Fundamental and Molecular Mechanisms of Mutagenesis ELSEVIER Mutation Research 333 (I 995) 13 I - I41

Mechanistic considerations for the relevance of animal data on thyroid neoplasia to human risk assessment

R. Michael McClain

Abstract

There are two basic mechanisms whereby chemicals produce thyroid gland neoplasia in rodents. The first involves chemicals that exert a direct carcinogenic effect in the thyroid gland and the other involves chemicals which, through a variety of mechanisms. disrupt thyroid function and produce thyroid gland neoplasia secondary to hormone imbalance. These secondary mechanisms predominantly involve effects on thyroid hormone synthesis or peripheral hormone disposi- tion. There are important species differences in thyroid gland physiology between rodents and humans that may account for a marked species difference in the inherent susceptibility for neoplasia to hormone imbalance. Thyroid gland neoplasia. secondary to chemically induced hormone imbalance, is mediated by thyroid-stimulating hormone (TSH) in response to altered thyroid gland function. The effect of TSH on cell proliferation and other aspects of thyroid gland function is a receptor mediated process and the plasma membrane surface of the follicular cell has receptors for TSH and other growth factors. Small organic molecules are not known to be direct TSH receptor agonists or antagonists; however, various antibodies found in autoimmune disease such as Graves’ disease can directly stimulate or inhibit the TSH receptor. Certain chemicals can modulate the TSH response for autoregulation of follicular cell function and thereby increase or decrease the response of the follicular cell to TSH. It is thus important to consider mechanisms for the evaluation of potential cancer risks. There would be little if any risk for non-genotoxic chemicals that act secondary to hormone imbalance at exposure levels that do not disrupt thyroid function. Furthermore, the degree of thyroid dysfunction produced by a chemical would present a significant toxicological problem before such exposure would increase the risk for neoplasia in humans.

Kqwords: Thyroid neoplasm: Chemical involvement

1. Introduction other aspects of thyroid gland function is a receptor mediated process. Many chemicals stimulate thyroid gland growth Small organic molecules are not known to be as a compensatory response to altered thyroid func- direct TSH receptor agonists or antagonists; how- tion. This response is mediated via thyroid-stimulat- ever, various antibodies found in autoimmune dis- ing hormone (TSH) released by the pituitary in ease such as Graves’ disease can directly stimulate response to decreased circulating levels of thyroid or inhibit the TSH receptor. Chemicals can, how- hormone. The plasma membrane surface of the fol- ever, modulate the TSH response by effects on com- licular cell has receptors for TSH and other growth ponents of the TSH receptor responsible for autoreg- factors. The effect of TSH on cell proliferation and ulation of follicular cell function and thereby in-

Elsevier Science B.V. SSDI 0027.5107(95)00139-S 132 R.M. M&lain / Muration Research 333 (lYY51 1316142 crease or decrease the response of the follicular cell pathogenic factor responsible for thyroid tumor pro- to TSH. duction. That excessive secretion of endogenous TSH alone (in the absence of any chemical treatment) will produce a high incidence of thyroid tumors has been 2. Altered thyroid gland function and thyroid clearly established by experiments in which rats neoplasia were fed diets deficient in (Axelrad and Leblond, 1955: Bielschowsky, 1953; Isler et al., An understanding of the mechanisms of chemical 1958: Leblond et al.. 1957) or in which TSH-secret- induction of thyroid neoplasia was obtained during ing pituitary tumors were transplanted into mice with experimentation in the 40s and 50s as a result of normal thyroids (Furth, 1954). Iodine deficient diets interest in thyroid cancer during this period. It was are goitrogenic and result in an increased TSH secre- recognized that two basic mechanisms were involved tion with concomitant high incidence of thyroid tu- in thyroid carcinogenesis. the first of which involves mors. These effects can be reversed by iodine sup- chemicals that exert a direct carcinogenic effect on plementation. thyroid hormone replacement, or hy- the thyroid gland. Thyroid tumors have been pro- pophysectomy. Goitrogenic substances or regimens duced by a variety of direct-acting carcinogenic sub- are also powerful promoters of thyroid gland neopla- stances such as the polycyclic hydrocarbons sia after the administration of direct acting carcino- (Esmarch. 1942; Gnatyshak, 1957). 2-acetylaminof- genic substances (Bielschowsky, 1944; Hiasa et al., luorene (Cox et al.. 1947). dichlorobenzidine (Pliss. 1982b; Morris, 1955). 1959). and a variety of nitrosamines. The second basic mechanism was the production of thyroid tu- mors by a variety of regimens that result in a hor- 3. Thyroid hormone synthesis mone imbalance. Kennedy and Purves, 1941 found thyroid adenomas in rats fed a diet containing bras- The functional unit of the thyroid gland is the sica seeds. a naturally occurring goitrogen. Subse- follicle which consists of an area of colloid, the quently, numerous studies have demonstrated that storage form of thyroid hormone. surrounded by a treatment with a variety of anti-thyroid substances single layer of follicular epithelium. The thyroid (, thiouracil and their derivatives, and 3- gland is unique among the endocrine glands in that amino-l,3,4-triazole) will result in a high incidence hormone synthesis and storage are essentially extra- of thyroid tumors in rats (Napalkov, 1976). Other cellular processes that occur at the apical surface of substances exerting anti-thyroid effects in rats such the cell membrane. Thyroid hormone synthesis in- as some sulfonamides will also produce thyroid tu- volves the active transport of iodine into the cell and mors (Swarm et al.. 1973). a peroxidase mediated iodination and coupling of A consistent mechanism, widely accepted by many residues on to form thyroid investigators, to explain the pathogenesis of thyroid hormone which is stored as colloid until released tumors induced in rats treated with anti-thyroid drugs into the circulation (Degroot and Niepomniszcze, has been described (Furth. 1959, 1968). Anti-thyroid 1977: Ekholm. 198 1). The thyroid follicle and the drugs initially produce a hormonal imbalance by follicular cell are highly organized and polarized interfering with thyroid hormone production. AS a structures. Thyroid hormone synthesis involves reac- result, a sustained increase in the synthesis and tive biochemical processes, and the organization of secretion of TSH occurs via the negative feedback its structure and function serves to protect the cell system of the pituitary gland to stimulate thyroid from accidental iodination and cytotoxicity from oxi- function. Increased TSH stimulation produces a vari- dation products. The thyroid peroxidases appear to ety of morphological and functional changes in the be active only at the apical surface of the cell follicular cell including follicular cell hypertrophy. membrane. hyperplasia, and ultimately neoplasia. The sustained Thyroid hormone release involves pinocytosis of excessive level of TSH is considered to be the colloid by microvilli and digestion by lysosomes to R.M. McClain/ Mutution Research 333 (19%) 131-142 133 release TJ and T, which are secreted into plasma and pituitary are released in response to decreased circu- bound to plasma proteins for transport to peripheral lating levels of thyroid hormone. Another important tissues. The most important aspect of the metabolism aspect in the control of thyroid function is autoregu- of thyroid hormone is monodeiodination to form T,. lation (Vanderlan and Caplan, 1954; Halmi and Spir- the physiologically active form of thyroid hormone. tos, 1954; Bray, 1968) via a modulation of the T, is also metabolized by glucuronidation and T, is receptor response to TSH. The TSH receptor has an predominantly sulfated and both are excreted in urine extracellular component which undergoes a confor- and bile (Robbins. 198 1). mational change after binding to TSH. The catalytic unit of the receptor (adenylcyclase) produces cyclic AMP (C-AMP) which functions as a second messen- 4. Regulation of thyroid function ger and mediates all the intracellular effects of TSH including the various differentiated functions of the Thyroid gland function is controlled by the hypo- follicular cell and cell proliferation (Kahn et al., thalamus and pituitary and is regulated via a negative 1985). The adenylcyclase activity of the TSH recep- feedback process in which thyrotropin-releasing fac- tor is subject to autoregulation and is controlled by a tor (TRH) from the hypothalamus and TSH from the class of iodolipids that attenuates the formation of

Fig. I. Photomicrograph (33 X ) of a thyroid gland from a control rat (A) and a control cynomolgus monkey (B). Note the ‘active’ appearance of the rat thyroid gland which consists of microfollicles with cuboidal follicular epithelium and small amounts of colloid. In contrast. the monkey thyroid gland consists of large follicles containing abundant colloid and surrounded by flattened follicular epithelium. C-AMP (Pisarev. 1985; Pisarev et al., 19881. When (Pisarev, 1985) or chemicals that can either supply the iodolipids are low as a result of either iodine iodine or perhaps be recognized as a modulator deficiency or an inhibition of the organification of similar to the endogenous iodolipids can attenuate iodine. the attenuation of the receptor response is the TSH response by inhibiting the formation of removed and the response to TSH is enhanced. The C-AMP. physiologic role of this autoregulatory process is most likely to modulate the response of the thyroid to TSH depending upon the supply of iodine. Ingbar. 5. Species differences in thyroid gland 1972, has referred to this process as the ‘iodostat’ biochemistry and physiology which is essentially a set point in the follicular cell for the response to TSH. Since autoregulation re- There are marked species differences in thyroid quires the enzymatic iodination of lipids, a process gland physiology that must be taken into account in similar to thyroid hormone synthesis, inhibitors of an evaluation of species differences in the induction iodination will decrease the iodolipid content and of thyroid gland neoplasia secondary to hormone thereby enhance the TSH response. Iodine deficiency imbalance. The most obvious species difference be- (Vanderlan and Caplan. 1954: Halmi, 1954; Bray. tween rodents and primates is the lack of thyroid 1968) and chemicals such as (PTU) binding globulin (TBG) in the rodent and other and methamidazole (Chazenbaik et al.. 19851 thus species including birds and reptiles (Dohler et al., enhance the TSH response beyond that due to in- 19791. TBG is the predominant plasma protein that creased TSH alone. In contrast. excess iodine binds and transports thyroid hormone in the blood. In

Fig. 2. Photomicrograph (66 x ) of a thyroid gland from a control rat (A), a rat supplemented with thyroid hormone in the diet for 8 wcrk\ (B). and a control cynomolgus monkey (Cl Note that in the rat supplemented with thyroid hormone (B) there is a marked change in histological appearance. The follicles are large with abundant colloid surrounded by flattened follicular epithelium. The follicle!. are more similar in appearance to the monkey (C). R.M. McChin / Mutution Research 333 (1995) 13/- I42 135 man, thyroxine binds to three plasma proteins (TBG, ante or simple hypothyroidism can be assessed by pre-albumin and albumin) with binding constants of comparing humans in iodine deficient areas of en- 10--‘O. IO-’ and IO-‘, respectively (Robbins and demic goiter to rats in these same areas or rats Rall. 1979). In rodents. the lack of TBG with a treated with iodine deficient diets. Over the years binding affinity of 3 and 5 orders of magnitude more there has been a slight, disputed association between than albumin and pre-albumin may be a factor re- endemic goiter and human thyroid cancer. Relatively sponsible for species differences in thyroid gland extensive epidemiologic studies in areas of endemic function. goiter have not shown a clear etiologic role (Doniach, The half-life (T,,?) of thyroxine (T,) is 12 h in 1970; Pendergrast et al., 1961; Saxen and Saxen, the rat compared to 5-9 days in humans: serum TSH 1954). In contrast, rodents in areas of endemic goiter is 25 times higher in the rodent as compared to man (Wegelin. 19281 or those treated with iodine defi- (Dohler et al.. 1979). These findings indicate a much cient diets (Hellwig, 1935; Bielschowsky, 1953; Ax- higher functional activity in the rodent thyroid gland elrad and Leblond, 1955; Isler et al.. 1958; Leblond as compared to the primate: a conclusion also sup- et al.. 1957) exhibit a high incidence of thyroid gland ported by the histological appearance of the thyroid neoplasia. gland. For example. in the cynomolgus monkey, the The species differences between rodents and pri- follicles are uniformly large with abundant colloid mates in thyroid gland physiology. the spontaneous and are lined by relatively flattened follicular epithe- incidence of thyroid gland neoplasia, and the appar- lial cells (Fig. I). In contrast, the rodent thyroid has ent susceptibility to neoplasia secondary to simple large follicles only in the periphery of the gland. The hypothyroidism support the conclusion that there is a interior is comprised of comparatively small follicles marked species difference in thyroid gland neoplasia with small amounts of colloid surrounded by a more secondary to hormone imbalance. The rodent will cuboidal follicular epithelium. Interestingly. when exhibit an increase in thyroid gland neoplasia in the rats are supplemented with exogenous thyroid hor- presence of mild to moderate increases in TSH. In mone, the follicles accumulate colloid and increase contrast, no clear etiologic role for hypothyroidism in size. The epithelial cells assume a more flattened in human thyroid cancer has been established even appearance, and the overall histological appearance though chronic hypothyroidism, in the moderate to resembles that of the monkey (Fig. 21. Thus, both the severe range, has occurred in humans in areas of physiologic parameters and the histologic appearance endemic goiter (Doniach. 19701. Thus, the contribu- indicate that the rodent thyroid gland is markedly tion of endemic goiter to human thyroid cancer is more active and operates at a considerably higher small at most. level with respect to thyroid hormone turnover as compared to the primate. The incidence of ‘spontaneous’ thyroid follicular 6. Mechanisms for altered thyroid function cell neoplasia is also markedly different between rats and humans. The Fischer-344 male rat exhibits about Thyroid hormone synthesis. release, transport, cel- a 2% incidence of thyroid follicular cell neoplasia lular uptake, conversion of T3 to T,, hormone (0.87~ carcinoma and 1% adenoma) (Haseman et al., metabolism and the regulation of these processes by 19841 as compared to average incidence of approxi- the hypothalamic-pituitary-thyroid axis and auto- mately 0.0040/c. with a range of 0.001% to 0.016% regulatory processes in the thyroid gland itself is a of carcinoma in humans (Sokal. 1953: Ron and complex process and provides many ways in which Modan, 1982). The prevalence of occult human car- chemicals can interfere with thyroid gland function. cinoma at autopsy has been high in some studies, all Regardless of mechanism, the response to hypo- or most tumors being papillary carcinoma (Ron and thyroidism is similar. The pituitary will release TSH Modan. 1982). a form rarely observed in the rat as a compensatory response to stimulate the thyroid (Napalkov, 1976). to produce more hormone. Chronic stimulation of the The relative susceptibility of rodents and humans thyroid gland by TSH in the rodent leads to the to thyroid neoplasia secondary to hormone imbal- progression of follicular cell hypertrophy. hyperpla- I.16 R. M. McClain /Mutation Rrsetrrc~h 33.? f I YY5) 13/- 142 sia and eventually neoplasia (Furth, 1959. 1968). thyroid gland neoplasia in ‘-year carcinogenicity Mechanisms for the action of chemicals can be either studies (McClain, 1989). in&a- or extrathyroidal or both. Intrathyroidal mecha- Studies have shown that small amounts of thyrox- nisms can involve iodine uptake or hormone synthe- ine will block the tumor promoting effect of a micro- sis, and extrathyroidal mechanisms can involve ef- somal enzyme inducer such as phenobarbital, thus fects on hormone metabolism or disposition. A few this effect and presumably those observed in 2-year important examples are discussed below. studies are considered to be secondary to hormone imbalance as opposed to a direct tumor promoting or direct carcinogenic effect in the thyroid gland (MC- Clain et al.. 1988. 1989; McClain. 1989). 7. Extrathyroidal mechanisms With respect to the chemicals mentioned, tetraiod- ofluoresceine produced only mild effects on thyroid The conversion of thyroxine (T,) into the more function in humans at very large multiples of the active hormone (T,) occurs in many allowable daily intake (Gardner et al., 1987). Chronic peripheral tissues and is mediated by a microsomal exposure to anticonvulsant drugs. many of which are 5’-monodeiodonase that removes one iodine at the 5’ enzyme inducing at therapeutic dosages. result in position in thyroxine (Robbins. 198 1). Various iodi- only mild changes in thyroid function [moderately nated organic compounds such as tetraiodofluores- decreased Ta with normal T, and TSH values and a ceine (Ruiz and Ingbar, 1982). amiodarone (Burger normal TSH response to administered thyrotropin et al., 1976) and various iodinated radio-contrast releasing hormone (TRH) (Ohnhaus and Studer. media (Burgi et al., 1976). will inhibit the 5’-mono- 198311. These changes are not clinically significant. deiodinase and disrupt the conversion of TJ to T,. The decrease in serum T, values results in a com- pensatory increase in pituitary TSH and prolonged 8. Intrathyroidal mechanisms treatment with high dosages of tetraiodofluoresceine will result in a moderate increase in thyroid follicular Two classes of chemicals known to inhibit thyroid neoplasia in rodents. PTU and methimidazole. in hormone synthesis are the thioureylenes (thiourea, addition to inhibiting TPO. will inhibit the mon- propylthiouracil, methimidazole) and the sulfo- odeiodinases. namides (sulfadiazine. sulfamethazine). The thioure- The effect of chemicals on various aspects of ylenes are considerably more potent in inhibiting thyroid hormone metabolism have an important im- thyroid hormone synthesis than the sulfonamides: pact on thyroid hormone economy in the rodent however. sufficiently high dosages of many sulfo- (Cavalieri and Pitt-Rivers, 198 I). The monodeiodi- namides are goitrogenic in the rodent (Mackenzie et nases are quantitatively the most important path in al., 1941; Mackenzie and Mackenzie, 1943; Swarm the disposition of thyroxine. In addition, thyroxine is et al.. 1973; Takayama et al., 1986). Although the glucuronidated and T, is sulfated and subsequently goitrogenicity of the sulfonamides has been known excreted in bile. Deamination. decarboxylation and since 194 1. only a few have been tested for carcino- cleavage of the ether link occur but are of quantita- genicity. Sulfamethoxazole at dosages of 50 mg/kg tively lesser importance (Robbins, 1981). Many per day or more is goitrogenic and produced thyroid chemicals will induce hepatic microsomal enzymes neoplasia in rats within 50 weeks of treatment. Sul- at high dosages and alter thyroid function in rodents famethazine produced goitrogenic effects and thyroid by increasing the hepatic disposition of thyroid hor- neoplasia in rats at doses of 600 ppm or more and in mone (Oppenheimer et al.. 1968: Comer et al.. 1985; mice at doses of 4800 ppm (Heath and Littlefield. Sanders et al.. 1988: McClain et al.. 1989; Hill et al.. 1984: Fullerton et al., 1987; Littlefield et al.. 1989. 1989). Decreased serum thyroid hormone results in a 1990). Sulfisoxazole which is only weakly goitro- compensatory increase in pituitary TSH which can genie in rodents did not produce thyroid gland neo- exert a tumor promoting effect in initiation-promo- plasia at doses up to 400 mg/kg per day in rats or tion models (Hiasa et al., 1982a) or an increase in 2000 mg/kg per day in mice treated for 2 years R.M. McClain/Mutation Research 333 (19951 131-142 137

(NCI, 1979). The various sulfonamides that have species differences in the goitrogenic effects of been tested had no apparent genotoxic effects, and sulfonamides is thus due to marked species differ- the observed thyroid gland neoplasia using goitro- ence in the inhibition of TPO. genie dosages were considered to be secondary to hormone imbalance. The goitrogenic effect of the sulfonamides is 9. Sulfamethazine known to be highly species-specific. Some species (including rats, mice, hamsters, dogs and swine) are As mentioned above, sulfamethazine is goitro- sensitive, whereas no goitrogenic effect is observed genie in rodents and produces thyroid gland neopla- in other species (including chickens, guinea pigs or sia in rats at doses of 600 ppm or more and in mice monkeys) (Mackenzie and Mackenzie, 1943: Swarm at a dosage of 4800 ppm. Since sulfamethazine is et al., 1973: Takayama et al., 1986). In the primate, used in food producing animals in the United States, no effect on thyroid function or morphology was it is subject to the provisions of the Delaney amend- observed in rhesus monkeys treated with sul- ments prohibiting the use of food or color additives famethoxazole (Swarm et al., 1973) for 1 year at and animal drugs found to ‘induce’ cancer in animal doses up to 300 mg/kg per day nor in cynomolgus or humans. An exemption known as the DES Proviso monkeys treated for 4 weeks with sulfamonomethox- was subsequently provided for animal drugs if no ine (Takayama et al., 1986) at doses up to 300 residues are present. Since modem analytical tech- mg/kg per day. In humans, no clinically significant nology can measure extremely small traces of drugs, effects on thyroid function have been observed with the FDA implemented a policy in 1973 known as the sulfonamides at therapeutic dose levels (Koch-Weser Sensitivity of the Method (SOM) Policy in which the et al., 197 1). Although Cohen et al. (1980) observed FDA reasoned that the ‘no residue’ provision was a small decrease in serum thyroid hormone values in fulfilled if an analytical procedure was available that patients treated with sulfamethoxazole in cotrimoxa- could measure a residue level considered to represent zole. no increase in TSH was observed. In addition, an ‘insignificant risk’ (FDA, 1973). The ‘insignifi- no abnormalities in thyroid function were observed cant risk’ was then defined as a residue level that in a group of young patients receiving chronic treat- would represent a risk of less than one additional ment with cotrimoxazole (Smellie et al.. 1982, 1983). case of cancer in 1 million lifetimes (< 1 X 10eh Takayama et al. (1986) studied the species differ- lifetime risk). This was determined by using a quan- ence between rats and monkeys using propylth- titative risk assessment procedure that assumes a iouracil (PTU) and a sulfonamide, sulfamonomethox- linear dose-response from high to low doses and the ine (SMM). SMM or PTU treated rats exhibited absence of a threshold (Gaylor and Kodell, 1980; decreased T, and TJ values and markedly elevated Farmer et al., 1982). This approach was further TSH values accompanied by follicular cell hyperpla- modified in 1987 in which the specific quantitative sia and increased thyroid gland weight at 30 or 300 risk assessment procedure was removed from regula- mg/kg per day. In contrast, no effect on thyroid tions and placed into guidelines (FDA, 1987). The function or morphology was observed in monkeys FDA will consider scientific data on mechanism of receiving 300 mg/kg per day of either compound. In action and the guidelines contain a waiver provision vitro there was a marked difference in the inhibition under which a sponsor may submit scientific data of microsomal thyroid peroxidase (TPO) depending and propose an alternative procedure to estimate risk upon the source of the enzyme. The concentration (e.g., no observed effect level (NOEL) and a safety (IC,,) of SMM required to inhibit monkey TPO was factor) to establish safe residue levels for an animal approx. 450 times greater than that required to in- drug. In the absence of data or when appropriate, the hibit TPO from rats which explains why the sulfo- linear extrapolation procedure with the 1 X IO-’ namides are goitrogenic in rats at relatively low lifetime risk benchmark will continue to be used as dosages but do not produce effects in monkeys at the default procedure. very high dosages or in humans at therapeutic In the case of sulfamethazine the residue level dosages. The biochemical basis for the observed determined by the linear extrapolation procedure was 138 R.M. McClnin /Mutution Rr.wmh 333 (1995~ I_?-142

well below the currently permitted level of 100 ppb operative in the non-human primate or under condi- and would preclude the use of sulfonamides in the tions or potential human exposure. The purpose of animal health market. In order to provide a scientific these studies was to develop a database which would basis to justify the use of an alternative risk assess- support the hypothesis that thyroid tumors are sec- ment procedure, studies were conducted to investi- ondary to hormone imbalance after treatment with gate the mode of action involved in the sulfametha- high doses of sulfamethazine. Such data would en- zine thyroid gland tumor response. These included a able the regulation of sulfamethazine on the basis of number of short-term tests to evaluate the genotoxic mechanistic considerations and dose-response char- potential of sulfamethazine, additional studies to de- acteristics for effects on thyroid gland function. termine the mechanism of tumor formation, studies Sulfamethazine was shown to be an inhibitor of to determine dose-response characteristics. and stud- thyroid gland microsomal peroxidase. an important ies to determine whether the proposed mechanism is enzyme in the biosynthesis of thyroid hormone.

-,m 1 T3 T4

TSH Thyroid Weight

I D..,’ , , I , 0 ?04osolllD aa ace I(0 3m Kmot2nmz~ Zb ?b 4k ia Ido &k&L&,&amolmm25m3-I - Log Do= @Pm Lw Do- Mm)

Fig. 3. Dose response characteristics for the effect of sulfametharine on thyroid function in CR/CD raw Rats were treated with sulfamethaaine as a dietary admix at dohager. of from 20 to 12000 ppm for 4 weeks. The data points represent the mean i SE for I5 rats per dose group. The response data are plotted as a function of the log dose. The horizontal line represent> the mean value for the control group and the vertical line represents the best fitting line for the four highest dose groups. There is a wide mnq of doses for which no effect on thyroid gland function ir: observed followed by a sharp relatively linear increase in response at approximately I60 ppm or more of sulfamethazine. Histologically. follicular cell hyperplasia paralleled the increase in plasma TSH. These data demonstrate that the responses to sulfamethazine are non-linear from high to low doses characteristic of a process that exhibits an apparent threshold. R.M. McClain /Mutation Research 333 (lW5l 131-132 139

Studies of thyroid gland function demonstrated that 11. Conclusion thyroid hormone imbalance occurs under the condi- tions of the rodent bioassay commensurate with a The results of the scientific studies are in accord minimal to moderate tumor response (decreased thy- with the conclusion that the mode of action of roid hormone, increased TSH, increased thyroid sulfamethazine is non-genotoxic and that all morpho- weights and follicular cell hypertrophy and hyperpla- logical effects are similar to and secondary to ele- sia). All the morphological effects on the thyroid vated TSH levels as a compensatory response to an gland were completely reversed after the withdrawal inhibition of thyroid hormone synthesis. A large of sulfamethazine treatment. These morphological body of evidence supports the conclusion that in- changes in the thyroid gland of sulfamethazine treated creased TSH results in follicular cell hypertrophy rats were similar in appearance to that observed in and hyperplasia and is the factor involved in the rats fed a low iodine diet. The supplemental dietary pathogenesis of thyroid gland neoplasia under these administration of thyroid hormone completely inhib- conditions. Thyroid hormone imbalance exists under ited the functional and morphological changes ob- the conditions of the sulfamethazine bioassay and served with sulfamethazine treatment at doses that this alone is sufficient to account for diffuse follicu- normalized but did not suppress TSH. Further, no lar cell hypertrophy and hyperplasia as well as focal detectable thyroid gland effects were observed in hyperplasia and neoplasia in sulfamethazine treated hypophysectomized rats treated with sulfamethazine. rats. These data thus support the conclusion that In vitro, sulfamethazine did not increase cell prolifer- thyroid gland neoplasia is secondary to hormone ation in FRTL-5 cells in the absence of TSH. In imbalance and that this does not represent a carcino- accord with the other studies with sulfonamides in genic effect of sulfamethazine. Furthermore, the monkeys. no effect on thyroid gland function in study of dose-response characteristics clearly demon- cynomolgus monkeys was observed at dosages of up strates that the effects of sulfamethazine on thyroid to 300 mg/kg per day for 13 weeks supporting the function are non-linear from high to low doses char- conclusion that the effect of sulfamethazine on the acteristic of a process with an apparent threshold. thyroid gland is species-specific. Thus, the use of a linear procedure to extrapolate risk would appear to be scientifically inappropriate in this case. The scientific data justify the use of an altema- tive risk assessment procedure to establish safe 10. Dose-response characteristics residues of sulfamethazine such as the no observed adverse effect level (NOAEL) for the effects of sulfamethazine on rat thyroid function in conjunction A 4-week study was conducted in rats to investi- with a safety factor of 100 for inter-and intraspecies gate the dose-response characteristics for the effects variation. The sulfamethazine example represents a of sulfamethazine on thyroid function at dose levels good case for the regulation of a chemical based on of sulfamethazine ranging from 15 to 12 000 ppm. A the mode of action in the rodent bioassay. characteristic log dose-response relationship was ob- served in all the parameters of thyroid function measured. There were no significant effects at lower dosages. followed by a sharp relatively linear rise at 12. Extrapolation higher dosages (Fig. 3). The results of this experi- ment clearly demonstrate that the effects of sulfamet- hazine on all aspects of thyroid function assessed, as Chemicals can alter thyroid function and result in well as the morphological changes in the thyroid hypothyroidism through a variety of intrathyroidal gland exhibit a non-linear response from high to low and extrathyroidal mechanisms. Many of these chem- doses. The dose-response characteristics observed in icals will cause thyroid hormone imbalance and neo- this experiment are consistent with a process that plasia in rodent carcinogenicity studies. It is thus exhibits an apparent threshold. important to consider mechanisms for the evaluation 140 R.M. McC&t / Mururion Rr.trurch 333 ( IYY5J 131-142

of potential cancer risks. There would be little if any as a model for the study of drug effects on thyroid function: risk for non-genotoxic chemicals that act secondary consideration of methodological problems. Pharmacol. Ther.. 5. 305-3 18. to hormone imbalance at exposure levels that do not Doniach. I. (1970) Aetiological consideration of thyroid carci- disrupt thyroid function. Furthermore, the degree of noma, in: D. Smithers (Ed.), Neoplastic Diseases at Various thyroid dysfunction produced by a chemical would Sites. Tumours of the Thyroid Gland. E&S Livingston. Edin- present a significant toxicological problem before burgh/London. such exposure would increase the risk for neoplasia Ekholm. R. ( 19X 1) Iodination of thyroglobulin, an extracellular or intracellular process? Mol. Cell. Endocinol., 24. 141~ 163. in humans. Esmarch. 0. (1943) Deposition of methylcholanthrene in some organs of the rats. Acta Pathol. Microbial. Stand.. 19, 79-99. Farmer, J.H., R.L. Kodell and D.W. Gaylor (1982) Estimation and References extrapolation of tumor probabilities from a mouse bioassay with survival/sacrifice components. Sot. Risk Anal.. 2, 27-34. Axelrad. A.A. and C.P. Leblond (1955) Induction of thyroid FDA (I 973) Compounds used in food-producing animals. Criteria tumors in rats by a low iodine diet. Cancer, 8, 339-367. and procedures for evaluating the safety of carcinogenic Bieischowsky. F. (1944) Tumours of the thyroid produced by residues. Fed. Reglst.. _38. 19226-19227. 2.acetyl-amino-fluorene and ally]-thiourea. Br. J. Exp, Pathol.. FDA (I 987) Sponsored compounds in food-producing animals: 2s. 90-95. Criteria and procedures for evaluating the safety of carcino- Bielschowsky. F. (1953) Chronic iodine deficiency as cause of genic residues: Final Rule, 53 CFR. 49572-49588. neoplasia in thyroid and pituitary of aged rats. Br. .I. Cancer, Fullerton. F.R.. R.J. Kushmaul. R.L. Suber and N.A. Littlefield 7, 203-213. ( 1987) Influence of oral administration of sulfamethazine on Bray. G.A. (1968) Increased sensitivity of the thyroid in iodine- thyroid hormone levels in Fisher-344 rats. J. Toxicol. Environ. depleted rats to the goitrogenic effects of thyrotropin. J. Clin. Health . 22 , 175-185. Invest., 47, 1640- 1647. Furth. J. (1953) Morphologic changes associated with thy- Burger, A.. D. Dinichert, P. Nicod, M. Jenny. T. Lemarchand- rotropine-secreting pituitary tumors, Am. J. Pathol.. 30, 321. Beraud and Vallotton. M.B. (1976) Effect of amiodarone on 463. serum triiodothyronine. reverse triiodothyronine, thyroxine and Furth, J. (1968) Pituitary Cybernetics and Neoplasia. Harvey thyrotropin. A drug influencing peripheral metabolism of thy- Lectures. Academic Press. New York/London. pp. 37-7 I. roid hormones. J. Clin. Invest., 58. 255-259. Furth, J. (1959) A meeting of ways in cancer research: Thoughts Burgi. H.. C. Wimpfheimer, A. Burger. W. Zaunbauer. H. Rosier on the evolution and nature on neoplasms. Cancer Res., 19. and T. Lemarchand-Beraud (1976) Changes of circulating 24 I-258. thyroxine, triiodothyronine and reverse triiodothyronine after Gaylor. D.W. and R.L. Kodell (1980) Linear interpolation alo- radiographic contrast agents, J. Clin. Endocrinol. Metab.. 33, gorithim for low dose risk assessment of toxic substances. J. 1203-1210. Environ. Pathol. Toxicol.. 3. 305-3 I?. Cavalieri. R.R. and R. Pitt-Rivers (1981) The effects of drug\ on Gardner. D.F.. R.D. Utiger. S.L. Schwartz. P. Witorsch. B. Mey- the distribution and metabolism of , Pharma- ers, L.E. Braverman and R.J. Witorsch (1987) Effects of oral co]. Rev.. 33, 55-80. erythrosine (2‘,5‘,5’.7’-tetraiodofluorescein) on thyroid func- Chazenbaik. G.D.. M.A. Pisarev, G.J. Juvenal, D.L. Kleiman de tion in normal men, Toxicol. Appl. Pharmacol.. 91, 299-304. Pisarev. H. Mercuri and M. De Tomaa (1985) Biosynthesis Gnatyshah. A.I. (19.57) Experimental thyroid tumors and accom- and regulation of iodolipids in calf thyroid, Acta Endocrinol., panying changes in some endocrine glands. Vop. Onkol.. 3. 10X.72-78. 659-665. Cohen. H.N.. G.H. Beastall. W.A. Ratcliffe. C. Gray, I.D. Watson Halmi. N.S. and B.N. Spirtos (1953) Analysis of the action of and J.A. Thompson (1980) Effects on human thyroid function propylthiouracil on the pituitary thyroid axis of rats. En- of sulphonamide and trimethoprim combination drugs, Br. docrinology. 55. 613. Med. J.. 281, 646-647. Halmi. N.S. (1954) Regulation of the rat thyroid in short term Comer C.P.. C.P. Chengelis, S. Levin and F.N. Kotsonia (1985) iodine deficiency. Endocrinology, 54. 216-224. Changes in thyroidal function and liver UDP-glucurono- Haaeman. J.K., J. Huff and G.A. Boorman (1984) Use of histori- syltransferase activity in rats following administration of a cal control data in carcinogenicity studies in rodents. Toxicol. novel imidarole. Toxicol. Appl. Pharmacol.. 95. 37X. Pathol., I?. 126-131. Cox. A.J.. R.H. Wilson and F. DeEds (1947) The carcinogenic Heath. J.E. and N.A. Littlefield (1983) Effect of subchronic oral activity of 2-acetylaminotluorene: Characteristics of the le- sulfamethazine administration on Fischer 344 rats and B6C3Fl sions in albino rats. Cancer Res.. 7. 647-657. mice. J. Environ. Pathol. Toxicol. Oncol.. 5, 201-213. Degroot. L.J. and H. Niepomniszcze (1977) Biosynthesis of thy- Hellwig. C.A. (1935) Thyroid adenoma in experimental animals. roid hormone: basic and clinical aspects. Metabolism, 26. 6. Am. J. Cancer. 23. 550-555. 665-718. Hiaaa. Y., M. Ohshimn. Y. Kitahori. T. Yuasa. T. Fujita and C. Dohler, K.D.. C.C. Wong and A. van zur Muhlen (1979) The rat Iwata (1983b) Promoting effects of 3-amino-l.3.4~tria7ole on R-M. McClain /Murarion Research 333 (lY95) 131-142 141

the development of thyroid tumors in rats treated with N- Napalkov. N.P. (1976) Tumors of the thyroid gland. in: V.S. bist2-hydroxypropylhritrosamine, Carcinogenesis. 3. 38 l-384. Turusov (Ed.). Pathology of Tumors in Laboratory Animals, Hiasa. Y., Y. Kitahori. M. Ohshima, T. Fujita. T. Yuasa, N. Vol. 1. Part 2. Tumors of the Rat, IARC Scientific Publication Konishi and A. Miyashiro (1982a) Promoting effects of No. 6. pp. 239-372, Lyon. France. phenobarbital and barbital on development of thyroid tumors NC1 (1979) National Cancer Institute. Bioassay of sulfisoxazole in rats treated with N-bist2-hydroxypropylhtitrosamine, Car- for possible carcinogenicity. Technical Report, NCI-CG-TR- cinogen&s. 3. 1187-I 190. 138, DHEW Publication No. 79-1393. Bethesda. MD. Na- Hill. R.N.. L.S. Erdreich. O.E. Paynter. P.A. Roberts, S.L. Rosen- tional Institute of Health. thal and C.F. Wilhinson t 1989) Review. Thyroid follicular cell Ohnhaus, E.E. and H. Studer (1983) A link between liver micro- carcinogenesis. Fund. Appl. Toxicol.. 12. 629-697. somal enzyme activity and thyroid hormone metabolism in Ingbar. S.H. (1972) Autoregulation of the thyroid. Response to man, Br. J. Clin. Pharmacol., 15. 71-76. iodide excess and depletion. Mayo Proc. 47, 814-823. Oppenheimer. J.H.. G. Bernstein and M.I. Surks (1968) Increased Isler. H.. C.P. Leblond and A.A. Axelrad (1958) Influence of age thyroxine turnover and thyroidal function after stimulation of and of iodine intake on the production of thyroid tumors in the hepatocellular binding of thyroxine by phenobarbital. J. Clin. rat. J. Natl. Cancer Inst.. 21. 1065-1081. Invest., 47. 1399- 1406. Kennedy. T.H. and H.D. Purves (1941) Studies on experimental Pendergrast. W.J., B.K. Milmore and S.C. Marcus (19611 Thyroid goitre. I. The effect of brassica seed diets on rats, Br. J. Exp. cancer and thyrotoxicosis in the United States. Their relation- Pathol., 22. 241-244. ship to endemic goiter. J. Chronic Dis., 13, 22-38. Koch-Weser. J., V.W. Side], M. Dexter. C. Parish, D.C. Finer and Pisarev, M.A. (1985) Thyroid autoregulation. J. Endocrinol. In- P. Kanarek (1971) Adverse reactions of sullisoxazole. sul- vest. 8. 475-484. famethoxazole and nitrofurantoin. Manifestations and specific Pisarev, M.A.. G.D. Chazenbalk, R.M. Valsecchi. G. Burton, L. reaction rates during. 2118 courses of therapy. Arch. Intern. Krawiec, E. Monteagu. G.J. Juvenal. R.J. Boado and H.A. Med., 128, 399-404. Chester t 1988) Thyroid autoregulation. Inhibition of goiter Kohn, L.D.. S.M. Aloj. D. Tombaccini, C.M. Rotella. R. Tocca- growth and of cyclic AMP formation in rat thyroid by iodi- fondi. C. Marcocci. D. Corda and E.F. Gullman (1985) The nated derivatives of arachidonic acid. J. Endoclinol. Invest.. thyrotropin receptor. In Biochemical actions of hormones. G. 1 I, 669-674. Litwack (Ed.), Academic Press, New York. pp. 457-512. Plisa. G.B. f 1959) The blastomogenic action of dichlorobenzidene. Leblond. C.P.. H. lsler and A. Axelrad t 1957) Induction of thyroid Vop. Onkol.. 5. 524-533. tumors by a low iodine diet. Can. Cancer Conf.. 2, 248-266. Robbins, J. and J.E. Rall (1979) The iodine containing hormones, Littlefield. N.A., W.G. Sheldon. R. Allen and D.W. Gaylor (19901 in: C.H. Gray and V.H.T. James (Eds.). Hormones in Blood. Chronic toxicity/carcinogenicity studies of sulphamethazine 3rd Edn.. Academic Press, London. p. 576. in Fischer-343/N rats: Two-generation exposure. Food Chem. Robbins, J. (19811 Factors altering thyroid hormone metabolism. Toxicol.. 28. 157-167. Environ. Health Perspect.. 38. 65-70. Littlefield. N.A.. D.W. Gaylor. B.N. Blackwell and R.R. Allen Ron. E. and B. Modan (1982) Cancer by tissue of origin: Thyroid. t 1989) Chronic toxicity/carcinogenicity studies of sulphamet- in: D. Schottenfeld and Framaueri tEds.1, Cancer Epidemiol- hazine in B6C3Fl mice. Food Chem. Toxicol.. 27. 455-463. ogy and Prevention, W.B. Sauders Co. Philadelphia, pp. 837- Mackenzie. C.G. and J.B. Mackenzie (1943) Effect of sulfo- 854. namides and on the thyroid gland and basal Ruiz. M. and S.H. Ingbar (1982) Effect of erythrosine (2’.4’,5’,7- metabolism. Endocrinology. 32, 1855209. tetraiodofluoresceinel on the metabolism of thyroxine in rat Mackenzie. J.B.. C.G. Mackenzie and E.V. McCollom (1941) liver, Endocrinology. 110. 1613-1617. Effect of sulfanilguanidine on thyroid of the rat. Science. 94. Sanders. J.E.. D.A. Elgenberg, L.J. Bracht, W.R. Wang, and M.J. 518. van Zwieten (19881 Thyroid and liver trophic changes in rats McClain. R.M., A. Levin A., R.C. Posch and J.C. Downing secondary to liver microsomal enzyme induction by an experi- (1989) The effect of phenobarbital on the metabolism and mental leukotriene antagonist (L-649.923). Toxicol. Appl. excretion of thyroxine in rats, Toxicol. Appl. Pharmacol.. 99. Pharmacol., 95, 378. 2 16-228. Saxen, E.A. and L.O. Saxen (1954) Mortality from thyroid dis- McClain, R.M., R.C. Posch. T. Bosakowski and J.M. Armstrong eases in an endemic goitre area. Studies in Finland. Dot. Med. t 19881 Studies on the mode of action for thyroid gland tumor Geogrdph. Trop.. 6. 335-34 1. promotion in rats by phenobarbital. Toxicol. Appl. Pharmacol.. Smellie. J.M.. H.M. Bantcok and B.D. Thompson (1982) Co-tri- 94. 254-265. moxazole and the thyroid. Lancet, 2, 96. McClain. R.M. (1989) The significance of hepatic microsomal Smellie. J.M.. M.A. Preece and A.M. Paton (1983) Normal so- induction and altered thyroid function in rats: Implications for matic growth in children receiving low-dose prophylactic co- thyroid gland neoplasia. Toxicol. Pathol.. 17. 294-306. trimoxazole. Eur. J. Pediatr.. 130, 30 I-304. Morris. H.P. (1955) The experimental development and Sokal. J.E. t 1953) Occurrence of thyroid cancer. N. Eng. J. Med., metabolism of thyroid gland tumors. Adv. Cancer Res.. 3, 249. 393-397. 51-l 15. Swarm, R.L.. G.K.S. Roberts, A.C. Levy and L.R. Hines (1973) Observations on the thyroid gland in rats following the admin- Vanderlan. W.P. and R. Caplan (19.5-I) Observations on a relation- istration of sulfamethoxazole and trimethoprim. Toxicol. Appl. ship between total iodine content and the iodine-concentrating Pharmacol.. 24. 3.5 I-363. mechanism of the thyroid gland of the rat. Endocrinology. S4. Taknyama, S., K. Aihara, T. Onodera and T. Akimoto (1986) 4i7-47. Antithyroid effects of propylthiouracil and sulfamonomethox- Wegelin. C. (1928) Malignant disease of the thyroid gland and it!. ine in rats and monkeys, Toxicol. Appl. Pharmacol.. 82. relationship to goitre in man and animals. Cancer Res.. 3. 191-199. 291-3 13.