal contami- vertebrates (see Hulbert & Else, 1999, 2000) suggests nation it is possible that T3 added in vitro could be that thyroid-induced changes in the composition of converted to 3,5-T2 by deiodination and exert membrane bilayers are a significant mechanism in stimulatory effects on mitochondrial respiration. thyroid hormone stimulation of metabolic activity. That thyroid hormones can have rapid in vivo The proposal that thyroid hormones directly effects on metabolic enzymes is illustrated by a series influence bilayer physical properties and that conse- of reports on liver and muscle from a diverse quent changes in bilayer acyl composition are collection of non-terrestrial vertebrates, including a primarily responsible for mediating the calorigenic cyclostome (Leary et al., 1997), holostean and teleost effect of thyroid hormones is compatible with many fish (Ballantyne et al., 1992) and an elasmobranch observations. It is compatible with the time lag (Battersby, McFarlane & Ballantyne, 1996). In all observed before there is an increase in organismal of these studies, tissues were isolated 3 h after in vivo oxygen consumption. Since bilayer acyl changes are injection of T3 and although T3 doses were generally mediated by various enzymes (especially those in the range 1–15 µg per 100 g, in some studies effects involved in deacylation\reacylation of phospholipids were observed after in vivo T3 doses as low as 0.001 as well as phosholipid synthesis) it is compatible with µg per 100 g. These researchers have shown a the effect of inhibitors of protein synthesis. It is also stimulation of pyruvate-fuelled state 3 mitochondrial compatible with the many studies that demonstrate respiration following a 5 min in vitro incubation with that membrane acyl composition influences the 0.3 n T3 or 3,5-T2 (Leary, Barton & Ballantyne, activity of various membrane proteins, as well as 1996). being compatible with the quantitative importance Although the effect of the thyroid hormones on of the maintenance of transmembrane gradients in basal metabolism has been known for over a century cellular energy consumption. Whilst the finding that there is still no general agreement on how this effect there is an increase in mRNA for a particular protein is initiated or mediated. Part of the reason for this is is evidence for the involvement of the nucleus in the that, until recently, we have not had much quan- effect, it is not evidence that the effect is initiated in titative understanding of the processes that constitute the nucleus by thyroid nuclear receptors. In view of basal metabolism. It is obvious from the discussion the concept of continual degradation\synthesis of above that the stimulation of oxygen consumption of proteins, an increase in protein synthesis may be due the whole organism is due to the stimulation of many to an increased information flow to the nucleus for individual processes in many different tissues. The the particular protein. However, not all effects can 570 A. J. Hulbert be mediated by changes in acyl composition. For been studied intensively in the rat and has been example, some are too rapid. The stimulation of shown to be mediated by an uncoupling protein, increased substrate uptake at the plasma membrane UCP1, that acts to uncouple mitochondrial res- appears to be receptor mediated and some effects at piration from ATP production by facilitating proton the mitochondrial membrane may also be due to a flux through the BAT mitochondrial inner mem- direct interaction with proteins in the mitochondrial brane (for a review see Nicholls & Locke, 1984). The membrane. selective advantage that favoured BAT thermo- The proposal that much of the stimulation of genesis in small eutherian mammals is presumably metabolic activity by thyroid hormones is mediated the fact that, unlike the tremor associated with by changes in membrane acyl composition is unusual severe shivering, this form of heat production in the in that it does not require a specific receptor and cold does not disturb the insulative air layer around therefore is not part of the current paradigm the animal and also allows muscle use for normal concerning hormone action. It is instead a response movement in the cold. Because of their size and of the cell to the physical properties of thyroid consequently high relative surface area, small mam- hormone. Although we know that thyroid hormones, mals cannot use the more energy-efficient strategy of (at physiological concentrations) have direct physi- increasing body insulation in the cold. Whilst BAT is cal effects on the behaviour of membranes, our an important energetic tissue in cold-adapted rats, knowledge of this area is still very rudimentary. the extrapolation of findings in the adult rat to other As well as effects on BMR, thyroid hormones have adult mammals (including humans) is generally not other effects on thermogenesis. For example, thermo- warranted. regulation is also influenced by thyroid status. The The role of the thyroid hormones in BAT lower resting body temperature of some hypothyroid thermogenesis has been reviewed by Silva (1995). mammals compared to euthyroid controls (e.g. Brown adipose tissue has a high amount of D2 Kaciuba-Uscilko et al., 1987) is often thought to be deiodinase and during cold exposure deiodinase due to a decreased metabolism and therefore activity is activated by norepinephrine resulting in thermogenesis. However, recent work using the substantial T3 production in the BAT cells. The laboratory rat suggests that rather than being a elevated T3 concentration appears to facilitate, via passive consequence of decreased thermogenesis this nuclear TRs, the expression of the BAT UCP1 gene. lower body temperature is instead a regulated Two TREs have been described for the rat UCP1 response (Yang & Gordon, 1997) and thus may be gene (Rabelo et al., 1995) and the interaction the result of a more direct effect on the neuronal between these TREs has been examined (Rabelo et system that determines the set point for body al., 1996). temperature regulation. UCP1 is a member of the mitochondrial mem- Exposure of endothermic vertebrates (mammals brane anion carrier family, and other proteins, and birds) to cold generally results in an increased related to UCP1, have recently been identified. thermogenesis to maintain body temperature homeo- Whilst these proteins have been called UCP2 and statically. The initial mechanism used by all en- UCP3 because of their relatedness to UCP1, there is dothermic vertebrates is the controlled microtremor little direct evidence that they are indeed mito- of antagonistic muscles, known as shivering. Some of chondrial uncoupling proteins. Both UCP2 and the small eutherian mammals (notably rodents) UCP3 mRNA can be induced by thyroid hormones have also evolved a specialized heat-production in some tissues of rats (Gong et al., 1997; Lanni et al., mechanism that is located in brown adipose tissue 1997) and it has been suggested that they have a (BAT) and controlled by the secretion of noradren- thermogenic function. Neither UCP2 nor UCP3 is aline from the sympathetic nervous system. Heat present in hepatocytes and can thus not be directly production is the only known function of BAT and responsible for the mitochondrial proton leak mea- this form of heat production is important in three sured in these cells. A thermogenic role for UCP2 situations: cold-adaptation in small eutherian mam- and UCP3 has yet to be demonstrated. They have mals, heat production in eutherian neonates and recently been discussed by Brand et al. (1999). during arousal from hibernation in small eutherian During acute cold exposure in many endotherms mammals. BAT thermogenesis is not present in there is an increase in both thyroid activity and monotreme, marsupial or medium to large-sized thyroid hormone metabolism. However, because eutherian mammals (e.g. see Hulbert, 1980a) and is cold exposure also often involves an increase in also absent in non-mammalian vertebrates. It has energy intake, it is not always obvious whether it is Thyroid hormones and their effects 571 cold exposure or the enhanced energy intake that is application. Concomitant with these effects, cells the primary influence on thyroid hormone. A series spontaneously firing action potentials also stopped of studies on young pigs has shown that elevated firing within minutes of T3 application (du Pont & energy intake has a greater effect on thyroid Israel, 1987). Other evidence of membrane effects is hormone metabolism than decreases in ambient the finding that cerebral cortex slices from hypo- temperature (for a review see Dauncey, 1990). thyroid mice show diminished endocytosis and uptake of amino acids and hexose compared to those from euthyroid mice. The fact that T3 can increase (5) Effects on excitable tissues these activities in nerve endings within 5 min of Before the advent of modern accurate hormone hormone exposure demonstrates that this stimu- assays, the Achilles tendon reflex time (ankle jerk lation does not involve the nuclear receptors (Iqbal, reflex) was proposed as a diagnostic for thyroid Koenig & Trout, 1984). status in humans although it was later found to be Both hypothyroidism and hyperthyroidism have unsatisfactory (Costin, Kaplan & Ling, 1970). The detrimental effects on skeletal muscle. Thyroid- Achilles tendon reflex time averaged 0.34 s in ectomy affects the neuromuscular junction in rats, euthyroid humans and 0.53 s in hypothyroids, with resulting in a decrease in both the number of the time between the tap on the tendon and the acetylcholine receptors (Kragie & Smiehorowski, muscle action potential being 0.03–0.04 s for hypo- 1993) and G4 isoform acetylcholinesterase activity in thyroid, euthyroid and thyrotoxic individuals (De- fast skeletal muscle (Kragie & Stock, 1994). How Long, 1996). This finding suggested that there was these changes are initiated is not known but it was no effect of thyroid hormones on nerve conduction suggested that they may involve changes in muscle velocity but that there were significant thyroid activity and\or alterations in the signal transduction influences on the processes involved in muscular systems regulating the G4 acetylcholinesterase iso- contraction. Measurement of peripheral and central form (Kragie & Stock, 1994). nerve conduction velocity in humans showed that Thyroid hormones affect transmembrane fluxes of hypothyroidism resulted in slower peripheral and both Na+ and Ca+ in skeletal muscle and thus have central nerve conduction velocity than in euthyroids dramatic effects on skeletal muscle function (for and that this difference was eliminated after T4 review see Everts, 1996). Thyroid status in rats does treatment, but there was no difference between not affect the peak force developed during either a thyrotoxic and euthyroid individuals (Abbott et al., single twitch or tetanus but has dramatic effects on 1983). These researchers showed that the slow the rate of force development and relaxation time of conduction of the hypothyroids was largely due to isolated muscles; this effect is more pronounced in the lower temperatures associated with hypothyroid- slow muscles than in fast muscles. In the rat soleus, ism. Central nerve conduction velocity was assessed such thyroid-hormone-induced changes take several by measuring the latency of visually evoked respon- days (Montgomery, 1992). ses. This technique has also been used to examine the The depolarization phase of action potentials at effects of thyroid hormones on the visual system of the sarcolemma is mediated by voltage-dependent rats. Whilst thyroidectomy resulted in an increased Na+ channels. Thyroid hormones increase both the latency of the response, there was no significant spontaneous electrical activity and the number of difference in the optic nerve conduction velocity. Na+ channels in cultured skeletal myotubes as well as The component of the rat visual system most sensitive influencing the affinity of these Na+ channels for the to thyroid status was located in front of the optic neurotoxin, saxitoxin (Brodie & Sampson, 1989). chiasma, probably in the retina (Takeda, Onoda & Thyroid hormones also increase Na+,K+-ATPase Suzuki, 1994). In view of the relative constancy of activity in skeletal muscle (see Clausen et al., 1991). brain T3 levels, irrespective of changes in organismal It has been suggested that the Na+ channels: Na+ thyroid status, it is understandable that conduction pumps ratio is an important determinant of the velocity is unaffected by thyroidectomy. Whether contractile endurance and rate of force recovery in peripheral nerves also maintain T3 homeostasis is muscle (Harrison, Nielsen & Clausen, 1997). In rat unknown. soleus muscle, T3 treatment increases the number of T3 has a direct effect on the cell membrane of Na+ channels and this precedes the increase in the GH3 cells in culture resulting in significant increases number of Na+ pumps; changes in contractile in membrane resistance and a hyperpolarization of endurance of the isolated muscle reflect the time the membrane potential within minutes of hormone course of changes in the leak\pump ratio. Such 572 A. J. Hulbert changes may be important determinants of the the inhibition of protein synthesis) does not necess- muscular fatigue associated with hyperthyroidism arily indicate a nuclear receptor mode of action. It (Harrison & Clausen, 1998). does imply that the nucleus is involved, but because The sarcolemmal action potential initiates transfer a thyroid effect is pre-translational does not necess- # of Ca + from the sarcoplasmic reticulum into the arily mean it is transcriptional. It may involve a pre- # cytosol via membrane-located Ca + channels. In transcriptional site of action. For example, the cultured skeletal muscle cells, T3 increases the increased manufacture of Na+ pumps, in response to # number of Ca + channels and influences their an increased Na+ leak, is due to a pre-transcriptional binding affinity for a specific channel antagonist increase in information flow to the nucleus that leads # (Brodie & Sampson, 1990). The Ca + released to increased transcription and a consequent increase during excitation is rapidly removed from the in the number of pumps. In this light, it is of interest # # sarcoplasm by a Ca +-ATPase located in the sarco- that cytosolic Ca + has been shown to be important # plasmic reticulum. The concentration of Ca +- in the control of the amount of acetylcholine ATPase is approximately six times higher in fast receptors and acetylcholinesterase (Birnbaum, Reis muscle fibres than slow muscle fibres and reflects & Shainberg, 1980) and voltage-sensitive Na+ their difference in speed of relaxation following channels (Sherman & Catterall, 1984). All three contraction (see Everts, 1996). Following T3 treat- effects are responsive to thyroid status (see above). ment, there is an increase in the rate of force Brodie & Sampson (1989) showed that factors that # # relaxation after a sustained contraction in slow influence intracellular Ca + levels (e.g. external Ca + muscle, correlated with an increase in the density of concentration and channel blockers) modified the # Ca +-ATPase in the sarcoplasmic reticulum (Dul- thyroid hormone effect on the number of Na+ hunty, 1990). The latter study showed that T3 channels. It is possible that other thyroid hormone effects on fast muscle fibres were small and insig- effects are mediated by such pre-transcriptional nificant, and furthermore that T3 treatment results changes in information flow. Whether the thyroid- in more rapid changes in factors regulating the rate hormone-induced changes in skeletal muscle sarco- of rise in tension than those affecting tension lemmal lipid composition and fluidity (e.g. Pilarska relaxation. et al., 1991) will affect such information flow is not Three genes encode different isoforms of the currently known. # sarcoplasmic reticulum Ca +-ATPase and in fast Exposure of primary cultures of vascular smooth muscle fibres SERCA1 is the predominant isoform muscle cells to T3 results in cellular relaxation whilst in slow skeletal muscle, SERCA2a is the main within 10 min of exposure (Ojamaa, Klemperer & isoform expressed. T3 specifically stimulates the Klein, 1996). The speed of the response suggests that production of SERCA1 isoform mRNA (Simonides, this is non-genomic. A decrease in vascular re- van der Linden & van Hardeveld, 1990, Sayen, sistance, which is compatible with this in vitro effect, Rohrer & Dillmann, 1992) resulting in increased is part of the enhanced cardiovascular hemo- # Ca +-ATPase activity and favouring the transform- dynamics associated with thyroid hormones in vivo. ation of slow muscle to the fast muscle phenotype. The association between the thyroid gland and The demonstration of TREs in the promoter region the heart was recognized over 200 years ago (Caleb, of the SERCA1 gene in a transfected system 1785, cited by Dillman, 1990). Indeed, cardiac demonstrates that this effect is mediated by thyroid hypertrophy is often used as an indication of nuclear receptors (Simonides et al., 1996). The hyperthyroidism. The in vivo effects of thyroid demonstration that thyroid hormones stimulate hormones on the heart are complex in that they # sarcoplasmic reticulum Ca +-ATPase activity in vitro involve a combination of direct hormone effects and suggests that non-genomic effects are also possible indirect consequences from peripheral hemodynamic (Warnick et al., 1993). Neither mode of thyroid changes (Klein, 1990). The use of heterotopic hormone action excludes the other. cardiac transplants, in which the transplanted heart The mode of thyroid hormone action involved in muscle receives blood from the recipient but is not the increase in skeletal muscle Na+ channels (Brodie hemodynamically loaded (in that it does not pump # & Sampson, 1989) and Ca + channels (Brodie & the recipients blood), has shown that cardiac Sampson, 1990) is unknown. Whilst thyroid hor- hypertrophy is not a direct thyroid hormone effect. mone effects observed during inhibition of protein For this reason among others, many effects have synthesis preclude a nuclear-receptor-initiated mode been studied in isolated cardiac myocytes. of action, the converse finding (i.e. no effect during As well as enlargement of the heart, changes in Thyroid hormones and their effects 573 thyroid status produce alterations in the electrical effects may mediate thyroid-hormone-induced ef- activity of the heart. Tachycardia (as well as fects on cardiac contractility (Wolska et al., 1997). arrhythmia) is a common sign of hyperthyroidism, Thyroid hormones also affect the contractile just as bradycardia is often observed during hypo- processes of cardiac muscle. In the heart, the # thyroidism. The duration of the cardiac action SERCA2a isoform of Ca +-ATPase predominates potential in isolated heart papillary muscle has been and T3 increases expression of this gene (Rohrer & shown to be inversely related to in vivo thyroid status Dillman, 1988). Three TREs have been described in in a variety of vertebrates: rats (Di Meo et al., 1997), the promoter region of the SERCA2 gene in the rat chickens (Di Meo et al., 1993), lizards (Venditti et al., (Hartong et al., 1994) and it has been suggested that 1996) and frogs (Di Meo et al., 1995). Resting heart this effect may be specific to TRβ-1 interacting with rate is directly related to in vivo thyroid status; the a myocyte-specific enhancing factor (Moriscot et al., use of blockers of the sympathetic and parasympa- 1997). Thyroid hormones appear also to have direct # thetic nervous systems has shown that changes in the effects on the sarcolemmal Ca +-ATPase in that very autonomic nervous system, whilst present, are not low concentrations of both T3 and T4 result in in the primary cause of thyroid-hormone-induced vitro stimulation of the sarcolemmal enzyme isolated tachycardia, and that the intrinsic heart rate is also from rabbits (Rudinger et al., 1984). Thyroid significantly increased (Di Meo et al., 1994). That hormones have been shown to have an acute thyroid-hormone-induced changes in the electrical transient effect on the contractile abilities of the activity of the heart may be associated with isolated rat heart that is plasma membrane mediated # membrane lipid changes is supported by the finding and Ca + dependent (Segal et al., 1996). Acute that the membrane antioxidant, vitamin E, atten- extranuclear actions of thyroid hormones on the uates such changes in the rat (Venditti, De Leo & Di heart have been reviewed by Davis & Davis (1993). Meo, 1997b). Both hypothyroidism and hyper- As well as influencing cardiac contraction by thyroidism have been shown to change the phos- affecting the proteins that regulate intracellular # pholipid fatty acyl composition of the sarcolemma of Ca + levels, thyroid hormones also influence the rabbit cardiac muscle (Szymanska et al., 1991). contractile proteins themselves. Myosin hydrolyses Studies on thyroid-hormone-induced shortening ATP and converts the chemical energy released into of action potential duration in the rabbit heart have mechanical movement. Each hexameric myosin suggested that changes in Na+ pump activity and molecule contains four light chains and two heavy intracellular Na+ levels can explain the changes in chains (MHCs). Myosin isolated from heart ventricle action potential duration (Doohan, Hool & Ras- was found to contain three isoforms that differed mussen, 1995). Membrane channel currents have only in their heavy chains (Hoh, McGrath & Hale, been shown to be influenced by thyroid hormones. 1978). The V" isoform consists of two αMHCs, the In neonatal rat cardiac myocytes, whole-cell-vol- V# consists of a single α combined with a single tage-clamp studies have shown that acute exposure βMHC, whilst the V$ isoform has two βMHCs. The to T3 promotes the slow inactivation of Na+ currents isoforms differ in their ATPase activity with the within 1–5 min after hormone addition (Craelius, αMHC having the higher activity. The control of Green & Harris, 1990; Harris, Green & Craelius, MHCs in the heart has been reviewed by Morkin 1991) which suggests a direct action on the cell (1993). Thyroid hormones induce the high-activity membrane. In rabbit cardiac myocytes, patch-clamp V" isoform (i.e. αMHC) and repress expression of studies have shown that acute exposure of the the low-activity V$ isoform (i.e. βMHC) in the extracellular side of the membrane to T3 modulated ventricle of the rat and the rabbit. This effect is the cardiac Na+ channels by increasing their pro- mediated by nuclear receptors and TREs have been pensity to enter a gating mode characterized by described for the promoter regions of both the ‘long events’ or ‘bursts’ (Dudley & Baumgarten, αMHC and βMHC genes in both the rat and the 1993). Thyroid hormones can affect the behaviour of human. The time course of in vivo thyroid hormone other membrane ion channels. For example, T3 effects on MHC isoform mRNA levels in the rat increases the ATP sensitivity of the ATP-dependent heart shows that small changes are evident after 12 K+ channel from rat heart (Light et al., 1998). h and complete by 72 h after hormone admin- Increased thyroid status is correlated with increased istration, a time course similar to that for mRNA # intracellular Na+ concentrations in rat ventricular observed for the sarcoplasmic reticulum Ca +- myocytes that is associated with increased intra- ATPase (Balkman, Ojamaa & Klein, 1992). Thyroid cellular acidity and it has been suggested that these hormones also affect the expression of MHC genes in 574 A. J. Hulbert rat skeletal muscles where the effects are highly tissue transected peripheral nerves has been shown to be and developmental stage specific (Izumo, Nadal- enhanced by local administration of thyroid hor- Ginard & Mahdavi, 1986). mone (Voinesco et al., 1998). Several other factors have been shown to influence MHC isoform expression in the heart including (6) Effects on growth pressure-volume overload which promotes βMHC and decreases αMHC expression. Although cardiac Hypothyroidism results in pronounced growth re- ventricle MHC isoforms are very responsive to in vivo tardation in both endothermic vertebrates, for manipulation of thyroid status in small mammals example, humans (Snyder, 1996), rats (Evans et al., like the rat, this is not true for larger mammals. For 1966), and possums (Buaboocha & Gemmell, 1996), example, in the euthyroid baboon, dog and calf as well as in ectothermic vertebrates such as lizards αMHC was not detectable in the heart ventricles, (Gerwien & John-Alder, 1992), turtles (Denver & and following 6 weeks of high daily doses of T4, Licht, 1991) and fish (Matty, 1985). In mammals, αMHC was not detected in the dog ventricle, was this growth retardation is the result of both reduced 17% in the baboon ventricle and 40% in the calf secretion of growth hormone (GH) from the anterior ventricle. In the euthyroid human, βMHC is the pituitary as well as impaired peripheral growth predominant form found in the ventricular myo- hormone action (Snyder, 1996) and non-GH-related cardium. Little is known concerning the thyroidal effects. Hypothyroidism results in lowered regulation of cardiac MHC genes in the human, but GHmRNA and GH content in the anterior pituitary it is unlikely to be the major mechanism underlying of rats, and a reduced serum GH concentration the inotropic action of the thyroid hormones (see (Samuels et al., 1989). In hypothyroid young Morkin, 1993). humans, the nocturnal secretion of GH is dra- Thyroid hormones also influence adrenergic recep- matically reduced (Chernausek & Turner, 1989). tors in cardiac muscle. In cultured ventricular The early cloning of the GH gene together with myocytes, they increase β1 receptor number via an the availability of rat-derived pituitary tumor cell increase in transcription of the β1 adrenergic cultures (GC and GH cell lines) allowed the first receptor gene (Bahouth, 1991). Thyroid status description of DNA sequences responsive to thyroid influences the action of catecholamines on cardiac receptors (Koenig et al., 1987). The rat growth tissue with hyperthyroidism potentiating and hypo- hormone gene has a number of positive TREs (in the thyroidism blunting catecholamine-sensitive res- upstream promoter region as well as in the third ponses. intron) and also a negative TRE (see Williams & Interestingly, the β-adrenergic blocker, propan- Brent, 1995). The GH response of GH3 cells to alol, has a favourable effect on many signs and thyroid hormones appears to be complex and symptoms associated with thyrotoxicosis. It de- thyroid-receptor specific, in that deletion of TRβ-1 creases plasma T3 levels and increases plasma rT3 results in an increase in cell TRβ-2 content and also concentration and this effect appears to be due to in the basal and T3-induced GH mRNA content in inhibition of 5h deiodination. This inhibition of these cells (Ball, Ikeda & Chin, 1997). Thyroid deiodination, with its consequent effects on plasma hormone control of the human GH gene is not as levels of T3 and rT3, is related to membrane well defined, in that whilst thyroid hormone in- stabilizing activity rather than β-blocking ability creases endogenous rat GH expression, it results in a (Wiersinga, 1991). decreased expression of transfected human GH gene The adult brain is often thought to be relatively (Cattini et al., 1986). Similarly, when the promoter non-responsive to thyroid status. A recent review, regions for both the human GH and bovine GH gene however, suggests that thyroid hormones influence were transfected into rat pituitary tumor cell cultures the expression of neuropeptides and growth factors together with a reporter gene, the promoters for rat in brain areas involved in cognitive processes in the and bovine GH were thyroid-hormone-responsive adult but the molecular mechanisms involved in whilst that for human GH was unresponsive to T3 these effects are unknown (Calza, Aloe & Giardino, (Brent et al., 1988). 1997). The spontaneous behavioural activity of adult That the relationship between thyroid status and rats has been shown to be influenced by T4 and it GH secretion in other vertebrates appears not to be was suggested that this effect was due to an enhanced as simple as it is in rats and cell cultures of rat sensitivity to noradrenaline (Emlen, Segal & Man- pituitary tumor cells is manifest from the observ- dell, 1972). In adult rats, the regeneration of ations of thyroidal inhibition of growth hormone Thyroid hormones and their effects 575 secretion in birds (Harvey, 1990). Age-related liferation in the growing rat is manifest by the changes in complex feedback mechanisms, as well as reduced tissue DNA content, which is most pro- appropriate experimental doses of hormones, need to nounced in those tissues undergoing proliferation be considered when examining the mechanisms of (Brasel & Winick, 1970). The stimulatory effects of thyroid hormone influence on growth in vertebrates. thyroid hormones on growth are not restricted to the High concentrations of thyroid hormones inhibit whole organism in that in vitro cell proliferation is growth in many tissues (Greenberg, Najjar & also stimulated by T4 in cell culture. Both T3 and Blizzard, 1974). T4 can stimulate cell growth and mitotic rate of GH3 As with some other thyroid hormone effects, 3,5,- cells in culture, as well as alter their morphological T2 has recently been shown to increase serum GH appearance (Kitagawa et al., 1987). In this respect, concentration in hypothyroid rats (Moreno et al., it is also of interest that cells in culture have a 1998). Whilst another diiodothyronine, 3,3h-T2, has requirement for polyunsaturated fatty acids for cell no influence on serum GH, the potency of 3,5-T2 division and that non-availability of polyunsaturates was similar to that of T3 in these experiments. As can lead to cessation of DNA synthesis and growth with the other reported effects of 3,5-T2, the (Holley, Baldwin & Kiernan, 1974; Hatten, Horwitz physiological relevance of these findings depends on & Burger, 1977; Doi et al., 1978). measurement of the plasma and tissue 3,5-T2 Whilst thyroid status influences GH secretion, concentrations. These results do however raise many of the anabolic and mitogenic effects of GH important questions regarding the uniqueness of T3 are mediated by insulin-like growth factor 1 (IGF-1) as the only active thyroid hormone, and regarding secreted by the liver into the blood in response to its mode of action, in view of the low affinity of GH. The low plasma GH concentration in hypo- nuclear receptors for T2. thyroid young humans is associated with decreased The amount of thyroid hormones required to plasma concentrations of IGF-1 and these are maintain normal growth in rats is relatively small. elevated following T4 treatment (Chernausek & For example, doses as low as 0.25 µgT4\day in Turner, 1989). In addition to hepatic production of thyroidectomised rats resulted in growth rates IGF-1, many other tissues also express IGF-1 mRNA similar to non-thyroidectomized controls (Evans et and the IGF-1 peptide, which is believed to have al., 1966). This dose is approximately 25–30% of the local functions (for review see LeRoith et al., 1995). normal daily T4 secretion rate in rats. Although this The actions of IGF-1 are mediated by the IGF-1 cell dose resulted in restoration of near normal growth surface receptors and are further modulated by a rates, it did not significantly stimulate the metabolic complex series of up to six non-receptor IGF binding- rate of these thyroidectomised rats which remained proteins (for review see Rechler, 1995). As well as considerably reduced compared to non-thyroid- affecting plasma IGF-1 concentrations, thyroid ectomized controls (Evans et al., 1966). This study status in humans also influences IGF-1 bioactivity also presented some fascinating results in that it and the concentrations of IGF-binding proteins demonstrated that daily injections of 5 mg iodide (Miell et al., 1993). were as effective in restoring normal growth as were Whilst many of the growth-retardation effects due T4 injections. This was proposed to be evidence for to hypothyroidism can be explained by the sec- extrathyroidal T4 formation. We have confirmed ondary reduction in hepatic IGF-1 secretion, the that daily injections of 0.25 µg T4 or 5 mg of sodium inability of GH administration alone to restore iodide were equivalent in restoring a normal growth normal growth to hypothyroid children (Greenberg rate to thyroidectomized rats, and extended the et al., 1974) shows that thyroid hormones also observations to show that both treatments resulted in stimulate growth by other pathways. This is also small increases in plasma T4 and T3 concentrations demonstrated by the fact that administration of GH (J.-A. Green & A. J. Hulbert, unpublished results). and T4 together has a greater effect on the reversal The plasma concentrations of both T4 and T3 of growth retardation of rats than does GH admin- following these two treatments, however, were istration alone (Burstein et al., 1979). Not all thyroid considerably lower than those in the euthyroid hormone effects on the IGF system in rats are controls and thus support the contention that normal secondary effects mediated by GH (Nanto-Salonen growth may not require normal thyroid hormone et al., 1993) and the administration of T4 alone to concentrations. Whether the iodide ion itself in- rats is capable of stimulating IGF-1 activity in the fluences growth is still unknown. absence of GH (Gaspard et al., 1978). Indeed, it has The influence of hypothyroidism on cell pro- been suggested that there is a feedback loop 576 A. J. Hulbert regulating peripheral thyroid hormone action in- other cells (e.g. see Table 3), then it is possible that volving the GH-IGF axis, in that administration of these hormones may initiate their effects on bone IGF-1 to rats results in a decrease in TRs (and their cells at the membrane (rather than the nucleus) by respective mRNAs) as well as the activities of two resulting in increases in the levels of the precursor of thyroid-hormone-responsive liver enzymes (Pellizas PGE#. Non-genomic effects on bone cells have been et al., 1998). The physiological relevance of this reported in that T3 stimulates the inositol second finding awaits demonstration that the doses of IGF- messenger system in rat bone rudiments, within 30 1 used are relevant to the euthyroid situation in rats. seconds (Lakatos & Stern, 1991). For a discussion of Although thyroid hormone replacement in hypo- thyroid effects on bone cell growth and differ- thyroid individuals restores normal growth, excess entiation the reader is referred to Klaushofer et al. thyroid hormone does not increase final body height (1995) and Williams, Robson & Shalet (1998). (Bernal & DeGroot, 1980). For reviews of thyroid A membrane site for thyroid hormone stimulation hormone effects on bone and mineral metabolism of cellular growth is compatible with the finding that the reader is referred to Auwerx & Bouillon (1986) elevated arachidonic acid (20:4) content of mem- and Klaushofer et al. (1995). At low concentrations, branes stimulates cell growth and increases mRNA thyroid hormones stimulate the growth and calci- levels of growth-related early response genes (c-fos fication of bone, whilst at high concentrations they and Egr-1) in non-bone cell cultures (Danesch, inhibit bone growth and stimulate resorption (Ren et Weber & Sellmayer, 1996). As they are mediated by al., 1990), and in adult humans, thyrotoxicosis increased PGE# production, such effects were re- results in bone loss (Fraser et al., 1971). It is thus of stricted to the n-6 PUFA, arachidonic acid and were interest that both T3 and T4 have biphasic effects on not observed following increases in the n-3 PUFAs. IGF-1 production by rat osteoblastic cells in culture Several other cell cultures have also been reported to and by fetal rat limb bones. Both these bone cell show increased cellular proliferation following alter- systems exhibit a basal secretion of IGF-1 that is ation of their membrane n-6 PUFA content (e.g. enhanced by thyroid hormones, however at high Holley et al., 1974; Murphy, 1986; Sylvester et al., concentrations both T3 and T4 dramatically inhibit 1994). IGF-1 secretion (Lakatos et al., 1993). Transplant studies have shown that skeletal and (7) Other effects other tissues from fetal rats grow equally well in hypothyroid and euthyroid rats whilst those from Effects to be discussed in this section are those juvenile rats show a growth retardation in hypo- related to: (i) reproduction, (ii) immune and thyroid hosts (Cooke, Yonemura & Nicoll, 1984). antiviral defence, and (iii) defence against free The potential importance of thyroid hormones in radicals (i.e. as antioxidants). skeletal development is illustrated by the finding In many vertebrate species, reproduction occurs that in serum-free culture conditions, chondrocytes only at certain times of the year and during the do not mature, but the addition of both T3 and T4 remainder of the year such species are refractory to can induce full expression of chondrocyte hypertro- reproduction. European starlings Sturnus vulgaris are phy leading to matrix calcification, the normal such seasonal breeders and following thyroidectomy process involved in the conversion of cartilage to remain in their breeding season indefinitely (Woit- bone (Alini et al., 1996). kewitsch, 1940). Many bird and mammal species The mode of thyroid hormone action on bone have since been shown to require thyroid hormones remodelling is not known. Work with bone organ for the cessation of reproductive activity in their cultures and cultured osteoblasts shows that thyroid normal seasonal breeding cycle (for review see hormones increase the production of the prosta- Karsch et al., 1995). Experiments with male Am- glandin PGE# and that inhibitors of PGE# pro- erican tree sparrows Spizella arborea, has shown them duction diminish the bone resorption caused by both to be programmed for photoperiodic gonadal growth T4 and T3. The fact that T3 binds to osteoblast and separately for post-nuptial moult. Both events nuclei from various sources has resulted in the are dependent on thyroid hormones and findings suggestion that thyroid hormones exert their bone suggest that separate neural control circuits are resorption effects via nuclear receptors. However, if organized early in the breeding season (Wilson & following thyroid hormone treatment the arachi- Reinert, 1996). In both this and another bird species, donic acid (20:4; the precursor of PGE#) content it has been demonstrated that T4 and not T3 is increases in the membranes of bone cells as it does in the ‘active’ thyroid hormone (Pant & Chandola- Thyroid hormones and their effects 577

Saklani, 1995; Reinert & Wilson, 1997). Experi- of cultured human granulosa cells to T3 stimulates ments with the ewe have shown that thyroid the in vitro production of tissue inhibitor of metallo- hormones are required for only a short ‘window’ of proteinases 1 (TIMP-1) which is suggested to be time during the breeding season for the development involved in ovulation, luteal development and of anoestrus at the end of the breeding period regression as well as trophoblast invasion (Goldman (Thrun et al., 1996; Karsch et al., 1995). et al., 1997). T3 has also been suggested to be a Such thyroid-hormone-dependent seasonal cess- negative modulator of steroidogenic function in the ation of reproductive activity is mediated by the avian adrenal gland (Carsia et al., 1997). hypothalamic-pituitary-gonad axis, involving both Thyroid status has limited influence on the gonadotropin-releasing hormone (GnRH) neurones immune system. For example, neither hypo- and secretion of luteinizing hormone (LH) and thyroidism nor hyperthyroidism had significant follicle-stimulating hormone (FSH) from the an- effects on the number of blood mononuclear cells or terior pituitary. The role of the GnRH system in immune function in the rat or guinea pig (Wall, seasonal breeding has been reviewed by Lehman et Twohig & Chartier, 1981). At supraphysiological al. (1997). In two seasonal breeding mammal species concentrations, thyroid hormones alter the response (sheep and hamsters), an antibody to both TRα1 of lymphocytes to mitogens but have no such effects and TRα2 co-localized with GnRH neurones in the at physiological levels (Ong, Malkin & Malkin, brain (Jansen et al., 1997). However, considering 1986). Thyroid hormones at physiological concen- that the effect of thyroid hormones is restricted to a trations do however potentiate the antiviral action of ‘window’ of time that precedes anoestrus by a interferon γ in cultured human cells, and T4 is considerable period, such nuclear thyroid receptors approximately ten times more potent than T3 (Lin may not be directly relevant to setting the seasonal et al., 1994). Thyroid hormone itself is not antiviral breeding timetable. Instead, it is tempting to and various analogues of thyroid hormones lack the speculate that the mode of this thyroid hormone ability to potentiate interferon activity. It has been effect involves T4-dependent interactions between suggested that there are two pathways for such an neurones and astrocytes and laminins in the extra- effect of thyroid hormones, one dependent on, and cellular matrix (see Leonard & Farwell, 1997). This the other independent of protein synthesis (Lin et al., is likely in view of the fact that there is considerable 1996). The protein-synthesis-dependent pathway is seasonal plasticity in the degree of synaptic inputs to blocked by inhibitors of protein kinases C and A and GnRH neurones, and GnRH neurones differ from appears to be independent of thyroid nuclear neighbouring neurones in that they are almost receptors (H.-Y. Lin et al., 1997). completely surrounded by glial processes from Although aerobic organisms cannot exist without adjacent astrocytes (Lehman et al., 1997). Experi- oxygen, the reductive nature of the cellular en- ments examining the relative potencies of rT3, T3 vironment provides ample opportunities for the and T4 in such situations will probably give creation of reactive oxygen species (ROS) which are additional insight into the mechanisms involved. extremely dangerous to living systems in that they Hypothyroidism has no influence on testicular can damage proteins, lipids, carbohydrates and development in prenatal rats but postnatal hypo- nucleic acids. This has been called the ‘oxygen thyroidism results in reduced levels of gonadotropins paradox’ (for review see Davies, 1995). Thus, and delays pubertal spermatogenesis (Francavilla et aerobic organisms both generate and accumulate a al., 1991). In the adult female rat, hypothyroidism variety of water- and lipid-soluble antioxidant results in dramatically extended oestrus cycles compounds, as well as producing a series of anti- (Evans et al., 1966). Whether this is an effect on the oxidant enzymes, to intercept and inactivate ROS. gonadotropin system or an effect at the ovarian level Membrane phospholipids are continually subject to is not known. Thyroid hormones are known to such oxidative damage and lipid peroxidation is an influence ovarian cells. For example, T3 exposure autocatalytic chain of reactions initiated at the influences steroid (androgen, estradiol and pro- double bond of an unsaturated acyl chain. Phos- gesterone) production by porcine ovarian follicle pholipids are more prone to peroxidation than cells. The effects are complex in that they depend on triglycerides, with the highly polyunsaturated acyl the particular type of follicle cell, the stage of chains (e.g. 20:4 and 22:6) being the most sensitive follicular development and also demonstrate inter- to damage (e.g. Catala & Cerruti, 1997). It is actions with human chorionic gonadotropin ex- estimated that approximately 1% of daily oxygen posure (Gregoraszczuk & Skalka, 1996). Exposure consumption goes to mitochondrial ROS generation. 578 A. J. Hulbert

Another major source are phagocytic cells which use Cells utilize a number of antioxidant scavengers an NADPH oxidase system to produce ROS as a and enzymes as a defence against damage by ROS. weapon against invading microrganisms. Such cells Vitamin E is a major membrane antioxidant whilst only produce ROS in large amounts when they have vitamin C is a major aqueous antioxidant. Other been activated to phagocytose (Davies, 1995). membrane-bound antioxidant compounds include Thyroid hormones themselves also pose a paradox. β-carotene, ubiquinone and, as will be discussed This ‘thyroid hormone paradox’ is that they both shortly, possibly thyroid hormones. Antioxidant promote damage by ROS, by stimulating aerobic enzymes include superoxide dismutase (SOD), gluta- metabolic activity, but they also act as antioxidants thione peroxidases and catalases (Davies, 1995). The themselves, as well as influencing other antioxidant influence of thyroid status on antioxidant defences is defences. mixed and there appears to be no consistent pattern. Hyperthyroidism increases lipid peroxidation in Some studies report that hyperthyroidism increases young, old and very old rats and hypothyroidism vitamin E content of cardiac muscle (Mano et al., results in lower levels of lipid peroxidation in young 1995; Venditti et al., 1997a) whilst others show no rats (Mooradian et al., 1994). One study reports that effect (Asayama et al., 1989). Hyperthyroidism hypothyroidism in rats significantly decreases lipid increases glutathione levels in erythrocytes and peroxidation in skeletal muscle but not thymus, serum (Morini et al., 1991; Seven et al., 1996) but spleen or lymph nodes (Pereira et al., 1994), whilst decreases glutathione levels in rat liver (Morini et al., another reports that it results in no reduction in lipid 1991). During hyperthyroidism, glutathione per- peroxidation in liver, cardiac or skeletal muscle oxidase is elevated in liver, erythrocytes, spleen and (Venditti et al., 1997a). In rats, hyperthyroidism skeletal muscle (Morini et al., 1991; Seven et al., results in increased levels of lipid peroxidation 1996; Pereira et al., 1994; Fernandez & Videla, products in liver, heart, soleus muscle and lymph 1995) diminished in thymus, heart and skeletal nodes (Asayama et al., 1989; Pereira et al., 1994; muscle (Seven et al., 1996; Pereira et al., 1994; Venditti et al., 1997a). Higher levels of mitochondrial Asayama et al., 1989; Fernandez & Videla, 1995; enzymes and increased ROS production by sub- Mano et al., 1995) and unaffected in liver and heart mitochondrial particles has also been reported for (Fernandez & Videla, 1995). Whilst in hypo- hyperthyroid rats (Fernandez & Videla, 1993). thyroidism it has been reported as both increased in Thyroid-induced increases in lipid peroxidation cardiac and skeletal muscle (Fernandez & Videla, are not restricted to enhanced mitochondrial ac- 1995; Mano et al., 1995) and reduced in thymus and tivity. For example, in rats T3 increases superoxide skeletal muscle (Venditti et al., 1997a; Pereira et al., production, NADPH oxidase activity and lipid 1994). Glutathione reductase activity appears rela- peroxidation in hepatic microsomes (Fernandez et tively unaffected by thyroid status (Venditti et al., al., 1985). Polymorphonuclear leucocytes from hy- 1997a) although it has been reported to be elevated perthyroid rats have an enhanced capacity to in liver during hyperthyroidism (Morini et al., 1991). produce superoxide when stimulated in vitro and this Catalase activity is reported both to increase in seems primarily related to thyroid-hormone-en- skeletal muscle (Pereira et al., 1994) and decrease in hanced NADPH oxidase activity (Fernandez & thymus, spleen, cardiac and skeletal muscle (Pereira Videla, 1995). A similar effect has been observed in et al., 1994; Asayama et al., 1989; Seven et al., 1996) # # phagocytes from hyperthyroid humans (Videla et al., during hyperthyroidism. The activity of Cu +,Zn +- 1993). Both T4 and T3 can also stimulate, in vitro, SOD is also reported to be increased in thymus, the activity of myeloperoxidase isolated from human leucocytes and skeletal muscle (Pereira et al., 1994) leukocytes (Van Zyl, Basson & Van der Walt, 1989). but decreased in cardiac and skeletal muscle (Asa- # Low-density lipoproteins isolated from hyperthyroid yama et al., 1989) whilst Mn +-SOD is reported to be humans also show an elevated lipid peroxidation increased in leucocytes, thymus, spleen, cardiac and compared to those from euthyroid individuals, skeletal muscle (Pereira et al., 1994; Asayama et al., intriguingly as do those from hypothyroid indiv- 1989) during hyperthyroidism. No general pattern iduals (Costantini et al., 1998). These results suggest can be discerned from the diverse responses of that, at least in humans, the euthyroid condition antioxidant defences to hyperthyroidism. Most represents that which results in the minimal peroxi- studies involving hypothyroidism report no sig- dative damage to low-density lipoproteins. Whether nificant influence on levels of either antioxidant this also applies to other molecules capable of scavengers or enzymes. oxidative damage is not known. The question as to whether thyroid hormones Thyroid hormones and their effects 579 themselves can act as antioxidant compounds was produced in Fig. 10. When human neutrophils are first raised after De Caro (1933) observed that T4 stimulated by pyrogenal-lipopolysaccharide from can decrease oxygen uptake by solutions of un- the Salmonella typhi cell wall, they initiate ROS saturated fatty acids. It was later shown that T4 generation that can be monitored by a chemi- could also prevent lipid peroxidation in isolated luminescence assay. Thyroid hormones at high erythrocytes (Bunyan et al., 1960) and liver homo- concentrations (5 µ) inhibit this chemilumin- genates (Bunyan et al., 1961), and that T4, T3 and escence as does another powerful antioxidant, ionol 3,5-T2 (but not the non-iodinated thyronine) could (4-methyl-2,6-di-isobutyl phenol). This inhibition inhibit lipid peroxidation in isolated rat liver decreases to approximately half as thyroid hormone mitochondria as well as dramatically inhibiting concentrations decrease to 1 n. Surprisingly, when carbon tetrachloride accelerated lipid peroxidation T4 and T3 concentrations are further decreased to (Cash et al., 1967). As with many early studies, the 10 p and 0.1 p, chemiluminescence inhibition T4 concentrations used in these experiments were increases. Such a bimodal response suggests that two high and non-physiological, the lowest being 1 µ. separate processes are operating. The effect was only Wynn (1968b) used an in vitro system to show that observed if T4 and T3 were added prior to # Fe + could activate lipid peroxidation of rat liver stimulation of ROS production and only when the microsomes and that T4 could terminate such lipid ROS generation stimulus was pyrogenal and not peroxidation by acting as an antioxidant. The lowest when it was 4β-phorbol 12-myristate 13-acetate T4 concentration examined was 100 n, and at this (PMA). Whilst T4 and T3 showed a bimodal concentration, T4, rT3 and T3 all showed anti- antioxidant response, the powerful antioxidant, ionol oxidant activity (in decreasing intensity) and their did not. These results suggest that T4 and T3, at antioxidant abilities were greater than that of either very low concentrations, have effects that influence vitamin E or ascorbic acid in the same system. He ROS production. They are more potent than ionol, also showed that pharmacological doses of T4 could which is in turn is a more potent antioxidant than reduce in vivo the peroxide content of epididymal fat vitamin E. These authors also discuss the proposal pads in the rat (Wynn, 1968a). Because of its relative that thyroid hormones are involved in other cellular metabolic inertness such a tissue is unlikely to have repair mechanisms. had significant thyroid-hormone-induced peroxi- That thyroid hormones possibly have antioxidant dation. In these lipid peroxidation experiments, T4 activity is also illustrated by the finding that was degraded when exerting its antioxidant effects. although resting neutrophils exhibit negligible de- This was consistent with the finding that the in vivo iodination of both T4 and T3, when they are rate of T4 degradation seemed to be related to its stimulated to phagocytose (and ROS production is clinical effect (Galton & Ingbar, 1962). Lipid consequently activated), the deiodination of both T4 peroxidation can also be catalyzed in vitro by and T3 dramatically increases severalfold. Neutro- hemoglobin, and iodothyronines inhibit such per- phils from individuals with chronic granulomatous oxidation, and are more potent antioxidants in these disease, which are deficient in the capacity to situations than vitamin E, glutathione or ascorbic produce ROS, degrade T4 and T3 poorly during acid (Tseng & Latham, 1984). phagocytosis (Klebanoff & Green, 1973). The physiological importance of the antioxidant Examination of the potential role of thyroid properties of thyroid hormones was questioned when hormones as potent natural membrane antioxidants the antioxidant capacities of both T3 and T4 in also deserves attention in light of other observations. retarding the auto-oxidation of rat brain homo- As can be seen from Fig. 4, a strong correlation has genate and free-radical-mediated oxidation of rat been observed in rats between the phospholipid erythrocyte membranes was evaluated and found to content of tissues and their T4 content. Analysis of be very low (1–2%) at hormonal concentrations of this relationship, assuming that all T4 is associated 50 n but substantial at micromolar hormonal levels with membranes, suggests a molar ratio of T4: (Faure, Lissi & Videla, 1991). phospholipid of approximately 1:1000. This is However, a more recent report of T4 and T3 similar to the molar ratio of vitamin E: lipid observed inhibition of ROS generation by activated human in biological membranes which is of the order of neutrophils (Antipenko & Antipenko, 1994) raises 1:2000–3000 (Packer & Landvik, 1989). Whilst questions with respect to the possible physiological obviously not all thyroid hormone molecules are importance of the antioxidant properties of the associated with membranes, measurements in hu- thyroid hormones. These findings have been re- man erythrocytes suggest that approximately half of 580 A. J. Hulbert

%

100 ! L-T3 Pyr ! Pyr L-T4

DL-T ! Pyr 50 4

! Ion Pyr ! Pyr DL-T4 chemiluminescence inhibition

–5 –6 –7 –9 –11 –13 [T4] [T3] [Ionol] log M

Fig. 10. The influence of the concentration of thyroid hormones (T4 & T3) and ionol (Ion) on the release of reactive oxygen species (monitored as chemiluminescence) by human neutrophils activated by pyrogenal-lipopolysaccharide (Pyr) from Salmonella typhi cell wall. The symbols on the right of the figure indicate the sequence of addition. Reprinted with permission from Antipenko & Antipenko (1994). cellular thyroid hormone molecules (34% of T4 and metamorphosed into frogs. Since that time it has 60% of T3) are associated with membranes (Bregen- become obvious that the thyroid is also involved in gaard et al., 1989). As the human erythrocytes lack normal development of vertebrates, ranging from intracellular membranes and in this particular study fish to mammals. they were washed several times before measurement, Indeed, thyroid hormones may be necessary from these are likely to be minimum values. conception for normal development. Interestingly, Another reason why this area is worthy of careful iodine deficiency in female rats results in a dramatic examination is that the resting oxygen consumption decrease in the size of their young even when such of mammals is proportional to approximately the iodine deficiency is ceased at the time they conceive 0.75 power of body mass (Kleiber, 1961). This is the their young (Sunitha, Udaykumar & Raghunath, same allometric exponent observed for whole- 1997). Pharoah, Buttfield & Hetzel (1971) have mammal ethane exhalation rate, which is indicative shown that iodized oil must be given prior to of lipid peroxidation (Topp et al., 1995) and for the conception to iodine-deficient mothers to prevent degradation rates of T4 and T3 in mammals neurological damage in their children. In sheep, (Tomasi, 1991). These findings of similar allometric maternal thyroidectomy prior to conception results exponents for these three processes means that, in in neurological damage (Potter et al., 1986). Al- mammals, irrespective of metabolic activity, there is though the plasma concentrations of thyroid hor- a relatively constant molar ratio between oxygen mones are relatively constant throughout adult life consumption, phospholipid peroxidation and de- of most vertebrates, all vertebrate classes examined iodination of T4 and T3. to date, during their early development, have a surge of plasma T4 (and T3) to levels that exceed those found in the adult plasma. Thyroid hormones are found in the eggs of fish IV. THYROID HORMONES AND VERTEBRATE (Tagawa & Hirano, 1987; Leatherland et al., 1989) DEVELOPMENT and birds (Sechman & Bobeck, 1988; Prati, Calvo & Morreale de Escobar, 1992; Wilson & McNabb, (1) Thyroid axis during vertebrate 1997). The detection of both T4 and T3 in development amphibian embryos one day after fertilization (at, The influence of thyroid hormones on vertebrate respectively, approximately 0.6 and approximately " development was first demonstrated by Gudernatsch 4.7 pmol g− ) suggests that they are also present in (1912) who showed that when tadpoles were fed amphibian eggs (Weber et al., 1994). Nothing is desiccated horse thyroid tissue they precociously known concerning reptilian eggs. Because mam- Thyroid hormones and their effects 581 malian eggs remain in the mothers body, it is also period and metamorphosis is associated with a likely that mammalian eggs contain thyroid hor- dramatic decrease in plasma thyroid hormone levels. mones. In fish eggs, the T4 content can either exceed The larvae lack a follicular thyroid gland but possess that of T3 (e.g. Leatherland et al., 1989) or be less instead an endostyle (a subpharyngeal gland) that than the T3 content (e.g Yamano & Miwa, 1998). develops into a follicular thyroid in the adult. In Partial depletion of both T4 and T3 in the eggs of the premetamorphic lampreys, inhibition of thyroid Medaka fish Oryzias latipes had no discernable effects function induces early metamorphosis whilst ex- on either hatching, survival or development of the ogenous thyroid hormones inhibit metamorphosis young fish (Tagawa & Hirano, 1991). In the (Manzon, Eales & Youson, 1998). The lampreys amphibian embryo, T3 concentration exceeded T4 appear to be unique among vertebrates in the concentration (Weber et al., 1994) whilst the opposite relationship between thyroid function and meta- was the case for avian eggs (Sechman & Bobeck, morphosis, whether they represent the original 1988; Prati et al., 1992; Wilson & McNabb, 1997). vertebrate condition or a derived condition is The deposition of thyroid hormones in avian eggs unknown. The other extant agnathans, the hagfishes, and their role has recently been reviewed by which are believed to be phylogenetically older than McNabb & Wilson (1997). lampreys, do not exhibit metamorphosis and possess There is a surge of T4 concentration during a follicular thyroid gland. The role of the thyroid in metamorphosis in the flounder Paralichthys olivaceus lamprey metamorphosis has been recently reviewed (Miwa et al., 1988). In tilapia fish Oreochromis by Youson (1997). mossambicus, there is a surge of T4 concentration Although earlier studies suggested that thyroid which corresponds in time to a morphogenic change hormones may not be present in the plasma of in body shape and is followed by an increase in T3 premetamorphic amphibians, recent work has shown concentration (Okimoto, Weber & Grau., 1993). that both T4 and T3 are measurable in the tissues Inhibition of thyroid function with goitrogens has of amphibian larvae from the time of hatching been shown to inhibit, and exogenous T4 to (Gancedo et al., 1997). In amphibians, there is a stimulate, the metamorphosis of larvae to juvenile developmental surge of T4 and T3 concentration fish in the flounder (Miwa & Inui, 1987), and the that peaks at metamorphic climax (Mondou & zebrafish Danio rerio (Brown, 1997), whilst both Kaltenbach, 1979; Suzuki & Suzuki, 1981; Weber et exogenous T4 and T3 promote early metamorphosis al., 1994). In Rana catesbeiana, the T4 surge precedes in grouper Epinephelus coioides larvae (de Jesus, the T3 surge (Suzuki & Suzuki, 1981) and there is a Toledo & Simpas, 1998). A developmental T4 surge surge in transthyretin (TTR) mRNA levels in liver has also been observed in the larvae of black sea during metamorphosis which is absent in the adult. bream (Tanaka et al., 1991) and red sea bream Transthyretin mRNA was not detectable in the Pagrus major (Kimura et al., 1992). tadpole choroid plexus (Yamauchi et al., 1998). Juvenile salmonid fish undergo a parr-smolt Mexican axolotls are neotenous amphibians that transformation that prepares them for the impending rarely undergo metamorphosis. Juvenile axolotls sea-water phase of their life cycle, and during this have no detectable thyroid hormones in their plasma process plasma T4 levels temporarily increase four- whilst adults have no detectable T3 and only low fold followed by a smaller T3 increase (Dickhoff, plasma levels of T4, and thus do not appear to have Darling & Gorbman, 1982; Specker et al., 1984). In such developmental surges of thyroid hormones coho salmon Oncorhynchus kisutch, a year before the (Galton, 1992b). However if they are given an parr-smolt change (which takes place in the second exogenous T4 surge they will undergo metamorphic year of freshwater life) there is a small surge in T4 changes (e.g. Brown, 1997). concentration followed by a small increase in plasma The egg, especially the yolk, is the source of T4 T3 levels. Whether this has any developmental and T3 found in the tissues of the avian embryo prior importance is unknown but it has been suggested to the onset of thyroid function (Prati et al., 1992; that it may be the basis for ‘phase differentiation’ Wilson & McNabb, 1997). Accumulation of T3 is measured in other salmon species (Dickhoff et al., especially high in the brain and is much greater than 1982). inferred from measurement of plasma levels, and in Lampreys are modern representatives of the this respect the chicken embryo in similar to the rat earliest vertebrates, the jawless agnathan fish, and fetus (Prati et al., 1992). Transthyretin mRNA is exhibit metamorphosis. In this group, plasma levels produced by the chick embryo liver well before the of thyroid hormones increase during the larval development of the thyroid gland and rapidly rises 582 A. J. Hulbert to adult levels, whilst the chick embryo’s choroid this species. The developmental profiles for T4 and plexus also produces TTR mRNA during incubation T3 are similar for both total and free hormone but at levels lower than that found in the adult concentrations. Plasma rT3 levels are higher (and chicken’s choroid plexus (Southwell et al., 1991). more variable) during pouch life than in adults and The role of the unidirectional secretion of TTR into show a peak at the same time as the T4 peak. Both cerebrospinal fluid as a brain thyroid hormone liver and kidney show no 5h deiodination activity at pump may be important in the accumulation of birth but such enzymatic abilities increase during thyroid hormones by the avian embryo’s brain. pouch life to reach adult levels shortly after the time In the chicken, there is a developmental surge in of pouch exit (Janssens et al., 1990). A similar plasma T4 concentration that peaks just before developmental surge in levels of thyroid hormones hatching, whilst the T3 level peaks at the time of has been recorded for another marsupial, the hatching and remains high thereafter (Darras et al., brushtail possum Trichosurus vulpecula (Buaboocha & 1992). In the precocial Japanese quail Coturnix Gemmell, 1995). japonica, there is a developmental surge of both T4 Eutherian mammals similarly show developmen- and T3 levels around the time of hatching, this surge tal surges in T4 and T3 levels. In the rat, both T4 in hormone concentration is relatively brief and and T3 are found in the embryo prior to onset of manifest in both the total and the free plasma fetal thyroid function (Obregon et al., 1984) and hormone levels (McNabb, Lyons & Hughes, 1984). maternal hypothyroidism at this time results in In the altricial ring doves Streptopelia risoria, starlings reduced tissue levels of both T4 and T3 in the fetus, Sturnus vulgaris and redwing blackbirds Agelaius which for some tissues persists throughout gestation phoeniceus there is a more extended developmental (Morreale de Escobar et al., 1985). The surge in surge of plasma T4 levels that begins after hatching plasma T4 levels precedes the increase in plasma T3 and lasts for approximately the first three weeks of concentration in the rat (Morreale de Escobar, post-hatching life (McNabb & Olson, 1996). In ring Obregon & Escobar del Rey, 1987), the horse doves, although total plasma T4 concentration does (Irvine & Evans, 1975), the ferret (Kastner, Kastner not demonstrate a pronounced peak, there is an & Apfelbach, 1987) and the lamb (Wrutniak, eightfold increase in free T4 levels peaking at Cabello & Bosc, 1985). In the rat, which is an approximately day 14 after hatching, that is sug- altricial eutherian, these hormone surges commence gested to be due to displacement from plasma before birth but predominantly occur in the first protein binding (McNabb, 1988). postnatal weeks, whilst in both the sheep and To my knowledge, nothing is known of the humans, the thyroid hormone surges are prenatal. developmental profile of thyroid hormones in reptiles Summarizing the accumulated data for verte- or the egg-laying mammals, the monotremes. In the brates, there is a severalfold increase in plasma T4 marsupials, the young are born at a very immature concentration during early vertebrate development phase and undergo a large amount of development which is either followed by, or coincident with, an in the pouch of the mother. Indeed, it is near the increase in plasma T3 levels. The plasma T4 surge is time of pouch exit, rather than birth, that the more prevalent in vertebrates than a developmental physiological development of marsupial young is surge in plasma T3 levels. similar to that of newborn eutherian mammals. Deiodinase activity also varies during vertebrate Marsupial pouch young have been likened to development. In larval lampreys, outer ring (5h) ‘exteriorised embryos’ and have been proposed as deiodination activity is high in the intestine and liver good model systems in which to study mammalian and low in kidney and muscle, whilst in adult development (see Tyndale-Biscoe & Janssens, 1988). lampreys the intestine, kidney and muscle retain 5h- In the tammar wallaby Macropus eugenii (Fig. 11) deiodination activity but it is completely absent from there is a developmental surge of T4 concentration adult liver tissue (Eales et al., 1997). Inner ring 5- that precedes a T3 surge. In very early pouch young deiodination activity was not detected in larval (well before histological development of the thyroid tissues but was observed in adult intestine and, to a gland), plasma T4 concentration is similar to adult lesser extent, in adult kidney (Eales et al., 1997). levels, rising to a peak at approximately 160 days of Manipulation of thyroid status in larval lampreys, age before returning to adult levels. By contrast, which influenced the timing of metamorphosis, had plasma T3 levels in newborn are considerably lower no influence on intestinal 5h-deiodinase activity than adult levels and peak later, at approximately (Manzon et al., 1998). 220 days of age, which is after initial pouch exit in In salmon, during the parr-smolt transformation, Thyroid hormones and their effects 583

Pouch Pouch Birthexit Adult Birthexit Adult 80 4 50 250 (a) (b)

40 200 60 3 ) ) ) ) M M M 30 150 M 40 Total T3 2 Free T3 20 100 Free T3 (p Total [T4] (n Total Total [T3] (n Total Free [T4] (p

20 1 10 50

Total T4 Free T4 0 0 0 0 0 100 200 300 400 Adult 0 100 200 300 400 Adult Age (days) Age (days)

Pouch Birth exit Adult 1.2 700 ) (c) Deiodinase (liver) –1 h

Deiodinase (kidney) –1 1 600 )

M 500 0.8 400 0.6 300 Total [rT3] (n Total 0.4 200 Total rT3 0.2 100 monodeiodinase activity (nmol g !

0 0 5 0 100 200 300 400 Adult Age (days) Fig. 11. Thyroid hormone levels and tissue deiodination activity during development of the tammar wallaby Macropus eugenii. (A) The developmental profiles for total plasma T4 and T3 concentrations. (B) The developmental profiles for free plasma T4 and T3 concentrations. (C) The developmental profile for total plasma rT3 concentration and the developmental profile of the in vitro 5h monodeiodination activity of liver and kidney. In all graphs, the time of initial pouch exit is marked. Young wallabies are at a stage physiologically similar to newborn eutherian mammals at this time. Reproduced using the data of Janssens et al. (1990). Values are means jS.E.M. (N l 2–8). changes in T4 production, distribution and metab- tivity is high in brain tissue, as well as heart and gill, olism are relatively independent of being in a fresh- but is low in liver and muscle (Morin, Hara & Eales, water or salt-water environment (Specker et al., 1993; Eales et al., 1993). These studies suggest that 1984). Tissue thyroid hormone levels do not echo tissue deiodinase activities are probably important in plasma concentrations during this period (Specker, determining different developmental profiles for Brown & Bern, 1992). During the parr-smolt intracellular thyroid hormones in individual tissues transformation, there is enhanced outer ring 5h- during the parr-smolt transformation in salmonid deiodination activity in liver and heart, but in gill fish. and skeletal muscle the relatively low deiodinase Galton & Cohen (1980) observed negligible in vivo activity remains unchanged. Plasma T3 levels are conversion of T4 to T3 in premetamorphic tadpoles correlated with liver 5h-deiodination activity and yet both exogenous T4 and T3 had effects in these brain 5h-deiodination activity increases at the end of premetamorphic tadpoles. Later it was demon- the transformation. Inner ring 5-deiodination ac- strated that outer ring 5h-deiodinase activity was 584 A. J. Hulbert absent in liver, heart, kidney and tail during derance of inner ring 5-deiodination prior to meta- premetamorphosis but was present in skin and morphic climax to predominantly outer ring 5h- exhibited minimal activity in brain and intestine deiodination after climax. whilst inner ring 5-deiodinase activity was detectable The ontogeny of deiodination in birds has been in most tissues during premetamorphosis, being reviewed by McNabb (1988). In the precocial highest in liver, intestine and kidney. At meta- Japanese quail, at the time of the T4 surge around morphic climax, induced by T4 exposure, 5h- the time of hatching there is also a dramatic increase deiodinase activity dramatically increased in skin in outer ring 5h-deiodination in the liver, which is and intestine and was present in tail but remained initially due an small increase in D2-type deiodinase absent in liver, heart and kidney, whilst 5-deiodinase but is rapidly replaced by a much larger increase in activity became barely detectable (Galton & Hie- D1-type deiodinase activity. After its initial surge in bert, 1988). Galton (1988) examined tissue de- activity, the D1 deiodinase declines to adult levels iodinase activities during spontaneous metamor- during the first few days after hatching. In the phosis and extended measurements to include pro- altricial ring dove, hepatic 5h-deiodinase activity is metamorphosis and adult tissues. She found that 5h- relatively constant in embryos and for a few days deiodinase activity was highest in skin and intestine after hatching after which it declines in activity. The and negligible in other tissues. In prometamorphic decline in 5h-deiodinase activity after hatching in tadpoles, deiodinase activity was comparable to that this species is the opposite of changes in the serum in premetamorphic tadpoles and it remained high in T3\T4 ratio which shows an increase during this adult tissues. In the adult frog, whilst 5h-deiodinase period (McNabb, 1988) This is probably due to the activity was only observed in intestine and skin, even more dramatic decline in 5-deiodination which inner ring 5-deiodinase activity was present in all will remove T3 formed from T4. During the last tissues examined. Unexpectedly, in view of the week of incubation, the chick embryo liver shows a observation that 5-deiodinase activity decreases substantial increase in outer ring 5h-deiodinase during metamorphic climax, it has been shown that activity, whilst there is an even more dramatic 5-deiodinase (D3) gene expression and D3 enzyme decrease in inner ring 5-deiodinase activity over the activity is increased by exposure to T3 (Becker et al., same period (Galton & Hiebert, 1987). Others have 1995). The measurement of both 5h-deiodinase (D2) confirmed that in the chick embryo, D1 deiodinase and 5-deiodinase (D3) enzyme activity in a variety activity increases whilst D3 deiodinase activity of tissues during metamorphosis in the amphibian decreases during the last days of incubation (Van Rana catesbeiana (Becker et al., 1997) has strongly der Geyten et al., 1997). A decline in 5h-deiodinase supported the concept that coordinated development activity following hatching is also reported for the of different tissues during the developmental surge of chicken and is unlike the situation after birth in some extracellular T4 levels could be determined by mammals. Although hepatic 5h-deiodination de- specific ontogenetic profiles for the deiodinases in creases after hatching in the chicken, the decrease in different tissues. In this excellent study, the authors 5-deiodination is even greater and begins approx- showed that some tissues (e.g. hindlimb) have high imately 4–5 days prior to hatching. In the newly D2 deiodinase activity prior to, and negligible hatched chicken, hepatic 5-deiodination is negligible activity after metamorphic climax, whilst other (Darras et al., 1992). The deiodinase activities of tissues (e.g. tail, intestine, forelimb) had the opposite avian tissues other than liver, and their importance, profile and still other tissues (e.g. skin, eye) had a is relatively unknown. In the chicken, kidney relatively constant activity throughout the various deiodinase activity is less than that found in liver and stages of amphibian development. The activity of is relatively constant before and after hatching the inner ring D3 deiodinase also showed a de- (Darras et al., 1992). The chicken brain is reported to velopmental profile that differed between tissues. have both D2 and D3 deiodinases and the deiodinase Thus, by tissue-specific changes in deiodinase ac- system in birds is reported to be similar to that in tivities, the common developmental surge in extra- mammals (Kuhn, Mol & Darras, 1993). cellular T4 levels can be modified in individual During early pouch life in the marsupial Macropus tissues. Prior to metamorphic climax there is a eugenii neither liver nor kidney demonstrate any in decline in liver 5-deiodinase activity whilst after vitro 5h-deiodinase activity. However, at approx- metamorphic climax there is an increase in 5h- imately the time of the plasma T4 surge, in vitro 5h- deiodinase activity in skin and intestinal tissue deiodinase activity increases in both liver and kidney (Galton, 1992a). There is a shift in the prepon- to levels found in the adult tissues (Fig. 11). The fact Thyroid hormones and their effects 585 that significant amounts of rT3 are measured in its relative absence in adult rat intestine, and the pouch young plasma prior to this period suggests importance of tissue deiodination for determining that 5-deiodination activity is present from a very normal tissue T4 and T3 levels in the adult rat, early developmental stage. The decline in rT3 suggests that there may be tissue-specific de- towards the end of pouch life supports the conclusion iodination profiles during the development of verte- that around the time of the developmental T4 surge brates. Indeed, in the rat, the development of 5h- there is a shift in emphasis from inner ring to outer deiodinase activity has different ontogenetic patterns ring deiodination in this marsupial species (Janssens in different tissues which consequently affect the et al., 1990). ontogenetic profiles for tissue T3 (Obregon et al., In the rat, 5h-deiodinase activity is absent in liver 1989). and intestine until day 18 of fetal life, after which it The concentrations of the two types of nuclear appears in both tissues and gradually increases in thyroid receptors and their respective mRNAs also activity. However, whereas 5h-deiodinase activity is vary during vertebrate development. The differ- much lower in the fetal rat than the adult mother, ential tissue expression of these receptors has been the opposite is the case for 5-deiodination which is elegantly demonstrated with in situ hybridization of especially high in fetal intestine and brain compared tissues from the developing flounder (Yamano & to adult intestine and brain and of approximately Miwa, 1998). This study showed that virtually every similar activity in fetal and adult rat liver (Galton, tissue expressed either an α-orβ-type TR gene, and McCarthy & St. Germain, 1991a). In the rat, there that both the intensity and the tissue distribution of is fourfold increase in the activity of brain D2 5h- the two types of TR mRNA varied. At metamorphic deiodinase between gestational days 17 and 22 climax, the TRα mRNA concentration was greatest (birth) that is correlated with an increase in brain in skeletal muscle and digestive epithelium, whereas T3 content. Inhibition of thyroid function during TRβ mRNA was strongest in cartilage cells and this period results in significant increases in D2 presumed osteoblasts, as well as in the myosepta activity (Ruiz de Ona et al., 1988). which separate the myotomes. At a whole-body Selenium deficiency in rats (which causes a near- level, TRα mRNA is more prevalent in the larval complete loss of hepatic D1 5h-deiodinase in adults) flounder than TRβ mRNA. During flounder meta- does not interfere with the postnatal plasma T3 morphosis TRα mRNA concentration increases and surge, and it has thus been suggested that peripheral peaks at climax, after which it decreases dramatically deiodination is not important in this T3 surge but in the postclimax flounder. TRβ mRNA concen- that it may predominantly be due to intrathyroidal tration increases during metamorphic climax and deiodination of T4 (Chanoine et al., 1993b). In peaks after the TRα mRNA peak, after which it sheep, there are increases in both D1 and D3 remains high. Unlike the situation in the larval deiodinase activities after the T4 surge which begins flounder, in the post-climax flounder TRβ mRNA is before birth. The dramatic decrease in rT3 con- more prevalent than TRα mRNA (Yamano & centration and the increase in T3 concentration in Miwa, 1998). sheep plasma at birth demonstrates that this species It has been reported that the TRα gene is too shows a switch from predominantly inner ring to expressed in eggs and embryonic stages that precede outer ring deiodination at this time (Fisher, Polk & the maturation of the thyroid gland in Xenopus laevis Wu, 1994). In humans, the increase in both T4 and and that TRα mRNA levels increase dramatically rT3 levels during the second half of gestation before metamorphic climax (Banker, Bigler & precedes the increase in T3 levels which does not Eisenman, 1991). That it may have some effects at commence until approximately three-quarters of the this early stage is suggested by the fact that way through gestation. This suggests that a similar precocious increase in synthesis of the α receptor switch occurs in humans (Fisher et al., 1977). leads to hormone-dependent developmental abnor- The switch over from inner ring to outer ring malities in Xenopus laevis embryos (Old et al., 1992). deiodination during development appears to be a In the larval amphibian, at the whole-animal level, general finding for many higher vertebrates. Whilst TRα mRNA predominates over TRβ mRNA and deiodinase activity has been predominantly studied reaches high levels prior to metamorphic climax. in liver tissue, the finding of tissue-specific profiles in The concentration of TRβ mRNA increases just intracellular deiodinase activities during develop- prior to metamorphic climax and reaches its peak ment in amphibians, together with the finding of level at climax after which the mRNA for both types high intestinal deiodinase activity in the rat fetus but of thyroid receptors decreases (Yaoita & Brown, 586 A. J. Hulbert

1990). There is also differential expression of TRs in postnatal day 6. In contrast to this pattern, receptor amphibian tissues, in that the nucleated amphibian numbers in liver and heart increased progressively erythrocytes express the gene for TRα but not TRβ throughout fetal and postnatal periods towards the (Schneider, Davey & Galton, 1993), whilst tran- level found in adult rat tissues (Perez-Castillo et al., scripts of the TRβ gene have been detected in 1985). The postnatal rise in receptor number in rat various tissues (tail, skin, kidney, leg, heart and liver is almost completely due to increases in TRβ intestine) of premetamorphic tadpoles but not in which is absent in the fetal rat liver but responsible their erythrocytes (Davey, Schneider & Galton, for approximately 80% of total binding in the liver 1994). In both cases, mRNA levels are highest at of the young adult rat. TRα binding remains metamorphic climax and are decreased in the adult relatively constant during this period and thus frog. Furthermore, the amount of transcript for both declines from being responsible for 100% of liver T3 types of receptors increases in the presence of T3. binding prior to birth to approximately 20% in the Kanamori & Brown (1992) have shown that TRβ adult rat (Rodd et al., 1992). Although there are mRNA is absent in unfertilized eggs, embryos and dramatic changes in mRNA for TRs during this young tadpoles but shows a dramatic developmental period, T3 binding capacity remains little changed surge in transcript levels peaking at metamorphic (Strait et al., 1990). climax and returning to negligible levels in the adult. They demonstrated that TRβ expression was stim- ulated by thyroid hormones and this stimulation (2) Effects of thyroid hormones on required protein synthesis. Schneider & Galton vertebrate development (1991) have similarly shown that expression of the TRα gene is also responsive to thyroid hormone level Thyroid hormones have effects on several tissues in tadpole erythrocytes. This thyroid hormone during development, the most notable being the responsiveness of thyroid receptors will result in the central nervous system. That their deficiency during autoinduction of receptors during metamorphosis development permanently impairs brain function in and has been reproduced in Xenopus laevis cell culture the adult has been known for some time and has as well as in tadpoles (Machuca & Tata, 1992). In been reviewed by a number of authors (e.g. Legrand, both whole tadpoles and the cultured cells, TRα 1986; Morreale de Escobar et al., 1987; Nunez et al., mRNA was present in the absence of thyroid 1991; Porterfield & Hendrich, 1993; Oppenheimer hormones whilst negligible TRβ mRNA was mea- & Schwartz, 1997). In humans, there are three sured. The degree of stimulation by thyroid hor- clinical situations where such effects are manifest. mones was greater for TRβ mRNA than for TRα They are, in order of decreasing severity of brain mRNA. The autoinduction of TRs is sensitive to damage, (a) substantial endemic goitre associated inhibition of protein synthesis (Machuca & Tata, with environmental iodine deficiency, (b) congenital 1992). Such autoinduction during metamorphosis is hypothyroidism in the newborn and (c) maternal likely to be due to the fact that the TR gene has a hypothyroxinemia during gestation. Such maldev- TRE in its own promoter region. This may be elopment is characterized by deaf-mutism, mental especially true for the TRβ gene which, in Xenopus retardation and motor disorders and can be rem- laevis, has been shown to have a number of TREs in edied by hormone replacement initiated soon after its own promoter region (see Tata, 1996). birth. If hormonal replacement is delayed, the effects In erythrocytes from the chick embryo, there is on the central nervous system can be irreversible. As dramatic decline in nuclear thyroid receptors during well as thyroid deficiency impairing development, development (Dasmahapatra, Thomas & Frieden, excess thyroid hormone levels are also detrimental to 1987). This is probably due to changes in TRα,as normal brain development (as they are also for TRα mRNA but not TRβ mRNA has been amphibian metamorphosis). Most of our knowledge identified in chick erythrocytes (Forrest, Sjoberg & concerning the effects of thyroid hormones on brain Vennstrom, 1990). development concerns the laboratory rat, in which Measurement of receptor binding has shown that the important periods are both the fetal and TRs are present in rat fetal tissues prior to the onset (particularly) the neonatal period. It was long of fetal thyroid function. Brain receptor number thought that maternal thyroid hormones did not increased threefold from day 14 to day 17, after cross the placenta of eutherian mammals and were which it remained constant until after birth when it thus not important. However, the extensive studies rapidly increased to approximately adult levels by of Morreale de Escobar and Escobar del Rey and Thyroid hormones and their effects 587 others have conclusively shown that thyroid hor- axons and dendrites in hypothyroid rats has been mones of maternal origin are found in the early fetus related to deficiencies of the cytoskeleton in per- before onset of fetal thyroid function and their ipheral and central neurons. The neuronal cyto- absence during early fetal life can have devel- skeleton includes microtubules and microfilaments. opmental effects (see Morreale de Escobar et al., Microtubules are built from dimers of tubulin and 1985, 1987). Maternal hypothyroidism during the also contain microtubule-associated proteins (known first half of gestation has been shown to compromise as MAPs) whilst microfilaments are composed of the normal development of rats during the second actin subunits. Actin is particularly important in half of their gestation (Bonet & Herrera, 1991). axonal growth cones. Normal neurite growth de- The rat brain at birth is at the same stage as the pends both on the synthesis of cytoskeletal proteins human brain at 5–6 months of gestation and and on their axonal transport. Axonal transport of consequently, stages of brain development that take tubulin has been measured in the optic nerve of the place during the last trimester of human devel- hypothyroid hyt\hyt mouse and found to be sig- opment occur postnatally in the rat. This period also nificantly slower than in euthyroid hyt\j littermates corresponds to the time of the developmental surge (Stein et al., 1991). of thyroid hormones in the rat. Recently, the Tubulin barely self-assembles into microtubules at culturing of fetal and neonatal brain cells has given physiological concentrations and two protein fam- substantial insight into the cellular mechanisms ilies, MAP2 and Tau, act as promoters of tubulin involved, but much still remains unknown. assembly to form microtubules. Many of these The early work of Eayrs and Legrand in the 1960s proteins have a number of isoforms that are expressed demonstrated that thyroid deficiency results in at different developmental times. The mRNA levels altered brain anatomy and dramatically decreased for many of these proteins change in diverse ways neuronal connectivity. This is largely due to a during the period of neuronal differentiation and drastic decrease in axonal and dendritic outgrowth maturation (e.g. Figueiredo et al., 1993). It has been of neurones and synaptogenesis. In the brain, thyroid suggested that thyroid hormones are biphasic (i.e. hormones not only influence the development of stimulating early, whilst inhibiting later) in their neurones but also that of glial cells. For example, influence on cytoskeletal proteins and their re- thyroid hormone deficiency results in reduced spective mRNAs (Biswas, Pal & Sarkar, 1997). myelination of neuronal axons. However, a better description of these findings is When rats are thyroidectomized at postnatal day that thyroid hormone deficiency delays the normal 10, there results a decrease in the density and a developmental program of expression of these cyto- changed distribution of spines from the apical shaft skeletal components. Whilst the mRNA levels for α1 of pyramidal neurons in their visual cortex, which is tubulin in cerebellum are elevated in hypothyroid a manifestation of reduced neuronal connectivity. If rats compared to euthyroid control rats during the T4 replacement is initiated at postnatal day 12, then first 3–4 postnatal weeks (Aniello et al., 1991; spine density and distribution are similar to the Figueiredo et al., 1993), β4 tubulin mRNA levels control (euthyroid) condition. If however, T4 re- show the opposite effect, being reduced in hypo- placement is not commenced until postnatal day 30, thyroid compared to euthyroid rats (Aniello et al., spine density and distribution remains identical to 1991). Both effects can be interpreted as hypo- that observed in the thyroidectomised group (Mor- thyroidism delaying the normal expression of reale de Escobar et al., 1987). tubulin isoforms. In a similar manner, thyroid Possibly the most dramatic visual manifestation of hormone deficiency delays the transition between reduced arborization during hypothyroidism is ob- immature and mature MAP2 and Tau variants. In served in the characteristic dendritic tree of Purkinje this case, the influence of thyroid hormones appears cells in the cerebellum (see Legrand, 1986, p. 518). to be at the level of the splicing mechanism, that is, Growing axons and dendrites must reach their post-transcriptional (Nunez et al., 1991). synaptic partners at an appropriate time and the The developmental profile for β-actin mRNA correct spatio-temporal timing of neurite outgrowth from normal (euthyroid) rat cerebra shows a peak at is an essential component of the development of postnatal day 5 whilst in hypothyroid rats the peak normal neuronal networks. Hypothyroidism retards occurs at postnatal day 10 (Poddar et al., 1996). In the rate of growth of neurite outgrowths and thus the vitro transcription measurements demonstrate a rate of migration of neurones towards their synaptic depressed actin gene transcription in hypothyroid targets. The impaired growth and arborization of cerebra that can be remedied by incubation of nuclei 588 A. J. Hulbert with T3. By contrast, in vitro tubulin gene tran- presence of T4 and the high levels of rT3 present scription rates are similar in hypothyroid and (largely due to the predominance of inner ring over euthyroid cerebra and incubation of nuclei with T3 outer ring deiodination) during early development has no effect on tubulin gene transcription. This has an important functional role in facilitating suggests that whereas transcription represents an interaction between body cells and the extracellular important level of thyroid hormone control of the matrix. Together with the recent confirmation that actin gene, thyroidal influence over tubulin ex- astrocytes have an important support role for pression is primarily post-transcriptional (Poddar et neurones and their function in the adult (see al., 1996). Levels of actin and its mRNA have also Magistretti & Pellerin, 1999) these findings make been measured in cultures of fetal human brain this an area very worthy of further investigation. during the second trimester of gestation and it has The expression of the different isoforms of nuclear been demonstrated that there is a peak in actin thyroid receptors varies between different types of mRNA levels at week 18 of gestation, which fetal brain cells. TRα-1 is found in both neuronal corresponds with both a significant increase in and astroglial cells, whilst TRβ-1 is localized to cytoskeletal actin levels as well as the period when neuronal cells and the non-hormone-binding TRα-2 thyroid hormones in the culture medium stimulate is restricted to astrocytes (Leonard et al., 1994). It actin production (Pal, Biswas & Sarkar, 1997). has recently also been demonstrated that D2 Thyroid hormones affect cytoskeletal actin and its deiodinase is primarily expressed in glial cells, rather mRNA during the maturation, in culture, of than neurones, in the neonatal rat brain (Guadano- astrocytes obtained from neonatal rat brain and also Ferraz et al., 1997). have a distinctive influence on the external mor- The regions of the brain that mature late are those phology of such astrocytes (Paul et al., 1996). most severely affected by hypothyroidism. This Neural networks are formed by neurite growth includes the cerebellum, where the external germinal cones migrating along predictable routes to form layer persists and remains the site of mitosis for an synapses with specific target cells. The pathways of extended period in hypothyroidism. In germinal this migration are determined by proteins of the cells, thyroid hormones affect cell maturation rather extracellular matrix, especially laminin, which is than cell replication (Legrand, 1986). Myelogenesis essential for synapse formation and cell survival as is also retarded in hypothyroidism. The retardation well as neurite migration. In the brain, astrocytes effect of hypothyroidism is also manifest in mRNA participate in the formation of these migration levels for four brain proteins during postnatal pathways by synthesizing and secreting laminin, development of the rat (Strait, Zou & Oppenheimer, which is a major component of the extracellular 1992). In this study, the mRNA levels for calbindin, matrix of the developing brain. Laminin is bound to myo-inositol-1,4,5-triphosphate(ICP$)receptor,Pur- the astrocyte cell surface by specific transmembrane kinje cell protein-2 (PCP-2), and myelin basic receptors known as integrins, which in turn are held protein (MBP) are all low at birth and show a in specific locations on the astrocyte surface by the similar postnatal profile in euthyroid rats. Hypo- actin microfilament network in the cytoplasm of the thyroidism did not change the final level of these four astrocyte (Reichardt & Tomaselli, 1991). In a mRNAs but did delay their increase, such that at potentially very important paper, Farwell et al. postnatal day 15 the mRNA levels for all four brain (1995) have shown that T4-treated astrocytes in proteins were much reduced in hypothyroid com- culture readily attach to laminin whereas attach- pared to euthyroid rats. Cotransfection experiments ment was delayed for thyroid-hormone-deficient in this study suggested that the thyroid hormone astrocytes. The T4-dependent laminin-integrin in- effect on PCP-2 was mediated by the TRβ-1 isoform. teractions provide a mechanism whereby T4 might Hypothyroidism had no effect on the mRNA levels influence neuronal migration and development. This for these four brain proteins after the first four effect has similarities to the previously described postnatal weeks of development in the rat, and effects of thyroid status on the D2 deiodinase in manipulation of thyroid status of late-fetal rats also astrocytes where specific T4-enzyme-actin inter- had no effect on the mRNA levels for PCP-2 and actions are necessary for the rapid internalization of MBP proteins, suggesting that these proteins are the D2 deiodinase (Farwell & Leonard, 1992). In unresponsive to thyroid hormones prior to birth in both cases, T4 and rT3 are considerably more rats (Schwartz, Ross & Oppenheimer, 1997). Two effective than T3. thyroid response elements have been identified for It is tempting to speculate that both the continual the PCP-2 gene, one in the 5h upstream region (A1) Thyroid hormones and their effects 589 and one in the first intron (B1) and it has been rat increased during the first postnatal month proposed that the function of the A1 TRE is (Valcana & Timiras, 1969). Hypothyroidism de- suppressed by a neighbouring silencer element layed this increase in brain Na+,K+-ATPase activity operating after the initial period of thyroid hormone and the later thyroid hormone replacement was responsiveness (Zou et al., 1994; Anderson et al., begun, the less was the remediation (Schmitt & 1997). A TRE has also been characterized for MBP McDonough, 1988). The effect of T3 on the mRNA (Farsetti et al., 1992) but examination of the 1100 levels of Na+,K+-ATPase subunit isoforms has been base pair region upstream of the calbindin gene has examined in a cell culture system of rat forebrain failed to find a TRE (see Oppenheimer & Schwartz, that reproduces in vitro the pattern of normal in vivo 1997). The situation has been complicated by the development. Although T3 influenced the levels of recent finding that a knockout strain of mice, devoid mRNA for some isoforms at a culture stage equiv- of TRβ-1, which was initially proposed to mediate alent to postnatal day 7 it had no influence at a the thyroid hormone effect, have a postnatal profile culture stage equivalent to birth. Transcription rates for PCP-2 mRNA and MBP mRNA that is identical for each isoform differed markedly but remained to that in normal mice that possess a TRβ (Sandhofer stable for the 12 days of culture and were not et al., 1998). These results indicate that either it is the affected by T3. The effects of T3 on mRNA levels for TRα isoform that mediates this developmental effect, the Na+,K+-ATPase subunit isoforms are likely to be or that nuclear receptors are not involved. They also post-transcriptional (Corthesy-Theulaz et al., 1991). suggest that translation of the findings of in vitro Few studies on the developmental effects of transfection experiments to the in vivo situation thyroid hormones on the brain and its function have should be made with caution. been carried out in non-mammalian vertebrates. Hypothyroidism during development also com- One such study, however, suggests that the thyroid monly results in deafness. Postnatal rats made hormones play a role in olfactory learning and hypothyroid from late-fetal stages show no auditory imprinting in salmonid fish during the parr-smolt evoked potentials and also show no postrotatory transformation (Morin, Dodson & Dore, 1989). nystagmus which demonstrates that vestibular func- Thyroid hormones influence the development of tion is also impaired in hypothyroidism (Meza, tissues other than brain. For example, in muscle, Acuna & Escobar, 1996). Little is known of the embryonic myosin isoforms are replaced by neonatal precise cellular targets of thyroid hormone effects on isoforms which in turn are replaced by adult-type the auditory system. Effects are manifest both in the myosin heavy chain isoforms during early devel- sensory organ itself as well as in the auditory cortex. opment. In rats, the postnatal surge in plasma T4 In normal rats, cochlear potentials following audi- concentration coincides with the transition from tory stimulation have been recorded from postnatal neonatal to mature fast myosin isoforms in fast day 9 but in hypothyroidism no cochlear potentials muscles and is retarded by hypothyroidism and could be elicited during the first 30 postnatal days of precociously induced in hyperthyroidism (Gambke life (Uziel, Rabie & Marot, 1980). Hypothyroidism et al., 1983; Butler-Browne, Herlicoviez & Whalen, results in abnormal synaptogenesis of efferent bou- 1984). The times of such transitions vary between tons at the level of the cochlear outer hair cells (Uziel muscles and hypothyroidism delays, whilst hyper- et al., 1983). In the brain itself, hypothyroidism thyroidism accelerates, such transitions (d’Albis et results in a decreased number and abnormal dis- al., 1990). Similarly, in humans, hyperthyroidism tribution of apical shaft spines in pyramidal cells of has been shown to result in precocious accumulation the auditory cortex (Ruiz-Marcos et al., 1983). of adult myosin heavy chains early in postnatal These changes were similar to those found in the development (Butler-Browne, Barbet & Thornell, visual cortex, by the same investigators, but were less 1990). Although hypothyroidism retards develop- responsive to thyroid hormone replacement and it mental changes in myosin expression in chick was suggested that the auditory cortex pyramids embryo skeletal muscles (Gardahaut et al., 1992), were either more sensitive to a minor degree of replacement of neonatal myosin isoforms with adult hypothyroidism, or have a shorter critical period isoforms is not influenced by thyroid hormones in the during which hormone replacement can remedy turkey (Maruyama, Kanemaki & May, 1995). In damage, than other cerebral or cerebellar neurones amphibians, a moderate increase in thyroid hormone and structures. levels is sufficient to induce differentiation of adult Early studies showed that Na+,K+-ATPase ac- fibre types and the production of adult myosin tivity of the both the cerebra and cerebellum of the isoforms in skeletal muscle (Chanoine et al., 1987). In 590 A. J. Hulbert

# a salmonid fish, the Arctic charr Salvelinus alpinus,T3 importance of the sarcolemmal Na+\Ca + exchanger accelerates the neonatal-adult isoform transition to a diminishes and the role of the sarcoplasmic reticulum # small extent, but is not as effective as increased Ca +-ATPase (SERCA2) increases. Inhibition of temperature (Martinez et al., 1995). In the meta- thyroid function during this time results in a lower morphosing flounder, T4 exposure accelerates level of both the SERCA2 protein and its mRNA # muscle developmental changes whilst exposure to and a higher level of the sarcolemmal Na+\Ca + thyroid inhibitors delays such changes (Yamano et exchanger and its mRNA, with even greater effects al., 1991). on the respective transport activities of these two During postnatal muscle development in the rat, a proteins (Cernohorsky et al., 1998). Administration dramatic increase in the ability of the sarcoplasmic of T3 during this period had no effect on the levels # reticulum to transport Ca + has been observed. This of either protein and their respective mRNAs, # increase is not observed in hypothyroid rats but can however, it did result in an increase in the Ca + be restored by T3 (Simonides & van Hardeveld, transport activity of SERCA2 and a decrease in that # 1989). Interestingly, this developmental response of of the sarcolemmal Na+\Ca + exchanger. The muscle is unlike brain, in that it does not appear to additional effect on the activity of SERCA2 is have a critical period during which thyroid hormone suggested to be due to either changes in phos- replacement is necessary to restore normal devel- pholamban or membrane lipids (or both), whilst the # opment. Hypothyroid rats given replacement T3 at effect on the activity of the Na+\Ca + exchanger is postnatal week 6 show the same response as those suggested to be due to effects of altered Na+,K+- given replacement T3 in the first postnatal week ATPase activity or membrane lipid changes (or (Simonides & van Hardeveld, 1989). both). Thus, the thyroid hormones are necessary for Mild hypothyroidism has also been shown to the reciprocal changes in expression of the cardiac # influence muscle development in the postnatal pig, sarcolemmal Na+\Ca + exchanger and SERCA2 in which it resulted in reductions in activities of both during normal postnatal development and also have # Na+,K+-ATPase and Ca +-ATPase in several influences on their respective activities that are muscles but only limited effects on the fibre independent of changes in the expression of their composition of skeletal muscles (Harrison et al., respective genes (Cernohorsky et al., 1998). 1996). Developmental regulation of cation pumps in It has been shown that T3 is a potent stimulant for skeletal and cardiac muscle has been reviewed by embryonic myoblasts to differentiate in culture. A Dauncey & Harrison (1996). In the rat, there are transient rise in cyclic AMP levels appears essential postnatal changes in the abundance of mRNAs for for the terminal differentiation of such myoblasts different subunit isoforms for Na+,K+-ATPase dur- and T3 causes both an earlier and a greater cyclic ing the first postnatal weeks that differ from the AMP concentration increase (Marchal et al., 1995). profile in the brain and other tissues (Orlowski & The influence of the thyroid hormones on cartilage Lingrel, 1988) but it is not known if these are and bone has been recently reviewed by Williams et influenced by the postnatal T4 surge. al. (1998). When various bones and cartilage are The importance of the developmental thyroid removed from one-day-old neonatal rats and trans- hormone surge for muscle maturation has also been planted under the kidney capsule of euthyroid and shown in the precocial sheep where this T4 surge hypothyroid rats it was shown that the growth and occurs prenatally. Thyroidectomy in utero resulted in differentiation of endochondral bones was more slower axon conduction in motor neurones (possibly dependent on thyroid hormone levels than was that due to thyroid hormone effects on myelination), and of membranous bones and cartilage. The former slower and weaker contractions of fast muscle, but develop their thyroid hormone dependence before had lesser effects on slow muscles (Finkelstein et al., postnatal day 4 whilst the latter develop such 1991). Such thyroid hormone effects on isolated dependence at the end of the first postnatal week muscle function are likely to be combinations of (Liu & Nicoll, 1986). effects on myosin composition as well as sarcoplasmic The transfer of immunoglobulins across the in- # Ca + and sarcolemmal Na+\K+ dynamics. testinal mucosa is a process that diminishes during In the heart of the newborn rat, the sarcoplasmic normal postnatal maturation of the mammalian reticulum is poorly developed and the cardiac cycle intestine. In two-week-old rats, T4 administration # is more dependent on Ca +fluxes across the sar- stimulated structural and functional maturation colemma than across the sarcoplasmic reticulum resulting in a decreased fluidity of the microvillus membrane. During the first postnatal weeks the membrane and diminishing uptake of immuno- Thyroid hormones and their effects 591 globulin (Israel et al., 1987). In the rat, there are scription factors, some are involved in metabolism significant changes in the activities of a number of whilst others are involved with extracellular sig- intestinal enzymes in the third postnatal week. Some nalling (Shi, 1996). The role of two of these latter enzymes decrease (e.g. lactase and acid β-galacto- genes has recently been discussed (Stolow et al., sidase), whilst others increase in activity (e.g. sucrase 1997) and these authors have stressed the important and glucoamylase). As well as plasma T4 con- role of the extracellular matrix and the transfer of centration, plasma corticosterone concentration also information between different cell types during increases during the second postnatal week and is an development. For example, the development of the important factor in this postnatal maturation of adult epithelium requires the presence of larval intestinal enzymes. Treatment with T4 increases connective tissue (Ishizuya-Oka & Shimozawa, serum corticosterone levels during this period and 1992) and without it, the only observed morpho- also synergistically heightens the effects of exogenous logical change in cultures of intestine treated with glucocorticoid on intestinal enzyme maturation thyroid hormones is apoptosis of the larval epi- although it does not in itself enhance maturation of thelium. The proliferation and differentiation of the intestinal enzymes (McDonald & Henning, 1992). adult epithelium is absolutely dependent on the That the thyroid is generally involved in the normal presence of larval connective tissue. This may present maturation of the vertebrate digestive system is an interesting example for other cases of devel- illustrated by the finding that exposure to T4 opmental changes. If cell to cell information transfer stimulates precocious development of the gastric is an important component of developmental gland in premetamorphic flounder whilst inhibition changes then thyroid hormone effects that enhance of thyroid function inhibits gastric development such transfer may speed up development; inhibition during flounder metamorphosis (Miwa, Yamano & of such information transfer would be likely to retard Inui, 1992). developmental change. This means that thyroid An interaction between corticosterone and thyroid hormone effects need not be genomic to influence hormones has also been observed with respect to development but could for instance act at a amphibian metamorphosis and has recently been membrane level to speed up information transfer reviewed by Hayes (1997). Exogenous corticoids in and consequently development. Interactions be- early development inhibit metamorphosis whilst in tween cells and the extracellular matrix can result in late development they accelerate metamorphosis. transcriptional regulation of cellular genes (see They interact with the thyroid axis in multiple ways. Stolow et al., 1997). For example, in erythrocytes from premetamorphic Apoptosis is also observed in other tissues (notably tadpoles, the normal thyroid hormone stimulation the tadpole tail) during amphibian metamorphosis of TRα number is inhibited by corticosterone (for review see Tata, 1994). Thyroid hormones play (Schneider & Galton, 1995), whilst it has been two distinctive roles during cell death in the tadpole shown that corticosterone stimulates 5h-deiodination tail. They induce and promote keratinization (pro- but inhibits 5-deiodination in premetamorphic tad- grammed cell death with terminal differentiation) in poles resulting in an increased conversion of T4 to epidermal cells of the tail and body, and influence T3 and thus an increase in body T3 levels (Galton, proliferation of these two types of cells in opposite 1990). Corticoids can also accelerate or inhibit directions. T3 inhibits DNA synthesis in tail cells, tadpole hindlimb growth and development de- but stimulates the same process in body cells pending on the stage of spontaneous development (Nishikawa, Kaiho & Yoshizato, 1989). Whilst T3 (Wright et al., 1994). Such hormonal interactions are accelerates keratinization of body cells, these cells the likely mediators of ecological and environmental will gradually keratinize in vitro without T3. Simi- effects on amphibian development. Cortisol and T3 larly, expression of the adult-type keratin gene in also interact in larval development of a fish species head skin during larval development of Xenopus laevis (Kim & Brown, 1997). has been demonstrated to be activated by two steps: Tissue remodelling during amphibian metamor- one independent and one dependent on T3 phosis can be extensive. Metamorphic changes in the (Mathisen & Miller, 1987). amphibian intestine involve the programmed cell One of the changes that occur during amphibian death (apoptosis) of tadpole epithelium and its metamorphosis is the development of the capacity replacement by newly formed adult epithelium. Of for urea biosynthesis in the adult amphibian liver. A several genes whose expression is increased early key enzyme for urea biosynthesis is carbamyl during amphibian metamorphosis, some are tran- phosphate synthetase (CPS); CPS mRNA levels and 592 A. J. Hulbert

CPS enzyme activity are increased when pre- where they predominantly speed up normal pro- metamorphic tadpoles are exposed to T4 (Morris, cesses when present in excess, and mostly slow down 1987) and T3 (Galton et al., 1991a). Thyroid normal processes when not present in adequate hormones have the opposite effects in rat liver in that amounts. urea biosynthesis is stimulated by hypothyroidism The influence of thyroid hormones on vertebrate (Marti et al., 1988). Interestingly, whilst T4 stim- development is likely to be via several different ulates synthesis of the gluconeogenic enzyme phos- modes of action. Many effects are likely to be direct phoenolpyruvate carboxykinase in rat liver, it has genomic effects, whilst several others are probably the opposite effect in tadpole liver (Morris, 1987). non-genomic. Differentiating between such modes Recently, it has been demonstrated that thyroid will involve several criteria. The observation that hormone induction of enzymes for urea synthesis thyroid hormone exposure results in elevated mRNA (including CPS) during both spontaneous and levels for a particular gene is, by itself, not adequate induced metamorphosis of Rana catesbeiana is a evidence for a direct genomic mode of action. Direct secondary response due to a transcription factor in genomic effects will be mediated by nuclear receptors the amphibian liver that is homologous to the binding to TREs in the genome and thus a TRE is mammalian transcription factor C\EBPα.The a mandatory requirement for a direct genomic mRNA for this transcription factor accumulates action. The role of thyroid effects on the membrane early in response to thyroid hormones, and the bilayers and changes in the rate of information product of this mRNA can bind to and transactivate transfer on development has not been investigated. the promoters of CPS and ornithine trans- This is presumably because of the widely held belief carbamylase (another urea biosynthesis enzyme) in that thyroid hormones only act via the nuclear the liver of Rana catesbeiana (Chen & Atkinson, 1997). receptor mode. These results suggest that there is a cascade of gene Membrane lipid compositional changes during activation induced by thyroid hormones during development of vertebrates are known but have been amphibian metamorphosis. Intriguingly, examin- little researched. For example, Hoch (1988) reports ation of thyroid hormone influence on the levels of an increase in the unsaturation index of lipids the C\EBPα and C\EBPβ proteins, as well as their (mainly due to an increased PUFA content) in the respective mRNAs, in the liver of the rat during liver of Rana catesbeiana during metamorphosis that is early postnatal development suggests that the in- coincident with the T4 surge. Changes in membrane fluence of thyroid hormones on these transcription acyl composition, cholesterol:phospholipid ratio and factors is either translational or post-translational membrane fluidity of various membranes have been (Menendez-Hurtado et al., 1997). Stimulation of observed during postnatal development of the rat malic enzyme in cultured chick embryo hepatocytes (e.g. Hubner et al., 1988; Rovinski & Hosein, 1983; by T3 has also been suggested to occur primarily at Kameyama et al., 1987) and the rabbit (e.g. a post-transcriptional level (Back et al., 1986). Nagatomo, Sasaki & Konishi, 1984; Ricardo et al., Many developmental changes that have been 1986; Schwartz, Lambert & Medow, 1992) as well examined appear to be associated with the de- as during embryonic development of the chicken velopmental surges of T4 and T3 levels. My (e.g. Schjeide, 1988; Fuhrmann & Sallmann, 1996; impressions from the literature concerning thyroid Rivas et al., 1996). For example, in the rat there is a hormone influences on development is that when postnatal peak in the unsaturation of liver mito- single time points in development are investigated chondrial phospholipids (Pollack & Harsas, 1981) the conclusions are that thyroid hormones act, in that is consistent with the developmental T4 surge. some cases, to increase and in others to decrease the Similarly, there is a postnatal peak in the 20:4 expression of genes, sometimes even the same genes content of plasma membrane phospholipids, as well at different times. However, when multiple time as microsomal and mitochondrial phospholipids points are examined, a more consistent pattern isolated from chick heart (Kutchai et al., 1978). emerges. This is that exposure to exogenous thyroid However, the role of thyroid hormones in these hormones generally accelerates development, whilst changes in membrane bilayer composition and their inhibition of thyroid function retards developmental functional consequences are not known. change. The acceleration of normal development One of the most interesting recent findings is with increased levels of thyroid hormone and thyroid hormone enhancement of the laminin- retardation with thyroid hormone deficiency is, in a integrin-actin interaction observed in astrocytes way, analogous to their effects in adult vertebrates (Farwell et al., 1995; Leonard & Farwell, 1997). Thyroid hormones and their effects 593

Whether this phenomenon is also applicable to other Binding to ALB is weak and insensitive to the cells and their interaction with the extracellular stereochemistry of the alanine side chain with the matrix, and whether thyroid-hormone-membrane major binding feature being the o-diiodophenolic interactions have a role in this effect is, in my structure in which the phenolic kOH group is opinion, one of the most fascinating and important ionized. Binding to TTR is highly sensitive to the questions regarding the developmental effects of stereochemistry of the alanine side chain with thyroid hormones yet to be answered. In this effect, binding being primarily to the carboxylate ion of the T4 is much more potent than T3, and rT3 is nearly side chain with the ammonium group weakening as potent as T4. In view of the importance of cell such binding. The o-diiodophenolic structure is also migration during development, this thyroid hor- important for binding to TTR with the 3h and 5h mone effect may be related to a number of findings iodine atoms being very important, together with a in vertebrates: (i) the presence of T4 in the eggs and hydrogen bond formed between the ionized form of embryos of vertebrates well before the development the phenolic kOH group and TTR N-terminal and maturation of thyroid gland function in the residues (which reside at the entrance to the TTR individual, (ii) the developmental surge of T4, and binding channel). In this light, it is of interest that (iii) the preponderance of inner ring 5-deiodination the N-termini are longer and more hydrophobic in early in development and the switch over to outer avian TTR compared to rat, sheep and human ring 5h-deiodination, with the consequence of high TTR, and that avian TTRs have a greater affinity levels of rT3 early in vertebrate development and its for T3 than the mammalian TTRs (Chang et al., decrease and replacement by T3 later at approx- 1998). Thyroid hormones bind to TBG with great imately the time of the developmental T4 surge. affinity because this protein associates with both the carboxylate and the ammonium groups of the alanine side chain. As with TTR, the 4hkOH in its ionized form together with the 3 and 5 iodine atoms V. SOME PERSPECTIVES h h are important in the binding of thyroid hormones to TBG (Jorgensen, 1981). (1) Analogues as a source of knowledge The in vitro binding of thyroid hormones and As well as natural analogues of T4 (i.e. T3, rT3, the analogues to thyroid nuclear receptors (TRs) has three T2s, the two T1s, and thyronine and other been examined for intact rat liver nuclei and its products of T4 metabolism) a substantial number of solubilized receptor. The two preparations closely other analogues have been synthesised and examined agree and both binding affinities are shown in Table for their thyroid-hormone-like potencies. A sub- 4. Whereas T3 binds to the plasma proteins with stantial amount of this work has been carried out by much less affinity than T4 (TBG, TTR and ALB Jorgensen and his colleagues and an excellent review have affinities for T3 that are respectively 9, 5 and of these findings is provided by Jorgensen (1981). 55% of their affinity for T4), the opposite is the case For additional reviews, see Cody (1996) and Craik, with binding to thyroid nuclear receptors where the Duggan & Munro (1996). Coupled with the study of affinity for T3 is approximately 800% of that for T4. the conformation of analogues, these studies have TRs have less affinity for the -isomer of the alanine resulted in considerable insight into: (i) important side chain than the -form. Binding is to the carboxyl- aspects of the molecular interactions during the ate ion of the alanine side chain, with the presence binding of thyroid hormones to various proteins, (ii) of an amino group reducing binding. Two aromatic the relative importance of the various parts of the rings separated by either an O, S or a C (with an thyroid hormone molecules in their physiological inter-ring bond angle of approximately 120m)is effects, and (iii) the mode of action for various effects also important. Non-polar groups on the 3- and of thyroid hormones. 5- positions on the inner ring are important for These analogues bind with different affinities to a binding to TRs. The greatest affinities are for iodine variety of proteins, which include both the extra- atoms in these positions, and their replacement with cellular distributor proteins and intracellular pro- smaller groups (e.g. Br atoms or kCH$ groups) teins (notably nuclear thyroid receptors). Although gives lower affinities. The role of these inner ring binding affinities vary between vertebrate species, iodine atoms appears to be to hinder rotation around detailed measurement has generally been restricted the ether linkage and thus restrain the two aromatic to the human or rat proteins. Such relative binding rings such that they are perpendicular to each other. affinities are presented in Table 4. This role can be easily observed in the molecular 594 A. J. Hulbert % & " $ ' . (1997) . (1976) # . (1996), . (1995) . (1964) . (1989) . (1989) et al C. .(1967) et al m et al et al et al et al et al et al Lanni Moreno Pitt-Rivers & Tata (1959) hormone at 36  µ -T1 T Reference h . (1993). et al -T2 3-T1 3 h ,5 h M hormone concentration. µ 100 Pitt-Rivers & Tata (1959), ,3-T2 3 h " g per 100 g mass injection. µ 50 0 Cash 10 Farias 100 -ATPase by Warnick " " + " # 70 Segal (1989) 100 " " 100), with relative potencies measured as less than 0.5 shown as zero. l 40 100 60 100 " " 96 100 0 108 0 0 18 Horst 70 32 1610 10060 3 100 35 1 50 13 0 0 0 Jorgensen (1981), Jorgensen (1981) 34 13 100 20 48 15 103 5 28 3 0 Goldfine 132 0 100 0 0 100 Davis 180 100 1200 100 400 Silva & Leonard (1985) " " " ) resulted in similar growth of several body organs as well as body mass and metabolic rate. " " − g day µ (St Germain, 1985). -ATPase + (rat) # # sugar in vitro in vitro (rat) (rat) (liposomes) in vivo The relative potencies of iodothyronines in binding to selected proteins and exerting selected thyroid hormone effects (rat) in vitro consumption (rat) in vivo in vitro uptake (rat) (by immersion) (by injection) in vitro uptake TBG (human serum)TTR (human serum)ALB (human serum)Intact nuclei (rat liver)Soluble TR (rat liver) 1100 2100Antigoitre 600 assay (rat) 100Body growth 56 180 (rat) 100 13 182 100 14 422 4 100 167 18 100 182 100 3 1 100 2 55 0 14 100 15 7 0 1 48 0 0 155 0 0 1 1 0 1 1 27 0 0 0 0 0 Jorgensen Jorgensen (1981) (1981) Jorgensen (1981) 0 Jorgensen (1981) Jorgensen (1981) Jorgensen (1981) Evans BMR(rat)Isolated liver O 70Brain 12 D2 deiodinase 100 Membrane antioxidant Muscle Amphibian metamorphosis Amphibian metamorphosis Membrane fluidity Thymocyte amino acid Erythrocyte Ca Values for effect on cortex D2 deiodinaseCalculated activity, from similar relative relative change potenciesCalculated in measured from membrane for relative fluidity pituitary % of D2 inhibition dimyristoyl-phosphatidylcholine deiodinase of liposomes activity liver due (Silva mitochondrial to & membrane 50 Leonard, lipids 1985) at 1 Both T4 and T3Calculated (at as 0.1 average values for diaphragm, atria and ventricles after 0.1 Similar relative potencies also reported for sarcoplasmic reticulum Ca % and GH3 cells & ' Table 4. IodothyronineRelative binding affinity L-T4Thyroid hormone D-T4 effect L-T3 D-T3 rT3 3,5-T2 3 All potencies are expressed" relative to L-T3# ( T, thyronine; TBG, thyroxine binding globulin; TTR, transthyretin; ALB, albumin; TR, thyroid nuclear receptor; BMR, basal metabolic rate. $ Thyroid hormones and their effects 595 models for T4, T3 and 3,5-T2 in Fig. 1 (compared to only 0.4% as potent as T4. Alteration of the alanine the model of rT3). On the phenolic ring, the 4hkOH side chain produced a series of analogues that were group is also important (in the non-ionized form), considerably more potent than T4 in stimulating and affinity is enhanced by a lipophilic halogen, metamorphosis but considerably less potent than T4 alkyl, or aryl group adjacent to this 4hkOH group. in preventing goitre or stimulating metabolic ac- Strongest binding is obtained when the 3h-position is tivity. Whilst some of the high relative potencies for occupied by either an iodine atom or the similar- stimulation of metamorphosis obtained for these sized isopropyl group. later analogues were measured by in vitro exposure, One of the in vivo tests of the thyromimetic activity and lesser relative potency values were obtained of thyroid hormone analogues is the rat goitre following in vivo injection into tadpoles (see Jor- prevention test, which is a measure of the amount of gensen, 1981), they still remained more potent than analogue required to inhibit thyroid gland en- T4 in this thyromimetic property but less potent largement in rats in which thyroid gland function is than T4 in the other assays. Whilst comparison of inhibited and a goitre is developing. It is a measure the stimulation of metamorphosis with the other two of the potency of the analogue to inhibit TSH release assays is complicated by the fact that different from the anterior pituitary and is also given in Table species are involved, this is not the case when 4. There is a strong correlation between the relative comparing the other two assay systems, which were affinity of nuclear TRs for an analogue and its both measured in the rat. The finding that analogues potency in the rat antigoitre assay (Dietrich et al., have different relative potencies when different 1977). Discrepancies between TR binding affinity thyroid hormone effects are measured is evidence for and the thyromimetic potency of particular that the effects themselves probably involve different analogues have been variously explained. For ex- modes of action. ample, the discrepancy between the low TR affinity Whilst the nuclear TR has an affinity for T3 that of analogues lacking the 4hkOH but their relatively is 800% that for T4, T3 has a thyromimetic potency greater thyromimetic potency has been explained by for stimulating growth and metabolism in rats that is the in vivo 4h-hydroxylation of such analogues similar to that for thyroxine (Table 4). This has been (Oppenheimer, 1983). The relatively low thyro- explained by assuming that T4 is converted to T3 to mimetic potency of Triac (the acetic acid analogue exert all its effects (see Oppenheimer, 1983). Another of T3) compared to the very high TR affinity for it, explanation is that thyroid hormone effects on has been explained by its more rapid metabolism metabolism and growth are not predominantly compared to T3 (Oppenheimer, 1983). mediated by thyroid nuclear receptors. In contrast The goitre prevention assay tests only one of the to their almost equivalent potency in stimulating myriad effects of thyroid hormones. Other assays of metabolism and growth, T3 is approximately four- thyromimetic potency that have been used in the fold more potent than T4 in suppression of rat serum past are the stimulation of metabolic rate of rats and TSH levels and the non-halogenated analogue, 3,5- the stimulation of amphibian metamorphosis. An dimethyl-3h-isopropyl-thyronine has approximately earlier compilation of analogue potency included 10% of the potency of T4 (Chopra et al., 1984). The these other two assays (Pitt-Rivers & Tata, 1959) relative potencies of these two analogues in the goitre and showed that the relative potency of some prevention assay are similar to the relative TR analogues varied between the types of assay (Fig. affinities (Jorgensen, 1981). That the thyroid stimu- 12). As can be seen from Fig. 12, there are some large lation of metabolic activity is not predominantly discrepancies among assays. For example, whilst T3 mediated by nuclear receptors is also supported by had a thyromimetic potency that was 400–500% the fact in humans that doses of Triac sufficient to that of T4 in the goitre prevention and the suppress TSH secretion had no effect on metabolic metamorphosis assay, it was only 67% more potent rate whilst doses of T4 sufficient to suppress TSH than T4 in the metabolic rate assay. Whilst 3,5,3h- secretion resulted in a significant stimulation of tribromothyronine had 20–28% of the T4 potency metabolic activity (Bracco et al., 1993). in goitre prevention and stimulation of metamor- The relative potencies of iodothyronines in several phosis it was 90% as potent as T4 in stimulating other effects are also presented in Table 4. These metabolic rate. Conversely, 3,5-diiodo-3h,5h-di- relative potencies differ markedly from the relative chlorothyronine was 22–48% as potent as T4 in binding affinity of nuclear receptors for the iodo- stimulating metamorphosis and preventing goitre thyronines and thus imply that these effects are not but had almost no effect on metabolic rate being mediated by nuclear receptors. Within this group 596 A. J. Hulbert

Goitre prevention Growth/differentiation Oxygen consumption

L-thyroxine (L-T4) ! 3,5,3 -triiodo-L-thyronine (L-T3) ! 3,5,3 -tribromo-DL-thyronine ! ! 3,5-diiodo-3 ,5 -dichloro-DL-thyronine ! ! 3,5,3 ,5 -tetraiodothyropropionic acid ! 3,5,3 -triiodothyropropionic acid

3,5-diiodothyropropionic acid ! ! 3,5,3 ,5 -tetraiodothyroacrylic acid ! ! 3,5,3 ,5 -tetraiodothyroacetic acid ! 3,5,3 -triiodothyroacetic acid

0.1 1 10 100 1000 10000 100000 Potency relative to L-thyroxine (= 100)

Fig. 12. A comparison of the potencies (relative to T4) of various thyroid hormone analogues on three different thyroid hormone effects. Data are from Pitt-Rivers & Tata (1959). The goitre prevention and oxygen consumption assays were performed in rats. The growth\differentiation assay involves induction of metamorphosis in amphibians.

there appear to be several different patterns. Meta- ronine yet it is the most potent stimulator of muscle # bolic rate and nutrient uptake may represent one sarcoplasmic Ca +-ATPase activity, and although # group, Ca +-ATPase stimulation another, and the there are substantial differences between TR af- effects on D2 deiodinase still another. This latter finities for T4 and T3, these two iodothyronines have # effect appears to involve the same mechanism as the almost equal influence on the activity of this Ca +- interaction of astrocytes with the extracellular ATPase. Whilst the acetic acid analogue of T3 has # matrix. T4 and rT3 may turn out to be more no significant influence on Ca +-ATPase activity, the important in the early development of vertebrates nuclear TR has a very high affinity for this analogue than T3. (Warnick et al., 1993). Such effects would not appear In Table 4, I have included two other effects of the to be initiated via nuclear TRs. thyroid hormones for which the relative potencies were obtained at high concentrations. These are the only data available for these effects and the relative (2) The early sixties revisited potencies for the various iodothyronines may change when they are measured at lower, more physio- It is the purpose of this section to use the wisdom of logically relevant, concentrations. hindsight to re-examine the seminal studies of Tata I have argued elsewhere in this review that the et al. (1962, 1963), Tata (1963) and Tata & Widnell effects due to 3,5-T2 may not be physiologically (1966). These detailed the time course of changes in relevant because plasma T2 levels are considerably a substantial number of variables following T3 lower than those of T4 and T3. However, the finding injection into thyroidectomised rats. Our under- that T2 is almost as active as T3 in exerting various standing of biological systems has increased greatly thyroid hormone effects, yet TRs show negligible since this time. For example, the beginning of the binding affinity for it, raises serious questions about modern understanding of fluid membranes and their the assumption that most thyroid hormone effects importance can be dated to almost a decade later are mediated by a TR mode of action. (Singer & Nicolson, 1972). Similarly, the often There are several other effects of thyroid hormones misleading but powerful concept of ‘rate-limiting where the relative potency of analogues does not enzymes’ permeated the understanding of the agree with their TR affinity. For example, TRs show control of metabolism at the time. The more negligible affinity for 3,5-dimethyl-3h-isopropyl-thy- sophisticated understandings from ‘metabolic con- Thyroid hormones and their effects 597 trol analysis’ which now guide our thinking also had ment, however, does not exclude other inter- their birth about a decade later (see Fell, 1997). pretations including changes in membrane per- Tata and colleagues conducted their experiments meability for these labelled substrates. It is quite within a paradigm that suggested that the primary feasible that thyroid hormone changes in the uptake action of thyroid hormones was on mitochondria dynamics of orotic acid into liver nuclei would result and that these hormones acted to uncouple ATP in increased measured incorporation into nuclear production from oxygen consumption and substrate RNA. Such an interpretation is supported by the utilization. They showed that when more physio- fact that although there was a 300% increase in "% logically relevant concentrations of thyroid hor- C-labelled orotic acid incorporation into nuclear mones were used in vivo that the mitochondria from RNA there was no change in the amount of nuclear both rat liver and muscle exhibited no change in RNA following T3 injection (see Fig. 7 in Tata & their P\O ratio. This was correctly interpreted to Widnell, 1966). In addition, the results of their study show that thyroid hormones did not result in showed that after in vivo T3 injection, ribo- uncoupled mitochondria. However, they also nucleoprotein particles, or the mRNA attached to showed that thyroid hormones stimulated liver and them, or both, were more firmly bound to micro- muscle mitochondrial respiration when a phosphate somal membranes. Thus, the results from these acceptor system is absent and thus provided early experiments can also be interpreted to support a evidence that thyroid hormones increased mitochon- membrane site of action for the thyroid hormones. drial proton leak, an effect studied extensively by Brand and others and discussed in Section III.4. Tata et al. (1963) also included the time course of (3) Knockouts from the nineties thyroid hormone influences on a number of non- mitochondrial enzymes presented in a series of One of the major advances made possible by the Tables and concluded: ‘it appears from our results revolution in molecular biology is the capacity to that enzymes or functions linked firmly to mem- interfere with specific genes. In particular, the ability branous subcellular structure are more markedly to produce ‘knockout mice’ has provided a new affected during the early phase of thyroid hormone experimental tool. Of relevance to the current review action than activities not dependent on structural are the recently published findings for a number of integrity’ (Tata et al., 1963, p. 426). Some of their knockout mice strains, in which the genes for either results are given in graphical form in Fig. 13. In this the TRα receptors, the TRβ receptors or the thyroid form, it becomes obvious that the injection of T3 hormone binding protein, TTR, have been made resulted in substantial stimulation of those enzymes non-functional. associated with microsomal membranes but had no The TRβ knockout mouse strain has a recessive consistent or substantial influence on those enzymes gene for nonfunctional TRβs (Forrest et al., 1996c). present in the cytosol. They also showed that T3 Heterozygous mice have functional TRβ receptors increased the efficiency of amino acid incorporation whilst homozygote recessives have neither functional into microsomally synthesised proteins in liver but TRβ-1 nor TRβ-2. There is no change in TRα that T3 resulted in no change in tissue RNA expression. Use of behavioural tests that had concentration. previously demonstrated learning disabilities in the The interpretation of the time course of thyroid hypothyroid (hyt) mouse (Anthony, Adams & Stein, hormone effects on RNA synthesis in rats (Tata & 1993) showed that there were no differences between Widnell, 1966) further focussed attention on the mice possessing and mice lacking functional TRβs nucleus as the site of thyroid hormone action. The (Forrest et al., 1996b). Histological and histo- most dramatic increases were in (i) the incorporation chemical analysis of the central nervous system "% of C-labelled orotic acid into nuclear RNA, (ii) an revealed no obvious abnormalities in brain anatomy, $# increase in [ P]phosphate into ribosomal RNA, (iii) including the cerebellum and hippocampus and "% an increased incorporation of C-labelled amino structures known to be thyroid hormone sensitive acids into microsomal protein. Other effects included (Forrest et al., 1996b). When homozygous mice were increased RNA polymerase activity that peaked tested for their hearing ability, they were observed to 30–45 h after T3 injection. The sequence of these have a permanent deficit in auditory function over a responses (especially the first) were very important wide range of frequencies, and as the auditory- evidence favouring a nuclear site of action for the evoked brainstem response had a normal waveform, thyroid hormones. The manner of their measure- although greatly diminished, it was suggested that 598 A. J. Hulbert

250 G-6-Pase NADPH cyt c reductase LDH (liver) 200 IsoCDH (liver) G6PDH (liver) CrPKase (muscle) RNA:DNA 150 amino acid incorporation

100 Per cent increase

50

0

050100 150 200 250 Time after T3 injection (h) Fig. 13. Time course following T3 injection, of the activities of microsomal (solid symbols and unbroken lines) and cytosolic (open symbols and broken lines) enzymic activities and liver RNA:DNA (open square and unbroken lines) in the rat. Data are from Tata et al. (1963). G-6-Pase, glucose-6-phosphatase; NADPH cyt c reductase, NADPH cyto- chrome C reductase; LDH, lactate dehydrogenase; IsoCDH, Isocitrate dehydrogenase; G6PDH, glucose-6-phosphate dehydrogenase; CrPKase, creatine phosphokinase.

the primary deficit was at the cochlear level. They Elevated TSH is normally inconsistent with high exhibited no other neurological defects (specifically serum thyroid hormone levels and similar to the no vestibular defects) and the developmental defect ‘resistance to thyroid hormone’ syndrome observed in hearing was not influenced by the maternal in humans. These mice also had enlarged thyroid uterine environment (Forrest et al., 1996a). Such glands (Forrest et al., 1966c), consistent with their TRβ-deficient mice appear to grow normally and high TSH levels. More recently, it has been shown become normally fertile adults. Litters were in that such TRβ-null mice respond normally to normal Mendelian ratios which showed there was no hypothyroidism (induced by dietary iodine de- differential prenatal mortality (Forrest et al., 1996b). ficiency) with an increased secretion of TSH but that TRβ-null mice have a postnatal profile for PCP-2 they are deficient in their response to injection of mRNA and MBP mRNA that is identical to that of thyroid hormones. Whereas mice with functional normal mice (Sandhofer et al., 1998). Other than the TRβ will completely suppress TSH secretion fol- hearing defect, the only other (so far) reported lowing thyroid hormone injection, TRβ-null mice abnormality relates to the thyroid hormone regu- are unable to suppress TSH levels below those lation axis, especially TSH responsiveness. observed in normal euthyroid mice (Weiss et al., Measurement of serum TSH levels in TRβ-null 1997). Thus, TRβ-null mice exist in a hyperthyroid mice showed them to have 3–4 times the serum TSH condition. concentration of mice with functional TRβs and Two different TRα knockout strains have been similar differences were observed in serum T4 and created. In one strain, only the TRα-1 isoform is T3 levels (both for total and free hormone levels). nonfunctional (Wikstrom et al., 1998), whilst in the Thyroid hormones and their effects 599 other strain both TRα–1 and the non-thyroid- type controls. They had very small thyroid glands in hormone binding c-erbAα-2 are both nonfunctional which some follicle cells exhibited vacuolar de- (Fraichard et al., 1997). I will refer to them here as generation. By postnatal week 5 this hypothyroidism TRα1-null and TRα-null mice, respectively. in the TRα-null mice had become extremely severe The TRα1-null mouse is a homozygous recessive in that total T4 and total T3 levels were respectively condition. Such mice are born in the expected 90% and 60% reduced compared to wild-type Mendelian ratio suggesting no differential prenatal controls and serum free T4 concentration was below mortality. They are viable and survive to at least 18 the limits of detection (Fraichard et al., 1997). months of age with both males and females being Although they had not exhibited any growth from fertile and producing litters of a normal size. The postnatal day 10, from postnatal week 4 onwards animals appear healthy with no overt abnormalities TRα-null mice actually lost 30–50% of their body detected at autopsy (Wikstrom et al., 1998). No mass and consequently died. Survival time varied compensatory changes in TRβ expression are ob- with litter size and was probably influenced by served in TRα1-null mice. Such TRα1-null mice, at lactational energy transfer from the mother. 2–4 months of age, show a reduction in serum TSH Intriguingly, a very small number of TRα-null levels with no significant reduction in free T3 mice (approximately 1%) spontaneously survived concentration but an approximately 30% reduction for up to 3–7 months. When TRα-null mice were in free T4 levels. They are thus slightly hypothyroid given T3 for a short period at 3 weeks of age, a (Wikstrom et al., 1998). substantial number (40%) survived for a further 3–6 The TRα1-null mice also exhibited significant months resuming a near-normal growth rate; when phenotypic variation in that they had heart rates serum was measured several weeks after the T3 that were nearly 20% lower, and a body tem- injections, they had the same levels of total T4 and perature that was on average 0.5 mC lower than T3 as wild-type control mice (Fraichard et al., 1997). wild-type controls (Wikstrom et al., 1998). However, These findings together with those of Weiss et al. whether this is due to the absence of TRα-1 or due (1997), suggest that TRα-1, but not TRβ,is to the hypothyroid state of these mice is unknown. responsible for the up-regulation of TSH levels and That injections of T3 into TRα1-null mice resulted that both TRα and TRβ are involved in the down- in an increase in both heart rate and body regulation of TSH levels, with TRβ being more temperature illustrates that they were reponsive to potent than TRα (in that it is able to inhibit TSH thyroid hormone despite the absence of TRα-1, secretion completely). suggesting that the second explanation is most likely. These results suggest the following scenario. The TRα-null mouse is also a homozygous Without both TRαs there is little likelihood that recessive condition. Young are born according to a serum thyroid hormones will reach a level where normal Mendelian ratio suggesting no differential TRβ alone can regulate their concentrations. If prenatal mortality. Similar to the TRα1-null mice, thyroid hormone levels can be elevated artificially to there was no compensatory change in TRβ ex- such a level, then TRβ alone can regulate thyroid pression. They exhibited no obvious external pheno- hormone levels by negative feedback. The difference type until postnatal week 2 when they ceased between TRα1-null mice and TRα-null mice sug- growing. At three weeks of age, although they had gests an important role for TRα2 in that it may ceased growing, TRα-null mice appeared healthy stimulate TSH release irrespective of serum thyroid and exhibited normal behaviour. Their internal hormone levels. It may be an important component tissues did not display any overt abnormalities and of the rising phase of the T4 surge during de- no cellular or morphological abnormalities could be velopment. The role of TRH in this feedback system detected. The cerebral cortex showed normal lami- is presently unknown. These findings give some nar organization, cytoarchitectonics and cortical insight into the developmental T4 surge, which in parcellation with the organization of subcortical normal mice occurs during these first postnatal nuclei being normal (Fraichard et al., 1997). How- weeks, and suggests that it may be an important ever, at three weeks of age they appeared more process involved in the ‘kicking in’ of the regulatory severely hypothyroid than the TRα1-null mice. system which then homeostatically regulates a They had mRNA levels for βTSH that were relatively constant plasma level of thyroid hormones approximately 70% reduced compared to wild-type during adult life. It also suggests that the two controls and serum levels of total T4, free T4 and different receptor isoforms may have different roles total T3 that were 40% lower than those of wild- during this surge; the TRα isoforms being re- 600 A. J. Hulbert sponsible for the rising phase whereas the TRβ siblings, there was no difference in the plasma levels isoform is more important in the falling phase of the of either free T4 or free T3 nor in plasma TSH surge. concentrations. Similarly, the activities of three Similarly, the TRα-null mice showed phenotypic thyroid-hormone-sensitive enzymes were unaffected effects. These were only manifest after the first in TTR-null mice (Palha et al., 1994). When the postnatal week and included delayed maturation of transfer kinetics of tissue uptake of labelled T4 was the small intestine as well as delayed bone de- examined in such mice it was observed that although velopment. That these were not direct effects due to the T4 clearance from plasma to brain was reduced the absence of TRαs but were more likely to be due by approximately 40%, and measured brain T4 to the postnatal hypothyroidism is illustrated by the content was reduced by approximately 30% in fact that TRα-null mice that were given T3 TTR-null mice, the T3 content of brain was not injections also had normal maturation of small significantly different compared to wild-type sib- intestine and bone following these injections (Frai- lings. Tissue T4 content was not significantly affected chard et al., 1997). in liver and kidney (Palha et al., 1997). Such results Differentiation between genomic and non-gen- are indicative of redundancy of plasma proteins and omic modes of thyroid hormone action could be emphasize the importance of both the ‘free hormone’ determined if mice in which both TRα and TRβ are concept and the brain D2 deiodinase. TTR also nonfunctional were examined. Because of the un- normally transports plasma retinol complexed with doubted role of both receptor isoforms in the plasma retinol-binding protein; the transport and hypothalamus-anterior-pituitary-thyroid axis, it metabolism of retinol in TTR-null mice is reviewed would be necessary to maintain plasma T4 levels by Wolf (1995). artificially (including its developmental surge) in such animals. The TRβ-1 receptor has also been deleted in a (4) Resistance to thyroid hormones GH3 cell culture system with the use of an antisense RNA vector directed against TRβ-1. Similar to the In 1964, a young deaf-mute girl was involved in a situation observed in knockout mice, there were no road accident in Los Angeles and when she was X- compensatory changes observed in TRα expression. rayed in hospital, it was observed that all major However, within the TRβ isoforms, there were secondary ossification centres in her bones were compensatory changes observed. In GH3 cells that stippled. This observation together with a slight had TRβ-1 isoform expression repressed, there was enlargement of her thyroid gland suggested neonatal an increase in TRβ-2 expression and also an increase hypothyroidism but her serum protein-bound iodine in basal expression, as well as the T3-stimulated was unexpectedly high and found to be in the expression of the growth hormone gene (Ball et al., hyperthyroid range. A follow-up examination of her 1997). Such a finding suggests that within cells there siblings and parents resulted in the first description are TR-isoform-specific expressions of thyroid hor- of the condition, now known as ‘resistance to thyroid mones and that one TR isoform may not compensate hormones’ or RTH (Refetoff, DeWind & DeGroot, for another. 1967; Refetoff, 1994). There have been several Another knockout mouse of relevance to this excellent recent reviews of this condition (Usala, review is the TTR-null mouse. This condition is also 1995; Refetoff, 1996; Chatterjee, 1997; Beck-Pecoz, a homozygous recessive condition and such mice are Asteria & Mannavola, 1997). Since the original born in the expected Mendelian ratio suggesting no description, several hundred patients have been differential influence on fetal development. TTR- recognised and categorised into two types of RTH; null mice display no obvious postnatal phenotypic generalized RTH (i.e. GRTH) in which the subjects abnormalities as determined morphologically and have (i) elevated levels of free thyroid hormones, (ii) by histopathological analysis. They exhibit the same inappropriately normal serum TSH levels, and (iii) longevity as wild-type siblings and both males and a mosaic of phenotypes but can generally be females have normal fertility. TTR was found regarded as clinically euthyroid. This is the pre- neither in the plasma of such mice nor was it dominant form representing approximately 85–90% produced by the choroid plexus (Episkopou et al., of described cases, with the other form known as 1993). Although total plasma T4 concentration was pituitary RTH (i.e. PRTH) exhibiting (i) elevated reduced by 50% and total T3 levels were reduced by free thyroid hormone levels, and (ii) inappropriately 15% in TTR-null mice compared to wild-type normal TSH, but (iii) showing signs and symptoms Thyroid hormones and their effects 601 of hyperthyroidism. The difference between these associated with non-functional mutant TRβ recep- two conditions is not distinct. tors and also generated the effects manifested by the The phenotype associated with RTH is extremely TRα receptors. This does not sit easily with the variable with the most common features being the dominant negative effects of such mutant TRβs. presence of goitre (95%), tachycardia (80%), However, it is compatible with many thyroid hyperkinetic behaviour (72%), and emotional dis- hormone effects, commonly assumed to be mediated turbance (65%). Other less frequently diagnosed by nuclear receptors, instead being mediated by features include cardiac disease (30%), learning non-nuclear-receptor mechanisms. The attentuated disabilities and speech impediments (both in 28% of response to exogenous thyroid hormones sometimes cases), growth retardation (19%), attention-deficit recorded in RTH subjects is compatible with the hyperactivity disorder (11%) and hearing loss in long-known attenuated response to exogenous thy- 8% of cases (Beck-Pecoz et al., 1997). The majority roid hormones associated with hyperthyroidism. of cases are inherited (80–90%) and are related to a Resistance to the action of thyroid hormones also variety of mutations in the TRβ hormone-binding occurs in situations other than the clinical RTH domain which generally occur in one of three specific syndrome. Galo, Unates & Farias (1981) reported ‘hot-spot‘ regions. All of the mutated TRβs for the novel finding that the influence of thyroid # which it has been measured show a diminished T3 hormones on rat erythrocyte Ca +-ATPase activity binding affinity (Usala, 1995). Most patients are differed when the rats were fed diets that differed heterozygous having both a mutant and a normal only in their fat composition. When the rats were fed TRβ but the mutant TRβ inhibits the activity of the a saturated fat diet, their basal (unstimulated) # normal TRβ as well as TRα. This condition is erythrocyte Ca +-ATPase activity was less than if the described as a ‘dominant negative‘ effect. One rats had been fed a polyunsaturated diet. The individual has been recorded as homozygous for a surprising finding was that both T4 and T3 # mutant TRβ and exhibited the most severe form of stimulated in vitro Ca +-ATPase activity in the the syndrome having a resting heart rate of 190 saturated-fat-fed rats but that both thyroid hor- beats\min. He died from cardiogenic shock com- mones inhibited enzyme activity in the poly- plicated by septicemia. Interestingly, the subjects unsaturated-fat-fed rats. At in vitro concentrations of # originally described were homozygous for a complete 10 p for T3 and 1 p for T4, the Ca +-ATPase deletion of both TRβ alleles (Beck-Pecoz et al., activity was similar in erythrocytes irrespective of 1997). TRβ-null mice exhibit both the deaf-mutism their type of dietary fat. These results showed that and the serum thyroid hormone and TSH profile not only did thyroid status influence membrane acyl typical of this form of RTH. No mutations have been composition (as discussed above) but that membrane recorded for TRα in RTH patients. acyl composition also has effects on thyroid hormone The variability of the phenotype in RTH may be action. related to the variety of mutations in TRβ but Other studies also suggest that membrane acyl individuals with the same mutation have been composition might influence thyroid hormone ac- classified differently and even within families with tion. Erschoff (1949) reported that soyabean meal the same mutation, phenotypic variation has been counteracted the thyrotoxic effects of feeding desic- recorded. In addition, significant temporal variation cated thyroid tissue to growing rats, that the anti- in clinical symptoms and thyroid hormone action thyrotoxic effects were related to the fat component has been recorded in affected individuals. Such a of the soyabean meal, and that the effect was not phenotypic spectrum has been attributed to as yet restricted to soyabean meal but was also present unknown environmental\genetic factors (see Beck- when other fats were used in the diet. His review of Pecoz & Chatterjee, 1994). Apart from serum the earlier literature (primarily German) indicated thyroid hormone levels, thyroid-hormone-responsive that this effect had been known since the 1920s and variables are often the same in affected and seemed to be associated with the relative unsat- unaffected individuals and most attempts to dem- uration of the fats. Whether these early reported onstrate tissue hyposensitivity to thyroid hormones influences of dietary fat on the effects of thyroid have given contradictory and nonreproducible re- hormones are mediated by changes in membrane sults (see Beck-Pecoz et al., 1997). Most explanations acyl composition is not known. However, later of the euthyroid-hyperthyroid phenotype associated studies have shown that alterations in the acyl with RTH have suggested that the high free T4 and composition of the nuclear envelope, by dietary-fat T3 levels have overcome the tissue resistance manipulation, also result in changes in T3 binding to 602 A. J. Hulbert the nuclear envelope (Venkatraman, Lefebvre & not necessarily mean that its expression is controlled Clandinin, 1986). by thyroid hormones unless the TRE is found in an Fatty acids have been demonstrated to inhibit the appropriate part of the gene. Some TREs are binding of thyroid hormones to plasma proteins and reported to be in exons and some in introns, whilst to nuclear receptors but such effects are not likely others are in promoter regions. Similarly, the finding normally to be of in vivo importance (Mendel, Frost of a TRE will not by itself mean it is a mediator of & Cavalieri, 1986; Mazzachi et al., 1992). In mice, thyroid hormone effects. TRs are also required for a diet enriched with omega-6 polyunsaturates in- such thyroid hormone effects to become manifest hibited the stimulatory effects of T4 and T3 on malic and therefore the cellular concentration of TRs will enzyme activity but showed no influence on some give additional insight into their relative importance other hormone effects (Deshpande & Hulbert, 1995). in different tissues. The fact that TR concentration The interaction between membrane acyl com- is greatest in anterior pituitary cells supports the position and the effects of thyroid hormones is proposal that they are especially important in effects worthy of more detailed examination. For example, related to the thyroid axis. ageing is associated with resistance to thyroid An intriguing finding was that the direct repeat hormone action (Mooradian & Wong, 1994) and it TRE (DRj4) was sometimes found associated with is possible that such changes may be mediated by another (DRj4) TRE on the other strand some changes in membrane acyl composition with age. distance away. This may be a type of long-distance palindrome. The complete sequencing of the human genome will allow many questions to be answered. (5) Future insights from the human genome Because nuclear receptors require the presence of VI. CONCLUSIONS TREs at appropriate places in the genome to exert their actions, the complete sequencing of the human (1) The thyroid hormones are very old molecules genome will allow determination of how many and appear to be omnipresent among vertebrates. TREs are located in the promoter region of genes The paradigm that T3 is the ‘active’ thyroid and in which genes. It will answer many questions. hormone and T4 is only a prohormone is inadequate. TREs have been known now for over ten years yet Several iodothyronines are active hormones (namely surprisingly few have been described (see Williams T4, T3, rT3 and 3,5-T2) and there are several & Brent, 1995). A preliminary search of the genome significant pathways of hormone action. databases in June 1998 for the three types of TRE (2) It is proposed that the physical chemistry of (Fig. 9) revealed some interesting results. As ex- these molecules is an important consideration and pected, the vast majority of matches were for the that several thyroid hormone effects are the conse- human genome but there were also TREs found in quence of such physical properties. This is additional many other types of organisms (including insects, to, but different from the current paradigm of effects plants and microorganisms). There were curious being mediated only by a nuclear receptor mode of repeats of palindrome TRE in the genome of action. It agrees with the proposition, recently raised Caenorhabditis elegans. Some of the matches appeared by a review of thyroid hormones in invertebrates, to be multiple reportings and thus this search may that these hormones may, in many ways, be better not represent a representative sample of the genome thought of in a ‘vitamin-like’ role than as a classical data. The majority (82%) of TREs found in the hormone. human genome were of the direct repeat DR-4 type, (3) In aqueous solution at physiological pH, with the inverted palindrome (IP-6) type being approximately 80% of T4 molecules, but only 10% 13% and the palindrome type 5% of total matches. of T3 molecules have an ionized phenolic kOH Most matches were not identified as being part of a group. Thus one-fifth of T4 and nine-tenths of T3 particular gene but rather on a particular chromo- molecules will be strongly amphipathic. This differ- some. TREs were reported for 17 of the 23 human ence in the relative ionization of the phenolic kOH chromosomes. Some of the findings were expected group is largely responsible for the greater hydro- (e.g. deiodinase promoter), whilst other genes such phobicity of T3 compared to T4 and also explains as a chloride channel and a phospholipase were the difference in their relative binding affinities to more suprising. plasma proteins (generally T4 " T3) and nuclear The finding of a TRE associated with a gene need receptors (T3 " T4). Thyroid hormones and their effects 603

(4) Iodothyronines are very hydrophobic mol- ify the fluid membrane bilayer. The mechanism ecules and their amphipathic nature results in them whereby this happens is unknown but it is im- associating with membranes with the iodine atoms mediate. The effects of T4 and T3 may be different located within the acyl chain region of biological in this respect and the physical state of the membrane membranes. is also a factor. (5) The various iodothyronines are formed by (11) Thyroid hormones result in an increased sequential monodeiodination of either or both the degree of unsaturation of membrane acyl chains outer and\or inner ring of T4. Such deiodinations (especially in the omega-6 PUFAs). The mechanism are membrane-associated process and there are three of this effect is unknown. It is proposed that the types of deiodinase enzymes (D1, D2 and D3) found normal phospholipid remodelling mechanisms in vertebrates. The T3 found in the plasma of (largely deacylation\reacylation) present in animal vertebrates is produced by deiodination of T4 both cells respond in a ‘homeoviscous’ manner to the in the thyroid follicles and in non-thyroid tissues. thyroid-hormone-induced membrane rigidification (6) In adult vertebrates, the free plasma concen- by altering the acyl composition of the membrane. trations of both T4 and T3 are maintained in the Nuclear receptors may be involved in stimulating picomolar range whilst the total concentrations of some parts of the process. Similarly, their effects on T4 and T3 are in the nanomolar range. The function membrane acyl composition may also be connected of the plasma proteins appears to be to ensure a to the antioxidant activities of thyroid hormones. reasonably even and constant distribution of the This is yet to be determined. thyroid hormones throughout the body. (12) T3 binds strongly to nuclear proteins which (7) The plasma concentrations of the other belong to a superfamily of nuclear receptors. These iodothyronines are considerably lower than those of (TRs) consist of two main types: a TRα and two T4 and T3 in those vertebrates in which they have TRβ receptors (TRβ-1 and TRβ-2) which have been measured. One of the main influences on different tissue distributions and developmental thyroid hormone concentrations and the various profiles. TRs act by binding to thyroid response deiodinative pathways appears to be the energy elements (TREs) in the genome and, in turn, status and food intake of the individual vertebrate activating or repressing transcription of associated being measured. genes. There are several types of TRE, which (8) Although there is considerable variation in generally have a consensus half-site nucleotide the resting metabolic rate of vertebrates there is no sequence of -AGGTCA-. TREs exist either as a single correlation between metabolic rate and either half-site, a direct repeat of two half-sites separated plasma T4 or T3 levels. However, the turnover of T4 by four bases, a palindrome separated by six bases or and T3 in mammals varies with body size in the same a palindrome not separated by any nucleotide bases. way that resting metabolic rate varies. This implies TREs are an evidential requirement for genes that that differences in the secretion rate of thyroid are modulated by thyroid nuclear receptors. hormones are a response to differences in the (13) The thyroid hormone axis consists of TRH metabolic activity of vertebrates rather than the release from hypothalamic cells, TSH secretion from cause of such variations in metabolic rate. anterior pituitary thyrotrophs, plasma binding pro- (9) Although thyroid hormones rapidly associate teins, cellular uptake mechanisms and intracellular with membranes, because of their amphipathic deiodinases as well as production of nuclear recep- nature they do not easily cross membrane bilayers. tors. This axis has been examined in most detail in They enter cells by various uptake mechanisms. adult vertebrates and appears organized as a series of They diffuse in aqueous solution at rates typical of hierarchical regulatory systems that function to amphipathic molecules and also diffuse laterally in maintain a relatively constant level of circulating T4 membranes at rates similar to those of other and relatively constant intracellular T3 levels. The membrane lipids and at a similar rate to their hierarchical nature includes the finding that in- aqueous diffusion rate. They are found at a variety tracellular homeostasis is maintained in some types of locations within the cell and are often located with of cells at the expense of other cells. The brain membranes. It is proposed that they are normal appears to be particularly protected in this regard. constituents of all membranes in vertebrates. The T4 Almost all the effects of thyroid hormones on the content of various tissues in the rat is proportional to thyroid hormone axis are mediated by nuclear the phospholipid content of the tissue. receptors with T3 being the most active intracellular (10) Once associated with membranes they rigid- thyroid hormone in this regard; TREs have been 604 A. J. Hulbert described for many of the important genes involved. thyronine suggests that other pathways may also be One exception is the effect of thyroid hormones on involved. Some of the non-GH-mediated effects may the D2 deiodinase enzyme system where the thyroid be associated with changes in membrane acyl hormone effect appears to be on the interaction composition, in that some effects appear to be between the membrane and the intracellular cyto- associated with prostaglandin production. skeleton, with T4 and rT3 being considerably more (17) Thyroid hormones are present from the egg potent than T3. stage through to adulthood and almost all verte- (14) Although the effects of thyroid hormones on brates appear to have a surge in the plasma T4 levels metabolic rate have been known for some time, the during development. This is often, but not in all mode of thyroid hormone action until recently species, associated with a metamorphic event. Dur- remained obscure. Thyroid hormones affect the ing development of vertebrates, there are also metabolic activity of cells and tissues by stimulating changes in the deiodinases and TR isoforms, the both the mitochondrial proton leak and ATP profiles of which may vary between individual turnover of cells. Many cellular processes are tissues. increased, including ion channel activity, substrate (18) The effect of thyroid status on development uptake mechanisms, and the activity of mito- is often to retard (in the case of hypothyroidism) or chondrial-membrane-bound enzymes. The amount accelerate (in the case of hyperthyroidism) particular of membranes within cells is also often increased. developmental processes rather than to initiate or The vast majority of the effects on metabolic rate can stop these changes altogether. This influence on the be explained by thyroid-hormone-induced changes rate of development of various parts of the vertebrate in both the amount of membranes and changes in can be particularly serious during the development membrane bilayer acyl composition with a conse- of the nervous system. The mode of action of most of quent increase in the molecular activities of mem- these processes is unknown but it is postulated that # brane proteins. Ca + fluxes are increased by thyroid the recent finding that T4 can influence the hormones and these effects are relatively immediate interaction between cells and the extracellular and appear to be membrane effects separate from matrix is likely to be very important during the changes in acyl composition. Relatively few effects development of tissues, especially the nervous system. on metabolism appear to be mediated via nuclear This finding may also explain the high rT3 and low receptors. Those that are, such as increased lipo- T3 concentrations that have been measured during genesis appear to be quantitatively unimportant as the early development of vertebrates (including sources of increased metabolism, but may be im- humans). portant not only in maintaining fat stores but also in (19) Other effects of thyroid hormones include remodelling and manufacture of more membranes. effects on reproduction, defence against viruses and (15) Thyroid hormones speed up several func- defence against free radicals. The defence against tions associated with the membranes of nerve and viruses appears to involve two mechanisms. Whilst muscle cells. They also stimulate the manufacture of the mechanisms of the effects on reproductive timing # specific isoforms of the sarcoplasmic reticulum Ca + and on learning are unknown, it is possible that they pump and myosin heavy chains in muscle. Some of too, involve thyroid hormone effects on the inter- these effects are mediated by nuclear receptors. actions between cells and the extracellular matrix. Many other effects, however, appear to be non- The role of thyroid hormones as membrane anti- nuclear-receptor mediated in that they are too rapid oxidants has received relatively little attention in for such a mode of action. The precise mode of action recent years but it too deserves more research in light of many of these effects is not known but may be of the findings that at very low concentrations T4 mediated either by direct effects on the membrane and T3 inhibit the respiratory burst of activated bilayer, associated with acyl changes in the mem- neutrophils. Whether the membrane antioxidant brane bilayer, or are effects due to interaction with role of thyroid hormones is also associated with its specific membrane receptors. direct effects on membrane fluidity or the thyroid- (16) Growth is stimulated by thyroid hormones hormone-induced changes in acyl composition is and part of this stimulation is due to increased currently unknown. secretion of growth hormone (GH) which is a (20) A comparison of the relative potencies of nuclear-receptor-mediated effect. However, the fact thyroid hormone analogues in various effects sup- that 3,5-T2 also stimulates GH secretion but that ports multiple modes of thyroid hormone actions. TRs have negligible binding affinity for this iodo- (21) The re-examination of results from the early Thyroid hormones and their effects 605

Thyroid follicles

Free extracellular Thyroid hormones bound Uptake thyroid hormones to extracellular proteins mechanisms

Free intracellular Thyroid hormones bound Tissue thyroid hormones to non-nuclear proteins deiodinases Membrane bilayers specific effects Membrane ? acyl composition Nuclear other receptors effects

excitable tissues

development metabolic thyroid growth activity axis Fig. 14. Schema illustrating the proposed pathways whereby thyroid hormones exert their effects in vertebrates. The effects are shown in italic type. The thickness of the arrows represents the relative importance of the pathway for the particular effects.

1960s that were very influential in the development these knockout mice strains suggests that there may of ideas concerning thyroid hormone action also be a certain amount of redundancy between TRα support a membrane site of action for thyroid and TRβ with one receptor isoform capable of hormones. substituting if the other is absent. However, there (22) ‘Knockout’ mice, in which the TR isoforms were no compensatory increases in expression of the have been inactivated, have been examined. TRβ other receptor in either TRα or TRβ knockout mice. knockout mice appear normal except for two factors, The results show the importance of both of these they exhibit a permanent auditory deficit and are receptors in the homeostatic system that maintains deficient in the thyroid hormone axis in that they are normal thyroid hormone levels in the adult. They unable to completely suppress TSH release and are also suggest that the non-thyroid-hormone-binding thus hyperthyroid. Two strains of TRα knockout c-erbAα-2 receptor may also have an important role mice were produced. Those that are only deficient in in initiating the developmental surge of T4 con- TRα-1 also appear normal except that they have centration. reduced TSH levels and are mildly hypothyroid. (23) The clinical syndrome of ‘resistance to Another strain of knockout mice in which both the thyroid hormone’ is due to mutations in the TRβ TRα-1 and the non-thyroid-hormone binding, but receptor in those humans affected. The considerable related, receptor c-erbAα-2 are non-functional shows phenotypic variation found for this syndrome is the most extreme abnormalities. These mice appear compatible with multiple pathways of thyroid normal until the time of the developmental T4 surge. hormone action. Results from animal studies suggest At this time they develop extreme hypothyroidism, that factors such as dietary fat composition (which fail to grow or develop normally and generally die. can influence membrane acyl composition) can also When affected individuals were given a short modulate thyroid hormone effects. temporary burst of exogenous thyroid hormone (24) In conclusion, this review suggests that the several became euthyroid, resuming normal de- effects of thyroid hormones can be mediated by a velopment and surviving for a considerable time number of pathways and proposes that there are without exogenous thyroid hormone but instead several modes of action: (i) the binding of T3 to now making their own. The relative lack of effects in thyroid nuclear receptors is an important pathway of 606 A. J. Hulbert hormone action, especially in the control system that VIII. REFERENCES maintains relatively constant hormone levels in adult vertebrates, it is also important in some other effects, A, R. J., O’M, B. P., B, D. B., T,L.& (ii) due to their physical properties, thyroid hor- R, F. D. (1983). Central and peripheral nerve conduction in thyroid dysfunction: the influence of - mones are normal constituents of membranes and thyroxine therapy compared with warming upon the con- correspondingly influence the properties of these duction abnormalities of primary hypothyroidism. Clinical membranes and their function; they also cause Science 64, 617–622. changes in the acyl composition of membranes A, M. G., H, P. J., T, T., R, S. J., which, in turn, influence the behaviour of proteins W,R.E.H.&S, G. (1992). Protein associated with such membranes and this is an synthesis at the blood-brain barrier. Journal of Biological important mode of stimulation of metabolic activity, Chemistry 267, 23170–23174. A,L.F.&I, S. H. (1967). Selective alteration by (iii) there are additional membrane effects about triiodothyronine of amino acid transport in embryonic chick which little is known, such as interaction with the bone. Endocrinology 81, 1362–1371. extracellular matrix and probable binding to pro- A, M. T., K, A., F, M. J., F, N. J. S., teins which can be regarded as receptors for non- MC, K. P., K, S., F, I. L., S, nuclear-mediated effects. The proposed pathways M. C. & R, D. B. (1993). The structure of the human whereby thyroid hormones exert their effects are binding globulin (TBG) gene. Biochimica et Biophysica Acta 1216, 446–454. summarised in Fig. 14. A-A, H., H,L.J.&S, J. E. (1997). Resting (25) Although study of the thyroid gland and the energy expenditure is sensitive to small dose changes in effects of its hormones has a relatively long history, patients on chronic thyroid hormone replacement. Journal of much still remains to be investigated before we know Clinical Endocrinology and Metabolism 82, 1118–1125. the full picture of the connections between these A, P., G,E.A.&L, P. R. (1986). Plasma T4 hormones and their observed effects. and T3 levels in naturally metamorphosing Eurycea bislineata (Amphibia; Plethodontidae). General & Comparative Endo- crinology 61, 151–163. A, M., K, Y., W, W., P,I.&P,A.R. VII. ACKNOWLEDGEMENTS (1996). In serum-free culture thyroid hormones can induce full expression of chondrocyte hypertrophy leading to matrix This extensive review was written during three periods of calcification. Journal of Bone & Mineral Research 11, 105–113. study leave from the University of Wollongong. It was A, G. W., H, S. G., L, R. J., S, K. A., begun in the laboratory of Dr Gabriella Morreale de S, H. L., M,C.N.&O,J.H. (1997). Purkinje cell protein-2 cis-elements mediate repression Escobar and Dr Francisco Escobar del Rey at the of T3-dependent transcriptional activation. Molecular & Instituto de Investigaciones Biomedicas, Consejo Superior Cellular Endocrinology 131, 79–87. de Investigaciones Cientificas (Spanish Research Council) A, R. C., B,J.A.&F, R. N. (1989). High in Madrid, Spain. It was further researched and written affinity -triiodothyronine binding to right-side-out and in the laboratory of Dr Martin Brand at the Department inside-out vesicles from rat and human erythrocyte mem- of Biochemistry, University of Cambridge, UK, whilst I brane. Journal of Biological Chemistry 264, 19143–19146. was gratefully in receipt of the Brenda Ryman Fellowship A, F., C, D., G,D.&N, J. (1991). from Girton College, Cambridge. It was completed in the Regulation of five tubulin isotypes by thyroid hormone during laboratory of Professor Gerhard Schreiber at the Russell brain development. Journal of Neurochemistry 57, 1781–1786. Grimwade School of Biochemistry and Molecular Biology, A, A., A,P.M.&S, S. A. (1993). The effects University of Melbourne, Australia and was partly of congenital hypothyroidism using the hyt\hyt mouse on supported by a grant from the University of Melbourne. locomotor activity and learned behaviour. Hormones & Professor Schreiber kindly brought the very early report of Behaviour 27, 418–433. T3 by Hird & Trikojus (1948) to my attention. I wish to A,A.Y.&A,Y. N. (1994). Thyroid thank all of these scientists and their laboratory colleagues hormones and regulation of cell reliability systems. Advances in for their hospitality, for sharing their wisdom with me and Enzyme Regulation 34, 173–198. A, J. E., G, M. A., S,C.A.&A,J.S. for their good humour during my time in their respective (1994). Absence of thyroid hormone resistance in vitamin D- laboratories. Especially, I would also like to thank Dr Jeff resistant new world primates. American Journal of Primatology Stuart at Cambridge for assistance in searching the 32, 215–222. genome databases for TREs in June 1998 as well as Drs A, S., G,F.&K, B. (1998). 3,5- Samantha Richardson and Linus Chang in Melbourne. diiodothyronine binds to sub-unit Va of cytochrome-c oxidase The molecular models for T4, T3, rT3 and 3,5-T2 and abolishes the allosteric inhibition of respiration by ATP. illustrated in Fig. 1 were kindly constructed by Dr Renate European Journal of Biochemistry 252, 325–330. Griffiths of the Department of Chemistry, University of A, K., D, K., H,H.&K, K. (1989). Wollongong. Vitamin E protects against thyroxine-induced acceleration of Thyroid hormones and their effects 607

lipid peroxidation in cardiac and skeletal muscles of rats. anterior pituitaries in vivo and in reaggregate cultures and Journal of Nutritional Science & Vitaminology 35, 407–418. GH3 cells in vitro. Endocrinology 138, 3242–3248. A, A., B, B., T, S., R,C.&R, B-P, P., A,C.&M, D. (1997). A. (1996). Low resting metabolic rate in subjects predisposed Resistance to thyroid hormone. In Diseases of the Thyroid (ed. to obesity: a role for thyroid status. American Journal of Clinical L. E. Braverman), pp. 199–239. Humana Press, Totowa, New Nutrition 63, 879–883. Jersey, USA. A,J.&B, R. (1986). Mineral and bone B-P,P.&C, V. K. K. (1994). The variable metabolism in thyroid disease a review. Quarterly Journal of clinical phenotype in thyroid hormone resistance syndrome. Medicine 60, 737–752. Thyroid 4, 225–232. B, D. W., W, S. B., M,S.M.J.&G, B, K. B., S, M. J., D,J.C.&G, A. G. (1986). Hormonal regulation of lipogenic enzymes in V. A. (1995). The type III 5-deiodinase in Rana catesbeiana chick embryo hepatocytes in culture. Journal of Biological tadpoles is encoded by a thyroid hormone-responsive gene. Chemistry 261, 12555–12561. Endocrinology 136, 4424–4431. B, S. W. (1991). Thyroid hormones transcriptionally B, K. B., S, K. C., D, J. C., S, regulate the β"-adrenergic receptor gene in cultured ven- M. J. & G, V. A. (1997). The type 2 and type 3 tricular myocytes. Journal of Biological Chemistry 266, iodothyronine deiodinases play important roles in co- 15863–15869. ordinating development in Rana catesbeiana tadpoles. Endo- B, C., O,K.&K, I. (1992). Time course of crinology 138, 2989–2997. the in vivo effects of thyroid hormone on cardiac gene B, Z., A, E., R, H., D,Z.& expression. Endocrinology 130, 2001–2006. J, V. (1989). Influence of thyroid status on the B, S. G., I,M.&C, W. W. (1997). Deletion of the membranes of rat liver mitochondria. FEBS Letters 243, thyroid hormone β1 receptor increases basal and tri- 247–250. iodothyronine-induced growth hormone messenger ribo- B,D.&P, M. (1987). A novel thyroid hormone nucleic acid in GH$ cells. Endocrinology 138, 3125–3132. receptor encoded by a cDNA clone from a human testis B, J. S., J, T. M., S,T.D.&O, library. Science 238, 788–791. O. V. (1992). Short-term effects of triiodothyronine on the B,S.&R, J. (1996). Altered thyroid hormone bowfin, Amia calva (Holostei), and the lake char, Salvelinus binding to plasma lipoproteins in hypothyroidism. Thyroid 6, namaycush (Teleostei). The Journal of Experimental Zoology 261, 595–600. 105–109. B,J.&DG, L. J. (1980). Mode of action of thyroid B, C. S., H,J.L.&K, S. S. (1995). hormones. In The Thyroid Gland (ed. M. DeVisscher), Thyroid hormone treatment alters phospholipid composition pp. 123–143. Raven Press, New York. and membrane fluidity of rat brain mitochondria. Biochemical B, C. D. (1988). Role of membrane lipids in Journal 305, 29–32. metabolic regulation. Nutrition Review 46, 145–149. B, D. E., B,J.&E, R. N. (1991). The B,M.J.&L, P. R. (1992). The role of selenium in thyroid hormone receptor gene (c-erbAα) is expressed in thyroid hormone action. Endocrine Reviews 13, 207–219. advance of thyroid gland maturation during early embryonic B, M. K., A,K.&C, S.-Y. (1994). Phos- development of Xenopus laevis. Molecular & Cellular Biology 11, phorylation enhances the target gene sequence-dependent 5079–5089. dimerization of thyroid hormone receptor with retinoid X B,S.B.&K, H. M. (1952). Metabolism of receptor. Proceedings of the National Academy of Sciences, USA 91, tissues excised from thyroxine injected rats. American Journal of 7927–7931. Physiology 170, 81–86. B, M., R,M.A.&S, A. (1980). Role of B, J. W., V, M. L. J., E, P. H., M- calcium in the regulation of acetylcholine receptor synthesis in H,M.&D N, P., E, I. V., B, cultured muscle cells. Pflugers Archivs 385, 37–43. A., M,J.A.&R, G. G. (1986). Thyroid B, S. C., P,U.&S, P. K. (1997). Regulation of hormone receptors bind to defined regions of the growth cytoskeletal proteins by thyroid hormone during neuronal hormone and placental lactogen genes. Proceedings of the maturation and differentiation. Brain Research 757, 245–253. National Academy of Science, U.S.A. 83, 9021–9025. B, P., N, C., O, C., A,A.&S,G. B-S, P., S,O.&G, J. (1973). Electron (1993). Transthyretin expression in the rat brain: effect of "#& microscopic radioautography of I-triiodothyronine in rat thyroid functional state and role in thyroxine transport. Brain posterior pituitary and median eminence. Endocrinology 93, Research 632, 114–120. 1410–1422. B, V., P, T. L., M, R., T, A., B,L.&R, J. (1992). Variations in thyroid S, S., C, A., F,C.&S, hormone transport proteins and their clinical implications. V. P. (1998). Decrease in mitochondrial energy coupling by Thyroid 2, 237–245. thyroid hormones: a physiological effect rather than a B, B. J., MF,W.J.&B,J.S. pathological hyperthyroidism consequence. FEBS Letters 430, (1996). Short-term effects of 3,5,3h-triiodothyronine on the 409–413. intermediary metabolism of the dogfish shark Squalus acanthias: B, F., B, L., B, S., B, A., G, evidence from enzyme activities. The Journal of Experimental L., D’U, E., M,L.&M, E. (1997). Zoology 274, 157–162. -thyroxine directly affects expression of thyroid hormone- B, A., B, K., J,H.&K, J. (1997). 3,5- sensitive genes: regulatory effect of RXRβ. Molecular & diiodothyronine stimulates type 1 5hdeiodinase activity in rat Cellular Endocrinology 134, 23–31. 608 A. J. Hulbert

B, P. (1996). Metabolic and hormonal changes during the regulate Ca-channels in cultured skeletal muscle of the rat. molt of captive gray seals (Halichoerus grypus). American Journal Neuroscience Letters 117, 325–330. of Physiology 270, R1051–R1058. B, P. S., B, J. A., T,A.M.H, B, P. A.,   B, C., P, N. J.,  Z, A. J. & B, M. D. (1998). The proton permeability of the H. J. D. &  J, J. M. B. V. (1983). Uptake of tri- inner membrane of liver mitochondria from ectothermic and iodothyronine into cultured human muscle cells. Acta Endo- endothermic vertebrates and from obese rats: correlations crinologica 103, 205–209. with standard metabolic rate and phospholipid fatty acid B,B.&H, E. (1991). Maternal hypothyroidism composition. Comparative Biochemistry & Physiology 119B, during the first half of gestation compromises normal catabolic 325–334. adaptations of late gestation in the rat. Endocrinology 129, B, D. D. (1997). The role of thyroid hormone in zebrafish 210–216. and axolotl development. Proceedings of the National Academy of B, D., M, O., S, Y., L, H., J,E.& Sciences, USA 94, 13011–13016. B, A. G. (1993). Comparison of the metabolic and B, S. B., M,M.&K, W. B. (1997). ‘Spot endocrine effects of 3,5,3h-triiodothyroacetic acid and thy- 14 ’ protein functions at the pretranslational level in the roxine. Endocrinology 77, 221–228. regulation of hepatic metabolism by thyroid hormone and B, M. D. (1997). Regulation analysis of energy metabolism. glucose. The Journal of Biological Chemistry 272, 2163–2166. The Journal of Experimental Biology 200, 193–202. B, V., V, D., R, R., P- B, M. D., B, K. M., B, J. A., H, ,D.&D, E. (1997). The effect of food intake J. A., R,D.F.S.&S, J. A. (1999). The significance from two weeks of age to sexual maturity on plasma growth and mechanism of mitochondrial proton conductance. Inter- hormone, insulin-like growth factor-1, insulin-like growth national Journal of Obesity 23 (suppl 6), S4–S11. factor-binding proteins, and thyroid hormones in female B, M. D., C,L-F,A, E. K., R,D.F.S.& broiler breeder chickens. Comparative & General Endocrinology P, R. K. (1994). The causes and functions of mito- 107, 212–220. chondrial proton leak. Biochimica et Biophysica Acta 1187, B,W.&G, R. T. (1995). Thyroid gland 132–139. development in the brushtail possum, Trichosurus vulpecula. The B, M. D., C, P., E, P. L., W,K.W.& Anatomical Record 243, 254–260. H, A. J. (1991). Evolution of energy metabolism: B,W.&G, R. T. (1996). The effect of proton permeability of the inner membrane of liver mito- methimazole on the growth of the developing brushtail chondria is greater in a mammal than in a reptile. Biochemical possum, Trichosurus vulpecula. Growth, Development & Aging 60, Journal 275, 81–86. 163–169. B, M. D., S, D., K, B., S, B, J., G, J., E,E.E.&D, A. T. (1960). P. M. & H, R. P. (1992). The mechanism of the increase Studies on Vitamin E. 5. Lipid peroxidation in dialuric acid- in mitochondrial proton permeability induced by thyroid induced haemolysis of vitamin E-deficient erythrocytes. hormones. European Journal of Biochemistry 206, 775–781. Biochemical Journal 77, 47–51. B,J.A.&W, M. (1970). Differential cellular growth B, J., G, J., E,E.E.&D, A. T. (1961). in the organs of hypothyroid rats. Growth 34, 874–882. Some antioxidant properties of -thyroxine in biological B,T.A.&D, P. K. (1988) Effect of hypo- systems. Biochimica et Biophysica Acta 47, 401–403. thyroidism on the lipid composition and fluidity of the rat B, A. (1986). Nondeiodinative pathways of thyroid colonic apical plasma membranes. Biochimica et Biophysica Acta hormone metabolism. In: Thyroid Hormone Metabolism (ed. G. 939, 189–196. Hennemann), pp. 255–276, Marcel Dekker, New York. B, C., K, C., F, J., P, S., B, A. G., O’C, M., S, K., W,R.& S-N,K.&F,T. (1989). Concentrations D,E.J., (1987). Monodeiodination of triiodo- of thyroxine, 3,5,3h-triiodothyronine, 3,3h,5h-triiodothyronine, thyronine and reverse triiodothyronine during low and high 3,3h-diiodothyronine, and 3h,5h-diiodothyronine in human red calorie diets. Journal of Clinical Endocrinology & Metabolism 65, blood cells. Scandinavian Journal of Clinical Laboratory Investi- 829–835. gation 49, 329–336. B, U., F,C.&G, A. U. (1986). Effects of acute B, R. R. (1981). Nutritional and hormonal factors bacterial infection on serum thyroid hormones and nuclear influencing desaturation of essential fatty acids. Progress in triiodothyronine receptors in mice. Endocrinology 119, 515–521. Lipid Research 19, 155–186. B, L. A., P,J.&S.G, D. L. (1997). B, R. R. (1984). Effect of unsaturated acids on mem- Thyroid hormones inhibit type 2 iodothyronine deiodinase in brane structure and enzyme kinetics. Progress in Lipid Research the rat cerebral cortex by both pre- and posttranslational 23, 69–96. mechanisms. Endocrinology 138, 5231–5237. B, G. A., H, J. W., M,D.D.&L,P.R. B, W. A., E, S. E., L, J., P,H.P.&R, (1988). Multihormonal regulation of the human, rat, and D. B. (1979). The ratio of thyroxine to thyroxine-binding bovine growth hormone promoters: differential effects of globulin in the assessment of thyroid function. Clinical 3h,5h-cyclic adenosine monophosphate, thyroid hormone, and Endocrinology 11, 333–342. glucocorticoids. Molecular Endocrinology 2, 792–798. B, P. J., D, B., J,C.J.&S,D.S. B,C.&S, S. R. (1989). Characterization of (1979). The effect of hypothyroidism on growth, serum thyroid hormone effects on Na channel synthesis in cultured growth hormone, the growth hormone dependent somato- # skeletal myotubes: role of Ca +. Endocrinology 125, 842–849. medin, insulin like growth factor and its carrier protein in B,C.&S, S. R. (1990). Thyroid hormones up- rats. Endocrinology 104, 1107–1111. Thyroid hormones and their effects 609

B-B, G. S., B,J.P.&T, L.-E. C-R, J. M., Z, S., C,M.&K,O. (1990). Myosin heavy and light chain expression during (1996). Development of enzyme immunoassays for 3,5,3h- human skeletal muscle development and precocious muscle triiodo--thyronine and -thyroxine: time course studies on maturation induced by thyroid hormone. Anatomy & Em- the effect of food deprivation on plasma thyroid hormones in bryology 181, 513–522. two marine teleosts, sea bass (Dicentrarchus labrax L.) and sea B-B, G. S., H,D.&W,R.G. bream (Sparus aurata L.). General & Comparative Endocrinology (1984). Effects of hypothyroidism on myosin isozyme tran- 103, 290–300. sitions in developing rat muscle. FEBS Letters 166, 7175. C, J., K, F., P, V., K,B.& V, R. (1998). Thyroid control of sarcolemmal B,F.&B, M. D. (1995). A hierarchy of ATP- + #+ #+ consuming processes in mammalian cells. Biochemical Journal Na \Ca exchanger and SR Ca -ATPase in developing 312, 163–167. rat heart. American Journal of Physiology 275, H264–H273. B, F., W,M.&B, G.-R. (1998). A C, L., M, S. L. A., R,S.J.&S, new hypothesis of modular glucocorticoid actions. Arthritus & G. (1998). Evolution of thyroid hormone binding by transthyretins in birds and mammals. European Journal of Rheumatism 41, 761–767. Biochemistry 259, 1–10. C, L., A,L.&G, L. (1997). Thyroid hormone- C, C., ’A, A., L-G, M., J,C. induced plasticity in the adult rat brain. Brain Research Bulletin &G, C. L. (1987). Regulation by thyroid hormones of 44, 549–557. terminal differentiation in the skeletal dorsal muscle. II C, G., L, J.-T., D S,N.&K, R. (1985). Urodelan amphibians. Developmental Biology 123, 33–42. Short term effect of low doses of triiodothyronine on proximal C, J.-P., A, S., F, S. L., L, J. L., K, tubular membrane Na-K-ATPase and potassium permeab- J. & B, L. E. (1992). Role of transthyretin in the ility in thyroidectomized rats. Pflugers Archiv 403, 90–96. transport of thyroxine from the blood to the choroid plexus, C, E. (1991). Calcium pump of the plasma membrane. the cerebrospinal fluid, and the brain. Endocrinology 130, Physiological Reviews 71, 129–153. 933–938. C, D. J., S, K.. A., S,H.L.&O- C, J.-P., B, L. E., F, A. P., S, , J. H. (1994). Immunofluorescent localization of M., A, S., D,S.&L, J. L. (1993a). The thyroid hormone receptor isoforms in glial cells of rat brain. thyroid gland is a major source of circulating T3 in the rat. Endocrinology 135, 1831–1836. Journal of Clinical Investigation 91, 2709–2713. C, D. J., S, K. A., S,H.L.&O- C, J.-P., V, I., A, S., S, S., F, S. L., , J. H. (1996). Thyroid hormone receptor isoform L,J.L.&B, L. E. (1993b). The postnatal content in cultured type 1 and type 2 astrocytes. Endocrinology serum 3,5,3h-triiodothyronine (T3) surge in the rat is largely 137, 911–917. independent of extrathyroidal 5h-deiodination of thyroxine to C, G., L, G., C,R.&M,L. T3. Endocrinology 133, 2604–2609. (1979). Response of rat hepatic fatty acid synthesis and C, V. K. K. (1997). Resistance to thyroid hormone. activities of related enzymes to changes in level of dietary fat. Hormone Research 48 (suppl. 4), 43–46. Journal of Nutrition 109, 162–170. C, V. K. K., C-B,R.J.&M,C. C, R. V., L, E. T., M, J. A., S,C.G.& (1997). The steroid hormone superfamily of receptors. In K, D. B. (1997). The thyroid hormone, 3,5,3h-tri- Molecular Endocrinology. Genetic analysis of hormones and their iodothyronine, is a negative modulator of domestic fowl receptors. (eds. G. Rumsby & S. M. Farrow). pp. 73–101. (Gallus gallus domesticus) adrenal steroidogenic function. General BIOS Scientific Publishers, Oxford. & Comparative Endocrinology 170, 251–261. C, R. N., R, M. R., M,R.D.&F, C, W. D., C, H. E., C, S. W., E, E. A., R. N. (1995). Differential effect of triiodothyronine and thy- E,J.S.&S, R. M. (1967). Studies on the roxine on liposomes containing cholesterol: physiological antioxidant action of thyroxine and related compounds. speculations. Journal of Membrane Biology 147, 217–221. Journal of Medicinal Chemistry 10, 1081–1085. C,H.J.&W, P. G. (1978). Effects of age and ovarian function on the pituitary-thyroid system in female rats. Journal C, L. W., W,H.C.&H, M. (1991). of Endocrinology 78, 225–232. Relationships between fatty acid synthesis and lipid secretion C,Y.&A, B. G. (1997). Role for the Rana catesbeiana in the isolated perfused rat liver: effects of hyperthyroidism, homologue of C\EBPa in the reprogramming of gene glucose and oleate. Biochimica et Biophysica Acta 1086, 197–208. expression in the liver of metamorphosing tadpoles. Develop-     C , M. I., A , S., Y , R. A., B ,L.E.& mental Genetics 20, 152–162. E, L. E. (1986). Total and free serum thyroid hormone C,Y-D.I.&H, F. L. (1977). Thyroid control over concentrations in fetal and adult pregnant and non pregnant biomembranes. Rat liver mitochondrial inner membranes. guinea pigs. Endocrinology 118, 533–537. Archives of Biochemistry & Biophysics 181, 470–483. C,A.&C, A. (1997). Non-enzymatic peroxidation C, S.-Y., M, F. R., R, J., W, of lipids isolated from rat liver microsomes, mitochondria and M. C. & P, I. H. (1980). Receptor-mediated uptake of nuclei. International Journal of Biochemistry & Cell Biology 29, 3,3h,5-triiodo--thyronine by cultured fibroblasts. Proceedings 541–546. of the National Academy of Sciences, USA 77, 3425–3429. C, P. A., A, T. R., B, J. D., M,P.& C,S.D.&T, R. (1989). Attenuation of E, N. L. (1986). The human growth hormone gene spontaneous, nocturnal growth hormone secretion in children is negatively regulated by triiodothyronine when transfected with hypothyroidism and its correlation with plasma insulin- into rat pituitary tumor cells. Journal of Biological Chemistry like growth factor 1 concentrations. Journal of Pediatrics 114, 261, 13367–13372. 968–972. 610 A. J. Hulbert

C, W. W. (1991). Thyroid nuclear hormone receptors. In C, F., P, S. D., D C, D., D Nuclear Hormone Receptors: Molecular Mechanisms, Cellular Func- R, P., B, T., B-B, G., C, tions, Clinical Abnormalities. (ed. M. G. Parker), pp. 79–102. G., N, G., G, M. T., S, S., C,F. Academic Press, London. &M, A. (1998). Effects of thyroid function on LDL C, I. J. (1996). Nature, source, and relative significance of oxidation. Arteriosclerosis, Thromosis & Vascular Biology 18, circulating thyroid hormones. In Werner and Ingbars. The 732–737. Thyroid. 7th edition (eds. L. E. Braverman & R. D. Utiger), C,G.K,S.A.&L, S. M. (1970). The Achilles pp. 111–121. Lippincott-Raven Publishers, Philadelphia. reflex time in thyroid disorders. Journal of Pediatrics 76, C, I. J., H, T. S., H,R.E.&S,D.H. 277–282 (1984). Thyromimetic effects of 3,5-dimethyl,3h-isopropyl C,P.&H, A. J. (1995a). Membrane fatty acid thyronine (DIMIT) and 3,5-diethyl,3h-isopropyl thyronine composition is related to body mass of mammals. Journal of (DIET) in various tissues of the rat. Metabolism 33, 622–628. Membrane Biology 148, 27–39 C, M., M, F., G, F., M,Y.&G,A. C,P.&H, A. J. (1995b). On the relationship (1996). Demonstration of in vivo metabolic effects of 3,5-di- between body mass, tissue metabolic rate and sodium pump iodothyronine. Journal of Endocrinology 149, 319–325. activity in mammalian liver and kidney. American Journal of C,S.D.&H, J. (1990). Inhibition of tri- Physiology 268, R641–R650. iodothyronine’s induction of lipogenic enzymes by dietary fat. C, W., G,W.L.&H, D. R. (1990). Acute Journal of Nutrition 120, 625–630. effects of thyroid hormone on sodium currents in neonatal C, T.,  H,C.&E, M. E. (1991). myocytes. Bioscience Reports 10, 309–315. Significance of cation transport in control of energy metab- C, D. J., D,B.M.&M, S. L. A. (1996). olism and thermogenesis. Physiological Reviews 71, 733–774. Conformations and binding interactions of thyroid hormones C, S., C,P.J.&M, V. A. (1980). and analogues. In Biological Inhibitors. Studies in Medicinal Hormones and liver mitochondria: influence of growth Chemistry (ed. M. I. Choudhary), vol 2, pp. 255–302. Harwood hormone, thyroxine, testosterone, and insulin on thermotropic Academic Publishers, Switzerland effects of respiration and fatty acid composition of membranes. C, D. G., I, D. R., A, C., ML,J.M.&E, Archives of Biochemistry & Biophysics 203, 744–752. J. G. (1998). Effects of long-term temperature acclimation on C, S., J, E., C, P. J., F,L.&M, thyroid hormone deiodinase function, plasma thyroid hor- mone levels, growth, and reproductive status of male atlantic V. A. (1981). Influence of growth hormone and thyroxine on cod, Gadus morhua. Comparative & General Endocrinology 109, thermotropic effects of respiration and 1-anilino-8-naph- 24–36. thalene sulfonate fluorescence and on lipid composition of D, G. E., E, N. P., T,L.A.&K,F.J. cardiac membranes. Biochimica et Biophysica Acta 678, 250–256. (1994). A central negative feedback action of thyroid C, V. (1996). Thyroid hormone structure-function relatio- hormones on thyrotropin-releasing hormone secretion. Endo- nships. In Werner and Ingbars The Thyroid. 7th edition (eds. crinology 135, 2392–2397. L. E. Braverman & R. D. Utiger), pp. 185–189. Lippincott- ’A, A., C, C., J, C., M, J.-C. & Raven Publishers, Philadelphia. C, R. (1990). Muscle-specific response to thyroid C, M., D, E., F, G., D,V.M.& hormone of myosin isoform transitions during rat postnatal K, E. R. (1996). Influence of feeding pattern on thyroid development. European Journal of Biochemistry 193, 155–161. hormones in long-term food-restricted rats. Hormone & D, U., W,P.C.&S, A. (1996). Differ- Metabolic Research 28, 315–318. ential effects of n-6 and n-3 polyunsaturated fatty acids on cell C, G. P., D, A., F, C., L,B.& growth and early gene expression in swiss 3T3 fibroblasts. L, A. (1983). Plasma concentrations of thyroid Journal of Cellular Physiology 168, 618–624. hormones in the domestic goat. Seasonal variations in relation D, E., J., (1986). Effects of fasting and altered to age. Bollettino della Societa Italiano di Biologia Sperimentale 59, nutrition on thyroid hormone metabolism in man. In Thyroid 779–785. Hormone Metabolism (ed. G. Hennemann), pp. 335–358. C,J.M.&H, G. A. (1980). Feedback effectiveness Marcel Dekker, New York. of periodic versus constant triiodothyronine replacement. D, A. Q., F,F.H.&C, W. J. (1985). Influence of Endocrinology 106, 911–917. hypo- and hyperthyroidism on rat liver glycerophospholipid C, P. S., Y,C.U.&N, C. S. (1984). metabolism. Lipids 20, 897–902. Development of thyroid hormone dependence for growth in D, V. M., V, T. J., B,L.R.&K,E.R. the rat: a study involving transplanted fetal, neonatal, and (1992). Ontogeny of type I and type III deiodinase activities juvenile tissues. Endocrinology 115, 2059–2064. in embryonic and posthatch chicks: relationship with changes C-T, I., M, A-M., H,P.& in plasma triiodothyronine and growth hormone levels. R, B. C. (1991). Differential regulation of Na-K- Comparative Biochemistry & Physiology 103A, 131–136. ATPase isoform gene expression by T3 during rat brain D, A. K., T,C.R.&F, E. (1987). development. American Journal of Physiology 261, C124–C131. Decrease in triiodothyronine binding sites in chick embryo C, A. R., B,K.&P, C. L. (1981). Homeo- erythrocytes during early development. Developmental Biology viscous adaptation and its effects upon membrane-bound 120, 412–417. proteins. Journal of Thermal Biology 6, 183–187. D, M. J. (1990). Thyroid hormones and thermogenesis. C, A. R., C,J.A.&P, C. L. (1978). Proceedings of the Nutrition Society 49, 203–215. Adaptation of biological membranes to temperature – the D,M.J.&H, A. P. (1996). Developmental lack of homeoviscous adaptation in the sarcoplasmic retic- regulation of cation pumps in skeletal and cardiac muscle. ulum. Biochimica et Biophysica Acta 511, 442–454. Acta Physiologica Scandinavia 156, 313–323. Thyroid hormones and their effects 611

D,M.J.&M, A. (1993). Investigation of sodium influx and calcium uptake in erythrocytes in hyperthy- mechanisms mediating the increase in plasma concentrations roidism: role of abnormal membrane lipid levels. Metabolism of thyroid hormones after a meal in young growing pigs. 45, 707–711. Journal of Endocrinology 139, 131–141. D,R.J.&L, P. (1991). Dependence of body growth D, M. J., R, D. B., K, A. L., M,M. on thyroid activity in turtles. Journal of Experimental Zoology &I, D. L. (1983). Increase in plasma concentrations of 258, 48–59. 3,5,3h-triiodothyronine and thyroxine after a meal, and its D,N.&H, A. J. (1995). Dietary ω6 fatty acids dependence on energy intake. Hormone & Metabolic Research 15, and the effects of hyperthyroidism in mice. Journal of 499–502. Endocrinology 144, 431–439. D, J. C., S,M.J.&G, V. A. (1994). D, S., N, F., G,F.&H, T. L. (1998). Cloning of a thyroid hormone-responsive Rana catesbeiana Fasting-induced increase in type II iodothyronine deiodinase c-erbA-β gene. Developmental Genetics 15, 339–346. activity and messenger ribonucleic acid levels is not reversed D, K. J. A. (1995). Oxidative stress: the paradox of aerobic by thyroxine in the rat hypothalamus. Endocrinology 139, life. Biochemical Society Symposium 61, 1–31. 2879–2884. D, F. B., D,P.J.&B, S. D. (1983). Role of D, W. W., D,D.S.&G, A. (1982). calmodulin in thyroid hormone stimulation in vitro of human Thyroid function during smoltification of salmonid fish. In: # erythrocyte Ca +-ATPase activity. Journal of Clinical Investiga- Phylogenic Aspects of Thyroid Hormone Actions (ed. Institute of tion 71, 579–586. Endocrinology, Gunma University), pp. 45–61. Centre for D, F. B., D, P. J., B,S.D.&S, M. (1987). Academic Publications Japan, Tokyo. # Action of long-chain fatty acids in vitro on Ca +-stimulatable, D, P. W., A, A. R., M, J. G. T., M, # Mg +-dependent ATPase activity in human red cell mem- P. D., S,W.H.&S, G. (1987). Thyroxine branes. Biochemical Journal 248, 511–516. transport in choroid plexus. Journal of Biological Chemistry 262, D,P.J.&D, F. B. (1993). Acute cellular actions of 13907–13915. thyroid hormone and myocardial function. Annals of Thoracic D, S. W., B, M. B., K,P.A.&J, Surgery 56, S16–23. E. C. (1977). Thyroxine analogues. 23. Quantitative struc- D,P.J.&D, F. B. (1996). Nongenomic actions of ture-activity correlation studies of in vivo and in vitro thyromimetic activities. Journal of Medicinal Chemistry 20, thyroid hormone. Thyroid 6, 497–504. 863–880. D, P. J., D,F.B.&L, W. D. (1989). Thyroid # D, V. M., C,F.&R, F. M. (1988). hormone regulation of membrane Ca +-ATPase activity. The effect of thyroid hormones on the lipid composition of rat Endocrine Research 15, 651–682. plasma and erythrocyte membranes. Bollettino della Societa D, A., D, D. C., D, A.C.K.& S,P.J. Italiano di Biologia Sperimentale 64, 979–984. (1996). Plasma thyroxine concentrations in farmed ostriches D, W. H. (1990). Biochemical basis of thyroid hormone in relation to age, body weight, and growth hormone. General action in the heart. The American Journal of Medicine 88, & Comparative Endocrinology 103, 308–315. 626–630. D C, L. (1933). Uber die antioxydative wirkung des D M, S.,  M R, P., P,M.C.& L, thyroxins und thyroxinahnlicher substanzen. Hoppe-Seyler’s T. (1993). Electrophysiological properties of papillary muscle Zeitschrift fur Physiologische Chemie 219, 257–266. fibres from euthyroid and hypothyroid chick. Comparative D G D, I. N. T.,  A,M.J.T.&B, Biochemistry & Physiology 105A, 719–724. R. R. (1977). Effect of thyroxine on ∆6 and ∆9 desaturation D M, S.,  M R, P., P,M.C.& L, activity. Advances in Experimental Medicine & Biology 83, T. (1994). Electrophysiological properties of the hyperthyroid 609–616. rat heart. Archives Internationales de Physiologie, de Biochimie et de DG, L. J., R, S., B, J., R,P.A.& Biophysique 102, 153–159. C, A. H. (1978). Nuclear receptors for thyroid hor- D M, S., P, M. C., V,P.& L, T. (1995). mone. Journal of Endocrinological Investigation 1, 79–88. Effect of thyroid state on cardiac electrical activity of the frog,  J, E. G. T., T,J.D.&S, M. S. (1998). Rana esculenta. General & Comparative Endocrinology 100, 162–169. Thyroid hormones promote early metamorphosis in grouper D M, S.,  M R, P., V, P., (Epinephelus coioides) larvae. General & Comparative Endocrinology B,M.& L, T. (1997). Action potential 112, 10–16. configuration in heart papillary muscles from female rats in DC, F., R, H., V, F., V, J., different thyroid states. Archives of Physiology & Biochemistry E,N.&L, C. D. (1993). Changes in body mass, 105, 58–65. hepatic and muscular cellular composition and serum thyroid DS,J.J..&F, D. (1988). Comparative aspects hormones in early weaned goat kids. Small Ruminant Research of the distribution, metabolism, and excretion of six iodo- 11, 45–56. thyronines in the rat. Endocrinology 123, 2514–2525. DL, G. R. (1996). The neuromuscular system and brain in DS, J. J. III., M, W. L., N, T. T.,  hypothyroidism. In Werner and Ingbars. The Thyroid.7th edition H,A.J.&F, W. (1993). Enterohepatic (eds. L. E. Braverman & R. D. Utiger), pp. 826–835. regulation and metabolism of 3.5.3h-triiodothyronine in Lippincott-Raven Publishers, Philadelphia. hypothyroid rats. Endocrinology 132, 1665–1670. D M, D., M, H., M,E.M,M,R.D. D, O., D, F., S, F., A,A.W.&V, &F, R. N. (1977). Thyroid hormone actions and P. R. (1978). Manipulation of fatty acid composition of membrane fluidity. FEBS Letters 84, 199–203. membrane phospholipid and its effects on cell growth in D R, C., V,F.&F, F. (1996). Increased mouse LM cells. Biochimica et Biophysica Acta 509, 239–250. 612 A. J. Hulbert

D,V.W.&H, G. M. (1998). Thyroxine stim- E, C. H., B, G., A, S., C, M. I., R,E. ulates the acylation of lysophosphatidylethanolamine in rat &B, L. E. (1988). The effect of thyroid dysfunction heart. Biochimica et Biophysica Acta 1391, 241–246. and fasting on the placental inner ring deiodinase activity in D, M. M., H,L.C.&R, H. H. (1995). the rat. Endocrinology 122, 809–816. Thyroid status and Na+-K+ pump current, intracellular E, W., S,D.S.&M, A. J. (1972). Thyroid sodium, and action potential duration in rabbit heart. state: effects on pre- and postsynaptic central noradrenergic American Journal of Physiology 268, H1838–H1846. mechanisms. Science 175, 79–82. D, M. D. (1974). On the mechanism of action of E, V., M, S., N, S., S, K., thyroxin, an amino acid analog of tyrosine. Journal of G, G. A., G,M.E.&R,E.J. Theoretical Biology 46, 255–270. (1993). Disruption of the transthyretin gene results in mice D, M. D., C, F. L., G,J.T.& with depressed levels of plasma retinol and thyroid hormone. J, A. S. (1983). Iodothyronine homeostasis in rat Proceedings of the National Academy of Sciences, USA 90, brain during hypo- and hyperthyroidism. American Journal of 2375–2379. Physiology 245, E185–E193. E-F, S., S,H.L.&O,J.H. D,M.D.&G, J. T. (1996). Thyroid hormones as (1996). Isoform-specific 3,5,3h-triiodothyronine receptor bind- neurotransmitters. Thyroid 6, 639–647. ing capacity and messenger ribonucleic acid content in rat D, M. P., D, F. B., D, P. J., S,M.&B, adenohypophysis: effect of thyroidal state and comparison S. D. (1986). Effects of hyperthyroidism and hypothyroidism with extrapituitary tissues. Endocrinology 137, 3228–3233. # on human red blood cell Ca +-ATPase activity. Journal of E, B. H. (1949). Protective effects of soybean meal for Clinical Endocrinology & Metabolism 62, 253–257. the immature hyperthyroid rat. Journal of Nutrition 39, DB, E. F. (1936). Basal Metabolism in Health and Disease. Lea 259–281. & Febiger, Philadelphia. E  R, F., R  O, C., B, J., O, D, S. C., J.&B, C. M. (1993). Bursting of M. J. & M  E, G. (1989). Generalized deficiency of 3,5,3 -triiodo-l-thyronine (T3) in tissues from cardiac sodium channels after acute exposure to 3,5,3h- h triiodo--thyronine. Circulation Research 73, 301–313. rats on a low iodine intake, despite normal circulating T3 D, A. F. (1990). The rate of tetanic relaxation is levels. Acta Endocrinologica 120, 490–498. correlated with the density of calcium ATPase in the terminal E-M, H. F., E  R, F., O, M. J. & M  E, G. (1996). Only the cisternae of thyrotoxic skeletal muscle. Pflugers Archivs 415, combined treatment with thyroxine and triiodothyronine 433–439. ensures euthyroidism in all tissues of the thyroidectomized rat. D P,J.S.&I, J. M. (1987). Evidence of a direct Endocrinology 137, 2490–2502. action of triiodothyronine (T3) on the cell membrane of GH3 E, A., C, F., L, A. M., G,J.M. cells: an electrophysiological approach. Experentia 43, &P, M. (1992). Induction of type III deiodinase 596–598. activity in astroglial cells by thyroid hormones. Endocrinology E, J. G. (1997). Iodine metabolism and thyroid-related 131, 1682–1688. functions in organisms lacking thyroid follicles: are thyroid E, K. (1993). Thyroxine-induced changes in metabolic hormones also vitamins? Proceedings of the Society for Experimental rate and cytochrome oxidase activity in Thamophilus sirtalis: Biology and Medicine 214, 302–317. effects of nutritional status. Comparative & General Endocrinology E, J. G., H, J. A., ML,J.M.&Y,J.H. 91, 66–73. (1997). Thyroid hormone deiodination in various tissues of E,W.&G, A. G. (1974). Evolution of thyroid function larval and upstream-migrant sea lampreys, Petromyzon marinus. in poikilothermic vertebrates. In Handbook of Physiology. Section General & Comparative Endocrinology 106, 202–210. 7, volume III, Thyroid (eds. M. A. Greer & D. H. Solomon), pp. E, J. G., H,M.&U, L. (1981). Effect of food intake 5–20. American Physiological Society, Washington, DC. on diel variation in plasma thyroid hormone levels in rainbow E, E. S., R, L. L., E,A.B.&K,A.A. trout, Salmo gairdneri. General & Comparative Endocrinology 45, (1964). Relative sensitivity of different biological responses to 167–174. small quantities of thyroxine and triiodothyronine. Endo- E, J. G., M, P-P., T,P.&H, T. J. (1993). crinology 74, 770–779. Thyroid hormone deiodination in brain, liver, gill, heart and E, E. S., S, R. A., E, A. B., J,C.A.& muscle of Atlantic salmon (Salmo salar) during photo- T, A. (1966). Biological evidence for extrathyroidal periodically-induced parr-smolt transformation. II. Outer- thyroxine formation Endocrinology 78, 983–1001. and inner-ring 3,5,3h-triiodo--thyronine and 3,3h,5h-triiodo- E,E.&F, Y.-C. (1972). Improved measurements of the -thyronine (reverse T3) deiodination. General & Comparative erythrocyte geometry. Microvascular Research 4, 335–347. Endocrinology 90, 157–167. E, R. M. (1988). The steroid and thyroid hormone receptor E,J.G.&S, S. (1985). Free T4 and T3 in relation superfamily. Science 240, 889–895. to total hormone, free hormone indices and protein in plasma E, M. E. (1996). Effects of thyroid hormones on con- of rainbow trout and arctic charr. General & Comparative tractility and cation transport in skeletal muscle. Acta Endocrinology 58, 291–302. Physiologica Scandinavia 156, 325–333. E,J.G.&S, S. (1987). Total and free thyroid E,M.E.&C, T. (1988). Effects of thyroid hormones in plasma of tropical marine teleost fish. Fish hormone on Na+ -K+transport in resting and stimulated rat Physiology & Biochemistry 3, 127–131. skeletal muscle. American Journal of Physiology 255, E604–E612. E, R. (1986). The free hormone concept. In Thyroid Hormone E, M. E., D, R.,  B, J. C. J.,  Metabolism (ed. G. Hennemann), pp. 76.77–106. Marcel K, P. M., H, L. J.,  J, M., K, Dekker, New York. E. P. & H, G. (1993). Effects of a furan fatty acid Thyroid hormones and their effects 613

and indoxyl sulfate on thyroid hormone uptake in cultured rat liver microsomal fractions in an experimental hyper- anterior pituitary cells. American Journal of Physiology 268, thyroid state: relation to lipid peroxidation. Endocrinology 117, E974–E979. 496–501. E, M. E., D, R., M, E. P. C. M.,  F,V.&V, L. A. (1993). Influence of hyper- K, P. M., V, T. J.,  J, M., K, thyroidism on superoxide radical and hydrogen peroxide E. P. & H, G. (1994). Uptake of thyroxine in production by rat liver submitochondrial particles. Free cultured anterior pituitary cells of euthyroid rats. Endocrinology Radical Research Communications 18, 329–335. 134, 2490–2497. F,V.&V, L. A. (1995). On the mechanism of E, M. E., V, F. A., B, K., M, thyroid hormone-induced respiratory burst activity in rat E. P. C. M., H, G., V,T.J.&L, polymorphonuclear leukocytes. Free Radical Biology & Medicine J. M. J. (1996). Uptake of thyroid hormones in neonatal rat 19, 359–363. cardiac myocytes. Endocrinology 137, 4235–4242. F, B. C., A, G., M, Y., T, W., E, M. E., V, T. J., M, E. P. C. M., T- M,F.D.&C, A. C. (1993). Gene expression in , A. M. P.,  T, H., D, R.,  J, M., the developing cerebellum during perinatal hypo- and K,E.P.&H, G. (1995). Uptake of hyperthyroidism. Molecular Brain Research 17, 258–268. 3,3h,5,5h-tetraiodothyroacetic acid and 3,3h,5h-triiodothyro- F, D. I., A, P., L,A.R.&W,D. nine in cultured rat anterior pituitary cells and their effects on (1991). Effects of thyroidectomy on development of skeletal thyrotropin secretion. Endocrinology 136, 4454–4461. muscle in fetal sheep. American Journal of Physiology 261, F,F.H.&C, W. J. (1981). Fatty-acid desaturation R1300–R1306. and microsomal lipid fatty-acid composition in experimental F, D. A. (1996). Physiological variations in thyroid hyperthyroidism. Biochemical Journal 193, 845–852. hormones: physiological and pathological considerations. F,F.H.&C, W. J. (1982). Fatty-acid desaturation Clinical Chemistry 42, 135–139. and microsomal lipid fatty-acid composition in experimental F, D. A., D, J. H., S,J.&C,I.J. hypothyroidism. Biochemical Journal 207, 29–35. (1977). Ontogenesis of hypothalamic-pituitary-thyroid func- F, J., K, C., L, I. B., S- tion and metabolism in man, sheep, and rat. Recent Progress in N,K.&F, T. (1982). Simultaneous measurement of Hormone Research 33, 59–116. 3,5-diiodothyronine and 3,5,3h-triiodothyronine turnover F, D. A., P,D.H.&W, S. Y. (1994). Fetal thyroid kinetics in euthyroid, hyperthyroid, and hypothyroid subjects. metabolism: a pluralistic system. Thyroid 4, 367–371. Journal of Clinical Endocrinology & Metabolism 55, 8–12. F, D., S,M.&V, B. (1990). Contrasting F, J., R, P., K,S-N,K. developmental and tissue-specific expression of alpha and &F, T. (1984). Serum free T4, T3, rT3, 3,3 -di- h beta thyroid hormone receptor genes. EMBO Journal 9, iodothyronine and 3h,5h-diiodothyronine measured by ultra- 1519–1528. filtration. Acta Endocrinologica 107, 357–365. F, D., E, L.C., N, L., A,R.&C, F, R. N., C, R. N., R,M.R.&M, T. (1996a). Thyroid hormone receptor β is essential for R. D. (1995). Differential effect of triiodothyronine and thy- development of auditory function. Nature Genetics 13, 354–357. roxine on the liposomal membrane in liquid-crystalline and F, D., G, G., C,J.&C, T. (1996b). gel state. Journal of Membrane Biology 143, 135–141. F, A., D, B., H, P., R,J.& Genetic analysis of thyroid hormone receptors in development N, V. M. (1992). Characterization of myelin basic and disease. Recent Progress in Hormone Research 51, 1–22 protein thyroid hormone response element and its function in F, D., H, E., S, R. J., E, N., the context of native and heterologous promoter. Journal of S, C. L., W, J.M. & C, T. (1996c). Biological Chemistry 267, 15784–15788. Recessive resistance to thyroid hormone in mice lacking F,A.P.&L, J. L. (1992). Dissociation of actin thyroid hormone receptor β: evidence for tissue specific polymerization and enzyme inactivation in the hormonal modulation of receptor function. The EMBO Journal 15, regulation of type II iodothyronine 5h-deiodinase activity in 3006–3015. astrocytes. Endocrinology 131, 721–728. F, P. A. (1988). Seasonal endocrine cycles in the F, A. P., S, M., D,S.&L,J.L. European hedgehog, Erinaceus europaeus. Journal of Reproduction (1996). Degradation and recycling of the substrate-binding & Fertility 84, 259–272. subunit of type II iodothyronine 5h-deiodinase in astrocytes. F, A., C, O., P, M., R,J.P, Journal of Biological Chemistry 271, 16369–16374. T, J., D, C., L, C., G, K., F, A. P., T,M.P.&L, J. L. (1995). K, M., M, L., R,B.&S,J. Thyroxine-dependent regulation of integrin-laminin inter- (1997). The T3Rα gene encoding a thyroid hormone receptor actions in astrocytes. Endocrinology 136, 3909–3915. is essential for post-natal development and thyroid hormone F, M., L,E.A.&V, L. A. (1991). Evaluation of production. The EMBO Journal 16, 4412–4420. the antioxidant properties of thyroid hormones and pro- F, S., C, G., P, G.,  C, L., pylthiouracil in the brain-homogenate autoxidation system J, E. A., P, S., F, E., L,B.& and in the free radical-mediated oxidation of erythrocyte D’A, M. (1991). Effect of thyroid hormone on the membranes. Chemico-Biological Interactions 77, 173–185. pre- and post-natal development of the rat testis. Journal of F, D. (1997). Understanding the Control of Metabolism. Portland Endocrinology 129, 35–42. Press, London. F, J. A., D, J. R., G, M. D., L, F, V., B, X., K, K., V,A. W. A., R,D.B.&S, M. C. (1985). Amioda- &V, L. A. (1985). Superoxide radical generation, rone and thyroid hormone action. Clinical Endocrinology 22, NADPH oxidase activity, and cytochrome P-450 content of 257–264. 614 A. J. Hulbert

F, S. A., A, J. B., S,D.A.&W,G.M. hormone concentrations and total contents through ontogeny (1971). Osteoporosis and fractures following thyrotoxicosis. in three anuran species: evidence for daily cycles. General & Lancet 1, 981–983. Comparative Endocrinology 107, 240–250. F,H.C.&O, J. H. (1995). Thermogenesis G, M. F., F-P, J., R, T., B- and thyroid function. Annual Review of Nutrition 15, 263–291. ,E.&F, R. (1992). Developmental modulation of F, G., L,J.M.&L, J. (1993). Thyroid myosin expression by thyroid hormone in avian skeletal hormones and putative nuclear T3 receptors in tissues of the muscle. Development 115, 1121–1131. ascidean, Phallusia mammillata Cuvier. Comparative & General G, T., W, R., H,M.&K, Endocrinology 92, 379–387. H. M. (1978). Serum somatomedin stimulation in thyroxine- F,H.&S, H. P. (1996). Phospholipid fatty treated hypophysectomized rats. Endocrinology 102, 606–611. acids of brain and liver are modified by α-tocopherol and G, A., M, M., L,J.F.&B, dietary fat in growing chicks. British Journal of Nutrition 76, T. (1996). Effect of feeding time on hepatic nucleic acid, 109–122. plasma T3, T4, and GH concentrations in rainbow trout. F, H., P, D. V., L,B.H.&K, F. A. (1996). Physiology & Behaviour 59, 1061–1067. Formation of vesicles by the action of acyl-CoA:1-acyllso- G, C. L. (1955). The apparent ionization constants of the phosphatidylcholine acyltransferase from rat liver micro- phenolic hydroxyl groups of thyroxine and related com- somes: optimal solubilization conditions and analysis of lipid pounds. Archives of Biochemistry & Biophysics 54, 359–367. composition and enzyme activity. Biochemistry 35, 2644–2650 G,J.C.&J, T. M. (1992). Flight behaviour and G, M. G., U,L.E.&F, R. N. (1981). Effect of thyroid hormone regulation in homing pigeons. Zoological membrane fatty acid composition on the action of thyroid Science 9, 675–677. # # hormones on (Ca +jMg +)-adenosine triphosphatase from G, M. C., G, D., F,S.H.&R,J. rat erythrocytes. Journal of Biological Chemistry 25, 7113–7114. (1976). Thyroxine-binding globulin synthesis by hepato- G, V. A. (1988). Iodothyronine 5h-deiodinase activity in carcinoma cells in continuous culture: effect of physiological the amphibian Rana catesbeiana at different stages of the life concentrations of thyroxine. Biochemical & Biophysical Research cycle. Endocrinology 122, 1746–1750. Communications 71, 76–82. G, V. A. (1990). Mechanisms underlying the acceleration G,R.W.&J-A, H. B. (1992). Growth and behaviour of thyroid deficient lizards (Sceloporus undulatus). of thyroid hormone-induced tadpole metamorphosis by Comparative & General Endocrinology 87, 312–324. corticosterone. Endocrinology 127, 2997–3002. G, C. K. (1996). Some new twists in the regulation of gene G, V. A. (1992a). The role of thyroid hormone in expression by thyroid hormone and retinoic acid receptors. amphibian metamorphosis. Trends in Endocrinology & Metab- Journal of Endocrinology 150, 349–357. olism 3, 96–100. G, F., L, A., H, C., M,M.&T,R. G, V. A. (1992b). Thyroid hormone receptors and (1994). In vitro binding of 3,5-di-iodo--thyronine to rat liver iodothyronine deiodinases in the developing Mexican axolotl, mitochondria. Journal of Molecular Endocrinology 13, 275–282. Ambystoma mexicanum. General & Comparative Endocrinology 85, G, I. D., S, G. J., S, C. G., I,S.H.& 62–70. J, E. C. (1976). Activities of thyroid hormones and G,V.A.&C, J. S. (1980). Action of thyroid related compounds in an in vitro thymocyte assay. Journal of hormones in premetamorphic tadpoles: an important role for Biological Chemistry 251, 4233–4238. thyroxine? Endocrinology 107, 1820–1826. G, S., D, M., A,H.&K,Z. G,V.A.&H, A. (1987). The ontogeny of the (1997). Triiodothyronine and follicle-stimulating hormone enzyme systems for the 5h- and 5-deiodination of thyroid alone and additively together, stimulate production of the hormones in chick embryo liver. Endocrinology 120, 2604–2610. tissue inhibitor of metalloproteinases-1 in cultured human G,V.A.&H, A. (1988). The ontogeny of luteinized granulosa cells. Journal of Clinical Endocrinology & iodothyronine 5h-monodeiodinase activity in Rana catesbeiana Metabolism 82, 1869–1873. tadpoles. Endocrinology 122, 640–645. G, J. M., B, T., B, G., S,A.&L B, G,V.A.&I, S. H. (1962). Observations on the P-Y. (1997). Individual diurnal plasma profiles of thyroid relation between the action and the degradation of thyroid hormones in rainbow trout (Oncorhynchus mykiss) in relation to hormones as indicated by studies in the tadpole and the frog. cortisol, growth hormone, and growth rate. Comparative & Endocrinology 70, 622–632. General Endocrinology 107, 74–83. G, V. A., MC,P.T.&S.G,D.L. G, D-W., H, Y., K,M.&R, M. (1997). (1991a). The ontogeny of iodothyronine deiodinase systems in Uncoupling protein-3 is a mediator of thermogenesis regu- liver and intestine of the rat. Endocrinology 128, 1717–1722. lated by thyroid hormone, β3-adrenergic agonists, and leptin. G, V. A., M, C. M., S,M.J.&Y, Journal of Biological Chemistry 272, 24129–24132. K. (1991b). The role of thyroid hormone in the regulation of G, O., G, E., S, C., P, J., G- hepatic carbamyl phosphate synthetase activity in Rana ,I.E, R., M, R., F,G.& catesbeiana. Endocrinology 129, 2298–2304. W, A. (1998). Effect of exercise on erythrocyte β- G, B., L, G. E., H, J., K,A.M.& adrenergic receptors and plasma concentrations of cate- R, N. A. (1983). Thyroidal and neural control of cholamines and thyroid hormones in thoroughbred horses. myosin transitions during development of rat fast and slow Equine Veterinary Journal 30, 72–78. muscles. FEBS Letters 156, 335–339. G, A. G. (1978). Regulation of malic enzyme synthesis G, B., A-G, A. L.,  P, N., D, by thyroid hormone and glucagon: inhibitor and kinetic M. J. & A-B, M. (1997). Changes in thyroid experiments. Molecular & Cellular Endocrinology 11, 19–29. Thyroid hormones and their effects 615

G, A., S, H., G, M., M, M., H,R.S.&L, J. N. (1984). Early enhancement of S,H.&G, J. (1994). The production of a passive potassium efflux from rat liver by thyroid hormone: monoclonal T3-antiidiotypic antibody (T3-MAAB) that relation to induction of Na,K-ATPase. Endocrinology 115, mimics the effects of T3 on 2-deoxy-D-glucose uptake in chick 291–297. embryo heart cells. Thyroid 4, 87–92. H, R. P., L,M.J.&B, M. D. (1989). Hypo- G, A. H., N,S.&B, R. M. (1974). thyroidism in rats decreases mitochondrial inner membrane Effects of thyroid hormone on growth, differentiation and cation permeability. FEBS Letters 248, 175–178. development. In Handbook of Physiology. Section 7. Volume III. H, R. P., N, C. D., MG,A.D.&B,M.D. Thyroid (eds. M. A. Greer & D. H. Solomon), pp. 377–389. (1988). Altered relationship between protonmotive force and Waverly Press, Baltimore. respiration rate in non-phosphorylating liver mitochondria G, R. I. (1963). Estimation of thyroxine secretion rate isolated from rats of different thyroid status. European Journal in the rat by radioactive thyroxine turnover technique: of Biochemistry 178, 511–518. influence of age, sex and exposure to cold. Endocrinology 72, H,M-E.&B, M. D. (1993). The quantitative 382–392. contributions of mitochondrial proton leak and ATP turnover G,E.L.&S, M. (1996). Thyroid hormone reactions to the changed respiration rates of hepatocytes from as a regulator of basal and human chorionic gonadotropin- rats of different thyroid status. Journal of Biological Chemistry stimulated steroidogenesis by cultured porcine theca and 268, 14850–14860. granulosa cells isolated at different stages of the follicular H,M-E.&B, M. D. (1995). Use of top-down phase. Reproduction, Fertility & Development 8, 961–967. G,R.B.&B, M. N. (1991). The administration of elasticity analysis to identify sites of thyroid-hormone-induced triiodothyronine to rats results in a lowering of the mito- thermogenesis. Proceedings of the Society for Experimental Biology chondrial membrane potential in isolated hepatocytes. Bio- & Medicine 208, 228–237. chimica et Biophysica Acta 1133, 89–94. H, D. R., G,W.L.&C, W. (1991). Acute G,R.B.&B, M. N. (1992). Stimulation by thyroid thyroid hormone promotes slow inactivation of sodium hormone of coupled respiration and of respiration apparently current in neonatal cardiac myocytes. Biochimica et Biophysica not coupled to synthesis of ATP in rat hepatocytes. Journal of Acta 1045, 175–181. Biological Chemistry 267, 8903–8908. H,A.P.&C, T. (1998). Thyroid hormone- + + + G, R. B., P, J. W., H,D.C.&B, induced upregulation of Na channels and Na -K pumps: M. N. (1996). Effects of thyroid status on glucose cycling by implications for contractility. American Journal of Physiology isolated rat hepatocytes. Metabolism 45, 101–108. 274, R864–R867.. G,R.&L, J. (1989). The effect of severe H, A. P., N,O.B.&C, T. (1997). Role of starvation and captivity stress on plasma thyroxine and Na+-K+ pump and Na+ channel concentrations in the triiodothyronine concentrations in an antarctic bird (emperor contractility of rat soleus muscle. American Journal of Physiology penguin). Comparative & General Endocrinology 73, 108–117. 274, R1402–R1408. G, J. (1993) The finding of 3h-3,5 triiodothyronine: personal H, A. P., T, D. R., C, T., D,C.& account. Thyroid 3, 161–169. D, M. J. (1996). Role of thyroid hormones in early G, J., L, C. P., F,A.E.&Q,J.H. postnatal development of skeletal muscle and its implications (1950). Presence of iodinated amino acids in unhydrolyzed for undernutrition. British Journal of Nutrition 76, 841–855. thyroid and plasma. Science 111, 605–608. H, R., W, N., K, R., L, M. A., G, G,E.&W, J. (1970). A molecular mechanism of C. K., A,J.W.&D, W. H. (1994). De- action of thyroxin: modification of membrane phospholipid lineation of three different thyroid hormone-response elements # by iodine. Journal of Theoretical Biology 26, 343–363. in promoter of rat sarcoplasmic reticulum Ca +ATPase gene. G-F, A., O, M. J., S.G,D.L.& Journal of Biological Chemistry 269, 13021–13029. B, J. (1997). The type 2 iodothyronine deiodinase is H, R., W, W. M., L, W. H., P, expressed primarily in glial cells in the neonatal brain. T. A., B,M.& B, M. H. (1987). Effects of Proceedings of the National Academy of Sciences, USA 94, amiodarone on thyroid hormone-responsive gene expression 10391–10396. in rat liver. Horm. Met. Res. Suppl. 17, 34–43. G, J. F. (1912). Feeding experiments on tadpoles. I. H, S. (1990). Thyroidal inhibition of growth hormone The influence of specific organs given as food on growth and secretion: negative feedback? In: Endocrinology of birds: differentiation. A contribution to the knowledge of organs Molecular to Behavioural. (eds. M. Wada, S. Ishii & C. G. with internal secretion. Wilhelm Roux Archiv fur Entwick- lungsmechanik Organismen 35, 457–483. Scanes), pp. 111–127. Springer-Verlag, Berlin. G,F.&P, A. (1997). Hepatic microsomal mem- H, M. E., H,A.F.&B, M. M. (1977). The brane lipidic composition and growth hormone effect in adult influence of membrane lipids on the proliferation of trans- male rat: evidence for a ‘feminization’ process of total formed and untransformed cell lines. Experimental Cell Research phospholipid fatty acid pattern. Biochimica et Biophysica Acta 107, 31–34. 1329, 97–110. H, T. B. (1997). Steroids as potential modulators of thyroid G, B.B.P.& T, J. P. (1991). Endocrine regu- hormone activity in anuran metamorphosis. American Zoologist lation of the oxidative metabolism in poikilothermic verte- 37, 185–194. brates. Zoological Science 8, 625–638. H,J.R.&W, E. E. (1990). The role of alterations H, R. S., I-B,F.&L, J. N. (1988). Time in membrane lipid composition in enabling physiological course of Na,K transport and other metabolic responses to adaptation of organisms to their physical environment. thyroid hormone in clone 9 cells. Endocrinology 123, 238–247. Progress in Lipid Research 29, 157–227. 616 A. J. Hulbert

H, M., J. O. O &W, H. G. (1985). H, R., C, S. Y., W,M.&P,I. Plasma lipoproteins and regulation of hepatic metabolism of (1982). Inhibition of the nuclear entry of 3,3h,5-triiodo-- fatty acids in altered thyroid states. Endocrine Reviews 6, thyronine by monodansylcadaverine in GH3 cells. Journal of 590–607. Biological Chemistry 257, 3139–3144. H, G. (1986). Thyroid hormone deiodination in H, D. J. (1977) Triiodothyronine and thyroxine levels in healthy man. In Thyroid Hormone Metabolism. (ed. G. Henne- the thyroid and serum of the sea lamprey Petromyzon marinus. mann) Marcel Dekker, New York, pp. 277–295. General & Comparative Endocrinology 31, 381–383. H, G., E, M. E.,  J, M., L, C.-F., H, B., H, C. B., Q, M., A, K. K., K,E.P.&D, R. (1998). The significance of P,D.&MD, A. A. (1990). Differential plasma membrane transport in the bioavailability of thyroid regulation of Na,K-ATPase α1, α2, and β subunit mRNA and hormone. Clinical Endocrinology 48, 1–8. protein levels by thyroid hormone. Journal of Biological H, B. S. (1989). The Story of Iodine Deficiency. Oxford Chemistry 265, 14308–14314. University Press, Oxford. H, C., R,H.&S, H. J. (1989). Rapid stimulation H, A. P. (1968a). The effect of serum on the uptake of of hepatic oxygen consumption by 3,5-di-iodo--thyronine. thyroid hormones by the perfused rat heart. Journal of Biochemical Journal 261, 945–950. Physiology 199, 161–168. H, C., H, A., S,H.J.&R, H. (1995). H, A. P. (1968b). The uptake and release of thyroxine and 3,5-di-iodo--thyronine suppresses TSH in rats in vivo and in triiodothyronine by the perfused rat heart. Journal of Physiology rat pituitary fragments in vitro. Journal of Endocrinology 145, 199, 151–160. 291–297. H, A. P. (1970). The binding of thyroid hormones to H, T., C, I. J., B, A., S,D.H.& phospholipid membranes. Journal of Physiology 211, 585–597. C T, G. N. (1985). Skin is an active site of inner ring H,F.J.R.&T, V. M. (1948). Paper partition monodeiodination of thyroxine to 3,3h,5h-triiodothyronine. chromatography with thyroxine and analogues. The Australian Endocrinology 117, 2106–2113. Journal of Science 10, 185–187. H, C., L, S. G., S, M., C,M.& Hoch, F. L. (1962). Biochemical actions of thyroid hormones. K, A. (1988). Membrane fluidity and lipid Physiological Reviews 42, 605–673. composition of rat small intestinal brush-border membranes during postnatal maturation. Biochimica et Biophysica Acta 939, H, F. L. (1982). Thyroid control over biomembranes. VII. 145–150. Heart muscle mitochondria from -triiodothyronine-injected H, A. J. (1978). The thyroid hormones: a thesis rats. Journal of Molecular & Cellular Cardiology 14, 81–90. concerning their action. Journal of Theoretical Biology 73, H, F. L. (1988). Lipids and thyroid hormones. Progress in 81–100. Lipid Research 27, 199–270. H, A. J. (1980a). Evolution from ectothermia towards H, F. L., S,C.D,G.A.& endothermia. In Advances in Physiological Sciences (eds. Z. M, J. F. (1981). Thyroid control over biomembranes: VI. Szelenyi & M. Szekely). Vol. 32. Pp. 237–247, Pergamon Lipids in liver mitochondria and microsomes of hypothyroid Press, London. rats. Lipids 16, 328–335. H, A. J. (1980b) The evolution of energy metabolism in H, R. A., L,M.A.&C, W. W. (1990). Differential mammals. In Comparative Physiology: Primitive Mammals (eds. and tissue specific regulation of multiple rat c-erbA messenger K. Schmidt-Nielsen, L. Bolis & C. R. Taylor), pp. 129–139. RNA species by thyroid hormone. Journal of Clinical Investi- Cambridge University Press, Cambridge. gation 85, 101–105. H, A. J., A,M.L.&R, J. K. (1976). The H, J. F., MG,P.A.&H, P. T. (1978). Electro- influence of thyroid hormones on the structure and function of phoretic analysis of multiple forms of rat cardiac myosin: mitochondrial membranes. Biochimica et Biophysica Acta 455, effects of hypophysectomy and thyroxine replacement. Journal 597–601. of Molecular & Cellular Cardiology 10, 1053–1076. H,A.J.&A, M. L. (1982). A comparative study of H, A. N., M, T., F, T. R., B, M-E., thyroid function in monotreme, marsupial and eutherian C,O.&W, F. E. (1995). The human mammals. Physiological Zoology 55, 220–228. thyrotropin-releasing hormone gene is regulated by thyroid H,A.J.&E, P. L. (1981). A comparison of the hormone through two distinct classes of negative thyroid ‘mammal machine’ and the ‘reptile machineh: energy use hormone response elements. Molecular Endocrinology 9, and thyroid activity. American Journal of Physiology 241, 540–550. R350–R356. H, R. W., B,J.H.&K, J. A. (1974). H,A.J.&E, P. L. (1999). Membranes as possible Control of growth of a tumor cell by linoleic acid. Proceedings pacemakers of metabolism. Journal of Theoretical Biology 199, of the National Academy of Sciences, USA 71, 3976–3978. 257–274. H, A. C., R, P. K., S,M.A.&L- H,A.J.&E, P. L. (2000). Mechanisms underlying , J. F. (1994). Diurnal rhythms of plasma growth the cost of living in animals. Annual Review of Physiology 62, hormone, somatostatin, thyroid hormones, cortisol and 207–235. glucose concentrations in rainbow trout, Oncorhynchus mykiss, H,A.J.&G, T. R. (1983). Thyroid hormone levels during progressive food deprivation. Biological Rhythm Research in an egg-laying mammal, the platypus Ornithorhynchus 25, 415–432. anatinus. General & Comparative Endocrinology 51, 401–405. H, A. C., W, K. Y., P, N. B., J,E.C.& H,A.J.&W, C. A. (1988). Thyroid function in G, I. D. (1980). Uptake of -triiodothyronine into a lizard, a tortoise and a crocodile, compared with mammals. human cultured lymphocytes. Acta Endocrinologica 95, 350–358. Comparative Biochemistry & Physiology 90A, 41–48. Thyroid hormones and their effects 617

H,H.&S, S. (1989). Rapid stimulation of J-A, H. B. (1983). Effects of thyroxine supplementation calcium uptake into rat liver by -tri-iodothyronine. Bio- on metabolic rate and aerobic capacity in a lizard. American chemical Journal 258, 363–367. Journal of Physiology 244, R659–R666. I, A., N. Y,Y.M,T.U,M. J-A, H. B. (1984a). Reduced aerobic capacity and Y &S.M (1989) Unesterified long-chain locomotory endurance in thyroid-deficient lizards. Journal of fatty acids inhibit thyroid hormone binding to the nuclear Experimental Biology 109, 175–189. receptor. European Journal of Biochemistry 183, 565–572. J-A, H. B. (1984b). Seasonal variations in activity, I, Z., K,H.&T J., J. J. (1984). Tri- aerobic energetic capacities, and plasma thyroid hormones iodothyronine (T$) regulates membrane transport processes (T3 and T4) in an iguanid lizard. Journal of Comparative in mature neurons and nerve endings. Neurology 34 (Suppl. 1), Physiology B 154, 409–419. 195. J-A,H.B.&J, B. (1991). Interactive effects of I,C.H.G.&E, M. J. (1975). Postnatal changes in thyroxine and experimental location on running endurance, total and free thyroxine and triiodothyronine in foal serum. tissue masses, and enzyme activities in captive versus field- Journal of Reproduction & Fertility, Supplement 23, 709–715. active lizards (Sceloporus undulatus) Comparative & General I, S., A,M.&L, I. (1989). Endocrinology 81, 120–132. Glutathione-S-transferases are major cytosolic thyroid hor- J, L. G. (1997). Thyroxine’s evolutionary roots. Per- mone binding proteins. Archives of Biochemistry & Biophysics spectives in Biology & Medicine 40, 529–535. 273, 265–272. J,B.&J-A, H. B. (1990). Effects of thyroxine on I-O,A.&S, A. (1992). Connective tissue standard and total metabolic rates in the lizard Sceloperus is involved in adult epithelial development of the small undulatus. Physiological Zoology 63, 873–885. intestine during anuran metamorphosis in vitro. Roux’s Archives J, E. C. (1981). Thyromimetic and antithyroid drugs. of Developmental Biology 201, 322–329. In Burger’s Medicinal Chemistry 4th edition. Part III (ed. M. E. I-B, F. (1993). Thyroid hormone regulation of Na,K- Wolff), pp. 103–145. John Wiley & Sons, New York. ATPase expression. Trends in Endocrinology & Metabolism 4, K-U, H., J, C., F,G.&B, 152–155. Z. (1987). Work performance, thermoregulation and muscle I-B,F.&E, I. S. (1970). Mechanism of thyroid metabolism in thyroidectomized goats (Capra hircus). Com- calorigenesis: role of active sodium transport. Proceedings of the parative Biochemistry & Physiology 87A, 915–921. National Academy of Sciences, USA 67, 1071–1078. K, Y., Y, K., M, M., O, A., I-B,F.&E, I. S. (1971). The mechanism of T,K.&Y, Y. (1987). Comparison of the calorigenic action of thyroid hormone: stimulation membrane phospholipid and its fatty acid compositions in of Na+ K+-activated adenosine-triphosphatase activity. j developing rat salivary glands. Comparative Biochemistry & Journal of General Physiology 57, 710–722. Physiology 87B, 741–746. I, G., B, V. T., G’,A.I.& K,A.&B, D. D. (1992). The regulation of S, T. S. (1986). Influence of various contents of thyroid thyroid hormone receptor β genes by thyroid hormone in hormones on the fatty acid composition of the lipids of rat Xenopus laevis. Journal of Biological Chemistry 267, 739–745. liver mitochondria. Biokhimiya 51, 80–83. K, M. M. (1986). Regulatory influences on iodothyronine I, E. J., P, K. Y., H,P.R.&W,W.A. (1987). Structural and functional maturation of rat gas- deiodination in animal tissues. In Thyroid Hormone Metabolism. trointestinal barrier with thyroxine. American Journal of (ed. G. Hennemann), pp. 231–253, Marcel Dekker, New Physiology 252, G762–G767. York. I, S., N-G,B.&M, V. (1986). All K,M.M.&U, R. D. (1978). Iodothyronine members of the MHC multigene family respond to thyroid metabolism in liver and kidney homogenates from hyper- hormone in a highly tissue-specific manner. Science 231, thyroid and hypothyroid rats. Endocrinology 103, 156–161. 597–600. K, E. M. (1986). Thyroid hormone metabolism in illness. J, S., S, H. H., G,N.J.&R,M. In Thyroid Hormone Metabolism (ed. G. Hennemann). pp. (1978). Biochemical and stereological analysis of rat liver 297–333. Marcel Dekker, New York. mitochondria in different thyroid states. Journal of Cell Biology K, F. J., D, G. E., H,T.M.&T,L.A. 77, 887–901. (1995). Involvement of thyroid hormones in seasonal re- J, H. T., L, L. S., M, E., DG,L.J.& production. Journal of Reproduction & Fertility Supplement 49, L, M. N. (1997). Thyroid hormone receptor (α) 409–422. distribution in hamster and sheep brain: colocalization in K, R., K,D.&A, R. (1987). De- gonadotropin-releasing hormone and other identified neur- velopmental patterns of thyroid hormones and testosterone in ons. Endocrinology 138, 5039–5047. ferrets. Hormone & Metabolism Research 19, 194–196. J, P. A., G, J., D,H.&H, A. J. (1990). K, M., P,E.L.&F, G. (1984). Regulation of Thyroid hormones during development of a marsupial, the membrane fluidity by lipid desaturases. In: Biomembranes. tammar wallaby, Macropus eugenii. Journal of Endocrinology 127, Vol. 12. Membrane Fluidity (eds. M. Kates & L. A. Manson), 427–436. pp. 379–395. Plenum Press, New York. J, E. V. (1991). Overview of the nuclear receptor family. K, H., F, T., P, C., MP,P.&C, In Nuclear Hormone Receptors pp. 1–13. Academic Press S.-Y. (1989a). Cytolsolic thyroid hormone-binding protein is J, T. M., G,J.C.&B, G. M. (1995). Effects of a monomer of pyruvate kinase. Proceedings of the National exogenous arginine vasotocin on circulating levels of thyroid Academy of Science, USA 86, 7861–7865. hormones and melatonin in the pigeon, Columbia livia. K, H., V,T.&C, S-Y. (1989b). High level Comparative Biochemistry & Physiology 112C, 345–351. expression of p55, a thyroid hormone binding protein which 618 A. J. Hulbert

is homologous to protein disulfide isomerase in a retroviral K, E. P., D, R., B, H. F., V,R.J.& vector. Biochemical & Biophysical Research Communication 164, H, G. (1978). Active transport of triiodothyronine 238–244. (T3) into isolated rat liver cells. FEBS Letters 91, 113–116. K,B.G.&B, C. L. (1997). Interaction of cortisol and K,E.P.&D, R. (1986). Plasma membrane thyroid hormone in the larval development of Pacific transport of thyroid hormone. In Thyroid Hormone Metabolism. threadfin. American Zoologist 37, 470–481. (ed. G. Hennemann), Marcel Dekker. New York, pp. 107–131. K, R., T, M., T,M.&H, T. (1992). K, D., L,D.C.&R, H. (1984). #+ Developmental changes in tissue thyroid hormone levels of The role of fatty acids in the regulation of brush-border Ca red sea bream Pagrus major. Nippon Suisan Gakkaishi 58, 975. transport. Biochimica et Biophysica Acta 777, 261–266. K, B. M., F,B.B.&G, I. D. (1981). Quantitative K,J.&S, B. (1985). Thyroid hormones increase "#& Na+ –H+ exchange activity in renal brush border membranes. electron microscopic autoradiograph of [ I]-triiodo- Proceedings of the National Academy of Sciences, USA 82, thyronine: evidence for multiple sites of hormone localization. 3606–3610. Clinical Research 29, 506A.    K, K. (1972a). Fatty acid composition of serum lipids K , E. R., M ,K.A.&D , V. M. (1993). Control strategies of thyroid hormone monodeiodination in verte- following a single intake of thyroxine. Acta Endocrinologica 71, brates. Zoological Science 10, 873–885. 73–79. K, H., R, T. F., D, D. M., M,M.& K, K. (1972b). Fatty acid composition of serum lipids in K, S. L. (1978). Developmental changes in the fatty acid hyper- and hypothyroidism. Acta Endocrinologica 71, 62–72. composition and cholesterol content of chicken heart plasma K, S., O, T., W,M.C.&C,S-Y. membrane. Journal of Comparative Physiology 125, 151–156. (1987). Thyroid hormone action: induction of morphological K, J. (1992). 3,5-T2 stimulates oxygen consumption, but changes and stimulation of cell growth in rat pituitary tumor not glucose uptake in human mononuclear blood cells. GH$ cells. Endocrinology 120, 2591–2596. Hormone & Metabolism Research 24, 322–325. K, K., V, F., G, H., F- K, A. H. M., T, M. F., C,R.R.&F,H. Z, N., C, E., L, H. J., K,K.& C. (1994). Serum free thyroxine in cats maintained on diets P, M. (1995). The regulatory role of thyroid hormones relatively high or low in iodine. New Zealand Veterinary Journal in bone cell growth and differentiation. Journal of Nutrition 42, 101–103. 125, 1996S–2003S. L,F.&P, V. (1960). Activite thyroidienne K,S.J.&G, W. L. (1973). Degradation of comparee de deux hibernants: le lerot (Eliomys quercinus), thyroid hormones by phagocytosing human leukocytes. The spermophile (Citellus citellus) et d’un homeotherme: le rat, au Journal of Clinical Investigation 52, 60–72. cours de l’exposition au froid pendant l’ete. Comptes Rendu 250, K, M. (1961). The Fire of Life. An Introduction to Animal 3883–3885. Energetics. Wiley, New York. L, C.-S. & C, S.-Y. (1982). Rotational and lateral Klein, I. (1990). Thyroid hormone and the cardiovascular diffusions of -thyroxine in phospholipid bilayers. Biochimica et system. American Journal of Medicine 88, 631–637. Biophysica Acta 692, 27–32. K, R. J., B,G.A.W, R. L., L,P.R.& L, C.-S. & C, S.-Y. (1984). Molecular dynamics of M, D. D. (1987). Thyroid hormone receptor binds to a 3,3h,5-triiodo--thyronine in model membranes: a spin label site in the rat growth hormone promoter required for study. Archives of Biochemistry and Biophysics 232, 477–481. induction by thyroid hormone. Proceedings of the National L, C.-S., K, W., N, C.-H. & C, S.-Y. Academy of Sciences, USA 84, 5670–5674. (1985). Transverse motion of spin-labelled 3,3h5-triiodo- K, J., A’, M., S, M., I, K., C, thyronine in phospholipid bilayers. Biochemical and Biophysical V. & H, R. D. (1986). Iodothyronine deiodinase is Research Communications 131, 408–412. inhibited by plant flavenoids. In Plant Flavenoids in Biology & L, P., C, M. D., K,V.&S,P.H. Medicine (eds. V. Cody, E. Middleton Jr. & J. B. Harborne), (1993). Thyroid hormones increase insulin-like growth factor pp. 359–371. Alan R. Liss Inc., New York. 1 content in the medium of rat bone tissue. Journal of Bone & Mineral Research 8, 1475–1481. K, J., B,G.&H, R-D. (1987). Metabolism of L,P.&S, P. H. (1991). Evidence for direct non- the thyroid hormones. Hormone Research 26, 58–78. genomic effects of triiodothyronine on bone rudiments in rats: K,L.&T, M. (1976). Thyroxine-protein stimulation of the inositol phosphate second messenger system. interactions. Interaction of thyroxine and triiodothyronine Acta Endocrinologica Copenhagen 125, 603–608. Journal of Biological with human thyroxine-binding globulin. L, L. C., G-P,T.&R- Chemistry 251, 3558–3562. F, C. (1987). Uptake of thyroid hormones (-T3 K, L. (1994). Membrane iodothyronine transporters part and -T4) by isolated rat adipocytes. Biochemical & Biophysical 1: review of physiology. Endocrine Research 20, 319–341. Research Communications 145, 105–110. K, L. (1996). Membrane iodothyronine transporters part L, M.-F., Z, K., M, T., S, J. D., 2: review of protein biochemistry. Endocrine Research 22, H,A.N.&W, F. E. (1997). A unique 95–119.. role of the β-2 thyroid hormone receptor isoform in negative K,L.&S, R. (1993). Measures of skeletal regulation by thyroid hormone. Journal of Biological Chemistry muscle calcium channels and acetylcholine receptors in 272, 24927–24933. thyroidectomized rats. Endocrine Research 19, 207–209.4 L, A., D F, M., L, A., M, M., K,L.&S, M. E. (1994). Acetylcholinesterase activity F, C., R,D.&G, F. (1997). Induction in fast-twitch skeletal muscle from thyroidectomized rats. of UCP2 mRNA by thyroid hormones in rat heart. FEBS Endocrine Research 247–57. Letters 418, 171–174. Thyroid hormones and their effects 619

L, A., M, M., L,A.&G, F. (1994a). L,J.L.&F, A. P. (1997). Thyroid hormone- Rapid stimulation in vitro of rat liver cytochrome oxidase regulated actin polymerization in brain. Thyroid 7, 147–151. activity by 3,5-diiodo--thyronine and by 3,3h-diiodo-- L, J. L., F, A. P., Y, P. M., C,W.W.& thyronine. Molecular & Cellular Endocrinology 99, 89–94. S, M. (1994). Differential expression of thyroid hormone L, A., M, M., H, C., L,A.&G, receptor isoforms in neurons and astroglial cells. Endocrinology F. (1994b). Specific binding sites for 3,3h-diiodo--thyronine 135, 548–555. (3,3h-T2) in rat liver mitochondria. FEBS Letters 351, 237–240. L,J.L.&K, J. (1996). Intracellular pathways L, A., M, M., L,A.&G, F. (1996). of iodothyronine metabolism. In Werner and Ingbars. The Calorigenic effect of diiodothyronines in the rat. Journal of Thyroid. 7th edition (eds. L. E. Braverman & R. D. Utiger), Physiology 494, 831–837. pp. 125–161. Lippincott-Raven Publishers, Philadelphia. L, P. R., S,J.E.&K, M. M. (1981). Re- L, J. L., S-K,C.A.&Z,C.J. lationship between circulating and intracellular thyroid (1990). Regulation of type II iodothyronine 5h-deiodinase by hormones: physiological and clinical implications. Endocrine thyroid hormone. Journal of Biological Chemistry 265, 940–946. Reviews 2, 87–102. LeRoith, D., Adamo, M., Werner, H. & Roberts, C. T., Jr. L, M., P,T.&C, A. (1985). Thyroid (1995). Molecular and cellular biology of insulin-like growth hormone binding in serum of 15 vertebrate species: isolation factors In Molecular Endocrinology: Basic Concepts and Clinical of thyroxine-binding globulin and prealbumin analogs. Correlations. (ed. B. D. Weintraub), pp. 181–193. Raven Press, General & Comparative Endocrinology 58, 360–375. New York. L, W. D., S,M.&D, P. J. (1989). L, P., W,J.F.&W, F. E. (1985). Annual and Stimulation in vitro of rabbit erythrocyte cytosol phos- diurnal cycles in plasma testosterone and thyroxine in the pholipid-dependent protein kinase activity. Journal of Bio- male green sea turtle Chelonia mydas. Comparative & General logical Chemistry 264, 4766–4768. Endocrinology 57, 335–344. L, V. J., C, A. J., M, V. P., S, L, P., D, R. J., P,S.&H, B. (1991). J. R., C,F.N.&B, J. W. (1996). Concen- Seasonality in plasma thyroxine binding in turtles. The trations of thyroid hormones and iodothyronine binding Journal of Experimental Zoology 260, 59–65. proteins in serum of the koala (Phascolarctus cinereus). Australian L, P., S, Y., H, S., G,W.&F, Veterinary Journal 74, 55–58. R. J. (1998). Hypothyroidism decreases the ATP sensitivity of L,M.A.&C, W. W. (1988). Regulation of two c-erbA KATP channels from rat heart. Journal of Membrane Biology 162, messenger ribonucleic acids in rat GH3 cells by thyroid 217–223. hormone. Molecular Endocrinology 2, 479–484. L, C-F., B, Y., T, D. J., B,J.W.&S, L, M. A. (1993). Thyroid hormone receptors: multiple J. R. (1988). Drug and fatty acid effects on serum thyroid forms, multiple possibilities. Endocrine Reviews 14, 184–193. hormone binding. Journal of Clinical Endocrinology & Metabolism L, S. C., B,K.N.&B, J. S. (1996). 67, 682–688. Direct effects of 3,5,3h-triiodothyronine and 3,5-diiodo- L, C-F., S,J.R.&H, G. (1995). Alter- thyronine on mitochondrial metabolism in the goldfish ations in hepatocyte uptake and plasma binding of thyroxine Carassius auratus. General & Comparative Endocrinology 104, in nonthyroidal illness and caloric deprivation. Trends in 61–66. Endocrinology & Metabolism 6, 17–20. L, S. C., R, K., B, B. J., D,S.& L, H.-Y., T, H. R., D,P.J.&D, F. B. (1994). B, J. S. (1997). Short-term effects of 3,5,3h- Thyroid hormone potentiates the antiviral action of inter- triiodothyronine on the activity of selected enzymes of feron-γ in cultured human cells. Journal of Clinical Endocrinology intermediary metabolism in the liver of the Atlantic Hagfish, & Metabolism 79, 62–65. Myxine glutinosa. The Journal of Experimental Zoology 278, L, H.-Y., T, H. R., D,F.B.&D, P. J. (1996). 333–337. Thyroid hormone analogues potentiate the antiviral action of L, J. F., L, L., D,N.E.&D,E.M. interferon-γ by two mechanisms. Journal of Cellular Physiology (1989). Thyroid hormone content of eggs and early de- 167, 269–276. velopmental stages of five Oncorhynchus species. Canadian Journal L, H.-Y., Y, P. M., D,F.B.&D, P. J. (1997). of Fisheries & Aquatic Science 46, 2140–2145. Protein synthesis-dependent potentiation by thyroxine of L,Y.A.&V, J. T. (1984). Charac- antiviral activity of interferon-γ. American Journal of Physiology terization of a thyroid-hormone-binding site on nuclear 273, C1225–C1232. envelopes and nuclear matrices of the male-rat liver. L, K.-H., A,K.&C, S.-Y. (1992). Phos- Biochemical Journal 219, 1001–1007. phorylation stimulates the transcriptional activity of the L, J. (1986). Thyroid hormone effects on growth and human β1 thyroid hormone nuclear receptor. Proceedings of the development. In Thyroid Hormone Metabolism. (ed. G. Henne- National Academy of Sciences, USA 89, 7737–7741. mann) pp. 503–534, Marcel Dekker, New York. L, K. H., Z, X.-G., H, H.-C., C, S.-L., S, H.-Y., L, M. N., G, R. L., K, F. J., J, C, S.-T., MP,P.&C, S.-Y. (1997). Dominant G. L., B,S.J.&J, H. T. (1997). The GnRH negative activity of mutant thyroid hormone α1 receptors system of seasonal breeders, anatomy and plasticity. Brain from patients with hepatocellular carcinoma. Endocrinology Research Bulletin 44, 445–457. 138, 5308–5315. L,A.L.&B, R. R. (1989). Fatty acid desaturase L,L.&N, C. S. (1986). Postnatal development of activities are modulated by phytosterol incorporation in dependence on thyroid hormones for growth and differ- microsomes. Biochimica et Biophysica Acta 1005, 187–191. entiation of rat skeletal structures. Growth 50, 472–485. 620 A. J. Hulbert

L, X-T., L, Q-F., H,C-X.&S, R-Y. (1997). Effects M, B. (1985). The isoenzymes of glutathione trans- of thyroid status on cold-adaptive thermogenesis in brandt’s ferase. Advances in Enzymology & Related Areas in Molecular vole, Microtus brandti. Physiological Zoology 70, 352–361. Biology 57, 357–417. L, H-C., & T, H. C. (1994). Functional synergism M, L. (1972). Studies with electron microscopic "#& between multiple thyroid hormone response elements regu- autoradiography of thyroxine I inorganotypic cultures of "#& lates hepatic expression of the rat S"% gene. Molecular the CNS. II. Sites of cellular localization of thyroxine I. Endocrinology 8, 1021–1037. Yale Journal of Biology & Medicine 45, 501–518. L, R. B., C, P. H., A, A. P., C, M, R. G., E,J.G.&Y, J. H. (1998). Blocking K. S. R. & R, W. R. (1991). Tri-iodothyronine of KClO%-induced metamorphosis in premetamorphic sea increases intra-cellular calcium levels in single rat myocytes. lampreys by exogenous thyroid hormones (TH); effects of Journal of Molecular Endocrinology 7, 77–79. KClO% and TH on serum TH concentrations and intestinal L, A., L, A., M, M., B,M.D.& thyroxine outer-ring deiodination. General & Comparative G, F. (1998). Effect of 3,5-di-iodo--thyronine on the Endocrinology 112, 54–62. mitochondrial energy transduction apparatus. Biochemical M, S., C-M, I., M, J. P., R, Journal 330, 521–526. J. P., W,C.&C, G. (1995). Stimulation of L,B.A.&W, R. A. (1992). A new method for avian myoblast differentiation by triiodothyronine: possible quantitating intracellular transport, application to the thyroid involvement of the cAMP pathway. Experimental Cell Research hormone 3,5,3h-triiodothyronine. American Journal of Physiology 220, 1–10. 263, G733–G741. M, M., D, M. J., R,D.B.&I,D.L. M, J., P, M., J-N, I., C,J.& (1983). Thyroid hormone metabolism after acclimatization to J, A. (1988). Effect of thyroid hormones on urea a warm or cold temperature under conditions of high or low biosynthesis and related processes in rat liver. Endocrinology energy intake. Quarterly Journal of Experimental Physiology 68, 123, 2167–2174. 709–718. M,E.&W, W. (1991). Characterization of M,I.&T, J. R. (1992). Autoinduction of thyroid hormone response elements. In Nuclear Hormone Receptors: hormone receptor during metamorphosis is reproduced in Molecular Mechanisms, Cellular Functions, Clinical Abnormalities. Xenopus XTC-2 cells. Molecular & Cellular Endocrinology 87, (ed. M. G. Parker), pp. 125–153. Academic Press, London. 105–113. M, I., D, B., A, A., L,A.& MK, D. S., S, M., P,R.E.&B, B, G. (1995). Effects of T3 and rearing temperature on B. (1987). Increased gonadotropin levels in goldfish do not growth and skeletal myosin heavy chain isoform transition result in alterations in circulating thyroid hormone levels. during early development in the salmonid Salvelinus alpinus General & Comparative Endocrinology 67, 202–213. (L.). Comparative Biochemistry & Physiology 112B, 717–725. M, M., M, K., H, D., P, D., M, K., K,N.&M, J. D. (1995). Thyroid D’H, P., R, R., D M,F.&S,S. hormones on embryo development and appearance of myosin (1997). Components of biological variation, including season- heavy chain isoforms in turkeys (Meleagris gallopavo). Com- ality, in blood concentrations of TSH, TT3, FT4, PRL, parative Biochemistry & Physiology 112C, 109–117. cortisol and testosterone in healthy volunteers. Clinical M,P.M.&M, L. (1987). Thyroid hormone Endocrinology 46, 587–598. induction of keratin genes: a two-step activation of gene M,P.J.&P, L. (1999). Astrocytes couple expression during development. Genes & Development 1, synaptic activity to glucose utilization in the brain. News in 1107–1117. Physiological Sciences 14, 177–182. M, S., T, B., H, A., K, C., B, M-L, A. (1895). Uber den respiratorischen gaswechsel H., G, H., H,W.&K, H. H. (1995). Effect unter dem einfluss der thyroidea sowie unter verschiedenen of in vivo thyroid hormone status on insulin signalling and pathalogischen zustanden. Berlin Klin. Wochenschr. 34, GLUT1 and GLUT4 glucose transport systems in rat 650. adipocytes. Journal of Endocrinology 144, 347–357. M, M. J. (1961). The effect of environmental temperature M, A. J. (1985). Fish Endocrinology. Croom Helm, London. on metabolic response to thyroxine in the lizard, Lacerta M, B. C., K, J. A., W,M.L.&E, muralis. American Zoologist 1, 461. A. M. (1992). Effect of fatty acids on rat liver nuclear T3- M, M. J. (1964). Metabolic response of isolated lizard receptor binding. Metabolism 41, 788–792. tissues to thyroxine administered in vivo. Endocrinology 74,    994–995. M D ,M.C.&H , S. J. (1992). Synergistic effects M, M. J. (1967). Response to thyroxine as a function of of thyroxine and dexamethasone on enzyme ontogeny in rat environmental temperature in the toad, Bufo woodhousii, and small intestine. Pediatric Research 32, 306–311. the frog, Rana pipiens. Copeia 2, 361–365. MG, M. E., W, R. L., A, J. W., W, M, A. L., B, M. J., S, R., H,J.W.& B. L., R, V., B,J.D.&F,R.J. L, P. R. (1995). Structural and functional differences in (1994). Preliminary crystallographic studies of ligand-binding the Dio1 gene in mice with inherited type 1 deiodinase domain of the throid hormone receptor complexed with deficiency. Molecular Endocrinology 9, 969–980. triiodothyronine. Journal of Molecular Biology 237, 236–239. M, T., S, R., S, Y., O, N., N, Y., MI,B.&G, C. A. (1997). Euthyroid sick syndrome: M, T., K, M., H, M., M, R., an overview. Thyroid 7, 125–132. N,A.&N, A. (1995). Effects of thyroid hormone MN, F. M. A. (1988). Peripheral thyroid hormone dy- on coenzyme Q and other free radical scavengers in rat heart namics in precocial and altricial avian development. American muscle. Journal of Endocrinology 145, 131–136. Zoologist 28, 427–440. Thyroid hormones and their effects 621

MN,F.M.A.&C, M.-F. (1985). Thyroid devel- M, S., T, M., I,Y.&H, T. (1988). opment in altricial ring doves, Streptopelia risoria. General & Thyroxine surge in metamorphosing flounder larvae. General Comparative Endocrinology 58, 243–251. & Comparative Endocrinology 78, 158–163. MN, F. M. A., L,L.J.&H, T. E. (1984). Free M, S., Y,K.&I, Y. (1992). Thyroid hormone thyroid hormones in altricial (ring doves) versus precocial stimulates gastric development in flounder larvae during (Japanese quail) development. Endocrinology 115, 2133–2136. metamorphosis. The Journal of Experimental Zoology 261, MN,F.M.A.&O, J. M. (1996). Development of 424–430. thermoregulation and its hormonal control in precocial and M,P.M.&K, J. C. (1979). Thyroxine altricial birds. Poultry and Avian Biology Reviews 7, 111–125. concentrations in blood serum and pericardial fluid of MN,F.M.A.&W, C. M. (1997). Thyroid hormone metamorphosing tadpoles and adult frogs. General & Com- deposition in avian eggs and effects on embryonic devel- parative Endocrinology 39, 343–349. opment. American Zoologist 37, 553–560. M, A. (1992). The time course of thyroid-hormone- M, C. M. (1989). The free hormone hypothesis: a induced changes in the isotonic and isometric properties of rat physiologically based mathematical model. Endocrine Reviews soleus muscle. Pflugers Archiv 421, 350–356. 10, 232–274. M, C. T., L,C.J.L.&M, A. W. (1996). M, C. M., F,P.H.&C, R. R. (1986). Effect of illness not associated with the thyroid gland on serum Effect of free fatty acids on the concentration of free thyroxine total and free thyroxine concentration in cats. Journal of in human serum: the role of albumin. Journal of Clinical American Veterinary Medical Association 208, 2004–2008. Endocrinology & Metabolism 63, 1394–1399. M, A. D., H, M. P., D,F.&Y- M, C. M., W, R. A., J,A.L.&C, , T. (1994). Effect of age on -3,5,3h-triiodo- R. R. (1987). Thyroid hormone-binding proteins in plasma thyronine-induced ethane exhalation. Journal of Applied facilitate uniform distribution of thyroxine within tissues: a Physiology 77, 160–164. perfused rat liver study. Endocrinology 120, 1742–1749. M,A.D.&W, N. C. W. (1994). Age-related M-H, A., V-N, E., S,A.& changes in thyroid hormone action. European Journal of P-C, A. (1997). Regulation by thyroid hormone Endocrinology 131, 451–461. and retinoic acid of the CCAAT\enhancer binding protein α M, M., L, A., L,A.&G, F. (1997). and β genes during liver development. Biochemical & Biophysical How the thyroid controls metabolism in the rat: different Research Communications 234, 605–610. roles for triiodothyronine and diiodothyronines. Journal of M, N. T., J, P. J., R, K. R., N, Physiology 505, 529–538. R. F., G,V.K.&K, G. F. (1995). Effect of food M, M., L, A., L, P., G,F.& deprivation on baseline iodothyronine and cortisol concen- L, A. (1998). Effect of 3,5,-diiodo--thyronine on thyroid trations in healthy, adult horses. American Journal of Veterinary stimulating hormone and growth hormone serum levels in Research 56, 116–121. hypothyroid rats. Life Science 62, 2369–2377. M, G., A,D.&E, C. (1996). Development of M, P-P., D,J.J.&D, F. Y. (1989). Thyroid vestibular and auditory function: effects of hypothyroidism activity concomitant with olfactory learning and heart rate and thyroxine replacement therapy on nystagmus and changes in Atlantic salmon, Salmo salar, during smoltification. auditory evoked potentials in the pigmented rat. International Canadian Journal of Fisheries and Aquatic Sciences 46, 131–136. Journal of Developmental Neuroscience 14, 515–522. M, P-P., H,T.J.&E, J. G. (1993). Thyroid M, J. P., T, A. M., Z, M., M, H. G., hormone deiodination in brain, liver, gill, heart and muscle of R,R.J.M.&V, R. (1993). Effects of hypo- Atlantic salmon (Salmo salar) during photoperiodically- thyroidism and hyperthyroidism on insulin-like growth factors induced parr-smolt transformation. I. Outer- and inner-ring (IGFs) and growth hormone- and IGF-binding proteins. thyroxine deiodination. General & Comparative Endocrinology 90, Journal of Clinical Endocrinology & Metabolism 76, 950–955. 142–156. M,R.V.&T, H. C. (1982). Changes in the rates M, P., C, E., S,C.&L,C. of synthesis and messenger RNA levels of hepatic glucose-6- (1991). The response of rat liver lipid peroxidation, anti- phosphate and 6-phosphogluconate dehydrogenases following oxidant enzyme activities and glutathione concentration to induction by diet or thyroid hormone. Journal of Biological the thyroid hormone. International Journal of Biochemistry 23, Chemistry 257, 11829–11835. 1025–1030. M, A. R., N, R. W., S-M, J. C., N, M, A. S., S, M. R., H,R.W.P.& I. & B, G. D. (1992). Serial thyroid hormone concen- D, W. H. (1997). Transcription of the rat sarco- # trations in healthy euthyroid dogs, dogs with hypothyroidism, plasmic reticulum Ca + adenosine triphosphatase gene is and euthyroid dogs with atopic dermatitis. British Veterinary increased by 3,5,3h-triiodothyronine receptor isoform-specific Journal 148, 451–458. interactions with the myocyte-specific enhancer factor-2a. M, J., K,E.R.& G, H. (1985). Plasma Endocrinology 138, 26–32. concentrations of thyroid hormones in dogs: influence of M, E. (1993). Regulation of myosin heavy chain genes in sampling hour, breed and age. Chronobiology International 2, the heart. Circulation 87, 1451–1460. 121–129. M  E, G., O,M.J.&E  M,S.&I, Y. (1987). Effects of various doses of thyroxine R, F. (1987). Fetal and maternal thyroid hormones. and triiodothyronine on the metamorphosis of flounder Hormone Research 26, 12–27. (Paralichthys olivaceus). General & Comparative Endocrinology 67, M  E, G., O, M. J., R  O,C. 356–363. &E  R, F. (1989). Comparison of maternal to 622 A. J. Hulbert

fetal transfer of 3,5,3h-triiodothyronine versus thyroxine in N, V., A,E.S.&C, E. (1981). Acidic rats, as assessed from 3,5,3h-triiodothyronine levels in fetal phospholipids, unsaturated fatty acids, and limited proteolysis tissues. Acta Endocrinologica 120, 20–30. mimic the effect of calmodulin on the purified erythrocyte # M  E, G., P, R., O,M.J.& Ca +-ATPase. Journal of Biological Chemistry 256, 8588–8592. E  R, F. (1985). Effects of maternal hypo- N, A., K,M.&Y, K. (1989). Cell death thyroidism on the weight and thyroid hormone content of rat in the anuran tadpole tail: thyroid hormone induces keratiniz- embryonic tissues, before and after the onset of fetal thyroid ation and tail-specific growth inhibition of epidermal cells. function. Endocrinology 117, 1890–1900. Developmental Biology 131, 337–344. M, S. M., J. (1987). Thyroxine elicits divergent changes N, J., C, D., A,F.&B, A. M. (1991). in mRNA levels for two urea cycle enzymes and one Regulation by thyroid hormone of microtubule assembly and gluconeogenic enzyme in tadpole liver. Archives of Biochemistry neuronal differentiation. Neurochemical Research 16, 975–982. & Biophysics 259, 144–148. O, M. J., M, J., E  R,F.& M, M. J., R, C. A., B, A. G., T, G., M  E, G. (1981). Presence of -thyroxine C,C.&F, E. (1995). Kinetic analysis of and 3,5,3h-triiodo--thyronine in tissues from thyro- thyroid hormone action on glucose metabolism in man. idectomized rats. Endocrinology 109, 908–913. European Journal of Endocrinology 132, 413–418. O, M. J., M, J., P, R., E  R,F. M,M.J.&S, H. J. (1980). Rapid and direct &M  E, G. (1984). -thyroxine and 3,5,3h- stimulation of hepatic gluconeogenesis by -triiodothyronine triiodo--thyronine in rat embryos before onset of fetal thyroid (T3) in the isolated-perfused rat liver. Life Sciences 27, 827–835. function. Endocrinology 114, 305–307. M,M.J.&S, H. J. (1984). Pleiotypic action of O, M. J., M  E,G.&E  thyroid hormones at the target cell level. Biochemical Phar- R, F. (1978). Concentrations of triiodo--thyronine in the macology 33, 1579–1584. plasma and tissues of normal rats, as determined by M, M. B., Z, N. D.. MC, N. L., MD, radioimmunoasssay: comparison with results obtained by an M. J. & T, H. C. (1988). Isolation and characterization isotopic equilibrium technique. Endocrinology 103, 2145–2153. of rat cDNA clones for two distinct thyroid hormone receptors. O, M. J., R  O, C., E  R,F.& Journal of Biological Chemistry 263, 12770–12777. M  E, G. (1989). Regulation of intracellular M, M. G. (1986). Studies of the regulation of basal thyroid hormone concentrations in the fetus. In Research in adenylate cyclase activity by membrane polyunsaturated Congenital Hypothyroidism (eds. F. Delange, D. A. Fisher, and fatty acids in cultured neuroblastoma. Journal of Neurochemistry D. Glinoer), pp. 79–94. Plenum Publ. Corp. New York. 47, 245–253. O, K., K,J.D.&K, I. (1996). Acute M, M. G. (1990). Dietary fatty acids and membrane effects of thyroid hormone on vascular smooth muscle. Thyroid protein function. Journal of Nutritional Biochemistry 1, 68–79. 6, 505–512. N, T., S,M.&K, T. (1984). Differences in O, D. K., W,G.M.&G, E. G. (1993). The lipid composition and fluidity of cardiac sarcolemma prepared effects of thyroxine and propylthiouracil treatment on changes from newborn and adult rabbits. Biochemical Medicine 32, in body form associated with a possible developmental 122–131. thyroxine surge during post-hatching development of the N, A., S, S., S, A., S, I., B,G.I.& tilapia, Oreochromis mossambicus. Zoological Science 10, 803–811. DG, L. J. (1988). Characterization of a novel thyroid O, R., J, E. A., S,G.&S, D. P. (1992). hormone receptor expressed in human kidney and other Precocious synthesis of a thyroid hormone receptor in Xenopus tissues. Proceedings of the National Academy of Sciences, USA 85, embryos causes hormone-dependent developmental abnor- 2781–2785. malities. Roux’s Archives in Developmental Biology 201, 312–321. N-S, K., M, H. L., H, A. R., V, O’L, P. C., B, P., A, G., M,K.J.& T. H. & R, R. G. (1993). Mechanisms of thyroid J, I. (1992). Longitudinal study of serum thyroid hormone action on the insulin-like growth factor system: all hormone levels during normal pregnancy. International Journal thyroid hormone effects are not growth hormone mediated. of Gynaecology & Obstetrics 38, 171–179 Endocrinology 132, 781–788. O, M. L., M,D.G.&M, A. (1986). Alteration of N, G., F,F.&B, J. (1987). Annual cycles lymphocyte reactivities by thyroid hormones. International in plasma testosterone and thyroxine in the male aspic viper Journal of Immunopharmacology 8, 755–762. Vipera aspis L., (Reptilia, Viperidae), in relation to the sexual O, J. H. (1983). The nuclear receptor-triiodo- cycle and hibernation. Comparative & General Endocrinology 65, thyronine complex: relationship to thyroid hormone dis- 254–263. tribution, metabolism, and biological action. In Molecular N, T. T., DS,J.J.,Y,H.&Y, Y.-M. Basis of Thyroid Hormone Action (eds. J. H. Oppenheimer & (1993). Steady state organ distribution and metabolism of H. H. Samuels) Academic Press, New York. thyroxine and 3,5,3h-triiodothyronine in intestines, liver, O, J. H., D. K,H.L.S &M.I. kidneys, blood, and residual carcass of the rat in vivo. S (1972). Specific nuclear triiodothyronine binding sites Endocrinology 133, 2973–2983. in rat liver and kidney. Journal of Clinical Endocrinology & N, D. G., & L, R. M. (1984). Thermogenic Metabolism 35, 330–333. mechanism in brown fat. Physiological Reviews 64, 1–16. O,J.H.&S, H. H. (eds) (1983). Molecular N, S. C., A,N.A.&T, T. E. (2000). Seasonal Basis of Thyroid Hormone Action. Academic Press, New York. variations in thyroid hormone levels in free-living echidnas O,J.H.&S, H. L. (1997). Molecular (Tachyglossus aculeatus). General & Comparative Endocrinology basis of thyroid hormone-dependent brain development. 117, 1–7. Endocrine Reviews 18, 462–475. Thyroid hormones and their effects 623

O, J. H., S,H.L.&S, M. I. (1974). thyroidism on the transport of pyruvate in rat-heart mito- Tissue differences in the concentration of triiodothyronine chondria. Biochimica et Biophysica Acta 935, 79–86. nuclear binding sites in the rat: liver, kidney, pituitary, heart, P,G.&R, F. M. (1989). Decreased activity of brain, spleen and testis. Endocrinology 95, 897–903. the pyruvate translocator and changes in the lipid composition O, J. H., H. L. S &M.I.S (1975). in heart mitochondria from hypothyroid rats. Archives of Nuclear binding capacity appears to limit the hepatic response Biochemistry and Biophysics 269, 595–602. to -triiodothyronine (T3). Endocrine Research Communications 2, P,G.&R, F. M. (1990a). Enhanced activity of 309–325. the tricarboxylic acid carrier and modification of lipids in O, J. H., S, H. L., M, C. N., K- hepatic mitochondria from hyperthyroid rats. Archives of , W. B., W,N.C.W.&F, H. C. (1987). Biochemistry and Biophysics 278, 425–430. Advances in our understanding of thyroid hormone action at P,G.&R, F. M. (1990b). Stimulation of the cellular level. Endocrine Reviews 8, 288–308 phosphate transport in rat-liver mitochondria by thyroid O, J. H., S, H. L., L,J.T.&T- hormones. Biochimica et Biophysica Acta 1019, 133–136. , M. P. (1991). Functional relationship of thyroid hor- P, G., R,F.M.&D, P. (1991). The mone-induced lipogenesis, lipolysis, and thermogenesis in the influence of hypothyroidism on the transport of phosphate rat. Journal of Clinical Investigation 87, 125–132. and on the lipid composition in rat-liver mitochondria. O, J. H., S,H.L.&S, K. A. (1995). Biochimica et Biophysica Acta 1070, 180–186. An integrated view of thyroid hormone actions in vivo. In P, G., R, F. M., D, P., P,G.& Molecular Endocrinology: Basic Concepts and Clinical Correlations Q, E. (1993). Decreased cytochrome oxidase (ed. B. D. Weintraub), pp. 249–268. Raven Press, New York. activity and changes in phospholipids in heart mitochondria O, J. H., S,H.L.&S, K. A. (1996). from hypothyroid rats. Archives of Biochemistry and Biophysics The molecular basis of thyroid hormone action. In Werner and 307, 91–95. Ingbars. The Thyroid. 7th edition (eds. L. E. Braverman & R. D. P, G., R, F. M., P,G.&Q- Utiger), pp. 162–184. Lippincott-Raven Publishers, Phil- , E. (1994). Enhanced cytochrome oxidase activity and adelphia. modification of lipids in heart mitochondria from hyper- O, J. H., S, H. C., S,H.L.& thyroid rats. Biochimica et Biophysica Acta 1225, 165–170. S, M. I. (1971). Dissociation between thyroxine metab- P,M.&P, N. (1996). Serum free thyroxine olism and hormonal action in phenobarbital-treated rats. concentrations measured by chemiluminescence in hyper- Endocrinology 88, 115–119. thyroid and euthyroid cats. Journal of the American Animal O’R,I.&M, M. (1992). Studies on the rapid Hospital Association 32, 489–494. stimulation of mitochondrial respiration by thyroid hormones. P, M., P, N., L,N.&F, M. (1996). Acta Endocrinologica 127, 542–546. Serum-free thyroxine concentrations, measured by chemi- O,J.&L, J. B. (1988). Tissue-specific and luminescence assay before and after thyrotropin admin- developmental regulation of rat Na,K-ATPase catalytic α istration in healthy dogs, hypothyroid dogs, and euthyroid isoform and β subunit mRNAs. Journal of Biological Chemistry dogs with dermathopathies. Canadian Veterinary Journal 37, 263, 10436–10442. 289–294. P,L.&L, S. (1989). Vitamin E: Introduction to P, F., K, L., S,J.S.&T, M. (1977). biochemistry and health benefits. In Vitamin E Biochemistry & Uptake of triiodothyronine and thyroxine by isolated rabbit Health Implications (eds. A. T. Diplock, L. J., Machlin, L. adipocytes. FEBS Letters 83, 145–147. Packer & W. A. Pryor), pp. 1–6. Annals New York Academy P, D. V., K, M. A., P, L., B,C.S. of Sciences, vol. 570, New York. &K, S. S. (1995). Thyroid hormones alter Arrhenius P, U., B,S.C.&S, P. K. (1997). Regulation of kinetics of succinate-2,6-dichloroindophenol reductase, and actin and its mRNA by thyroid hormones in cultures of fetal the lipid composition and membrane fluidity of rat liver human brain during second trimester of gestation. Journal of mitochondria. European Journal of Biochemistry 230, 576–581. Neurochemistry 69, 1170–1176. P,J.F.&P, W. S. (1970). Cold acclimation and P, J. A., E, V., M, S., S, K., thyroxine effects on liver and liver mitochondrial fatty acids. G,M.E.&S, M. J. M. (1994). Thyroid American Journal of Physiology 218, 1417–1422. hormone metabolism in a transthyretin-null mouse strain. P, S., D, S., P,R.&S, P. K. (1996). Role of Journal of Biological Chemistry 30, 33135–33139. thyroid hormone in the maturation of astrocytes: temporal; P, J. A., H, M. T., M  E, G., correlation with synthesis and organization of actin. European E, V., G,M.E.&S,M.J.M. Journal of Neuroscience 8, 2361–2370. (1997). Transthyretin is not essential for thyroxine to reach P, M., B, M., D,P.&K,M. the brain and other tissues in transthyretin-null mice. American (1997). Diel pattern of haematocrit, serum metabolites, Journal of Physiology 272, E485–E493. osmotic pressure, electrolytes and thyroid hormones in sea P, D. A., H,A.J.&S, L. H. (1994). Dietary bass and sea bream. Aquaculture International 5, 237–247. lipids, membrane phospholipids and obesity. Journal of P, D. J. (1995). Hypothyroid state and membrane fatty Nutrition 124, 1555–1565. acid composition influence cardiac mitochondrial pyruvate P,K.&C-S, A. (1995). T3 fails to mimic oxidation. Biochimica et Biophysica Acta 1235, 231–238. certain effects of T4 in munia birds: physiological implications P,D.J.&A, E. M. K. (1995). Influence of for seasonal timing. Comparative Biochemistry & Physiology 111C, hypothyroid state on cardiac sarcolemmal incorporation of 157–164. dietary ω–6 and ω–3 fatty acids. Nutrition Research 15, P,G.&R, F. M. (1988). Effect of hyper- 1211–1222. 624 A. J. Hulbert

P, C. G., C, A. H., C, M. E., D development current perspectives. Endocrine Reviews 14, F,M.&M-R, A. M. (1998). Insulin-like 94–106. growth factor 1 reduces thyroid hormone receptors in the rat P, B. J., MI, G. H., M, M. T., B, liver. Evidence for a feed-back loop regulating the peripheral P. A., C, J., H, C. H., C,B.G.&H, thyroid hormone action. Journal of Endocrinology 158, 87–95. B. S. (1986). The effect of maternal thyroidectomy prior to P, B., C R, L. F. B. P., S, D. A., B, conception on foetal brain development in sheep. Acta E. J. H. C, R. (1994). Control of superoxide dismutase, Endocrinologica 112, 93–99. catalase and glutathione peroxidase activities in rat lymphoid P, M., C,R.&M  E, G. (1992). organs by thyroid hormones. Journal of Endocrinology 140, -thyroxine and 3,5,3h-triiodothyronine concentrations in the 73–77. chicken egg and in the embryo before and after the onset of P-C, A., B, J., F,B.&P,T. thyroid function. Endocrinology 130, 2651–2659. (1985). The early ontogenesis of thyroid hormone receptor in P,E.L.&K, M. (1979). Membrane-bound phos- the rat fetus. Endocrinology 117, 2457–2461. pholipid desaturases. Lipids 14, 159–165. P, M. E., M,C.&N, R. (1997). Measure- P, G. L., B, C. P., P,D.M.&S,K.R. ment of serum total thyroxine, triiodothyronine, free thy- (1994). The appearance, distribution, and longevity of roxine and thyrotropin concentrations for diagnosis of receptor-[125I]T3 complexes within the nuclei of isolated rat hypothyroidism in dogs. Journal of American Veterinary Medical hepatocytes. Thyroid 4, 305–312. Association 211, 1396–14021. R, R., R, C., S,A.&S, J. E. (1996). P, K. J., D, B., M,T.&N, Interactions among receptors, thyroid hormone response V. M. (1990). Identification of a thyroid hormone response elements, and ligands in the regulation of the rat uncoupling element in the malic enzyme gene. Journal of Biological protein gene expression by thyroid hormone. Endocrinology Chemistry 265, 7395–7400. 137, 3478–3487. P, P. O. D., B,I.H.&H, B. S. (1971). R, R., S, A., R, A., S,X.&S, Neurological damage to the fetus resulting from severe iodine J. E. (1995). Delineation of thyroid hormone-responsive deficiency during pregnancy. The Lancet 1, 308–310. sequences within a critical enhancer in the rat uncoupling P,K.D.&E, I. S. (1977). Thyroid hormone protein gene. Endocrinology 136, 1003–1013. control of Na+jK+ -ATPase and K+-dependent phosphatase. R, D., M, C. A., M,U.&W,P. American Journal of Physiology 232, C196–C201. (1991). Hypothyroidism and thyroxin substitution affect the P, M., W, A., P,S.&F,K.S. n-3 fatty acid composition of rat liver mitochondria. Lipids 26, (1991). Thyroid hormones control lipid composition and 781–787. membrane fluidity of skeletal muscle sarcolemma. Biochimica et R, G. S., E, J., R,M.L.&B, H. (1976). Uptake Biophysica Acta 1068, 167–173. of thyroid hormones by isolated rat liver cells. Biochemical & P, G., M, H., H, L., T, R., H, T., Biophysical Research Communications 73, 98–104. G, K-J., S-D, G., E, M., P- R, R. E., C, P. W., R,P.J.&T, , H., B, O., F,R.&B, A. (1997). B. C. (1998). Seasonal differences in resting oxygen con- Elevated 3,5-diiodothyronine concentrations in the sera of sumption, respiratory quotient, and free thyroxine in wood- patients with nonthyroidal illnesses and brain tumors. Journal chucks. American Journal of Physiology 274, R963–R969. of Clinical Endocrinology & Metabolism 82, 1535–1543. R, M. M. (1995). Non-receptor-binding proteins for P-R,R.&T, J. R. (1959). The Thyroid Hormones. insulin-like growth factors and other cytokines: modulators of Pergamon Press, London. peptide action. In Molecular Endocrinology: Basic Concepts and P, W. S., P,B.C.&C, R. R. J. (1972). Clinical Correlations. (ed. B. D. Weintraub), pp. 155–180. A comparison of magnesium deficiency, cold acclimation and Raven Press, New York. thyroxine administration on mitochondrial fatty acid com- R,P.K.&L, J. F. (1995). Influence of the position. Proceedings of Society for Experimental Biology and combination of time of feeding and ration level on the diurnal Medicine 150, 857–861. hormone rhythms in rainbow trout. Fish Physiology & P, R., P, S., C,S.&S, P. K. (1996). Biochemistry 14, 25–36. Regulation of actin and tubulin gene expression by thyroid R, H. L., B, K. D., S, K. M. M. & O’B, hormone during rat brain development. Molecular Brain J. T. (1986). Alterations in the hypothalamic-pituitary- Research 35, 111–118. thyroid axis after prolonged residence in Antarctica. Clinical P,J.K.&H, W. (1981). Changes in the Endocrinology 25, 55–65. phospholipid catabolism of mitochondria and microsomes R, S. (1994). Resistance to thyroid hormone: an during development of rat liver. Biochemical Journal 200, historical overview. Thyroid 4, 345–349. 521–528. R, S. (1996). Resistance to thyroid hormone. In Werner P, R. K., H,A.J.&B, M. D. (1996). and Ingbars. The Thyroid. 7th edition (eds. L. E. Braverman & Allometry of mitochondrial proton leak: influence of mem- R. D. Utiger), pp. 1032–1048. Lippincott-Raven Publishers, brane surface area and fatty acid composition. American Philadelphia. Journal of Physiology 271, R1550–R1560. R, S., DW,L.T.&DG, L. J. (1967). P,S.P.&H, C. E. (1992). Tissue iodo- Familial syndrome combining deaf-mutism, stippled epi- thyronine levels in fetuses of control and hypothyroid rats at physes, goitre and abnormally high PBI: possible target organ 13 and 16 days gestation. Endocrinology 131, 195–200. refractoriness to thyroid hormone. Journal of Clinical Endo- P,S.P.&H, C. E. (1993). The role of crinology & Metabolism 27, 279–294 thyroid hormones in prenatal and neonatal neurological R,L.F.&T, K. J. (1991). Regulation of Thyroid hormones and their effects 625

neural development by the extracellular matrix In Receptors R, D. F. S., H,A.J.&B, M. D. (1994). for Extracellular Matrix (eds. J. A. McDonald & R. P. Characteristics of mitochondrial proton leak and control of Mecham), pp. 157–193. Academic Press, San Diego. oxidative phosphorylation in the major oxygen-consuming R,B.D.&W, F. E. (1997). Effects of thyroxine tissues of the rat. Biochimica et Biophysica Acta 1118, 405–416. (T4) or triiodothyronine (T3) replacement therapy on the R,D.F.S.&B, M. D. (1997). The physiological programming of seasonal reproduction and postnuptial molt significance of mitochondrial proton leak in animal cells and in thyroidectomized male American tree sparrows (Spizella tissues. Bioscience Reports 17, 9–16. arborea) exposed to long days. The Journal of Experimental R,D.F.S.&B, G. C. (1997). Cellular energy Zoology 279, 367–376. utilization and molecular origin of standard metabolic rate in R, S. G., H, Z., S, D. E., M,S.& mammals. Physiological Reviews 77, 731–758. C, F. (1990). Biphasic response of rat tibial growth to R, J. M. M., DG, W. J.,   S, P., thyroxine administration. Acta Endocrinologica 122, 336–340. K, B., K,W.&V, T. J. (1988). Effect of R, A., M,M.&T, J. (1998). Nongenomic thyroid status and paraventricular area lesions on the release actions of steroid hormones in reproductive tissues. Endocrine of thyrotropin-releasing hormone and catecholamines into Reviews 19, 3–17. hypophysial portal blood. Endocrinology 123, 523–527. R, E. B., O  N, C. M., A, I. S., R,B.&H, E. A. (1983). Adaptive changes in lipid H,A.E.&D, M. S. (1997). Hormonal and composition of rat liver plasma membrane during postnatal metabolic adaptations to fasting in monosodium glutamate- development following maternal ethanol ingestion. Biochimica obese rats. Journal of Comparative Physiology B 167, 430–437. et Biophysica Acta 735, 407–417. R, R. C. J., A, J. W., W, B. L., W, R,P.&P, G. (1986). Acute changes of the conversion R. L., F, R. J., S,F.&B,J.D. of thyroxine to triiodothyronine in hypophysectomized and (1995). The molecular biology of thyroid hormone action. thyroidectomized chickens exposed to mild cold (10m). General Annals of New York Academy of Science 758, 366–389. & Comparative Endocrinology 63, 408–413. R, M. J., J., S, G. W., M,Q.N.&K, R, A., M, K. M., D, P. J., D,F.B.& # L. S. (1986). Lipid composition of alveolar macrophage B, S. D. (1984). Rabbit myocardial membrane Ca +- plasma membrane during postnatal development. Journal of ATPase activity stimulation in vitro by thyroid hormone. Immunology 136, 1054–1060. Archives of Biochemistry & Biophysics 229, 379–385. R, S. J., B, A. J., D, W., W, R, F. M., L, C., G,G.V.&Q- R. E. H., H, P. J., B, J. J., S, B. R., , E. (1984). Lipid composition of liver mitochondria N, S., D,S.C.&S, G. (1994). and microsomes in hyperthyroid rats. Lipids 19, 171–178. Evolution of marsupial and other vertebrate thyroxine- R, F. M., G,G.V.&Q, E. (1989). binding plasma proteins. American Journal of Physiology 266, Lipid composition of brown adipose tissue mitochondria and R1359–R1370. microsomes in hyperthyroid rats. International Journal of R, D., M,G.&H, P. (1975). Alterations of Biochemistry 21, 327–332. mitochondrial phospholipids in the rat brown adipose tissue R  O, C., O, M. J., E  R,F.& after chronic treatment with cold or thyroxine. FEBS Letters M  E, G. (1988). Developmental changes 53, 342–346. in rat brain 5h-deiodinase and thyroid hormones during the R, W.W.J.&E, J. G. (1993). Characterization of fetal period: the effects of fetal hypothyroidism and maternal -thyroxine transport into hepatocytes isolated from juvenile thyroid hormones. Pediatric Research 24, 588–594. rainbow trout (Oncorhynchus mykiss). General & Comparative R-M, A., S, J., S-T, F., E Endocrinology 90, 31–42.  R,F.&M  E, G. (1983). Effect of R, E. A., F-T, M. D. C., B, J. C., neonatal and adult-onset hypothyroidism on pyramidal cells S-S,N.B.&D P, S. F. (1996). Onto- of the rat auditory cortex. Developmental Brain Research 9, genic development of membrane lipids in the chick optic lobe. 205–213. International Journal of Developmental Neuroscience 14, 93–104. S, H. H., S,I.&S, L. E. (1977). R, J. (1991). Thyroid hormone transport proteins and the Modulation of thyroid hormone nuclear receptor levels by physiology of hormone binding. In Werner & Ingbars. The 3,5,3h-triiodothyronine in GH1 cells. Journal of Biological Thyroid 6th Edition (eds. L. E. Braverman & R. D. Utiger), Chemistry 252, 6052–6060. pp. 111–125. J. B. Lippincott, Philadelphia, U.S.A. S, H. H., F, B. M., H,Z.D.&Y, Z.-S. R,J.&R, J. E. (1957). The interactions of thyroid (1988). Regulation of gene expression by thyroid hormone. hormones and protein in biological fluids. Recent Progress in Journal of Clinical Investigation 81, 957–967. Hormone Research 13, 161–208. S,H.H.&T, J. S. (1973). Thyroid hormone action in R, C., S, H. L., S,K.A.&O, cell culture: demonstration of nuclear receptors in intact cells J. H. (1992). Ontogeny of hepatic nuclear triiodothyronine and isolated nuclei. Proceedings of National Academy of Sciences of receptor isoforms in the rat. Endocrinology 131, 2559–2564. the United States of America 70, 3488–3492. R, L. M., H, I. B., K, J. R., H,L. S, M. H., W, M. E., W,C.&R,E.C. &T, J. T. (1988). Thyroid hormone action on glucose (1989). The effect of altered thyroid status on pituitary transporter activity in astrocytes. Biochemical & Biophysical hormone messenger ribonucleic acid concentrations in the rat. Research Communications 156, 275–281. Endocrinology 124, 2277–2282. R,D.&D, W. H. (1988). Thyroid hormone S, J. P.,   G, S., K, E., D, markedly increases the mRNA coding for sarcoplasmic V. M., K, E. R., L,J.L.&V, T. J. (1997). # reticulum Ca +-ATPase in the rat heart. Journal of Biological Characterization of a propylthiouracil-insensitive type I Chemistry 263, 6941–6944. iodothyronine deiodinase. Endocrinology 138, 5153–5160. 626 A. J. Hulbert

"#& S, C., S, H. L., M, C. N., F, distribution of I-thyroxine in rat liver: importance of D. & O, J. H. (1998). Beta receptor isoforms are microsomal binding. Endocrinology 84, 270–276. not essential for thyroid hormone-dependent acceleration of S, H. L., L,M.A.&O, J. H. (1994). PCP-2 and myelin basic protein gene expression in the Widespread distribution of immunoreactive thyroid hormone developing brains of neonatal mice. Molecular & Cellular β2 receptor (TRβ2) in the nuclei of extrapituitary rat tissues. Endocrinology 137, 109–115. Journal of Biological Chemistry 269, 24777–24782. S, F., C, I. J., S,D.H.&C T, S, H. L., R,M.E.&O, J. H. (1997). G. N. (1992). Evidence that the human placental 5-mono- Lack of effect of thyroid horrmone on late fetal rat brain deiodinase is a phospholipid-requiring enzyme. Journal of development. Endocrinology 138, 3119–3124. Clinical Endocrinology & Metabolism 74, 1366–1371. S, S. M., L,A.S.&M, M. S. (1992). S,C.R.A.&P, D. M. (1999). Identification of Ontogeny of proximal colon basolateral membrane lipid transthyretin in fish (Sparus aurata): cDNA cloning and charac- composition and fluidity in the rabbit. Biochimica et Biophysica terisation. Endocrinology 140, 2430–2433. Acta 1107, 70–76. S, J., M, A., D, K., G, Y., G, J., S,A.&B, S. (1988). Presence of iodothyronines in L, A., B,H.&V, B. (1986). The c-erb-A the yolk of the hen’s egg. General & Comparative Endocrinology protein is a high-affinity receptor for thyroid hormone. Nature 69, 99–105. 324, 635–640. S, J. (1989a). A rapid, extranuclear effect of 3,5,3h- S, M. R., R,D.K.&D, W. H. (1992). triiodothyronine on sugar uptake by several tissues in the rat Thyroid hormone response of slow and fast sarcoplasmic in vivo. Evidence for a physiological role for the thyroid # reticulum Ca + ATPase mRNA in striated muscle. Molecular hormone action at the level of the plasma membrane. & Cellular Endocrinology 87, 87–93. Endocrinology 124, 2755–2764. S, O. A. (1988). Changes in fatty acid compositions of S, J. (1989b). Action of the thyroid hormone at the level of mitochondria during embryonic development. Comparative the plasma membrane. Endocrine Research 15, 619–649. Biochemistry & Physiology 89B, 747–753. S, J. (1990a). Calcium is the first messenger for the action S,P. C., D,E.&S, H. H. (1995). Generation and of thyroid hormone at the level of the plasma membrane: first remodeling of phospholipid molecular species in rat hepato- evidence for an acute effect of thyroid hormone on calcium cytes. Archives of Biochemistry & Biophysics 319, 168–176. uptake in the heart. Endocrinology 126, 2693–2702. S,C.A.&MD, A. A. (1988). Thyroid hormone S, J. (1990b). Thyroid hormone action at the level of the regulates α and αj isoforms of Na,K-ATPase during plasma membrane. Thyroid 1, 83–87. development in neonatal rat brain. Journal of Biological S, J., S,H.&G, A. (1977). The effect of $ Chemistry 263, 17643–17649. triiodothyronine on 2-deoxy-D-[1- H]glucose uptake in cul- S, M. J., D,J.C.&G, V. A. (1993). Rana tured chick embryo heart cells. Endocrinology 101, 143–149. catesbeiana tadpole red blood cells express an α, but not a β, S,J.&I, S. H. (1979). Stimulation by tri- c-erbA gene. Endocrinology 133, 2488–2495. iodothyronine of the in vitro uptake of sugars by rat S,M.J.&G, V. A. (1991). Regulation of thymocytes. Journal of Clinical Investigation 63, 507–515. c-erbA-α messenger RNA species in tadpole erythrocytes by S,J.&I, S. H. (1984). An immediate increase in thyroid hormone. Molecular Endocrinology 5, 201–208. calcium accumulation by rat thymocytes induced by triiodo- S,M.J.&G, V. A. (1995). Effect of gluco- thyronine: its role in the subsequent metabolic responses. corticoids on thyroid hormone action in cultured red blood Endocrinology 115, 160–166. cells from Rana catesbeiana tadpoles. Endocrinology 136, 1435– S, J., M, S., S, H., M, G., B, 1440. J. B. & U, G. (1996). Acute effect of thyroid hormone S,G.&R, S. J. (1997). The evolution of in the rat heart: role of calcium. Journal of Endocrinology 149, gene expression, structure and function of transthyretin. 73–80. Comparative Biochemistry & Physiology 116B, 137–160. S, H., A, M., Y, K., M,N.&T, S, F. (1982). Hormonal effects on fatty acid binding O. (1989). Plasma thyroxine-binding proteins and thyroid and physical properties of rat liver plasma membranes. hormone levels in primate species; Is callithricidae thyroid Journal of Membrane Biology 68, 1–10. hormone resistant? Endocrinologica Japonica 36, 665–673. S, F., W,W.G.&K, A. B. (1998). The S, A., S, O., H, S., H, H., Y,G.& biological membrane and lipid domains. In Cell Physiology C, G. (1996). Antioxidant status in experimental Source Book (2nd edition). (ed. N. Sperelakis) pp. 61–74, hyperthyroidism: effect of vitamin E supplementation. Clinica Academic Press, San Diego, USA. Chimica Acta 256, 65–74 S, U., H, G., S, K., S- S,M.J.&H, F. L. (1976). Thyroid control over ,W.&P, R. M. (1995). Impact of captivity biomembranes. II. rat liver nuclei. Life Science 19, 1359–1364. and contamination level on blood parameters of harbour seals S,M.J.&H, F. L. (1977). Thyroid control over (Phoca vitulina). Comparative Biochemistry & Physiology 112A, biomembranes. IV. rat heart muscle mitochondria. Journal of 455–462. Molecular & Cellular Cardiology 9, 749–761. S, U., R, G., P,J.&W, U. (1992). S,S.J.&C, W. A. (1984). Electrical activity Basic biochemical data on blood from antarctic weddell seals and cytosolic calcium regulate levels of tetrodotoxin-sensitive (Leptonychotes weddelli): ions, lipids, enzymes, serum proteins sodium channels in culture rat muscle cells. Proceedings of the and thyroid hormones. Comparative Biochemistry & Physiology National Academy of Sciences, USA 81, 262–266. 102A, 449–451. S, Y.-B. (1996). Thyroid hormone-regulated early and late S, H. L., B,G.&O, J. H. (1969). genes during amphibian metamorphosis. In Metamorphosis: Effect of phenobarbital administration on the subcellular Post-embryonic reprogramming of gene expression in amphibian and Thyroid hormones and their effects 627

insect cells (eds. L. I. Gilbert, J. R. Tata & B. G. Atkinson), Thyroxine transport to the brain: role of protein synthesis by pp. 505–538. Academic Press, New York. the choroid plexus. Endocrinology 133, 2116–2126. S, M. M., P,D.G.&L, J. B. (1989). Character- S, B. R., D,W.T, G.-F. & S,G. ization of the human Na,K-ATPase α2 gene and identification (1991). Ontogenesis of transthyretin gene expression in of intragenic restriction fragment length polymorphisms. chicken choroid plexus and liver. Comparative Biochemistry & Journal of Biological Chemistry 264, 17532–17543. Physiology 100B, 329–338. S, M. A., R,E.C.&C, W. W. (1989). S, S. A., P,E.&G, A. (1985). Changes Molecular biology of thyrotropin. Endocrine Reviews 10, in plasma steroid and thyroid hormones and insulin during 459–475. final maturation and spawning of the sea lamprey, Petromyzon S, J. E. (1995). Thyroid control of thermogenesis and marinus. General & Comparative Endocrinology 58, 259–269. energy balance. Thyroid 5, 481–492. S, J. L., B,C.L.&B, H. A. (1992). Asyn- S,J.E,&L, J. L. (1985). Regulation of rat chrony of changes in tissue and plasma thyroid hormones cerebrocortical and adenohypophyseal type II 5h-deiodinase during the parr-smolt transformation of Coho salmon. General by thyroxine, triiodothyronine and reverse triiodothyronine. & Comparative Endocrinology 88, 397–405. Endocrinology 116, 1627–1635. S, J. L., DS III, J. J., G, E. G., N, S, W. S., B, G. A., T, M. H. M.,   R. S. & B, H. A. (1984). Development-associated changes L, C. G., L,P.R.& H, C. (1996). in thyroxine kinetics in juvenile salmon. Endocrinology 115, Characterization of the promoter of the rat sarcoplasmic #+ 399–406. endoplasmic reticulum Ca -ATPase 1 gene and analysis of S,A.A.&Y, M. A. (1985). Membrane lipid thyroid hormone responsiveness. Journal of Biological Chemistry composition and cellular function. Journal of Lipid Research 26, 271, 32048–32056. 1015–1035. S, W. S.,   L,G.C.& H,C. S, B. J., ML, K. M., R,J.&B,J.D. (1990). Thyroid hormone differentially affects mRNA levels (1975). Thyroid hormone receptors: binding characteristics of Ca-ATPase isozymes of sarcoplasmic reticulum in fast and and lack of hormonal dependency for nuclear localization. slow skeletal muscle. FEBS Letters 274, 73–76. Journal of Biological Chemistry 250, 4113–4119. S,W.S.& H, C. (1987). Effects of S.G, D. L. (1985). Metabolic effect of 3,3h,5h-tri- hypothyroidism on the distribution and fatty acyl composition iodothyronine in cultured growth hormone-producing rat of phospholipids in sarcoplasmic reticulum of fast skeletal pituitary tumor cells. Journal of Clinical Investigation 76, muscle of the rat. Biochimica et Biophysica Acta 924, 204–209. 890–893. S,W.S.& H, C. (1989). The postnatal # S.G,D.L.&G, V. A. (1997). The deiodinase development of sarcoplasmic reticulum Ca + transport activity family of selenoproteins. Thyroid 7, 655–668. in skeletal muscle of the rat is critically dependent on thyroid S.G, D. L., S, R. A., C, W., hormone. Endocrinology 124, 1145–1153. K, A., W, Z., B,D.D.&G,V.A. S-M,M.W.&S, R. L. (1976). The (1994). A thyroid hormone-regulated gene in Xenopus laevis transcapillary exchange of thyroid hormones and thyroxine- encodes a type III iodothyronine 5-deiodinase. Proceedings of binding proteins between blood and tissue fluids. Journal of the National Academy of Sciences USA 91, 7767–7771. Physiology 257, 123–136. S,S.J.&N, G. L. (1972). The fluid mosaic model S,D.G.&P, W. S. (1976). Subcellular mem- of the structure of cell membranes. Science 175, 720–731. brane fatty acids of rat heart after cold acclimation or S, M. (1974). Homeoviscous adaptation – a homeostatic thyroxine. American Journal of Physiology 231, 650–654. process that regulates viscosity of membrane lipids in S, S. A., K, L. L., S, D. R., A, Escherichia coli. Proceedings of the National Academy of Sciences USA P. M. & B, S. T. (1991). Hypothyroidism reduces the 71, 522–525. rate of slow component A (Sca) axonal transport and the S, T. J., D,F.B.&D, P. J. (1989). Retinoic acid is amount of transported tubulin in the hyt\hyt mouse optic # a modulator of thyroid hormone activation of Ca +-ATPase in nerve. Journal of Neuroscience Research 28, 121–133. the human erythrocyte membrane. Journal of Biological S, M. B., F, A. L., G,R.W.&J- Chemistry 264, 687–689. A, H. B. (1993). Behavioral effects of thyroxine in a Smith, T. J., Davis, F. B., Deziel, M. R., Davis, P. J., Ramsden, lizard (Ameiva undulata: Teiidae). Physiological Zoology 66, D. B. & Schoenl, M. (1994). Retinoic acid inhibition of 148–165. thyroxine binding to human transthyretin. Biochimica et S, K., S, L., F, B., N,E.& Biophysica Acta 1199, 76–80. L, G. (1993). Thyroid hormones, procollagen III S, P. J. (1996). The pituitary in hypothyroidism. In peptide, body composition and basal metabolic rate in Werner and Ingbars. The Thyroid. 7th edition (eds. L. E. euthyroid individuals. Scandinavian Journal of Clinical Laboratory Braverman & R. D. Utiger), pp. 836–840. Lippincott-Raven Investigation 53, 793–803. Publishers, Philadelphia. S, K. (1979). Thyroid hormone action at the cell level. S, S. (1993). Thyroid hormone action on mitochondrial New England Journal of Medicine 300, 117–123 & 173–177. energy transfer. Biochimica et Biophysica Acta 1144, 1–16. S,K.&B, M. A. (1995). Thyroid hormone S,K.&R, J. (1977). Metabolism of thyroid action: effect of triiodothyronine on mitochondrial adenine hormones by cultured hepatocarcinoma cells. Nonphenolic nucleotide translocase in vivo and in vitro. Metabolism 44, ring deiodination and sulfation. Journal of Biological Chemistry 193–199. 252, 4458–4463. S, K., C, G. A., T,G.S.&N, S, B. R., D, W., A, D., B, C., E. A. (1984). Mitochondrial binding of triiodothyronine R, S. J., K,J.&S, G. (1993). (T3): demonstration by electron-microscopic radioautog- 628 A. J. Hulbert

raphy of dispersed liver cells. Cell and Tissue Research 236, T,M.&H, T. (1987). Presence of thyroxine in eggs 321–325. and changes in its content during early development of chum S, J. R. (1996). Serum thyrotropin and thyroid hormone salmon, Oncorhynchus keta. General & Comparative Endocrinology measurements and assessment of thyroid hormone transport. 68, 129–135. In Werner and Ingbars. The Thyroid. 7th edition (eds. L. E. T,M.&H, T. (1991). Effects of thyroid hormone Braverman & R. D. Utiger), pp. 377–396. Lippincott-Raven deficiency in eggs on early development of the Medaka, Publishers, Philadelphia. Oryzias latipes. The Journal of Experimental Zoology 257, 360–366. S, M. A., I-O, A., S,Y.&S, Y.-B. (1997). T, M., O,N.&S, M. (1994). Characterization Gene regulation by thyroid hormone during amphibian of thyroid hormone effect on the visual system of the adult rat. metamorphosis: implications on the role of cell-cell and cell- Thyroid 4, 467–474. extracellular matrix interactions. American Zoologist 37, 195– T, T., N, T., M, T., H,K.& 207. DG, L. J. (1997). The function of retinoid X receptors S, K. A., S, H. L., P-C,A.& on negative thyroid hormone response elements. Molecular and O, J. H. (1990). Relationship of c-erbA mRNA Cellular Endocrinology 128, 85–96. content to tissue triiodothyronine nuclear binding capacity T, H., M, T., T,K.&S,M. and function in developing and adult rats. Journal of Biological (1995). Serum triiodothyronine concentration and Na+,K+- Chemistry 265, 10514–10521. ATPase activity in liver and skeletal muscle are influenced by S, K. A., Z,L.&O, J. H. (1992). β1 dietary fat type in rats. Journal of Nutrition 125, 2364–2369. isoform-specific regulation of a triiodothyronine-induced gene T, M., K, R., T,M.&H, T. (1991). during cerebellar development. Molecular Endocrinology 6, A thyroxine surge during development of black sea bream 1874–1880. larvae and its ecological implication in inshore migration. S,C.D.&S, A. D. (1984). The modification of Nippon Suisan Gakkaishi 57, 1827–1832. mammalian membrane polyunsaturated fatty acid compo- T,J.P.&F, D. M. (1990). Pre-translational sition in relation to membrane fluidity and function. Biochimica + + regulation of the (Na jK )-ATPase in response to demand et Biophysica Acta 779, 89–137. for ion transport in cultured chicken skeletal muscle. Journal of S, A., Y, P. M., A, J. W., R, R. C. J., Biological Chemistry 265, 4116–4123. S, D. B., B,J.D.&C, W. W. (1994). Phos- T, Y., I,M.&I, S. (1986). Annual cycle of plasma phorylation selectively increases triiodothyronine receptor thyroid hormone levels in toad, Bufo japonicus. General & homodimer binding to DNA. Journal of Biological Chemistry Comparative Endocrinology 62, 404–410. 269, 433–437. T, J. R. (1963). Inhibition of the biological action of thyroid S, C. V., I,R.N.&D, W. W. (1987). hormones by actinomycin D and puromycin. Nature 197, Thyroid hormones in blood plasma of developing salmon 1168–1169. embryos. General & Comparative Endocrinology 65, 337–345. T, J. R. (1994). Hormonal regulation of programmed cell S, Y., U,P.&R, M. (1997). death during amphibian metamorphosis. Biochemistry & Cell Changes in blood-brain barrier nutrient transport in the Biology 72, 581–588. offspring of iodine-deficient rats and their preventability. T, J. R. (1996). Metamorphosis: an exquisite model for Neurochemical Research 22, 785–790. S,S.&S, M. (1981). Changes in thyroidal and hormonal regulation of post-embryonic development. Bio- plasma iodine compounds during and after metamorphosis of chemical Society Transactions 62, 123–136 the bullfrog, Rana catesbeiana. General & Comparative Endo- T, J. R., E,L.&L, O. (1962). Control of crinology 45, 74–81. basal metabolic rate by thyroid hormones and cellular S, S., M, T., O, A., S,A.& function. Nature 193, 1058–1060. DG, L. J. (1994). Two thyroid hormone response T, J. R., E, L., L, O., A, E., elements are present in the promoter of human thyroid P,S.&H, R. (1963). The action of thyroid hormone receptor β1. Molecular Endocrinology 8, 305–314. hormones at the cell level. Biochemical Journal 86, 408–428. S,P.&D, W. W. (1987). Variation in thyroid T,J.R.&W, R. R. (1966). Ribonucleic acid response to thyroid stimulating hormone in juvenile coho synthesis during the early action of thyroid hormones. salmon (Oncorhynchus kisutch). General & Comparative Endo- Biochemical Journal 98, 604–620. crinology 68, 473–485. T, C. C., W, C., L,R.&E,R.M. S, P. W., B, H. P., H,H.L.&B, (1987). Identification of a novel thyroid hormone receptor K. P. (1994). Fatty acid modulation of epidermal growth expressed in the mammalian central nervous system. Science factor-induced mouse mammary epithelial cell proliferation in 237, 1610–1614. vitro. Experimental Cell Research 214, 145–153. T,G.A.J.&N, Y. (1984). The regulation of S, G., P,S.&Z, J. (1991). Effect of membrane fluidity in Tetrahymena. In Biomembranes vol.12 hyper- and hypothyroidism on phospholipid fatty acid Membrane Fluidity (eds. M. Kates & L. A. Manson) Plenum composition and phospholipases activity in sarcolemma of Press, New York, pp. 397–432. rabbit cardiac muscle. Biochimica et Biophysica Acta 1083, T, L. A., D, G. E., E,N.P.&K,F.J. 265–270. (1996). Time-course of thyroid hormone involvement in the T, M. T., C, G., F, G., S,P.S.& development of anestrus in the ewe. Biology of Reproduction 55, S, M. (1991). Effect of hypothyroidism induced in 833–837. adult rats on brain membrane fluidity and lipid content and T, T. (1997). Thyroxine, triiodothyronine and reverse- composition. Research Communications in Chemical Pathology & triiodothyronine concentrations in blood plasma in relation to Pharmacology 71, 85–103. lactational stage, milk yield, energy and dietary protein Thyroid hormones and their effects 629

intake in Estonian dairy cows. Acta Veterinaria Scandinavia 38, V,T.&T, P. S. (1969). Effect of hypothyroidism 339–348. on ionic metabolism and Na-K activated ATP phos- T, T. (1984). Shrew metabolic rates and thyroxine phohydrolase activity in the developing rat brain. Journal of utilization. Comparative Biochemistry & Physiology 78A, 431–435. Neurochemistry 16, 935–943. T, T. E. (1991). Utilization rates of thyroid hormones in V,S.&G, A. (1988). Fatty acid mammals. Comparative Biochemistry & Physiology 100A, 503–516. composition in plasma and platelet phospholipid in hypo- T, T. E., H,E.C.&T, T. J. (1998). thyroid patients before and after l-thyroxine substitution. Thyroid hormone concentrations in black bears (Ursus Scandinavian Journal of Clinical & Laboratory Investigation 48, americanus): hibernation and pregnancy effects. Comparative & 21–25. General Endocrinology 109, 192–199. V,S.&M, M. (1994). Increased ratio T,T.E.&H, B. A. (1987). Thyroid function and between anaerobic and aerobic metabolism in lymphocytes cold acclimation in the hamster, Mesocricetus auratus. American from hyperthyroid patients. European Journal of Endocrinology Journal of Physiology 252, E260–E267. 130, 276–280. T, H., V, M., K,K.&S, G. (1995). V,S.&M, M. (1995). Beta-adrenoceptor Assessment of lipid peroxidation in rats of different body antagonist therapy in hyperthyroidism: effects on lymphocyte weight by determining expired ethane. Biological Chemistry metabolic activity and the relation between aerobic and Hoppe-Seyler 376, 691–694. anaerobic metabolism in lymphocytes. Thermochimica Acta T, S., S, T. I. A., R, B., C,N.J. 251, 191–197. &A, A. (1996). Twenty-four-hour energy expenditure: V  G, S., S, J. P., K, E., D, the role of body composition, thyroid status, sympathetic V. M., K, E. R., L,J.L.&V, T. J. (1997). activity and family membership. Journal of Clinical Endo- Expression of chicken hepatic type I and type III iodo- crinology & Metabolism 81, 2670–2674. thyronine deiodinases during embryonic development. Endo- T, H. C., M,C.N.&O, J. H. (1980). crinology 138, 5144–5152. Changes in the hepatic levels of messenger ribonucleic acid for   K, F. R. M., W,W.M.& V, malic enzyme during induction by thyroid hormone or diet. J. J. M. (1989) Studies on the mechanism of inhibition of Biochemistry 19, 579–585. nuclear triiodothyronine binding by fatty acids. FEBS Letters T, N., Z, A. M., M, A. L., H,J.W.& 246, 6–12. L, P. R. (1995). A novel retinoid X receptor-inde-  D, J. J., M, F. A. J., M,I.A.& pendent thyroid hormone response element is present in the D, H. (1986). Changes in fatty acid profiles of human type 1 deiodinase gene. Molecular & Cellular Biology 15, plasma, erythrocytes and polymorphonuclear leukocytes in 5100–5112. induced hypothyroidism in man: indirect evidence for altered T, Y.-C. L. & L, K. R. (1984). Iodothyronines: ∆6 desaturase activity. Clinica Chimica Acta 156, 299–314. oxidative deiodination by hemoglobin and inhibition of lipid  D, J., R,F.&  H, D. (1985). peroxidation. Lipids 19, 96–102. Concentrations of thyroxine and 3,5,3 -triiodothyronine at 34 T, H. M., K, S.-W., S, D., B, T., L, h different sites in euthyroid rats as determined by an isotopic G., L,P.R.&L, R. M. (1997). Regional equilibrium technique. Endocrinology 117, 1201–1208. distribution of type 2 thyroxine deiodinase messenger ribo- nucleic acid in rat hypothalamus and pituitary and its  D, J., R,F.&  H, D. (1986). regulation by thyroid hormone. Endocrinology 138, 3359–3368. Conversion of thyroxine to 3,5,3h-triiodothyronine in several T, K., K, F. A., B,D.N.&D,P. rat tissues in vivo: the effect of hypothyroidism. Acta (1987). Effect of short term triiodothyronine administration Endocrinologica (Copenhagen) 113, 59–64. on human leukocyte Rb(K) influx and Na efflux. Journal of V, G., J,G.F.M.&K, E. R. (1987). Clinical Endocrinology & Metabolism 65, 1031–1034. Seasonal changes of the 5h-monodeiodination activity in T-B,C.H.&J, P. A. (editors) (1988). The kidney and skin homogenates of male Rana ridibinda: relation Developing Marsupial. Models for Biomedical Research. Springer- to plasma thyroxine (T4) and testosterone. General & Verlag, Berlin. Comparative Endocrinology 68, 163–169. U,K.&E, R. M. (1989). Determinants of target V Z, J. M., B,K.&V  W, B. J. (1989). gene specificity for steroid\thyroid hormone receptors. Cell 57, Stimulation of the chlorinating activity of human myelo- 1139–1146. peroxidase by thyroid hormones and analogues. Hormone & U, K., H,A.&Y, K. (1994). Serum thyroid Metabolism Research 21, 441–444. hormone, guanine and protein profiles during smoltification V, P., B, S., D M,T.&D L,T. and after thyroxine treatment in the masu salmon, Onco- (1997a). Effect of thyroid state on lipid peroxidation, rhynchus masou. Comparative Biochemistry & Physiology 107A, antioxidant defences, and susceptibility to oxidative stress in 607–612. rat tissues. Journal of Endocrinology 155, 151–157. U, S. J. (1995). Thyroid hormone resistance. In Molecular V, P., D L,T.&D M, S. (1997b). Vitamin E Endocrinology. Basic Concepts and Clinical Correlations (ed. B. D. administration attenuates the tri-iodothyronine-induced mod- Weintraub), pp. 393–409. Raven Press, New York. ification of heart electrical activity in the rat. The Journal of U, A., P, R., L,C.&L, J. (1983). Experimental Biology 200, 909–914. Cochlear synaptogenesis in the hypothyroid rat. Developmental V, P., D M, S.,  M R,P.&D L, Brain Research 7, 295–301. T. (1996). Effect of T3 administration on electrophysiological U, A., R,A.&M, M. (1980). The effect of properties of lizard ventricular muscle fibres. Journal of hypothyroidism on the onset of cochlear potentials in Comparative Physiology B 165, 552–557. developing rats. Brain Research 182, 172–175. V, J. T., L,Y.A.&C,M.T. 630 A. J. Hulbert

(1986). Diet fat alters the structure and function of the nuclear rat: evolutionary considerations. Endocrinology 120, 664–677. envelope: modulation of membrane fatty acid composition, W, R. A. (1996). Cytoplasmic transport of lipids: role of NTPase activity and binding of triiodothyronine. Biochemical binding proteins. Comparative Biochemistry and Physiology 115B, & Biophysical Research Communications 135, 655–661. 319–331. V, A. J., K, P., G,A.K.&T,J.M. W, R. E., F, D., P, J., C, K., C,T.& (1985). Effects of thyroid hormone on mitochondrial oxidative R, S. (1997). Thyrotropin regulation by thyroid phosphorylation. Biochemical Journal 226, 183–192. hormone in thyroid hormone receptor β-deficient mice. V, L. A., C, L., R,M.&S, T. (1993). Endocrinology 138, 3624–3629. Zymosan-induced luminol-amplified chemiluminescence of W, T. A. (1992). Plasma thyroid hormone levels in whole blood phagocytes in experimental and human hyper- migratory and lake-resident coho salmon juveniles from the thyroidism. Free Radicals in Biology & Medicine 14, 669–675. Karluk River system, Alaska. Transactions of the American V, F., G, L., K,R.&B-W, Fisheries Society 121, 199–205. I. (1998). Local administration of thyroid hormones in W, W. M. (1991). Propanolol and thyroid hormone silicone chamber increases regeneration of rat transected metabolism. Thyroid 1, 273–277. sciatic nerve. Experimental Neurology 150, 69–81. W, W. M., C,I.J.&T, G. N. C. (1988). V, R., S,L,M, M., R-R,M.& Inhibition of nuclear T3 binding by fatty acids. Metabolism 37, N, E. A. (1990). Immunological quantitation of rat and 996–1002. mouse thyroxine-binding globulins. Ontogenesis and sex- W, L., J, C., S, C., B, C., B, dependence of the circulating levels of the thyroxine-binding A. C., B, F., F, D., T,P.&V,B. globulins. Acta Endocrinologica (Copenhagen) 123, 649–656. (1998). Abnormal heart rate and body temperature in mice V, R., S,L.&N, E. A. (1989). The micro- lacking thyroid hormone receptor α1. The EMBO Journal 17, heterogeneity of rat TBG. FEBS Letters 244, 343–346. 455–461. W, R. L., A, J. W., MG, M. E., W, W,G.R.&B, G. A. (1995). Thyroid hormone B. L., B,J.D.&F, R. J. (1995). A structural response elements. In Molecular Endocrinology: Basic Concepts role for hormone in the thyroid hormone receptor. Nature 378, and Clinical Correlations (ed. B. D. Weintraub), pp. 217–239. 690–697. Raven Press, New York. W, J. R., T,P.C, B. (1981). Effects of W, G. R., R,H.&S, S. M. (1998). Thyroid experimental hyper- and hypothyroidism on numbers of hormone actions on cartilage and bone: interactions with blood mononuclear cells and immune function in rats and other hormones at the epiphysial plate and effects on linear guinea-pigs. Journal of Endocrinology 91, 61–67. growth. Journal of Endocrinology 157, 391–403. W,C.P.&B, J. A. (1997). Plasma and tissue W,R.A.&D, T. F. (1987) Effect of increased thyroxine and triiodothyronine contents in sublethally circulating corticosterone on serum and thyroidal conce- stressed, aluminium-exposed brown trout (Salmo trutta). ntrations of iodothyronines and the responses to thyrotrophin General & Comparative Endocrinology 106, 120–126. in the immature fowl (Gallus domesticus). General & Comparative W, P. R., D, P. J., D, F. B., C, V., G, Endocrinology 65, 65–72. J., J.&B, S. D. (1993). Rabbit skeletal muscle W,C.M.&MN, F. M. A. (1997). Maternal thyroid # sarcoplasmic reticulum Ca +-ATPase activity: stimulation in hormones in Japanese quail eggs and their influence on vitro by thyroid hormone analogues and bipyridines. Bio- embryonic development. General & Comparative Endocrinology chimica et Biophysica Acta 1153, 184–190. 107, 153–165. W, G. M., F, E. S., T,C.K.F.&G,E.G. W,F.E.&R, B. D. (1996). The timing of thyroid- (1994). Changes in whole-body thyroxine and tri- dependent programming in seasonally breeding male Am- iodothyronine concentrations and total content during early erican tree sparrows (Spizella arborea). General & Comparative development and metamorphosis of the toad Bufo marinus. Endocrinology 103, 82–92. General & Comparative Endocrinology 94, 62–71. W,K.W.&H, A. J. (1987). The influence of W, M. (1994). Nongenomic actions of steroid hormones. dietary fatty acids and hypothyroidism on mitochondrial fatty Trends in Endocrinology & Metabolism 5, 347–353. acid composition. Nutrition Research 7, 1139–1150. W, M. R. (1986). Changes in plasma thyroxine levels during W, A. A. (1940). Dependence of seasonal period- and after spontaneous metamorphosis in a natural population icity in gonadal changes on the thyroid gland in Sturnus of the green tree frog, Rana clamitans. General & Comparative vulgaris L. Comptes Rendus (Doklady) de l’Academie des Sciences de Endocrinology 62, 8–12. l’URSS 27, 741–745. W, C., T, C. C., O, E. S., L, R., W, G. (1995). Retinol transport and metabolism in G,D.J.&E, R. M. (1986). The c-erb-A gene transthyretin-‘knockout’ mice. Nutrition Reviews 53, 98–99. encodes a thyroid hormone receptor. Nature 324, 641–646. W,B.A.&F, D. M. (1986). Regulation of W,S.P.&H, R. S. (1992). Differential regulation the (Na+jK+)-ATPase in cultured chick skeletal muscle. of glucose transporter isoforms by thyroid hormone in rat Modulation of expression by the demand for ion transport. heart. Biochimica et Biophysica Acta 1136, 302–308. Journal of Biological Chemistry 261, 9990–9999. W,S.P.&H, R. S. (1993). Glucose transport W, B. M., A-F, V., O, A., S- stimulation by thyroid hormone in ARL 15 cells: partial role , M. O., K,R.G.&S, R. J. (1997). + of increased GLUT1 glucose transporter gene transcription. Changes in thyroid state affect pHi and Nai homeostasis in Thyroid 3, 135–142. rat ventricular myocytes. Journal of Molecular & Cellular W, R. T., S,H.L.&O,J.H. Cardiology 29, 2653–2663. (1987). An analysis of the interrelationship of nuclear and W, C. J., W, S. L., L, H. E., I-W, plasma triiodothyronine in the sea lamprey, lake trout and V., A-A,P.&S, B. A. (1998). The effects Thyroid hormones and their effects 631

of chronic exercise on metabolic and reproductive functions in teristics of bullfrog (Rana catesbeiana) transthyretin and its male rats. Life Sciences 62, 327–332. cDNA. European Journal of Biochemistry 256, 287–296. W, M. L., C, L. J., L, L., H, L., Y,Y.&G, C. J. (1997). Regulated hypothermia in K, D. M., F,E.E.&A,C.M. the hypothyroid rat induced by administration of propyl- (1994). Anterior pituitary and adrenal cortical hormones thiouracil. American Journal of Physiology 272, R1390–R1395. accelerate or inhibit tadpole hindlimb growth and devel- Y,Y.&B, D. D. (1990). A correlation of thyroid opment depending on stage of spontaneous development or hormone receptor gene expression with amphibian meta- thyroxine concentration in induced metamorphosis. The morphosis. Genes & Development 4, 1917–1924. Journal of Experimental Zoology 270, 175–188. Y,P.M.&C, W. W. (1994). New advances in under- W, C., C,G.&B, M. (1985). Plasma free and standing the molecular mechanisms of thyroid hormone total iodothyronine levels in hypophysectomized and intact action. Trends in Endocrinology & Metabolism 5, 65–72. lamb foetuses during the last third of gestation. Acta Y, J. H. (1997). Is lamprey metamorphosis regulated by Endocrinologica 110, 388–394. thyroid hormones? American Zoologist 37, 441–460. W, J. (1968a). Antioxidant function of thyroxine in vivo. Y, J. H., L, J. F., B,R.A.& P, E. M. (1994). Systemic levels of thyroid Endocrinology 83, 376–378. hormones and insulin in landlocked sea lampreys, Petromyzon W, J. (1968b). Thyroxine degradation. Antioxidant function marinus, during the juvenile feeding period. Comparative & and nonenzymatic degradation during microsomal lipid General Endocrinology 94, 237–243. peroxidation. Archives of Biochemistry & Biophysics 126, 880–891. Y, B., O-C, J., P,A.&A, A. (1988). Y, N., I, A., T, K. P., L, Q., N- Mechanism of -triiodothyronine (T3) uptake by glial C6 , H., T, T.,S, S.,I,H.&M,S. cells: regulation by butyrate. Molecular & Cellular Endocrinology (1992). Loss of thyroid hormone receptor activity in primary 58, 191–198. cultured rat hepatocytes is reversed by 2-mercaptoethanol. Z,J.&L, M. A. (2000). The mechanism of action of Biochemical Journal 281, 669–673. thyroid hormones. Annual Review of Physiology 62, 439–466. Y,K.&M, S. (1998). Differential gene expression of Z, N. D., M,M.B.&T, H. C. (1990). Iden- thyroid hormone receptor α and β in fish development. General tification of multiple thyroid response elements located far & Comparative Endocrinology 109, 75–85. upstream from the rat S"% promoter. Journal of Biological Y, K., M, S., O,T.&I, Y. (1991). Thyroid Chemistry 265, 8136–8143. hormone regulates developmental changes in muscle during Z, L., H, S. G., S,K.A.&O,J.H. flounder metamorphosis. General & Comparative Endocrinology (1994). Identification of thyroid hormone response elements 81, 464–472. in rodent Pcp-2, a developmentally regulated gene of Y, K., T, H., O, M., D, M., cerebellar purkinje cells. Journal of Biological Chemistry 269, M,S.L.A.&S, G. (1998). Structural charac- 13346–13352.