Axolotl Paedomorphosis: a Comparison of Juvenile, Metamorphic, and Paedomorphic Ambystoma Mexicanum Brain Gene Transcription

University of Kentucky UKnowledge

Theses and Dissertations--Biology Biology

2013

AXOLOTL PAEDOMORPHOSIS: A COMPARISON OF JUVENILE, METAMORPHIC, AND PAEDOMORPHIC AMBYSTOMA MEXICANUM BRAIN GENE TRANSCRIPTION

Carlena Johnson [email protected]

Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Recommended Citation Johnson, Carlena, "AXOLOTL PAEDOMORPHOSIS: A COMPARISON OF JUVENILE, METAMORPHIC, AND PAEDOMORPHIC AMBYSTOMA MEXICANUM BRAIN GENE TRANSCRIPTION" (2013). Theses and Dissertations--Biology. 13. https://uknowledge.uky.edu/biology_etds/13

This Master's Thesis is brought to you for free and open access by the Biology at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Biology by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT:

I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained and attached hereto needed written permission statements(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine).

I hereby grant to The University of Kentucky and its agents the non-exclusive license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known. I agree that the document mentioned above may be made available immediately for worldwide access unless a preapproved embargo applies.

I retain all other ownership rights to the copyright of my work. I also retain the right to use in future works (such as articles or books) all or part of my work. I understand that I am free to register the copyright to my work.

REVIEW, APPROVAL AND ACCEPTANCE

The document mentioned above has been reviewed and accepted by the student’s advisor, on behalf of the advisory committee, and by the Director of Graduate Studies (DGS), on behalf of the program; we verify that this is the final, approved version of the student’s dissertation including all changes required by the advisory committee. The undersigned agree to abide by the statements above.

Carlena Johnson, Student

Dr. S. Randal Voss, Major Professor

Dr. Brian Rymond, Director of Graduate Studies

AXOLOTL PAEDOMORPHOSIS: A COMPARISON OF JUVENILE, METAMORPHIC, AND PAEDOMORPHIC AMBYSTOMA MEXICANUM BRAIN GENE TRANSCRIPTION

______

THESIS ______

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in the College of Arts and Sciences at the University of Kentucky

Carlena K. Johnson

Lexington, Kentucky

Director: Dr. S. Randal Voss, Professor of Biology

Lexington, Kentucky

2013

ABSTRACT OF THESIS

AXOLOTL PAEDOMORPHOSIS: A COMPARISON OF JUVENILE, METAMORPHIC, AND PAEDOMORPHIC AMBYSTOMA MEXICANUM BRAIN GENE TRANSCRIPTION

Unlike many amphibians, the paedomorphic axolotl (Ambystoma mexicanum) rarely undergoes external morphological changes indicative of metamorphosis. However, internally, some axolotl tissues undergo cryptic metamorphic changes. A previous study examined interspecific patterns of larval brain gene expression and found that these species exhibited unique temporal expression patterns that were hypothesized to be morph specific. This thesis tested this hypothesis by examining differences in brain gene expression between juvenile (JUV), paedomorphic (PAED), and metamorphic (MET) axolotls. I identified 828 genes that were expressed differently between JUV, PAED, and MET. Expression estimates from JUV were compared to estimates from PAED and MET brains to identify genes that changed significantly during development. Genes that showed statistically equivalent expression changes across MET and PAED brains provide a glimpse at aging and maturation in an amphibian. The genes that showed statistically different expression estimates between metamorphic and paedomorphic brains provide new functional insights into the maintenance and regulation of paedomorphosis. For genes that were not commonly regulated due to aging, paedomorphs exhibited greater transcriptional similarity to juvenile than metamorphs did to juvenile. Overall, gene expression differences between metamorphic and paedomorphic development exhibit a mosaic pattern of expression as a function of aging and metamorphosis in axolotls.

KEYWORDS: Ambystoma; metamorphosis; paedomorphosis; microarray; gene expression

Carlena K. Johnson (July 1, 2013)

AXOLOTL PAEDOMORPHOSIS: A COMPARISON OF JUVENILE, METAMORPHIC AND PAEDOMORPHIC AMBYSTOMA MEXICANUM BRAIN GENE TRANSCRIPTION

Carlena K. Johnson

DR. S. RANDAL VOSS DIRECTOR OF THESIS

DR. DAVID WESTNEAT DIRECTOR OF GRADUATE STUDIES

(JULY 1, 2013)

ACKNOWLEDGEMENTS

I thank the many people who have helped me with this project. I expressly thank Randal

Voss for all of his time and patience in teaching me about genomics and the writing process. I am immensely grateful for your help and assistance during this project. I thank Antony Athippozhy for his assistance and guidance on how to manipulate and analyze microarray data. I also thank all the members of the Voss lab, especially, John

Walker, Kevin Kump, Sri Putta, and Tyler Frye for their support and encouragement. I thank my committee members, Jeramiah Smith and Vinnie Cassone for their support in this project. Thank you to the faculty and staff of the Biology Department. This project was supported through grants received by Randal Voss from NSF and NIH. I would like to thank my friends and family for their unwavering support through this project. Thank you to my mom and fiancé, Tom, for their encouragement and reassurance.

iii

TABLE OF CONTENTS Acknowledgements.…………………………………………………………………… iii List of Tables.………………………………………………………………………….. vi List of Figures.………………………………………………………………………… vii Chapter One: Literature Review.……………………………………………………….. 1 Why paedomorphosis and why salamanders?………………………………..… 1 Paedomorphosis and salamander life history variation…….…………………... 4 Hormonal basis of salamander metamorphosis………………………..……….. 6 Hormonal basis of paedomorphosis……………………………….……..…..… 8 Paedomorphosis in the Mexican Axolotl……………………….…….….…… 11 Transcriptional analysis of salamander metamorphosis and paedomorphosis.………………………………………………………….. 13 Genetic analysis of paedomorphosis………………………………….….….... 15 The link between met1 and thyroid hormone regulation of metamorphosis…………………………………………………………...… 17 What genes map to metamorphic timing QTL?………………………….…… 18 Synthesis: evolution of salamander paedomorphosis……………………..…... 19 Chapter Two: Microarray comparison of juvenile, paedomorphic and metamorphic brain gene transcription……………………………..…….... 28 Introduction ………………………………………………………………....… 28 Methods……………………………………………………………..…………. 29 Animals and tissue sampling…………………………………...…….... 29 RNA isolation and microarray analysis………………………...…….... 29 Global analyses of microarray data……………………………...…….. 30 Identification of differently expressed genes (DEGs) using microarray……………………………………………………….. 30 Identification of statistically enriched biological processes………....… 30 Bioinformatic validation of morph specific gene expression……...…... 31

Results ………………………………………………………………………….. 31 Global analysis of gene expression……………………………………... 31 Identification of differently expressed genes among juvenile and adult brains………………………………………….…..… 32 Classification of adult gene expression patterns relative to juvenile…… 32 Identification of statistically enriched biological processes………….… 33 Bioinformatic analysis of morph specific gene expression ……………………………………………………….... 35 Discussion………………………………………………………………………. 36 Aging and transcription in the A. mexicanum brain………………….…. 37 Genes associated with metamorphosis and paedomorphosis…………… 39 Aging, species, and morph specific differences in hemoglobin expression…………………………………………………. 43 References…………………………………………………………………………...…. 61

LIST OF TABLES Table 1.1, General characteristics of salamander life history………………………..… 22 Table 2.1, List of genes that show greater than three standard deviations from predicted values………….………………………..……………….…. 46 Table 2.2, List of over-represented ontology terms………………………….………… 47 Table 2.3, Pattern of gene expression comparison with DEGs from Page et al. (2010)……………………………………………………………….… 51

LIST OF FIGURES Figure 1.1, Representative salamander larva, adult metamorph (Ambystoma tigrinum) and adult paedomorph (A. mexicanum)……………………………………………...… 23 Figure 1.2, Timeline showing timing of fore and hindlimb development and the metamorphic period of laboratory reared A. t. tigrinum………………………. 24 Figure 1.3, Photo of Ambystoma dumerilii……………………………………………. 25 Figure 1.4, Gene expression profiles……………………...…………………………... 26 Figure 1.5, Model showing relationships among salamander life history variation, physiology, and habitat stability…………………………………...……….. 27 Figure 2.1, Box plots of GeneChips……………………….………………………...... 54 Figure 2.2, Principle components analysis………………….……………………….… 55 Figure 2.3, Hierarchical clustering analysis………………………………….………... 56 Figure 2.4, Venn diagram of significant DEGs…………………………..……………. 57 Figure 2.5, Patterns of gene expression change for PAED and MET relative to JUV..……………………………………………………………………….. 58 Figure 2.6, Bivariate plot of PAED and MET expression relative to JUV……….…… 59 Figure 2.7, Bar graph of expression for significant collagen genes…………………… 60

vii

CHAPTER ONE: LITERATURE REVIEW Salamanders are one of three primary groups of amphibians, the other two being caecilians and anurans. Ancestrally, all three groups trace their origins to ancestors that present biphasic life cycles with an aquatic larval phase and a more terrestrial adult phase (Duellman & Trueb, 1986). However, alternate modes of development subsequently evolved within all three groups. Interestingly, ancestral vestiges of metamorphosis are observed during early development of direct developing anurans and salamanders (Callery & Elinson, 2000; Kerney et al., 2012), suggesting shared evolutionary potential for radical and early shifts in the timing of metamorphosis. Radical, later shifts in metamorphic timing are only observed in salamanders. In the most extreme cases, metamorphosis has been abolished completely, yielding bizarre larval- form adults with completely aquatic life cycles (Gould, 1977; Shaffer & Voss, 1996). These unique paedomorphic forms are ideal for investigating mechanisms of thyroid hormone (TH) regulation that are associated with adaptive delays in metamorphic timing and the evolution of novel life histories.

What is paedomorphosis and why salamanders?

Paedomorphosis is a heterochronic term that describes a specific pattern of evolution, the retention of ancestral juvenile traits in the adult stage of a derived species (Gould, 1977). In this meaning, paedomorphosis references an evolutionary change in developmental timing between an ancestor and descendant species. However, paedomorphosis is often used in a more general sense to describe the retention of larval morphological traits in adult salamanders, irrespective of phylogeny. The evolution or expression of paedomorphosis is clearly associated with TH, the primary metamorphic hormone in anurans, salamanders, and some fish. In species that undergo a metamorphosis, TH induces the regression of larval traits and the development of traits typical of a more terrestrial adult (Figure 1.1). As discussed in more detail below, many paedomorphic salamanders can be induced to undergo partial or complete metamorphosis by simply placing them in a bath of TH. This suggests the possibility that paedomorphosis evolves “simply” by blocking the synthesis, secretion, or reception of TH in target cells (Page et al., 2010).

However, TH secretion is regulated by complex, neuroendocrine pathways that must first develop during the larval period before these systems become competent to signal TH release in response to intrinsic and extrinsic cues (Denver et al., 2002). Thus, paedomorphosis could conceivably evolve by altering the development or function of a number of different cells, tissues, and organs that regulate the release or reception of TH. Given this complexity and the broad pleiotropic role that TH assumes in amphibian development and physiology, it seems likely that paedomorph evolution requires multiple changes across many genes.

The specific heterochronic process most commonly associated with salamander paedomorphosis is neoteny (Gould, 1977) – relative to the ancestor or metamorphic condition, somatic development is delayed but rate or timing of reproductive maturation is the same. However, paedomorphosis in some species is associated with an earlier time to first reproduction, achieved by accelerating the rate of gonadal development (progenesis in Triturus alpestris; Denoel & Joly, 2000) or by initiating reproductive maturation earlier in the life cycle (peramorphosis in Ambystoma talpoideum; Ryan & Semlitsch, 1998). Thus, very different patterns of growth and development may be associated with the expression of paedomorphosis within and between species (Whiteman, 1994). Among species that occur in stable aquatic habitats, paedomorphosis is a fixed trait with little or no incidence of metamorphosis. Within species that use less permanent bodies of water for breeding, individuals may reproduce initially as paedomorphs but undergo metamorphosis in subsequent years (Denoel et al., 2007). The different patterns of growth and development observed among paedomorphs yield different adult morphologies, and this suggests different mechanistic bases for the evolution of paedomorphosis within and among salamander families. However, exactly how patterns of development and morphology map onto mechanisms that block TH signaling and metamorphosis, or affect the rate of gonadal development, is essentially unknown. Without such mechanistic information, it is not possible to distinguish among examples of paedomorphosis that evolve convergently among lineages, which clearly has occurred during salamander phylogenesis (Shaffer & Voss, 1996; Weins & Hoverman, 2008). Thus, all terms that describe the evolutionary origin or developmental expression of paedomorphic salamanders are potentially useful metaphors, but if they are

not based upon mechanistic insight and genetically based models, they are not likely to resolve outstanding questions concerning salamander life history variation and life cycle evolution.

It is interesting to consider why paedomorphosis is unique to salamanders and not other amphibians. In as much as paedomorphosis evolves as a prolongation of the larval period, it is notable that salamanders with biphasic life cycles typically have longer larval periods than caecilians or anurans (Duellman & Trueb, 1986). In biphasic salamander species, metamorphosis may occur within the same season or year that eggs are laid or larvae may overwinter in permanent habitats and metamorphose after one or more additional years of development (e.g., Beachy, 1995; Castanet et al., 1996; Voss et al., 2003). Although there are examples of anurans with multiyear larval periods, salamander larval periods are typically longer. So, there are fundamental differences between salamanders and other amphibians; salamanders have greater evolutionary potential to protract (and contract in cases of direct development) the length of the larval period. This greater evolutionary potential is associated with at least five aspects of salamander development and life history. First, salamanders exhibit slow rates of growth and development. Some anurans can complete metamorphosis in just a couple of weeks, while the fastest developing salamander larvae require more than a month, and typically several months (Duellman & Trueb, 1986). In general, salamander metamorphosis is developmentally less radical and more protracted than that of anurans. For example, salamander forelimb development does not coincide with other morphological metamorphic changes, as it does in anurans (Figure 1.2). Instead, it occurs shortly after hatching and before gills and tailfins are resorbed at metamorphosis. This suggests either a decoupling of forelimb development from metamorphosis or an incredibly gradual metamorphosis in salamanders that essentially spans the entirety of the larval period (Duellman & Trueb, 1986). Second, aspects of TH signaling may differ between salamanders and other amphibians. While it is clear that TH is required for salamander metamorphosis, Ambystoma mexicanum tolerates levels of TH that cause abnormal development and mortality in some anurans (Page et al., 2008). Also, several orthologous genes, including thyroid hormone receptor beta (thrb), do not show the same pattern of expression between TH-induced A. mexicanum and metamorphosing

anurans (Page et al., 2007, 2008). Third, salamanders often utilize low temperature habitats, either because they occur at relatively high elevations or latitudes (Bizer, 1978; Sexton & Bizer, 1978). Low temperature slows growth and development, depresses activity of neurons and glands that regulate metamorphosis, and reduces the responsiveness of tissues to TH (Uhlenhuth, 1919; Moriya, 1983a, 1983b). Fourth, salamander larvae present a highly similar body design to adults. Thus, unlike anuran tadpoles that must reorganize the body cavity at metamorphosis to allow space for gonads to develop, salamanders can potentially allocate resources to developing gonads and associated fat bodies during the larval period. Fifth, salamander larvae are carnivorous like adults, and thus unlike herbivorous anuran tadpoles, do not rely upon seasonal, primary productivity of ponds for food. Many anuran species have tadpoles that are highly specialized for rapid development, utilizing seasonally abundant resources in ephemeral ponds that select for finite larval periods (Wilbur, 1980). Multiyear larval periods are simply not possible within the context of anuran life history strategies that depend upon seasonally limiting resources. In contrast, salamander larvae present greater developmental flexibility to utilize more permanent aquatic habitats by extending the length of the larval period, indefinitely in the case of some paedomorphic species.

Paedomorphosis and salamander life history variation

Salamander life histories can be broadly classified into three categories: biphasic, paedomorphic, and facultative (Table 1.1). These are simply categories of convenience and do not adequately treat the continuum of life histories represented by each class, and the possibility of different genetic, developmental, and physiological bases within classes. Species with biphasic life histories almost invariably undergo a metamorphosis. This life history strategy is viewed as an adaptation to exploit transient opportunities for larval growth and is often found where aquatic habitats are ephemeral (Wilbur & Collins, 1973; Wilbur, 1980). At the opposite extreme are salamanders that almost invariably remain paedomorphic throughout their lives. Paedomorphic salamanders are associated with isolated and stable aquatic habitats, such as large closed-basin lakes, spring-fed lakes, caldera lakes, and river systems (Sprules, 1974; Shaffer, 1984). In facultative

species, metamorphs and paedomorphs are observed at varying frequencies within the same population. In such species, it is thought that the ability to express paedomorphosis and metamorphosis is advantageous when a landscape contains a mixture of ephemeral and stable aquatic habitats (Wilbur & Collins, 1973; Sexton & Bizer, 1978; Wilbur, 1980; Semlitsch et al., 1990; Whiteman, 1994; Denoel et al., 2005;). Facultative life histories are sometimes associated with aquatic habitats that may only be permanent for a single year of reproduction; the fitness advantage must be considerable because it is not uncommon to observe a paedomorphic brood in a cattle-watering tank on an arid landscape with no permanent ponds (Randal Voss, personal observation). When facultative species lay eggs in habitats that decline in quality during the larval period, metamorphosis is initiated in all or some proportion of individuals in the population (Semlitsch, 1987; Semlitsch et al., 1990).

Whereas metamorphic and facultative taxa are capable of dispersal among aquatic habitats, obligatorily paedomorphic taxa are confined to isolated bodies of water. This suggests that biphasic or facultative ancestors colonized aquatic habitats and ecological conditions (e.g., permanent aquatic habitat and/or arid terrestrial conditions) selected for paedomorphic life histories. The expression of paedomorphic life histories is known to affect population structure and may affect the probability of speciation (Shaffer, 1984; Shaffer & Breeden, 1989). The fitness benefits of paedomorphosis are an earlier time to first reproduction, more than one breeding event per year, larger clutch size, and a higher probability of mating success (Scott, 1993; Krenz & Server, 1995; Ryan & Semlitsch, 1998). It is important to note that paedomorphosis is not a pathology or default life history strategy – a failure to undergo metamorphosis. Paedomorphic species are highly adapted to their aquatic habitats. For example, Ambystoma dumerilii has secondarily evolved webbed-feet and extra gill filaments that develop at the time metamorphosis occurs in related metamorphic species (Figure 1.3). Thus, evolution of paedomorphosis may release tissues from the constraint of TH remodeling, allowing new functions to evolve that better adapt adults to aquatic habitats. The combination of selection and genetic drift acting on isolated paedomorphic lineages may explain rapid divergence of morphology and gene expression within some species groups (Shaffer & Voss, 1996; Page et al., 2010).

Paedomorphic life history strategies are widespread among salamanders, occurring in nine out of 10 families with a total of 57 species. Studies show that paedomorphic salamanders have evolved multiple times in different lineages (Shaffer & Voss, 1996; Denoel et al., 2005; Weins & Hoverman, 2008). Four salamander families (Amphiumidae, Sirenidae, Proteidae, and Cryptobranchidae) are comprised entirely of paedomorphic species, having no extant biphasic species. Some salamanders, such as species in the families Cryptobranchidae and Amphiumidae, are said to undergo an incomplete metamorphosis – meaning they undergo some anatomical changes, but not others, and remain aquatic throughout their lives. For example, adult Hellbenders (Cryptobranchus alleganiensis) lose their larval external gills, but do not develop eyelids and retain a single gill-slit that appears as a circular opening on the neck (Larson et al., 2003). The most species-rich salamander family, Plethodontidae, includes many biphasic and direct developing species, with the Tribe Hemidactyliini (Subfamily Plethodontinae) containing numerous paedomorphic species that are cave dwelling and exhibit reduced pigmentation and vision (Larson et al., 2003). Perhaps the best-studied salamander families with respect to life history variation are Salamandridae and Ambystomatidae. Studies of paedomorphic and facultative species within these families have better elucidated environmental and broad-sense genetic correlates of metamorph and paedomorph expression, and these species continue to provide exceptional models for investigating ecological aspects of life history evolution (Denoel et al., 2005; Doyle & Whitemann, 2008; Takahashi & Parris, 2008; Denoel et al., 2009; Whiteman et al., 2012).

Hormonal basis of salamander metamorphosis

Biphasic, paedomorphic, and facultative life histories begin the same way, with embryonic and larval development occurring in an aquatic habitat. The primary difference is when and if metamorphosis occurs. Although not as dramatic as metamorphosis in anurans, multiple tissues undergo hormonally regulated changes in cellular composition, biochemistry, and morphology during salamander metamorphosis (Dodd & Dodd, 1976). These changes include degeneration of external gills and tailfins; remodeling of skeletal elements and integument; shortening of the intestine;

modification of pigmentation; muscle degeneration and differentiation; and changes in nitrogen excretion, urogenital function, and oxygen transport (Duellman & Trueb, 1986). Also, new structures arise de novo, including eyelids and skin glands. All of these changes adapt aquatic larvae for survival and reproduction as adults in more terrestrial environments.

Although salamanders have a more protracted and gradual metamorphosis than anurans, TH is the primary metamorphic hormone in both amphibian groups. Serum TH in the form of tetraiodothyroine (T4) is low during embryonic and early larval development, but as larvae approach metamorphosis, T4 levels increase (Norman et al., 1987). This pattern of hormone secretion is not unlike that observed for other examples of hormonally regulated, postembryonic development (e.g., insect and fish metamorphosis; Gilbert et al., 1996; Power et al., 2001; Laudet, 2011). The rise in T4 during amphibian larval development is associated with maturation and stimulation of brain regions that regulate thyroid activity, in response to extrinsic and intrinsic cues. These brain regions include the hypothalamus and pituitary gland. Increasing levels of T4 during larval development stimulate corticotropin-releasing hormone (CRH) release from modified nerve endings within the median eminence. CRH travels through capillaries to the anterior pituitary where it stimulates thyrotropes to release thyrotropin (TSH) into the blood stream. TSH in turn stimulates the thyroid gland to release T4. The TSH-releasing factor in amphibians is CRH, which also stimulates the release of corticotropin (ACTH) and subsequently, glucocorticoids (Denver & Licht, 1989). The potent, synergistic action of corticoids and TH in promoting anuran and salamander metamorphosis clearly implicates the hypothalamic-pituitary- thyroid (HPT) and HP- interrenal (HPI) axes in the physiological regulation of metamorphosis (Dodd & Dodd, 1976; Buscaglia et al., 1985; Galton, 1991, 1992; Gancedo et al., 1992; Hayes et al., 1993; Kuhn et al., 2005). Exactly how T4 levels initially increase at appropriate times during amphibian development to trigger the onset of metamorphosis, and sustain T4 at high levels during metamorphosis, is not fully understood (Manzon & Denver, 2004).

Peripherally, the availability and activity of TH is precisely regulated within cells. In mammals, TH is transported into target cells via TH transporters, which actively

and preferentially import the hormone across the plasma membrane (Hennemann et al., 2001). TH transporters have been studied in X. laevis (Ritchie et al., 2003; Connors et al., 2010), but there have been no studies using salamanders. Deiodinases have received considerable attention because they directly activate or inactivate intracellular forms of TH in amphibians (Buscaglia et al., 1985; Galton, 1991; Kuhn et al., 2005). The low activity T4 form of TH that is delivered to cells is converted intracellularly by deiodinase-type 2 (DIO2) into highly active 3,5,3’-triiodothyronine (T3). A second deiodinase, DIO3, inactivates intracellular T3. Although T4 may participate in cellular signaling (Caria et al., 2008), T3 is the primary TH signaling molecule in anurans (Denver et al., 2002) and presumably this is also true for salamanders, acting through nuclear receptors (thra, thrb) that function as transcription factors. The higher activity of T3 is associated with a higher affinity for TH receptor (TR) binding. The binding of T3 activates transcriptional regulatory networks in cell-specific manners (Shi, 2000). TH is also known to cause cellular changes via nongenomic mechanisms, including membrane receptors, transporter molecules, and cytoplasmic receptors that elicit changes in cells through signaling pathways (Caria et al., 2008; Davis et al., 2008; Furuya et al., 2009; Cheng et al., 2010). The effects of these nongenomic mechanisms of TH action and their role in amphibian metamorphosis have yet to be studied in detail, however, Buchholz et al. (2004) used transgenic analysis to show that TR is sufficient to mediate TH signaling during metamorphosis in X. laevis. Regulation of deiodinases, T4/T3 concentrations, and developmental expression of TRs help explain how a single hormone can coordinate responses among different cell types, and more generally regulate the temporal sequence of remodeling events during amphibian metamorphosis (Brown & Cai, 2007).

Hormonal basis of paedomorphosis

Only a year after, Gudernatsch (1912) established TH as the anuran metamorphic hormone; TH was shown by Laufberger to induce metamorphosis in A. mexicanum (Mexican axolotl) (reviewed by Huxley & Hogben, 1922). Over the next few decades, the responsiveness of many different paedomorphic salamanders to TH was investigated. Studies showed that morphologically similar paedomorphic salamanders (with external gills or gill slits, no eyelids, larval pigmentation, etc.) responded differently to TH.

Members of the family Proteidae, such as the American mudpuppy (Necturus maculosus), are known to be refractory to TH, such that treating them with exogenous or injected TH does not induce morphological metamorphosis (Swingle, 1922). Larsen (1968) proposed that peripheral tissues of N. maculosus had lost sensitivity to TH, given that it exhibits a morphologically normal thyroid gland that had been tested for functionality via trans- plant experiments (Swingle, 1922; Charipper, 1929; Grant, 1930). This prompted investigations into the functionality of the animal’s deiodinase activity and TR expression. Galton (1985) found that deiodinase activity capable of converting T4 to T3 was present within N. maculosus, and Safi et al. (2006) showed that N. maculosus expresses functional TRs that bind to DNA and activate transcription in response to TH treatment, including transcriptional changes of classical TR target genes like stomelysin 3 (mmp11) and thrb (Safi et al., 2006; Vlaeminck-Guillem et al., 2006). Additionally, in situ hybridization has been used to show expression of thra and thrb in central nervous system, epithelia of the digestive tract, myocardium, and skeletal muscle (Vlaeminck-Guillem et al., 2006). These studies suggest that the paedomorphic phenotype exhibited by N. maculosus may be due to the loss of key TH response genes needed to induce morphological changes, even though juveniles experience a postembryonic period of high TH sensitivity. Apparently, individuals undergo natural transcriptional and biochemical changes during a “cryptic metamorphosis” that does not result in gross morphological alterations (Vlaeminck-Guillem et al., 2006; Laudet, 2011). Species of the families Cryptobranchidae and Amphiumidae are responsive to THs in some respects, but not others. They often reabsorb external gills and develop adult characteristics of some features, but do not develop eyelids and retain gill slits as mature adults, indicating that some metamorphic processes may have been decoupled from TH regulation (Larson et al., 2003). When species of these families are treated with high doses of TH, often the only result is skin shedding, a trait that is regulated by TH in adults of metamorphic species, though otherwise the individual continues to exhibit paedomorphic characteristics (Dent, 1968; Fox, 1984). It was originally thought that the Texas blind salamander (Eurycea rathbuni, formerly assigned to the genus Typhlomolge) was the only known vertebrate to lack a thyroid gland (Emerson, 1905). However, Gorbman (1957) later showed the presence of a thyroid gland in E. rathbuni and Dundee

(1957) showed that an individual of this species underwent some morphological changes (gill and tail reabsorption) in response to T4, but other larval features were resistant (histological sections of skin revealed only larval characteristics). However, related paedomorphic species (E. tynerensis and E. neotenes) complete metamorphosis after T4 treatment in less than 2 weeks (Kezer, 1952). Paedomorphic species of Ambystoma also complete metamorphosis when treated with T4 but show species-specific variation in metamorphic timing (Voss et al., 2012). Collectively, these studies establish that paedomorphosis is associated with TH regulation and tissue responsiveness, and these characteristics vary among closely and distantly related species.

Based on results of early TH induction experiments, it was proposed that paedomorphic species evolve via several mechanisms, including disruption of peripheral and central mechanisms of TH regulation (Kezer 1952; Dent, 1968). Species that were shown to be refractory to TH but had active thyroid glands were presumed to be deficient in peripheral mechanisms that receive and transduce the T4 signal, such as TRs. Species that responded to high levels of TH but did not apparently metamorphose in nature were assumed to be deficient in some aspect of central regulation within the hypothalamus or pituitary. Such a deficiency would normally prevent T4 release but could artificially be over-ridden in the lab by T4 treatment. It also seems likely that central regulatory mechanisms have been altered in the evolution of facultative species from biphasic ancestors. This is because the brain and pituitary regulate appropriate physiological and developmental responses in response to environmental cues. In the case of facultative paedomorphosis, the HPT axis is not sufficiently activated if cues favor a paedomorphic life history, or environmental conditions suppress hormone activity, or if larval development and maturation is delayed (Whiteman, 1994). Several environmental conditions that are associated with habitat quality affect the probability that an individual will express paedomorphosis, including low density of conspecifics (Harris, 1987; Semlitsch, 1987), food availability (Semlitsch, 1987; Denoel & Poncin, 2001; Ryan & Semlitsch, 2003), pond permanence (Semlitsch, 1987; Semlitsch et al., 1990), and low temperature (Snyder, 1956; Sprules, 1974; Eagleson & McKeown, 1980).

Paedomorphosis in the Mexican Axolotl

The exemplar of all paedomorphic salamander is the Mexican axolotl (A. mexicanum). The Mexican axolotl is native to the Xochimilco lake system, near present- day Mexico City. Whereas historical populations were sufficiently large to provide an important food source to native people for thousands of years (Smith, 1989), the current A. mexicanum population at Xochimilco is on the verge of extinction (Graue, 1998; Zambrano et al., 2007). Smith (1989) provides an enjoyable historical account of the original collection of A. mexicanum from Xochimilco in 1863. Six A. mexicanum were transferred to Paris where Dumeril (1870) subsequently reported that individuals reproduced in an aquatic, larval state and that some of the resulting offspring underwent a metamorphosis. Smith (1989) suggests that some of the original stock that arrived in Paris included closely related members of the Tiger salamander species complex, that are capable of expressing a paedomorphic or metamorphic life history. However, it seems likely that domestication altered the penetrance for expressing paedomorphosis and the original axolotl stock was pure A. mexicanum that maintained a higher propensity to express metamorphosis in nature (Voss & Shaffer, 2000). Even though metamorphic forms have been culled from the Ambystoma Genetic Stock Center axolotl collection for decades, the frequency of spontaneous metamorphosis is only 1–2%, with a 10% frequency observed if A. mexicanum experience stressful conditions (Randal Voss, unpublished data).

The simplest explanation for paedomorphosis in A. mexicanum is a faulty thyroid gland. However, the thyroid gland is functional; it can synthesize T4 and release T4 after TSH stimulation (Prahlad, 1968; Taurog et al., 1974). Numerous studies, including our own (Page et al., 2007, 2008), have demonstrated that individuals can be stimulated to undergo morphological metamorphosis by administering T4. This suggests that intracellular receptors and deiodinase enzymes are functional. Putative T3 receptors have been identified in red blood cells (Galton, 1991) and TR function has been shown in vitro using mammalian cells (Safi et al., 2004). Also, the activity of deiodinases has been demonstrated (Galton, 1991; Darras et al., 2002). Blount (1950) showed that metamorphosis is inhibited if the pituitary of A. mexicanum is transplanted into the

metamorph A. tigrinum, while the reciprocal transplantation experiment resulted in metamorphosis. Tassava (1969) pointed out that these grafts were probably not precise enough to rule out a hypothalamic contribution. TSH is present in the A. mexicanum pituitary (Taurog et al., 1974) at a quantity sufficient for inducing metamorphosis (Darras & Kuhn, 1983). Thus, the pituitary is capable of producing sufficient functional TSH for metamorphosis, but for some reason, it is not released. There are at least four explanations for this observation in A. mexicanum: (1) The pituitary does not receive appropriate stimulation from the hypothalamus, (2) the pituitary is insensitive to hypothalamic stimulation, (3) the pituitary is defective in releasing TSH, and (4) the pituitary is functional but stimulated to release TSH at the wrong time during ontogeny (Rosenkilde & Ussing, 1996). Galton (1992) suggested that T4 levels might never eclipse a threshold during development to stimulate brain regions to increase activity of the thyroid gland. Kuhn et al. (2005) similarly proposed that brain T3 availability is the limiting factor for metamorphic onset. They found that DIO2 activity in the brain, which is required to generate T3, requires appropriate doses of corticoids and T4. Thus, stimulation of both the HPA and HPI axes is required for metamorphosis, and in fact, it is possible to induce metamorphosis in A. mexicanum if sub-metamorphic levels of T4 and corticoids are administered together. Because injection of amniote CRH stimulates corticotropin but not thyrotropin release, Kuhn et al. (2005) implicated the thyrotroph CRH receptor as a likely candidate gene for paedomorphosis in A. mexicanum. However, researchers have demonstrated T4 increases after stimulating hypothalamic regions in neotenic tiger salamanders (A. tigrinum) and A. mexicanum; this could seemingly only occur if downstream thyrotrophin signaling pathways are functional (Norris & Gern, 1976; Rosenkilde & Ussing, 1996).

It is also possible that a recently discovered TSH signaling pathway may be implicated in salamander paedomorphosis. Studies of neuroendocrine regulation of amniote reproductive cycles showed that TSH from the pars tuberalis (PT) of the anterior pituitary gland is released to signal DIO2 transcription in the hypothalamus (Hanon et al., 2008, Nakao et al., 2008). This increases local T3 concentrations and thus stimulates canonical TH signaling pathways. Surprisingly, these studies show that the flow of information may be bidirectional between the pituitary and hypothalamus. If a similar PT

signaling pathway is necessary for metamorphosis, this pathway would not be activated in A. mexicanum because of a block in TSH release. This might also explain why A. mexicanum does not show strong seasonality in breeding within lab populations and can breed multiple times per year. Whether paedomorphosis involves a PT signaling mechanism or a classical TSH signaling mechanism, the metamorphic block in A. mexicanum is likely associated with hypothalamus and pituitary function (Prahlad & DeLanney, 1965; Norris et al., 1973; Rosenkilde & Ussing, 1996; Kuhn et al., 2005; Laudet, 2011).

Transcriptional analysis of salamander metamorphosis and paedomorphosis

Most mechanistic studies of metamorphosis and paedomorphosis have focused on physiology and to a lesser extent biochemistry. Recently, microarrays developed for A. mexicanum have been used to characterize global transcriptional responses of metamorphic and paedomorphic salamanders. These studies have focused on the skin and brain, the former because it is a peripheral target tissue of TH and much is known about anatomical changes that occur during metamorphosis (Fahrmann, 1971a, 1971b, 1971c; Fox, 1983; Page et al., 2009). The brain has also been targeted because it is where HPT and HPI axes are centrally regulated. Reviewed below are three important findings from these studies.

T4 precisely activates metamorphic transcriptional programs in a concentration- dependent manner

Page et al. (2008) varied T4 concentration by an order of magnitude and examined transcriptional patterns in A. mexicanum skin after 0, 2, 12, and 28 days of T4 treatment. Individuals exposed to 50 nM T4 initiated metamorphosis approximately 1 week prior to those in 5 nM, however, the sequence of transcriptional and morphological changes during metamorphosis, and the length of the metamorphic period did not differ (Figure 1.4A). Over 1000 genes were identified as differently expressed in both treatments, showing the same directional trends, but the patterns were temporally shifted by T4 concentration. The study showed that the timing of metamorphic onset is T4 concentration dependent, and surprisingly, high T4 concentration did not disassociate

subsequent transcriptional programs. It remains to be determined if the T4 concentration effect in A. mexicanum is manifested at the level of target cells, the HPT axis, or possibly both (Rosenkilde & Ussing, 1996).

Patterns of gene expression are complex, reflecting tissue remodeling, changes in cellular metabolism, and loss and gain of larval and adult cell types

One of the advantages in working with A. mexicanum is that they present a natural, hypothyroid condition throughout life (Huggins et al., 2012). Thus, transcriptional programs can be activated precisely at juvenile or adult stages of the life cycle. When T4 is administered, thousands of genes show significant changes in transcript abundance. In a very general sense, these genes show four temporal gene expression patterns: they increase or decrease in abundance, or they transiently increase or decrease during metamorphosis (Figure 1.4B). Linear, quadratic, and piece-wise regression models have been used to identify groups of genes that similarly increase or decrease in abundance as a function of days of T4 treatment. The earliest gene expression changes precede tissue-remodeling events and are associated with cells that are specific to larvae or cells that differentiate to form adult tissues. For example, after 2 days of T4 treatment, genes that are specifically expressed in apical and Leydig cells of the larval skin show decreases in transcript abundances (Page et al., 2007, 2008, 2009). At this time, apical and Leydig cell numbers are normal for larval skin and there is no histological evidence of remodeling. However, after 12 days of T4 treatment, apical cells are no longer present in histological sections and Leydig cells are significantly reduced in number (Page et al., 2009); presumably, both cell types undergo TH-induced, programmed cell death, as has been demonstrated for apical cells in metamorphosing anurans (Schreiber & Brown, 2002). As larval cells and their gene expression products decrease in abundance during metamorphosis, TH induces the proliferation and differentiation of adult progenitor cells. As a result, transcripts associated with adult cells increase in abundance. The most dramatic changes are seen for basal keratinocytes that proliferate and differentiate during metamorphosis and transcribe adult specific keratins that generate a more cornified epithelial layer to limit desiccation during the adult phase. Genes that transiently increase in abundance code for matrix-remodeling proteins (e.g.,

MMPs) and cellular processes like metabolism, transcription, and translation, while genes that decrease in abundance include negative regulators of cellular processes (Page et al., 2008). These gene expression patterns illuminate tissue remodeling and macromolecular biosynthetic processes that occur during metamorphic climax.

Although some genes are expressed similarly between metamorphic and paedomorphic individuals during larval development, the majority of the differentially expressed genes are unique

For example, brain transcription has been compared between A. mexicanum and A. t. tigrinum individuals during larval development and the period spanning the metamorphic period of the later species (Page et al., 2010). As A. t. tigrinum individuals near the onset of anatomical metamorphosis, more gene expression differences are observed between larvae of these species. Presumably, many of these expression differences trace to increasing levels of glucocorticoids and THs in A. t. tigrinum individuals as they initiate early metamorphic changes. In support of this idea, a glucocorticoid receptor (nr3c2) showed significantly higher expression in A. t. tigrinum. Also, some genes that were more highly expressed during natural metamorphosis in A. t. tigrinum were also similarly upregulated in A. mexicanum brains when individuals were induced to metamorphose with T4 (Huggins et al., 2012). However, hundreds of genes were expressed more highly in A. t. tigrinum than A. mexicanum throughout the larval period. At this point, it is not clear if all of these differences are associated with metamorphic and paedomorphic life histories; however, the observed systematic bias in transcription suggests more than species-specific divergence of gene expression.

Genetic analysis of paedomorphosis

The idea that paedomorphic salamanders simply need a dose of TH to induce the second half of the ancestral ontogeny influenced the way paedomorphic life histories were thought to originate and evolve. For example, Goldschmidt (1940) cited A. mexicanum as the “hopeful monster”, an example of evolution waiting around for the right macromutation – simply block a single physiological step in TH regulation and instantly a novel form is originated. The idea that paedomorphosis results from a simple

mechanistic change gained further support when genetic studies showed that metamorphic and paedomorphic forms could be segregated according to Mendelian expectations. In particular, Humphrey (1967) showed that when A. mexicanum/A. t. tigrinum hybrids were backcrossed to A. mexicanum, there is 1:1 segregation of phenotypes. Later, Tompkins (1978) and Gould (1981) argued from Humphrey’s single cross that A. mexicanum was an example of single gene, macromutational evolution. To further test the single gene hypothesis, Voss (1995) crossed domesticated A. mexicanum and A. t. tigrinum to create F1 hybrids and like Humphrey (1967), backcrossed these to A. mexicanum. The ratios of metamorphs and paedomorphs arising from two relatively large backcrosses were consistent with single locus control, thus supporting the classical idea of a single mutation underlying the evolution of paeodomorphosis. In these crosses, the A. t. tirginum and A. mexicanum alleles were dominant and recessive in effect, respectively. Subsequently, DNA was isolated from individuals of these crosses, amplified fragment length polymorphisms (AFPLs) were typed, and a genetic linkage map was constructed. This map was used to roughly locate the position of the major gene (met1) for paedomorphosis (Voss & Shaffer, 1996). To determine if met1 arose independently in the domesticated stock, the back-cross design was repeated using A. mexicanum collected from Lake Xochimilco (Voss & Shaffer, 2000). The segregation of phenotypes did not fit a single gene model, and the segregation of AFLPs marking the met1 region exhibited segregation distortion, suggesting an epistatic effect between met1 and other loci on viability. Still, met1 genotypes did associate with metamorphic and paedomorphic phenotypes, indicating a difference in genetic background between wild- caught A. mexicanum and the domesticated stock.

Voss and Smith (2005) made additional A. mexicanum/A. t. tigrinum backcrosses and reared over 500 offspring to rigorously quantify the genetic effect of met1 from wild-caught A. mexicanum. These crosses showed that met1 not only explained discrete variation in the expression of metamorphosis and paedomorphosis but also continuous variation in metamorphic timing. Approximately 20% of the individuals that inherited two met1 alleles from wild A. mexicanum were paedomorphic. The remainder of these homozygotes delayed metamorphic timing by an average of 36 days, relative to heterozygotes. These results showed that met1 determines the expression of

metamorphosis and paedomorphosis by altering metamorphic timing. Within the context of hybrid crosses, alleles from metamorphic and paedomorphic species decrease and delay the time to metamorphosis, respectively. Variation in the number of paedomorphic individuals generated from wild-caught versus domesticated A. mexicanum crosses reflect differences in genetic background. Here, genetic background is defined as the summative effect of loci whose independent effects are too small to identify and quantify statistically. Within the genetic background of the domesticated stock, which has undergone strong artificial selection for paedomorphosis, the penetrance of met1 is greater, and this results in a higher proportion of paedomorphic individuals in hybrid crosses (Voss & Smith, 2005).

The link between met1 and thyroid hormone regulation of metamorphosis

The link between met1 and TH was recently established by combining quantitative trait locus (QTL) analysis with TH induction. Recognizing that paedomorphic species show variation in time to complete metamorphosis when treated with TH, Voss et al. (2012) crossed two paedomorphic species (A. mexicanum and A. andersoni) and then performed a backcross to segregate TH response alleles among second generation offspring. At 120 days postfertilization (dpf), backcross offspring were treated with 2.5 nM T4 and then scored for time to complete metamorphosis. Essentially, all offspring initiated and completed metamorphosis over an ~ 160 day interval of time. A QTL screen identified the genomic position of met1 and two additional QTL (met2, met3), each explaining approximately 10% of the variation in metamorphic timing; the effects of all alleles were additive with A. andersoni alleles decreasing the time to metamorphosis. On average, individuals that inherited met1-3 alleles from A. andersoni metamorphosed earlier, just as did hybrid A. mexicanum that inherited met1 alleles from metamorphic A. t. tigrinum (Voss & Smith, 2005). The decrease in metamorphic timing was approximately 19 days for A. andersoni met1 versus 36 days for A. t. tigrinum met1. These results demonstrate that variation in metamorphic timing is determined in part by alleles that segregate at TH-responsive QTL, and alleles from paedomorphic species may vary in effect.

In natural populations of amphibians, variation in metamorphic timing can affect the values of other life history traits. Life history theories predict that individuals should metamorphose early if conditions for larval growth are poor, but delay metamorphosis and attain large larval body sizes if growth conditions are optimal (Gould, 1977; Wilbur & Collins, 1973). This is because larger body sizes at metamorphosis translate into increased reproductive success in the adult phase (Semlitsch et al., 1988). From a physiological perspective, metamorphic timing and body size are expected to covary as a function of resource allocation. From an evolutionary perspective, metamorphic timing and body size are expected to covary and coevolve if they share a common genetic basis. This prediction holds in the case of met2. In the A. andersoni/A. mexicanum backcross study described above, met2 not only explained significant variation in metamorphic timing but it also explained significant variation in adult body size (Voss et al., 2012). Individuals that delayed metamorphosis significantly increased body length at 300 dpf and weight at 400 dpf, and the effect was approximately equivalent for sexually dimorphic males and females. Overall, the genetic results reviewed here show that TH- responsive QTL account for variation in metamorphic timing, expression of metamorphosis versus paedomorphosis, and adult fitness traits.

What genes map to metamorphic timing QTL?

The A. mexicanum genome is in early stages of genome sequencing and thus comparative mapping has been used to more finely resolve the positions of met1-3 and identify candidate genes. The most promising candidate that has been identified is pou1f1, a transcription factor associated with combined pituitary hormone deficiency (CPHD) in humans. Pou1f1 is predicted to map to the position of met3 on linkage group (LG) 7. CPHD is associated with deficiencies in the secretion of pituitary hormones that regulate growth and development, but not hormones associated with reproductive physiology. Thus, pou1f1 is a candidate gene for evolutionary developmental changes that are independent of reproductive maturation (neoteny), as is seen in the example of salamander paedomorphosis.

In contrast to met3, identifying candidates for met1 and met2 has proven more difficult. For example, the 2 cm interval defining the location of met1 on LG2 marks a

unique synteny disruption in Ambystoma and perhaps other salamanders (Voss et al., 2011). This is a bit surprising and unlucky because the Ambystoma genome has undergone relatively few chromosomal rearrangements relative to other vertebrates and especially amniotes (Smith & Voss, 2006). Genes (e.g., rai1, shmt1, drg2, med9) from the Smith–Magenis disease syndrome region in the human genome flank one side of met1, while several genes associated with neural development and function (e.g., setd2, ngfr, ccm2) flank the other side. In no other vertebrate genome are these groups of flanking markers observed in synteny. Recently, microarray analysis was performed to identify genes that are expressed differently during larval development as a function of met1 genotype (Robert Page and Randal Voss, unpublished data). It was reasoned that this combined genetic-genomic approach would identify differentially expressed genes that map to the position of met1, potentially identifying candidate genes. As predicted, the majority of differentially expressed genes from LG2 mapped within 10 cm of met1, and six of the 14 genes that located within 2 cm of met1 were differentially expressed as a function of genotype during larval development and/or metamorphosis. At this point, it is not clear if the met1 genomic region is more enriched for differentially expressed genes than other regions of the Ambystoma genome, and perhaps genes that are TH responsive and associated with brain development. Regardless, it is clear from microarray studies that many genes are differentially expressed between paedomorphic A. mexicanum and metamorphic A. t. tigrinum, and met1 locates to a uniquely structured genomic region with several candidate genes for paedomorphosis.

Synthesis: evolution of salamander paedomorphosis

Many times during salamander evolution, mechanisms that regulate TH induction of metamorphosis were altered, resulting in paedomorphic salamanders that retain larval traits into the adult, breeding phase. Over the past 100 years, researchers have treated paedomorphic salamanders with TH to investigate the physiological and evolutionary basis of this unique salamander trait. Studies show that tissues within some species are entirely resistant to TH, while others show partial or complete metamorphosis. These varied responses are not explained by phylogeny or morphology. Salamanders are a morphologically conservative group and paedomorphosis has evolved independently

during salamander evolution. It is not likely that investigations of convergent morphological differences among deeply divergent paedomorphic species will yield a synthetic model for the origin of salamander paedomorphosis. Instead, an understanding of the origins of paedomorphosis requires knowledge of physiological and genetic mechanism, environmental context, and life history traits associated with fitness (Figure 1.5). In the case of ambystomatid salamanders, paedomorph expression is associated with TH-responsive loci with pleiotropic affects on metamorphic timing and adult fitness traits. According to our model, paedomorphosis and metamorphosis are mechanistically connected by loci that segregate allelic variation for responsiveness to T4. Phenotypes arising from hybrid ambystomatid crosses clearly show that it is possible to artificially phenocopy the complete spectrum of metamorphic timing variation observed among species in nature. Given variation in habitat stability, selection should favor low TH response alleles in biphasic populations when habitats allow for extensions of the larval period, as these alleles will delay metamorphic timing and increase adult body size. Under directional selection for mutations that increase the length of the larval period, central mechanisms of TH-release become progressively desensitized to intrinsic and extrinsic signals that provide reliable cues for initiating metamorphosis within the context of a biphasic life history. It is hypothesized that central mechanisms are desensitized because TH-response loci become fixed for alleles that are less responsive to positive TH feedback. As genetic background changes, the likelihood of paedomorph expression increases. In support of this model, studies of facultative paedomorphic populations (A. talpoideum) have demonstrated the efficacy of selecting for higher paedomorph expression in relatively few generations (Semlitsch & Wilbur, 1989). If directional selection occurs over many generations and within the context of large and stable aquatic habitats, as has happened independently in the case of Tiger salamanders species in Mexico (Shaffer, 1984; Shaffer & McKnight, 1996; Shaffer & Voss, 1996), further changes in genetic background are predicted to yield individuals that are no longer capable of initiating metamorphosis in nature.

Many studies have shown that environmental factors, and especially low temperature, are often associated with paedomorphic populations (Snyder, 1956; Sprules, 1974; Eagleson & McKeown, 1980). It is possible that some examples of

paedomorphosis may be pathological, reflecting extreme physiological conditions that may be exacerbated by inbreeding depression within small, isolated populations. However, Matsuda (1982; 1987) suggested that low temperature is pivotal in the adaptive evolution of paedomorphic salamanders; low temperature initially causes the expression of paedomorphosis in a population, and then through genetic assimilation (Waddington, 1953) the paedomorphic phenotype becomes fixed. The model proposed above is consistent with genetic assimilation because it also assumes directional selection of TH-response alleles that increase an individual’s likelihood of expressing a paedomorphic phenotype. The selection process and allelic effects depend upon environmental context.

The experimental approaches and evolutionary model outlined above provide a possible framework from which to initiate studies of the genetic basis of metamorphic timing in natural amphibian populations. At this point, it is not clear how many TH- response genes contribute to adaptive differences in metamorphic timing within natural populations and among species, or whether the same loci and physiological mechanisms are implicated in independent examples of paedomorph evolution. Also, much remains to be learned about facultative paedomorphosis at a mechanistic level. For example, does facultative expression of paedomorphosis mechanistically present an intermediated condition between metamorphosis and paedomorphosis, and how is it regulated? It is exciting to think that these and other questions can now be resolved using genetic and genomic methods that are applicable to nongenetic model organisms. The future looks bright for investigating evolutionary, developmental, and physiological aspects of salamander life history and life cycle evolution.

Table 1.1: General characteristics of salamander life history. This table uses common salamander life history terminology and is meant to provide an overview of common characteristics within these categories. Metamorphic Paedomorphic Facultative

• Habitat often • Aquatic habitat is • Habitat often contains a contains ephemeral often stable (6) mixture of temporary ponds (1) • Rarely/never and stable aquatic • Almost invariably metamorphose in features (3) metamorphose in nature (6) • Populations of the same nature (1) • Reach sexual species often vary in the • Generally reach maturity while proportion of individual sexual maturity after remaining in larval that metamorphose and metamorphosis (2,3) aquatic habitat (2,7) the time to • Thyroid hormone • May or may not metamorphosis (3) causes respond to thyroid • Timing of sexual morphological hormone with maturity may vary changes (e.g. gill morphological signs between metamorphic reduction, larval to of metamorphosis and paedomorphic adult skin changes, (4) individuals of the same tail fin reduction, • In rare cases, population (but not etc.) (4) stressful always) (10,11,12) • Species may environmental • Hormonal regulation has respond to stressful factors may cause not been studied in great 3 environmental some species to detail . Species are factors with metamorphose (for thyroid hormone precocious example, A. responsive (13,14,15) metamorphosis (5) mexicanum) (8,9) • Environmental factors affect the probability of an individual to express metamorphosis or paedomorphosis (6,16,17) (1) Wilbur, 1980 (2) Duellman and Trueb, 1986 (3) Denoël et al., 2005 (4) Bruce, 2003 (5) Wilbur and Collins, 1973 (6) Sprules, 1974 (7) Armstrong et al., 1989 (8) Smith, 1969 (9) Voss, unpublished observations (10) Scott, 1993 (11) Ryan and Semlitsch, 1998 (12) Denoël and Joly 2000 (13) Lynn and Wachowski, 1951 (14) Snyder, 1956 (15) Voss et al., 2012, (16) Eagleson and McKeown, 1980 (17) Semlitsch, 1987

Figure 1.1: A representative salamander larva, adult metamorph (Ambystoma tigrinum), and adult paedomorph (A. mexicanum). Critical levels of thyroid hormone induce metamorphosis in salamanders. Environmental and genetic factors may cause thyroid hormone levels to be low, resulting in paedomorphic salamanders. The pictures were taken by Jeramiah Smith.

Figure 1.2: Timeline showing timing of fore and hindlimb development and the metamorphic period of laboratory reared A. t. tigrinum. The dashed line indicates variation among individuals in presenting early signatures of anatomical metamorphosis (bulging eyes, reduced tail fins). In the anuran X. laevis, hindlimb and forelimb development initiates 2 weeks after hatching. Hindlimbs complete development prior to metamorphosis, while forelimbs emerge during metamorphosis.

Figure 1.3: Ambystoma dumerilii is a paedomorph that has secondarily evolved webbed- feet and extra filaments that develop on the dorsal surfaces of each gill. The bottom panel is a magnified image of the dorsal gill surface. The arrow indicates dorsal filaments (<1.0 mm in length) that are sprouting on the gill surface. These extra fibers form late in the larval period, at the time metamorphosis normally occurs in related metamorphic species. The A. dumerilii picture was taken by Brad Shaffer.

Figure 1.4: (A) Representative expression profile for a gene that shows a linear increase in mRNA abundance after TH induction of A. mexicanum using 5 or 50 nM T4. (B) Four general types of gene expression response observed after TH induction of metamorphosis in A. mexicanum. See text for details.

Figure 1.5: Model showing relationships among salamander life history variation, physiology, and habitat stability. Paedomorphosis is associated with low thyroid hormone responsiveness, permanent aquatic habitats, delayed metamorphic timing, and large body size. Metamorphosis is associated with contrary values for these variables.

CHAPTER TWO: MICROARRAY COMPARISON OF JUVENILE, PAEDOMORPHIC, AND METAMORPHIC BRAIN GENE TRANSCRIPTION

INTRODUCTION

Mexican axolotls (Ambystoma mexicanum) have long intrigued scientists because they present a paedomorphic mode of development (Gould, 1977). Unlike closely related species with aquatic and terrestrial phases of development, axolotls do not initiate metamorphosis and as a result, retain larval characteristics throughout life (Johnson and Voss, 2013). Thus, in comparison to species that undergo metamorphosis, axolotls seemingly do not mature several of their external anatomical features, including tail fins, gills, eyes, and epidermis. However, internally, gonads and brain mature similarly to metamorphic species. For example, axolotls and terrestrial A. tigrinum share many temporal patterns of brain gene expression at the time A. tigrinum larvae initiate metamorphosis (Page et al., 2010). Conservation of gene expression among metamorphic and paedomorphic salamanders suggests axolotls retain cryptic maturational changes of the metamorphic ancestor. Additionally, Page et al. (2010) showed that axolotls also present unique, temporal gene expression patterns that differ from A. tigrinum; this suggests that these differences may correlate with paedomorphic development. However, the above speculations are tenuous because an interspecific comparative approach cannot differentiate paedomorphic and metamorphic characteristics from characteristics that arise independently among species during evolution. In this study, I used an intraspecific comparative approach to identify gene expression changes that correlate with metamorphic and paedomorphic modes of development. Although axolotl almost always exhibit paedomorphosis, some individuals spontaneously undergo metamorphosis, the frequency of which is < 1-2% in the Ambystoma Genetic Stock Center (AGSC) (Johnson and Voss, 2013). I obtained juvenile, paedomorphic, and metamorphic A. mexicanum brains, isolated RNA, and performed microarray analysis. Using estimates of gene expression, I performed statistical contrasts to accomplish two objectives: (1) To determine how transcription differs between juvenile and adult brains, and (2) To identify genes and associated ontologies that change significantly as a function of paedomorphic and metamorphic development.

METHODS Animals and tissue sampling

Three adult male axolotls (four years old) that had spontaneously metamorphosed and three paedomorphic axolotls of the same age (two males and one female) were obtained from the Ambystoma Genetic Stock Center (AGSC) at the University of Kentucky. Additionally, three juvenile axolotl siblings were obtained from the AGSC and reared under the same laboratory conditions until 130 days post hatching (dph). The juveniles were immature with respect to gonad maturation, so it was not possible to determine genetic sex. All individuals were anesthetized in 0.02% benzocaine, and the brains and pituitaries were removed, flash frozen in liquid nitrogen, and stored individually at -80 C until RNA isolation was performed. Samples were divided into three groups: juvenile (JUV), adult paedomorph (PAED), and adult metamorph (MET). Animal care and use was carried out under University of Kentucky IACUC protocol #1087L2006.

RNA isolation and microarray analysis

Total RNA was isolated independently from whole brain, including pituitary, of each of the three individuals in the JUV, PAED, and MET groups using TRIzol (Invitrogen, Carlsbad, CA). Samples were further purified using Qiagen RNeasy mini- columns following manufacture’s guidelines. All RNA samples were quantified using a NanoDrop ND-1000 and an Agilent BioAnalyzer. RNA expression profiling was conducted using custom Amby_002 microarrays (Huggins et al. 2012). The nine isolated RNA samples were labeled and hybridized to independent microarray GeneChips and scanned by the University of Kentucky Microarray Core Facility according to standard Affymetrix protocols. A total of nine GeneChips were used in the experiment – one chip per individual, three individuals per group, and three groups (JUV, MET, PAED). Background correction, normalization, and expression summaries were accomplished using the robust multi-array average (RMA) algorithm (Irizarry et al. 2003)

Global analyses of microarray data

The distribution of raw hybridization intensities for each array was evaluated using box plots to identify any abnormal arrays. Microarray chips were then processed using RMA and grouped using principal components and hierarchical cluster analysis in JMP genomics version 5.0 (SAS Institute, Inc.) For the principal components analysis, arrays were mean centered and scaled to a variance of one. For the cluster analysis, “Fast Ward’s” method was used (Ward, 1963).

Identification of differently expressed genes (DEGs) using microarray

Only probesets that measured signal intensities > 6.25 (the average value of the midpoint of the saddle region from a signal intensity histogram for each individual array) (Buechel et al., 2011) on at least two chips per group, for at least one group, were retained (N = 12,785). The retained probesets were log2 transformed and analyzed using a one-way ANOVA with three groups. A fold-change cutoff of 1.5 and an FDR of q = 0.05 (the p-value, corresponding to a q-value of 0.05, is 0.009) were used to create a list of differently expressed genes (DEGs) (N = 828). Then, t-tests were applied to the list of DEGs, and a protected Fisher’s least significant difference (LSD) was used to detect significantly different changes in transcript abundance among the three pairwise contrasts: JUV and PAED, JUV and MET, and PAED and MET. DEGs identified from t- tests were classified into groups to investigate how gene expression changed between juvenile and adult brains. MET and PAED expression values were evaluated relative to JUV because it is likely that metamorphic and paedomorphic axolotls exhibit similar patterns of gene expression as juveniles, but diverge after paedomorphic adults undergo spontaneous metamorphosis. Operationally, this grouped genes according to statistical significance and direction of change.

Identification of statistically enriched biological processes

To identify biological processes that were statistically enriched in the lists of DEGs, overrepresentation analyses were conducted using PANTHER (Thomas et al. 2003). To generate a reference list, probesets were filtered to identify non-redundant human NCBI gene symbols. Of the 12,785 probesets that exceeded the low expression

cutoff, 10,054 probesets annotated to human NCBI gene symbols; the list was further reduced by removing redundant human NCBI gene symbols (N = 8,088). DEGs were analyzed according to statistical significance and direction of change (see above). A Bonferroni correction of α = 0.05 was used to identify overrepresented ontology terms.

Bioinformatic analysis of morph specific gene expression

DEGs identified from the JUV vs. MET and JUV vs. PAED contrasts were compared to genes identified by Page et al. (2010). In that study, genes were identified as commonly and differently expressed between A. mexicanum and A. tigrinum during larval development (42-84 dph). Because the Page et al. (2010) study used a different Affy GeneChip (Amby_001), a BLAST search was performed to identify probesets between Amby_001 and Amby_002 that were designed to target the same transcript, and thus could be reliably compared between the studies.

RESULTS

Global analysis of gene expression

Gene expression variation at the GeneChip level was examined using box plots, principal components analysis (PCA), and hierarchical clustering. Box-plots were generated to examine the overall quality of the data. These plots showed the distribution of raw intensity values to be consistent among all GeneChips (Figure 2.1); thus, all GeneChips were retained for RMA normalization. The normalized gene expression values were then subjected to PCA (Figure 2.2) and hierarchical clustering (Figure 2.3). Both of these analyses yielded the same result: replicate GeneChips within experimental treatments grouped together, forming JUV, PAED, and MET groups. The first principal component explained 98.6% of the variation in gene expression and the first and second principal components resolved distinct groupings for JUV, PAED, and MET replicate GeneChips. There was no evidence that female and males of the PAED group yielded variable gene expression estimates. These results show that juveniles, metamorphs, and paedomorphs are distinct in terms of overall gene expression.

Identification of differently expressed genes among juvenile and adult brains

A total of 828 probesets (i.e. DEGs) yielded statistically significant transcript abundance estimates among the three brain types, and exceeded a 1.5 fold-change threshold between any of the three pairwise contrasts (Figure 2.4). Approximately twice as many DEGs were differently expressed between JUV and MET (N = 722) than between JUV and PAED (N = 364). This indicates that metamorphosis is associated with more unique brain transcriptional changes than paedomorphosis. There was considerable overlap of DEGs between the JUV vs. MET and JUV vs. PAED contrasts; 271 of the 364 DEGs identified between JUV and PAED were also differently expressed between JUV and MET. Thus, 74% of the differences observed between JUV and PAED can be attributable to aging because they were also identified as significant between JUV and MET. These results show that transcriptional patterns within the brains of adult axolotls are associated with metamorphosis, paedomorphosis, and aging.

Classification of adult gene expression patterns relative to juvenile

To more finely resolve expression patterns, DEGs were classified into groups (Figure 2.5) according to statistical significance and direction of change relative to JUV. No DEGs exhibited significantly lower expression for MET and significantly higher expression for PAED (LMHP) or significantly higher expression for MET and significantly lower expression for PAED (HMLP); in other words, no genes showed opposite directional changes in transcript abundance between MET and PAED. More of the DEGs unique to MET decreased in transcript abundance (LM, N = 240; HM, N = 193). This pattern was also observed for genes that were differently expressed within MET and PAED brains (LMP, N = 191; HMP, N = 98) and DEGs unique to PAED (LP, N = 43; HP, N=32). Thus, overall, DEGs tended to show lower transcript abundances in adults than juveniles, indicating that lower transcript abundance correlates with aging and maturation. To determine if PAED and MET expression showed the same pattern and magnitude of divergence from JUV expression values, the 797 DEGs that were significant relative to JUV were graphed as a bivariate plot and fitted using least squares regression (Figure 2.6). Under a null hypothesis that PAED and MET expression is

equivalent for all DEGs, a best-fitting line through these data would have a slope of 1.0. However, the line with the best fit has a positive slope, the value of which is not included within the 95% confidence interval of the null model. The positive slope indicates that on average, MET expression estimates deviated more from JUV expression estimates, than did PAED. Indeed, 12 of 14 DEGs that deviated by three or more standard deviations from predicted values of the best fitting line were classified as LM (Table 2.1). Thus, for genes that were differently expressed between juveniles and adults, transcriptional patterns were more similar between juveniles and paedomorphs, than between juveniles and metamorphs.

Identification of statistically enriched biological processes

Overrepresentation analyses were conducted using PANTHER to identify statistically enriched gene ontologies within groups listed in Figure 2.5. Pathway, protein class, cellular component, molecular function, and biological process gene ontologies were enriched for LM and LMP (Table 2.2). Only a few ontologies were identified as enriched for HM (molecular function, biological process) and HMP (protein class, biological process), and no significantly over-represented ontologies were identified for LP and HP. The number of significantly enriched ontologies identified within groups correlated with group size. This suggests that the over-representation analyses may have been too statistically underpowered to identified DEGs within LP and HP groups. Thus, for these groups, genes were screened by eye to identify candidates with functions potentially associated with metamorphosis and paedomorphosis.

Over-representation of ontologies for LMP and HMP

To identify similarities in gene expression between paedomorphic and metamorphic axolotls, the gene lists for LMP (lower expression) and HMP (higher expression) were analyzed. Over-representation analysis for the LMP group identified ontologies associated with DNA replication, cell cycle, and mitosis as significantly enriched (Table 2.2). The genes that enriched these terms encode histones (e.g. cenpa, hist1h2bj, hist2h2be) and other chromosome associated proteins (e.g. hmgn2, incep, ncapd2, ncapd3, ncapg, smc2, tox3), cell cycle regulators (e.g. bub1, bub1b, ccna2,

ccnb1, ccnb3, ccni, cdc20), kinesins (e.g. aurka, aurkb, cdc2, cdkn1c, cks2, kif4a, kif11, kif15, kif22, kifc1, plk1), DNA replication associated genes (e.g. mcm2, mcm3, mcm4, mcm5, mcm7, pcna, pola), genes associated with cell fate (e.g. notch1, notch2), and transcription factors (e.g hes6). Six extracellular matrix (ECM) genes (col2a1, col4a1, col4a6, col5a1, col11a1, fn1) enriched ontologies associated with antibacterial response and two embryonic hemoglobin genes (hbe1, hbz) enriched the blood circulation ontology. Thus in comparison to juveniles, metamorphic and paedomorphic salamanders expressed relatively fewer transcripts for genes associated with cell division, extracellular matrix, and oxygen transport. Genes that were similarly up regulated in metamorphs and paedomorphs (HMP) yielded lipid metabolism, hydroxylase, and oxygenase terms that were enriched for cytochrome p450 genes (cyp2a6, cyp2a7, cyp2a13, cyp2f1, cyp7b1, cyp39a1). These patterns suggest that aging in A. mexicanum is associated with moderation of cell division, extracellular matrix restructuring, hemoglobin switching, and lipid metabolism.

Over-representation of ontologies for LM and HM

The ontologies identified for genes that uniquely decreased in metamorphs (LM) partially overlapped the list described above for LMP. Genes with extracellular functions, and especially collagens (col1a1, col1a2, col3a1, col4a2, col4a5, col5a2, col6a1, col6a2), enriched several gene ontologies in LM. Although this group of collagens was not identified as significant between JUV and PAED, transcript abundances were generally lower in PAED and thus showed the same direction of change as MET (Figure 2.6). Ontology terms associated with blood circulation, immune function, and cellular and developmental processes were identified as enriched in LM. The blood circulation term identified a fetal hemoglobin paralog (hbg2) as enriched within this group. Only two ontologies were identified as enriched for genes that uniquely increased in metamorphic brains (HM): system process and enzyme inhibitor activity. The system processes included voltage gated potassium (kcnc2, kcng1) and calcium (cacna1d) ion channels, neurotransmitter binding (gabra4, npy1r), synapse remodeling (nlgn4x), and two somatostatin receptors involved in neurotransmission and endocrine signaling (sstr1, sstr5). The enzyme inhibitor term was enriched by a group of protease inhibitors (cst6,

pkib, ppp1r14a, spint1), including a regulator of extracellular matrix remodeling (timp2). Overall, the gene functions identified for LM suggest relatively lower expression of extracellular matrix components and immune related genes in metamorphic brains relative to juveniles and paedomorphs. While collagens generally decreased in abundance as a function of aging, a greater diversity of collagens decreased significantly in metamorphic brains. The genes that showed significantly higher abundances in metamorphic brains (HM) suggest active inhibition of extracellular matrix remodeling relative to juveniles and paedomorphs, and fundamental differences in neural and endocrine signaling.

Genes that were significantly differently expressed in paedomorphs

The DEGs that showed significantly higher or lower expression in paedomorphs (HP and LP) were screened by eye to identify functions associated with paedomorphic development, as these gene lists did not yield significantly enriched terms. The LP list contained genes associated with regulation of gene transcription (znf207, dot1l, and nr2e1), nuclear transport (nup107), mRNA processing (hnrnpk), and ribosome biogenesis (urb2). The HP list included genes associated with blood circulation and iron uptake (fth1, hba2), transcriptional regulation (crem), and homocysteine metabolism (bhmt, cubn). Additionally, genes associated with human disease were identified, including Fanconi anemia (fancl), type II diabetes (iapp), neurodegenerative diseases - Alzheimer’s disease (aplp2, pcmt1) and Sandhoff disease (hexb).

Bioinformatic analysis of morph specific gene expression

A previous study (Page et al., 2010) examined brain gene expression between larval axolotls and a closely related metamorphic species, Ambystoma tigrinum. Genes that were commonly and differently expressed between the species were identified during the larval period, using regression modeling. Although Page et al. (2010) considered the differently expressed genes to be associated with metamorphic and paedomorphic development, they could not rule out the possibility that expression differences were species-specific and not morph-specific. In contrast, the DEGs identified in this study are more likely to be associated with metamorphic and paedomorphic development because

gene expression was examined within a common genetic background. Considering 59 probesets that could be reliably compared between the studies, 59% of these were classified as paedomorphic, metamorphic, or expressed similarly between metamorphs and paedomorphs. This included 21 genes that showed the same pattern of expression between metamorphs and paedomorphs, 13 that were metamorph specific, and one that was paedomorph specific. These results validate 14 genes as exhibiting morph specific expression patterns that are detectable in larvae and adults (Table 4.3). The metamorph specific genes have functions relating to alcohol metabolism (adh1b), chromatin remodeling (anp32e), DNA repair (apex1), extracellular matrix (col1a2, nid2), RNA binding (msi1), lipid and cholesterol metabolism (pltp, psap), ribosomal binding (rrbp1), sulfatase (sulf1), T-cell lymphoma (tiam1), and astrocytes (gfap). The paedomorph specific gene is a homocystine methyltransferase (bhmt). Of the probeset pairs that did not agree, 18 of 24 were metamorph specific in Page et al. (2010), but expressed similarly among metamorphs and paedomorphs in this study. These genes showed functions related to protease inhibition (a2m), cell cycle regulation and mitosis (aurka, aurkb, ccnb3, kif11, kif22, nusap1, pbk. zwint), basement membrane (col4a1), nuclear transport (kpna2), DNA replication (rfc2), ribonuclease activity (rnaseh2a), and chromatin structure (smarca5).

DISCUSSION

In this study, a custom microarray was used to compare gene expression among the brains of juvenile, paedomorphic and metamorphic A. mexicanum. I compared expression estimates from juveniles to estimates from paedomorphs and metamorphs to identify genes that changed significantly during development. The genes that showed statistically equivalent expression changes across metamorphic and paedomorphic brains provide the first glimpse of aging in an amphibian. The genes that showed statistically different expression estimates between metamorphic and paedomorphic brains provide new functional insights about the maintenance and regulation of these different modes of development. Below, I discuss the significance of the results, highlighting specific DEGs and gene functions associated with aging, metamorphosis, and paedomorphosis.

Aging and transcription in the A. mexicanum brain.

Pairwise contrasts among the three groups revealed metamorphic and paedomorphic axolotls to express many genes in common. These genes are discussed within the context of aging below, though genes that were expressed in common between metamorphic and paedomorphic axolotls relative to juvenile could be related to maturational changes within the brain. For example, gonadotropin-releasing hormone (gnrh1) showed a similar increase in expression in both adults relative to juveniles, suggesting equivalence of reproductive physiology between metamorphs and paedomorphs. If so, this supports the idea that axolotl paedomorphosis is an example of neoteny and not progenesis (Gould, 1976; Voss and Shaffer, 1996).

Genes that showed increased expression in metamorphs and paedomorphs relative to juveniles were enriched for cytochrome p450 genes involved in lipid and steroid metabolism. It may be that the increase in cytochrome p450 genes in adult axolotls is related to age-related changes in lipid metabolism, as has been described for mammals. Changes in lipid concentration and distribution in the brain has been linked to neurodegenerative diseases in humans (Ledesma, et al., 2012; van Vilet, 2012) and increased expression of genes associated with lipid metabolism in the mouse brain are associated with aging (Zahn et al., 2007). Additionally, cyp7b1, a cytochrome p450 gene involved in neurosteroid metabolism within mammalian brain shows decreased enzymatic activity with aging; this gene’s product decreases levels of active glucocorticoids in vitro, though the in vivo effects are unknown (Yau & Seckl, 2012). The neurosteroids produced by enzymatic activity of cyp7b1 and another cytochrome p450 gene involved in neurosteroid metabolism – cyp39a1 – are involved in the regulation of the immune response, cell proliferation and apoptosis in mouse and rat (Lathe, 2002). Thus, changes in lipid metabolism and neurosteroid synthesis, relating to specific cytochrome p450 transcript abundance, may be related to changes in the immune system, as well as in cell growth and cell death.

Genes that showed lower expression in metamorphs and paedomorphs relative to juveniles were associated with DNA metabolism, cell cycle regulation, and mitosis. In humans, aging is associated with a global decrease in gene expression across all regions

of the brain and an increase in expression of genes related to cell death (Berchtold et al., 2008); aging in mouse brain tissues is associated with an increase in genes involved in cell cycle arrest (Zahn et al., 2007). Though metamorphs and paedomorphs did not show a similar increase in genes directly related to cell death or cell cycle arrest, a general decrease in genes related to cell proliferation was observed.

Berchtold et al. (2008) showed an increase in expression of genes associated with angiogenesis, oxygen transport, and immune function due to aging in humans, and Zahn et al. (2007) noted an increase in expression of genes associated with iron ion transport and inflammatory responses associated with aging in mouse brain tissues. While blood circulation and immune function were enriched in axolotls, the genes that enriched these categories showed lower transcript abundances in adults. However, a B-cell associated receptor (bcap29) showed higher expression in both morphs and a T-cell associated gene (tiam1) showed higher expression in metamorphs only. Thus, while immune related genes were generally lower in axolotl adults, morph specificity in expression varied for specific genes. Morph specificity of gene expression related to immune function may be related to differences in pathogens that a terrestrial metamorph would be likely to encounter compared to an aquatic paedomorph. Additionally, paedomorphs showed higher expression of ferritin (fth1), which is involved in iron transport and metamorphs showed lower expression of collagen genes associated with angiogenesis, indicating that the aging related gene expression noted in mammals may vary among axolotl morphs. Such variation may correlate with specializations necessary for paedomorphs to live in aquatic habitats and metamorphs to transition to more terrestrial habitats, in the adult phase of life (Dodd & Dodd, 1976; Duellman & Trueb, 1986).

The immune related genes that showed lower expression in metamorphs and paedomorphs included five collagen genes (col2a1, col4a1, col4a6, col5a1, col11a1) and an additional extracellular matrix protein (fn1). This group included fibrillar (col2a1, col5a1, col11a1) and basement membrane (col4a1, col4a6) collagens. Additionally, one microfibrillar collagen (col6a3) also showed the same pattern of expression. Collagens are one of the most abundant components of the extracellular matrix (ECM) and are commonly known to aid in cell attachment, cell migration, and structural integrity of a

tissue, equipping tissues with specific mechanical and biochemical properties (Bateman et al., 1996). However, cell-matrix interactions can be mediated by cell receptors on ECM molecules and can also stimulate cell differentiation and regulate gene expression levels (reviewed by Gelse et al., 2003; Heino, 2007). Collagens can bind some growth factors and cytokines, and in particular, type I collagens have been shown to indirectly block TGF-beta-action (Yamaguchi et al., 1990) and to bind integrins (Tulla et al., 2001). Additionally, changes in collagen expression within brain vasculature are associated with aging in humans (del Zoppo, 2012). Neurodegeneration and gene expression changes due to aging have mostly been considered in humans and other mammals, and have not been well studied in non-mammalian vertebrates (Jiang et al., 2001; Berchtold et al., 2008; Cookson, 2012). My findings suggest that some collagen and extracellular matrix proteins are similarly regulated in salamanders as a function of aging.

Genes associated with metamorphosis and paedomorphosis

Prior to my work, paedomorphic and metamorphic brain gene expression was compared between two species – A. mexicanum and A. tigrinum. That study examined changes in gene expression over time during larval development and metamorphosis in A. tigrinum. It identified many significant DEGs that were unique to larval A. mexicanum and larval A. tigrinum, as well as many commonly expressed genes between the two species (Page et al., 2010). The results suggested axolotl development to be a mosaic of ancestral and derived gene expression patterns, where some genes were expressed similarly with A. tigrinum, though with respect to other genes, there were morph-specific differences. The Page et al. (2010) study, however, could not differentiate genes that showed unique expression due to species-specific differences from those that showed unique expression due to metamorphic and paedomorphic development. My study has better resolved candidate genes for metamorphic and paedomorphic development because it related the early gene expression changes noted by Page et al (2010) between metamorphic A. tigrinum and axolotl larvae, to metamorphic and paedomorphic axolotls. The majority of the comparable genes between Page et al. (2010) and my study showed agreement, with the same pattern of expression observed in both studies.

Eighteen genes that Page et al. (2010) classified as metamorph specific during the larval period were expressed similarly among metamorphic and paedomorphic axolotls in my study. Many of these genes are related to cell cycle regulation and mitosis (aurka, aurkb, ccnb3, kif11, kif22, nusap1, pbk. zwint), chromatin structure and transcriptional regulation (smarca5), DNA replication (rfc2), and ribonuclease activity (rnaseh2a). These results suggest that some genes with similar functions in paedomorphs and metamorphs are transiently, differently expressed during metamorphosis, when processes associated with cell proliferation, transcription, and protein synthesis are necessary for remodeling events (Page et al., 2009, 2010; Huggins et al., 2012).

Genes with extracellular matrix functions, and especially collagens (col1a1, col1a2, col3a1, col4a2, col4a5, col5a2, col6a1, col6a2), were more lowly expressed in metamorphs than juveniles and adults (Figure 4.7). Similar to the collagen genes that showed significantly lower expression in both metamorphs and paedomorphs, these collagens are a mix of fibrillar (col1a1, col1a2, col3a1, col5a2), basement membrane (col4a2, col4a5), and microfibrillar (col6a1, col6a2) collagens. Relative to juvenile, both morphs exhibit lower expression of collagen genes, but the deviation was greater for metamorphs. Two extracellular matrix (ECM) proteins that were classified as metamorph specific in Page et al. (2010) and my study are collagen type 1 alpha 2 (col1a2) and a basement membrane protein, nidogen (nid2), suggesting that these two ECM proteins are similarly regulated as a function of metamorphosis. Interestingly, metamorphs also showed a higher expression of TIMP metallopeptidase inhibitor 2 (timp2), a peptidase involved in the degradation of the extracellular matrix. These differences in gene expression suggest changes in ECM structure and regulation as functions of both aging and morph specific development.

Hormone receptor genes showed significantly higher expression in metamorphs but not paedomorphs; these included somatostatin receptor 1 (sstr1), somatostatin receptor 5 (sstr5), and nuclear receptor corepressor 1 (ncor1). Somatostatin receptors are known to inhibit the release of hormones and other secretory proteins. Ncor1 is a ligand- independent repressor of TR and has been shown in mouse and rat to be involved in CRH and TSH secretion in the pituitary, as well as the recognition of the HPT set point (van

der Laan et al., 2008; Astapova et al., 2011; Costa-e-Sousa et al., 2012). Additionally, a metamorph specific gene validated between both Page et al. (2010) and my study, ribosomal binding protein 1 (rrbp1), also showed significantly higher expression in metamorphs only. This gene functions in the terminal differentiation of secretory cells (Benyamini et al., 2009) and is also known to enhance in vitro collagen biosynthesis (Ueno et al., 2010). The higher expression of these neuroendocrine related receptors and a gene involved in secretory cell differentiation appear to be correlated with metamorphosis and may influence differences in circulating hormone levels between paedomorphs and metamorphs.

As previously mentioned, cytochrome p450 genes associated with lipid metabolism appear to be similarly regulated due to aging, however, other genes associated with lipid metabolism appear to be regulated differently between metamorphs and paedomorphs. For example, two genes that are morph specific in both Page et al. (2010) and my study are a phospholipid transfer protein (pltp) and a homocystine methyltransferase (bhmt). Pltp is metamorph specific and shows higher expression in metamorphs, but not paedomorphs. It is found in blood plasma and involved in cholesterol metabolism (Albers et al., 2012). Bhmt is paedomorph specific and shows higher expression in paedomorphs relative to juvenile. It has been shown to be involved in lipid metabolism (Teng et al., 2011, 2012). Additionally, apolipoprotein H (apoh) is expressed significantly lower in metamorphs relative to juvenile and is known to be involved in cholesterol and lipid metabolism. These results indicate that some genes involved in brain lipid metabolism appear to be regulated differently as a function of aging and metamorphosis. Interestingly, Hervant et al. (2001) found that the obligate cave dwelling, paedomorphic species Proteus anguinus exhibits a consistently lower standing metabolic rate (SMR) over a 24-hour cycle than the facultative cave dwelling, metamorphic species Euproctus asper. Additionally, two paedomorphic salamanders that are not cave dwelling, Necturus maculosus and Amphium means (Light & Lowcock, 1991), exhibit similar SMR to P. anguinus, though the ability to compare these species was called into question by Hervant et al. (2001). This suggests that there may be differences in SMR between paedomorphic and metamorphic salamanders, though differences between facultative and obligate cave dwelling salamanders may also play a

role. Therefore, an investigation of the possible differences between juvenile, metamorphic, and paedomorphic axolotls, as well as between axolotl and tiger salamanders, may help to elucidate differences in metabolism that are associated with metamorphosis.

Metamorphs and paedomorphs also appear to regulate some genes related to neurological processes differently as a function of metamorphosis. Recently, apoh has been implicated in neurodegenerative processes, including Alzheimer’s disease, and low levels of apoh have been correlated with cognitive decline in humans (Thambisetty et al., 2010; Katzav et al., 2011; Song et al., 2012). Other neuronal associated genes that showed significantly higher expression in metamorphs but not paedomorphs, included GABA receptor (gabra4), neuropeptide Y receptor (npy1r), and neuroligin 4 (nlgn4x), which is involved in nervous system remodeling. A nuclear phosphoprotein (anp32e) was classified as metamorph specific in Page et al. (2010) and my study. It showed significantly higher expression in metamorphs and is commonly involved in chromatin remodeling, However, anp32e is thought to be involved in protein trafficking during synaptogenesis through its regulatory effect on protein phosphatase 2A activity (Costanzo et al., 2006). Metamorphs also show significantly lower expression of glial fibrillary acidic protein (gfap), which is an intermediate filament protein of mature astrocytes and may indicate that metamorphs have fewer astrocytes than similarly aged paedomorphs. Paedomorphs exhibited significantly higher expression of genes associated with neurodegenerative disorders in humans as well, including aplp2, hexb, iapp, and pcmt1. Aplp2 is an amyloid precursor protein that is implicated in the pathogenesis of Alzheimer’s disease and pcmt1 has been shown to negatively regulate amyloid beta formation and may play a protective role in the pathogenesis of Alzheimer’s disease (Bae et al., 2011). Additionaly, iapp encodes an amyloid-like protein. Hexb is associated with Sandhoff disease and degrades GM2 ganglioside, an overabundance of which can lead to neurodegeneration. Metamorphs appear to show higher expression of genes relating to neural signaling and nervous system remodeling, while both metamorphs and paedomorphs show morph specific regulation of genes related to neurodegeneration in humans, possibly due to differences in the maintenance of neurological function related to aging and metamorphosis. These expression results indicate that some genes associated

with neurological function and neurodegenerative disorders are not regulated similarly due to aging only, but instead show expression that is both age and morph specific. These expression differences between morphs may be related to physiological remodeling of neural tissues as a result of metamorphosis.

Uromodulin (umod) and surfactant protein D (sftpd) both showed significantly lower expression in metamorphs. Page et al. (2009) found that expression of umod decreased in axolotl head skin during T4 induced metamorphosis. Surfactant genes were also enriched among genes specific to metamorphs, and in particular, sftpd, showed down regulation in the present study and the Page et al. (2010) study. Sftpd is expressed in human brain, though it is primarily expressed in lung tissue, and may play a role in the activated immune system (Madsen et al., 2000; Davé et al., 2004). Additionally, the sftpd promoter can be activated by thyroid transcription factor-1 (ttf1) and abnormal regulation of surfactant proteins has been implicated in brain-lung-thyroid syndrome, which involves congenital hypothyroidism, neurological symptoms and respiratory distress in humans (Guillot et al., 2010). These genes both appear to be good indicators of metamorphosis – umod shows decreased expression in both brain and head skin of metamorphic axolotls, and sftpd shows significant down regulation that correlates with metamorphosis across axolotls and A. tigrinum.

Aging, species, and morph specific differences in hemoglobin expression

Mammals and other vertebrates exhibit altered expression of hemoglobin subunits at different stages of development (Palis, 2008; Dzierzak & Philipsen, 2013). Likewise, axolotls are known to undergo a change in hemoglobin expression between 100-150 dph, which is thought to be due to a cryptic metamorphosis (Tompkins & Townsend, 1977). This shift in hemoglobin expression can also occur prematurely with T4 treatment that is too low to result in morphological metamorphosis. The transcriptional analysis resulted in the identification of morph specific hemoglobin expression changes that are most likely due to gene expression changes within erythrocytes that were isolated from vasculature within the brain tissue. Page et al. (2010) also identified differences in expression of hemoglobin genes during larval and metamorphic development of A. tigrinum compared to A. mexicanum, though species and morph specific differences in expression could not

be disentangled. My findings help to resolve differences in hemoglobin expression that are associated with morph and species-specific variation.

In the Page et al. (2010) study, hemoglobin zeta (hbz) showed similar regulation between A. tigrinum and axolotl larvae, and in my study hbz was regulated similarly among metamorphs and paedomorphs. Thus, hbz expression is conserved among species and morphs. Hemoglobin gamma (hbg) exhibited the same pattern of expression over time in A. tigrinum compared to A. mexicanum (Page et al., 2010), though its expression was consistently higher in A. tigrinum. Similarly, my study found hbg2 to exhibit lower expression in both morphs relative to juveniles, though the difference in expression was significant for metamorphs only. This indicates that lower expression of hbg2 is associated with aging, but metamorphs and paedomorphs show differences in the magnitude of expression during development. Paralogs of hemoglobin alpha (hba) showed both increasing and decreasing patterns of expression in A. tigrinum, but in larval A. mexicanum, these genes showed increasing or flat patterns of expression (Page et al., 2010). I also observed different patterns of expression between paralogs. One hba2 paralog showed significantly higher expression in paedomorphs but not metamorphs (compared to juveniles), while the other hba2 paralog showed significantly decreased expression in both morphs relative to juveniles. Thus, one paralog showed aging related changes and the other showed changes attributable to aging, species, and morph. Hemoglobin epsilon (hbe) showed higher levels of expression in A. mexicanum throughout larval development compared to A. tigrinum (Page et al., 2010), but hbe1 was similarly down regulated relative to juveniles in both metamorphic and paedomorphic individuals in my study. The expression difference in this case seemingly implicates species-specific variation in the timing of hemoglobin switching events that are independent of metamorphosis. Overall, the complex regulation of hemoglobin loci during development can only be deciphered using longitudinal experiments that are capable of parsing variation attributable to species and morph. These differences in hemoglobin expression, as well as enrichment of genes associated with angiogenesis that showed lower expression in metamorphs relative to juvenile, may indicate that brain vasculature differs with age and metamorphosis. An investigation of differences between morph and age specific vascularization could indicate if this is indeed the case.

In summary, my study shows that metamorphosis and aging strongly affect the transcriptional program of the brain in A. mexicanum. My study provides an initial glimpse of gene expression related to aging in an amphibian. Biological processes associated with aging were identified, including changes in lipid and neurosteroid metabolism, cell proliferation, blood circulation, and immune function. My study also identified morph specific gene expression patterns that are detectable during early development and continue into adulthood, including differences in expression of collagen, hemoglobin, hormone pathways and neurodegenerative associated gene regulation. More significant DEGs exhibited higher or lower expression that were unique to metamorphs (compared to juvenile), than those that were unique to paedomorphs. This suggests that paedomorph brain gene expression is more similar to juvenile than metamorph expression for genes are not commonly regulated due to aging. Future studies should address possible differences in brain vasculature and metabolism between morphs, as well as include longitudinal studies that would better clarify transient transcriptional changes during metamorphosis across species from transcriptional changes due to aging and maturation.

Table 2.1: Genes that deviated by three or more standard deviations (residual absolute value > 1.72) from values predicted by a linear model fit to a bivariate plot of Paed and Met DEGs. LM = Lower Met, HM = Higher Met, LMP = Lower Met and Paed. Gene Probe Set ID Gene Name Group Symbol

axo20166-r_s_at HBG2 G-gamma globin [Homo sapiens] LM axo20348-f_s_at COL1A1 alpha 1 type I collagen preproprotein [Homo sapiens] LM axo15318-f_at UMOD uromodulin precursor [Homo sapiens] LM axo22229-f_at --- LM axo05346-f_at APOH apolipoprotein H precursor [Homo sapiens] LM axo20347-f_s_at COL1A1 alpha 1 type I collagen preproprotein [Homo sapiens] LM axo10397-f_at COL3A1 collagen type III alpha 1 preproprotein [Homo sapiens] LM axo27355-f_at --- LM

46 pulmonary surfactant-associated protein D precursor [Homo

axo08766-f_at SFTPD sapiens] LM axo02853-f_at ZNF605 zinc finger protein 605 [Homo sapiens] LM serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), axo08393-r_at SERPINB5 member 5 [Homo sapiens] LM axo10212-f_at COL1A2 alpha 2 type I collagen [Homo sapiens] LM axo02258-f_at KRTCAP2 keratinocyte associated protein 2 [Homo sapiens] HM baculoviral IAP repeat-containing protein 5 isoform 1 [Homo axo06982-f_at BIRC5 sapiens] LMP

Table 2.2: List of over-represented ontology terms, including the number of genes present for each term, and the associated p-value for each term. Significant Ontology Significant Group Over-represented Term Gene List Category Gene List (P-value) cell cycle 48 6.10E-14 cellular process 76 4.99E-10 mitosis 25 5.93E-09 DNA replication 16 1.62E-07 chromosome segregation 14 5.96E-07 DNA metabolic process 17 5.20E-04 blood circulation 8 8.23E-04 cellular component organization 25 1.88E-03

47 regulation of liquid surface tension 5 2.19E-03

Biological cytokinesis 9 6.94E-03 LMP Process developmental process 35 7.66E-03 ectoderm development 22 1.94E-02 defense response to bacterium 6 2.46E-02 nucleobase, nucleoside, nucleotide and nucleic acid metabolic 45 2.73E-02 process establishment or maintenance of chromatin architecture 10 2.74E-02 asymmetric protein localization 5 2.81E-02 protein localization 5 2.81E-02 system development 25 3.18E-02 meiosis 8 3.89E-02 extracellular matrix 12 3.33E-03 Cellular LMP extracellular region 12 3.49E-03 Component ribonucleoprotein complex 8 5.78E-03

Table 2.2 Continued extracellular matrix structural constituent 8 1.27E-04 microtubule motor activity 6 1.62E-03 Molecular LMP structural molecule activity 26 2.95E-03 Function DNA helicase activity 6 9.14E-03 motor activity 6 1.92E-02 LMP Pathway Integrin signalling pathway 8 2.84E-02 DNA binding protein 22 5.09E-05 extracellular matrix structural protein 8 1.54E-04 microtubule binding motor protein 6 1.96E-03 surfactant 5 2.31E-03 LMP Protein Class DNA helicase 6 8.39E-03

48 extracellular matrix protein 12 1.67E-02

antibacterial response protein 6 2.59E-02 ribonucleoprotein 8 2.89E-02 regulation of liquid surface tension 11 1.73E-10 blood circulation 14 2.26E-08 defense response to bacterium 12 4.03E-07 cell-cell adhesion 25 1.59E-06 cell adhesion 33 2.30E-06 Biological macrophage activation 13 3.91E-06 LM Process homeostatic process 12 6.51E-05 asymmetric protein localization 8 1.58E-04 protein localization 8 1.58E-04 ectoderm development 33 2.23E-04 localization 11 7.41E-04 mesoderm development 28 8.64E-04

Table 2.2 Continued skeletal system development 12 2.63E-03 receptor-mediated endocytosis 10 4.10E-03 immune system process 38 7.90E-03 developmental process 46 9.35E-03 Biological angiogenesis 10 1.31E-02 LM Process (Con't) response to stimulus 27 1.67E-02 endocytosis 16 1.83E-02 system process 37 2.21E-02 system development 33 2.45E-02 cellular process 79 3.59E-02 Cellular extracellular matrix 19 9.57E-06 LM Component extracellular region 19 1.03E-05 49 extracellular matrix structural constituent 9 2.29E-04 Molecular LM structural molecule activity 32 1.20E-02 Function receptor activity 27 1.65E-02 LM Pathway Integrin signalling pathway 11 3.03E-03 surfactant 11 1.82E-10 defense/immunity protein 23 1.93E-09 cell adhesion molecule 24 9.47E-08 LM Protein Class antibacterial response protein 12 4.24E-07 extracellular matrix protein 19 4.79E-05 extracellular matrix structural protein 9 2.77E-04 receptor 27 1.94E-02 Biological HM system process 30 1.20E-02 Process Molecular HM enzyme inhibitor activity 10 3.77E-02 Function

Table 2.2 Continued Biological HMP lipid metabolic process 14 4.81E-03 Process oxygenase 8 1.76E-06 HMP Protein Class hydroxylase 5 4.11E-04

Table 2.3: Comparison of DEGs from this study and Page et al. (2010). M = metamorph specific, P = paedomorph specific, MP = shared between metamorph and paedomorph. Expression Expression Official Gene Amby_001 Amby_002 pattern in Pattern in Sal-site V0.3 Contig Symbol Probeset Probeset Page et al. current (2010) study C6orf115 SRV_05942_at axo18117-r_at contig79311 MP MP CDC2 SRV_01180_at axo07587-r_at contig24691 MP MP CDC20 SRV_00806_a_at axo07048-f_at contig07123 MP MP CENPV SRV_08376_at axo02633-f_at contig21590 MP MP COL2A1 SRV_01205_at axo07638-f_at contig81592 MP MP FAAH SRV_00949_at axo07273-r_at contig45884 MP MP FXYD3 SRV_02740_a_at axo06274-f_at contig32491 MP P

51 GDA SRV_02137_at axo09923-r_at contig79334 MP MP

KIAA0101 SRV_03588_at axo21059-f_at Mex_NM_014736_Contig_2 MP MP KIFC1 SRV_06485_a_at axo08033-f_at contig08742 MP MP LAMA1 SRV_02580_a_at axo11181-f_at contig61801 MP MP MCM2 SRV_09459_at axo10169-r_at contig03405 MP MP MCM3 SRV_01414_at axo08153-f_at contig07339 MP MP PLK1 SRV_02420_at axo10701-f_at contig64493 MP MP POLR2L SRV_04514_at axo18021-f_at contig23009 MP MP RCC1 SRV_00812_a_at axo04428-f_at contig32974 MP M RCC2 SRV_04287_at axo06321-f_at contig84321 MP MP RRM1 SRV_00703_a_at axo00677-r_at contig04610 MP MP RRM2 SRV_00706_a_at axo00701-r_at contig86260 MP MP SEPT8 SRV_04195_at axo05007-r_at contig97712 MP M SFRP2 SRV_09416_at axo08757-f_at contig94308 MP MP TMEM176B SRV_03419_at axo05154-f_at contig44794 MP MP

Table 2.3 Continued TMPO SRV_01795_at axo01770-f_at contig40054 MP M TYROBP SRV_01821_at axo02980-f_at contig03910 MP MP UHRF1 SRV_03362_at axo13318-f_at contig63025 MP MP A2M SRV_00116_at axo00006-f_at contig97275 M MP ADH1B SRV_00457_at axo07742-r_at contig08448 M M ANP32E SRV_04811_a_at axo19462-f_at contig77943 M M APEX1 SRV_05165_a_at axo00384-f_at contig40552 M M AURKA SRV_01907_at axo03050-f_at contig06048 M MP AURKB SRV_02115_at axo09856-f_at contig88180 M MP CCNB3 SRV_05031_at axo20258-f_at contig86908 M MP COL1A2 SRV_09970_at axo10212-f_at contig81602 M M COL4A1 SRV_01207_s_at axo07643-f_at contig84121 M MP

52 CRYAB SRV_01225_at axo07684-f_at contig02235 M MP

DLGAP5 SRV_03593_at axo06865-f_at contig08508 M MP GFAP SRV_01304_at axo07863-r_at contig21475 M M KIF11 SRV_02235_at axo10165-f_at contig84891 M MP KIF22 SRV_03220_at axo12862-f_at contig02837 M MP KPNA2 SRV_01354_a_at axo08037-f_at contig93141 M MP MSI1 SRV_10183_at axo08202-f_at contig04157 M M NID2 SRV_03227_at axo12902-f_at contig82626 M M NUP107 SRV_04423_a_at axo17535-f_at contig11105 M P NUSAP1 SRV_04253_a_at axo16931-r_at contig05609 M MP PBK SRV_04270_at axo16972-r_at contig61944 M MP PLTP SRV_07538_at axo11843-r_at contig08790 M M PSAP SRV_01558_at axo04317-f_at contig64611 M M RFC2 SRV_05574_at axo02595-r_at contig05592 M MP RNASEH2A SRV_02906_a_at axo12035-f_at contig28679 M MP

Table 2.3 Continued RRBP1 SRV_02269_a_at axo04353-f_at contig03012 M M SFTPD SRV_07632_at axo08766-f_at contig45505 M M SMARCA5 SRV_01909_at axo09291-f_at contig20796 M MP SULF1 SRV_03675_at axo05681-f_at contig38682 M M TIAM1 SRV_01791_at axo08977-r_at contig28568 M M TM6SF1 SRV_04650_a_at axo18613-r_at contig07019 M MP TPX2 SRV_03256_at axo12960-f_at contig13550 M MP ZWINT SRV_05024_at axo12663-r_at contig09898 M MP BHMT SRV_01149_at axo07527-f_at contig38970 P P CAPNS1 SRV_01164_a_at axo07565-r_at contig74383 P MP 53

Figure 2.1: Box plots, generated prior to RMA correction, of log2 microarray signal intensity. Juvenile microarray chips: 454-11, 454-12, 454-13. Adult Metamorph microarray chips: 454-1, 454-2, 454-3. Adult Paedomorph microarray chips: 454-6 (female), 454-7 (male), 454-8 (male).

C13 C11 C12

B6 B8 B7 A3

A2 A1

Figure 2.2: Results from principle components analysis. Juvenile microarray chips are blue (C11 = 454.11, C12 = 454.12, C13 = 454.13), Metamorph microarray chips are red (A1 = 454.1, A2 = 454.2, A3 = 454.3), and Paedomorph microarray chips are green (B6 = 454.6 (female), B7 = 454.7 (male), B8 = 454.8 (male)).

55 5HSRUW8QWLWOHG 3DJHRI +LHUDUFKLFDO&OXVWHULQJ 0HWKRG )DVW:DUG 'HQGURJUDP

& & & $ $ 5HSRUW8QWLWOHG$ 3DJHRI % +LHUDUFKLFDO&OXVWHULQJ% 0HWKRG% )DVW:DUG 'HQGURJUDP

& & & $ $ $ % % % &OXVWHULQJ+LVWRU\ 1XPEHURI &OXVWHUV 'LVWDQFH /HDGHU -RLQHU % % & & % % & & $ $ $ $ $ % & $ &OXVWHULQJ+LVWRU\ Figure 2.31XPEHURI: Results of hierarchical clustering analysis. Microarray chips cluster according to their sample&OXVWHUV groups. Juvenile'LVWDQFH microarray/HDGHU chips are-RLQHU C11 = 454.11, C12 = 454.12, C13 = 454.13. Metamorph microarray chips are% A1 =% 454.1, A2 = 454.2, A3 = 454.3. Paedomorph microarray chips are B6 = 454.6& (female),& B7 = 454.7 (male), B8 = 454.8 (male). The clustering history states the distance% between% the leader and the joiner (i.e. between two GeneChips) where distance& is represented& as a function of the squared Euclidian distance. $ $ $ $ $ % & $

Figure 2.4: Venn diagram showing overlap of the 828 DEGs between the three different pair-wise comparisons – JUV and MET, JUV and PAED, and PAED and MET.

%#"" %!"" &'"(")*+,"" &'"("*-,"" $" !" $" #" #" !"

.#"" .!" &'"("/0-,"" &'"("-),"" #" !" $" !" $" #"

.#!" %#!" &'"("01," &'"("/0/,"" !" $" #" #" $" !"

%#.!" .#%!" &'"("+,"" !" &'"("+,"" #"

$" $" #" !"

Figure 2.5: Patterns of gene expression changes for PAED and MET relative to JUV expression. LM = significantly lower expression for Met, but not Paed; LP = significantly lower expression for Paed, but not Met; HM = significantly higher expression for Met, but not Paed; HP = significantly higher expression for Paed, but not Met; HMP = significantly higher expression for Met and Paed; LMP = significantly lower expression for Met and Paed; LMHP = significantly lower expression for Met and significantly higher expression for Paed; HMLP = significantly higher expression for Met and significantly lower expression for Paed.

Figure 2.6: Bivariate plot of the mean PAED and mean MET expression relative to JUV for the 797 DEGs significant compared to JUV. The green line is the expected best fitting line (y = x) and the red line is the best-fit line (y = 1.2640142x – 0.086525) for the data. The colored data points correspond to the seven groups as follows, purple is LMP; dark blue is LM; light blue is LP; yellow is HP; orange is HM; and red is HMP. Data points represented by r deviate by > three standard deviations from the best-fit line for the data (red line). The fourteen corresponding probesets are listed in Table 4.1.

-1

-2

-3 juv paed -4 met 60 Log2 Fold Change (relative to Juvenile) to Juvenile) ChangeLog2 Fold (relative

-5

-6

Gene Figure 2.7: Bar graph of significant collagen gene expression relative to JUV for seventeen significant probsets that annotate to fourteen collagen genes. Probesets that annotate to the same gene show the same pattern of expression between JUV, PAED, and MET. LM = significantly lower in MET only; LMP = significantly lower in MET and PAED.

REFERENCES

Albers, J. J., Vuletic, S., Cheung, M. C. (2012). Role of plasma phospholipid transfer protein in lipid and lipoprotein metabolism. Biochimica et Biophysica Acta, 1821, 345-357.

Arbeitman, M. N., Furlong, E. E. M., Imam, F., Johnson, E., Null, B. H., Baker, B. S., et al. (2002). Gene expression during the life cycle of Drosophila melanogaster. Science, 297, 2270-2275

Astapova, I., Vella, K. R., Ramadoss, P., Holtz, K. A., Rodwin, B. A. Liao, X. H., et al. (2011). The nuclear receptor corepressor (NCoR) controls thyroid hormone sensitivity and the set point of the hypothalamic-pituitary-thyroid axis. Molecular Endocrinology, 25, 212-224.

Armstrong, J. B., Duhon, S. T., & Malacinski, G. M. (1989). Raising the axolotl in captivity. In J. B. Armstrong & G. M. Malacinski (Eds.), Developmental biology of the axolotl (pp. 220–227). New York: Oxford University Press.

Bae, N., Byeon, S. E., Song, J., Lee, S. J., Kwon, M., Mook-Jung, I., et al. (2011). Knock-down of protein L-isoaspartyl O-methyltransferase increases β–amyloid production by decreasing ADAM10 and ADAM17 levels. Acta Pharmacologica Sinica. 32, 288-294.

Bateman, J. F., Lamande, S. R., Ramshaw, J. A. M. (1996). Collagen superfamily. In W. D. Comper (Ed.), Extracellular Matrix (pp. 22-67). Melbourne: Harwood Academic Press.

Benyamini, P., Webster, P., Meyer, D. I. (2009). Knockdown of p180 eliminates the terminal differentiation of a secretory cell line. Molecular Biology of the Cell, 20, 732-744.

Berchtold, N. C., Cribbs, D. H., Coleman, P. D., Rogers, J., Head, E. Kim, R. et al. (2008). Gene expression changes in the course of normal brain aging are sexually dimorphic. Proceedings of the National Academy of Science of the United States of America, 105, 15605-15610.

Blount, R. F. (1950). The effects of heteroplastic hypophyseal grafts upon the axolotl, Ambystoma mexicanum. The Journal of Experimental Zoology, 113, 717–739.

Brown, D. D., & Cai, L. (2007). Amphibian metamorphosis. Developmental Biology, 20–33.

Bruce, R. C. (2003). Life histories. In D. M. Sever (Ed.), Reproductive biology and phylogeny of Urodela (pp. 477–525). Enfield, NH: Science Publishers, Inc

Buchholz, D. R., Tomita, A., Fu, L., Paul, B. D., & Shi, Y. B. (2004). Transgenic analysis reveals that thyroid hormone receptor is sufficient to mediate the thyroid hormone signal in frog metamorphosis. Molecular and Cellular Biology, 24, 9026–9037.

Buchholz, D. R., Paul, B. D., Fu, L. Z., Shi, Y.B. (2006). Molecular and developmental analyses of thyroid hormone receptor function in Xenopus laevis, the African clawed frog. General and Comparative Endocrinology, 145, 1-19.

Buechel, H.M., Popovic, J., Searcy, J.L., Porter, N.M., Thibault, O., et al. (2011). Deep Sleep and pariental cortex gene expression changes are related to cognitive deficits with age. PLoS ONE 6, e18387.

Buscaglia, M., Leloup, J., & De Luze, A. (1985). The role and regulation of monodeiodination of thyroxine to 3,5,3’-triiodothyronine during amphibian metamorphosis. In M. Balls & M. Bownes (Eds.) Metamorphosis (pp. 273–293). Oxford, UK: Oxford University Press.

Caria, M. A., Dratman, M. B., Kow, L. M., Mameli, O., & Pavlides, C. J. (2008). Thyroid hormone action: Non-genomic modulation of neuronal excitability in the hippocampus. Journal of Neuroendocrinology, 21, 98–107.

Charipper, H. A. (1929). Studies on amphibian endocrines. V. Accelerated metamorphosis of Rana clamitans larvae by means of implants of fresh anterior pituitary from adult Necturus maculosus. The Anatomical Record, 44, 117–141.

Cheng, S. Y., Leonard, J. L., & Davis, P. J. (2010). Molecular aspects of thyroid hormone actions. Endocrine Reviews, 31, 139–170.

Connors, K. A., Korte, J. J., Anderson, G. W., & Degitz, S. J. (2010). Characterization of thyroid hormone transporter expression during tissue-specific metamorphic events in Xenopus tropicalis. General and Comparative Endocrinology, 168, 149–159.

Cookson, M. R. (2012). Evolution of neurodegeneration. Current Biology, 22, R753-761.

Costa-e-Sousa, R. H., Astapova, I., Ye, F., Wondisford, F. E., Hollenberf, A. N. (2012). The thyroid axis is regulated by NCoR1 via its actions in the pituitary. Endocrinology, 153, 5049-5057.

Costanzo, R. V., Vila-Ortiz, G. J., Perandones, C., Carminatti, H., Mantilla, A., Randrizzani, M. (2006). Anp32e/Cpd1 regulates protein phosphatase 2A activity at synapses during synaptogenesis. European Journal of Neuroscience, 32, 309- 324.

Darras, V. M., & Kuhn, E. R. (1983). Effects of TRH, bovine TSH, and pituitary extracts on thyroidal T4 release in Ambystoma mexicanum. General and Comparative Endocrinology, 51, 286–291.

Darras, V. M., Van der Geyten, S., Cox, C., Segers, I. B., De Groef, B., & Ku¨ hn, E. R. (2002). Effects of dexamethasone treatment on iodothyronine deiodinase activities and on metamorphosis-related morphological changes in the axolotl (Ambystoma mexicanum). General and Comparative Endocrinology, 127, 157– 164.

Davé, V., Childs, T., Whitsett, J. A. (2004). Nuclear factor of activated T cells regulates transcription of the surfactant protein D gene (sftpd) via direct interaction with thyroid transcription factor-1 in lung epithelial cells. Journal of Biological Chemisty, 279, 34578-34588.

Davis, P. J., Leonard, J. L., & Davis, F. B. (2008). Mechanisms of nongenomic actions of thyroid hormone. Frontiers in Neuroendocrinology, 29, 211–218.

Del Zoppo, G. J. (2012). Aging and the neurovascular unit. Annals of the New York Academy of Sciences, 1269, 127-133.

Denoel, M., Ivanovic, A., Dzukic, G., & Kalezic, M. L. (2009). Sexual size dimorphism in the evolutionary context of facultative paedomorphosis: Insights from European newts. BMC Evolutionary Biology, 9, 278.

Denoel, M., & Joly, P. (2000). Neoteny and progenesis as two heterochronic processes involved in paedomorphosis in Triturus alpestris (Amphibia: Caudata). Proceedings of the Royal Society of London, Series B: Biological Sciences, 267, 1481–1485.

Denoel, M., Joly, P., & Whiteman, H. H. (2005). Evolutionary ecology of facultative paedomorphosis in newts and salamanders. Biological Reviews, 80, 663–671.

Denoel, M., & Poncin, P. (2001). The effect of food on growth and metamorphosis of paedomorphs in Triturus alpestris apuanus. Archiv fu¨r Hydrobiologie, 152, 661– 670.

Dent, J. N. (1968). Survey of amphibian metamorphosis. In W. Etkin & L. I. Gilbert (Eds.), Metamorphosis: A problem in developmental biology (pp. 271–311). New York: Appleton-Century-Crofts.

Denver, R. J., Glennemeier, K. A., & Boorse, G. C. (2002). Endocrinology of complex life cycles: Amphibians. In D. W. Pfaff, A. P. Arnld, A. M. Etgen, S. E. Fahrbach & R. T. Rubin (Eds.), Hormones, brain and behavior (pp. 469–513). San Diego: Elsevier.

Denver, R. J., & Licht, P. (1989). Neuropeptide stimulation of thyrotropin secretion in the larval bullfrog: Evidence for a common neuroregulator of thyroid and interrenal activity in metamorphosis. The Journal of Experimental Zoology, 252, 101–104.

Dodd, M. H. I., & Dodd, J. M. (1976). The biology of metamorphosis. In B. Lofts (Ed.), Physiology of amphibia, Vol. 3, (pp. 467–599). New York: Academic Press.

Doyle, J. M., & Whitemann, H. H. (2008). Paedomorphosis in Ambystoma talpoideum: Effects of initial body size variation and density. Oecologia, 156, 87–94.

Duellman, W. E., & Trueb, L. (1986). Biology of amphibians. New York: McGraw-Hill.

Dumeril, A. (1870). Quatrieme notice sur la menagerie des reptiles du Museum d’Histoire Naturelle. Nouvelles Archives du Museum D’Histoire Naturelle De Paris, 5, 47–60.

Dundee, H. A. (1957). Partial metamorphosis induced in Typhlomolge rathbuni. Copeia, 1957, 52–53.

Eagleson, G. W., & McKeown, B. A. (1980). T4-induced metamorphosis in Ambystoma gracile (Baird) from two populations: The effects of aging and temperature. Biochemistry and Experimental Biology, 16, 111–126.

Emerson, E. T. (1905). General anatomy of Typhlomolge rathbuni. Proceedings of the Boston Society of Natural History, 32, 43–76.

Fox, H. (1984). Amphibian morphogenesis. Clifton, NJ: The Humana Press, Inc.

Fujiwara, K., Jindatip, D., Kikuchi, M., Yashiro, T. (2010). In situ hybridization reveals that type I and III collagens are produced by pericytes in the anterior pituitary gland of rats. Cell and Tissue Research, 342, 491-495.

Furuya, F., Lu, C., Guigon, C. J., & Cheng, S. Y. (2009). Nongenomic activation of pho- sphatidylinositol 3-kinase signaling by thyroid hormone receptors. Steroids, 74, 628–634.

Galton, V. A. (1985). 3,5,3’-triiodothyronine receptors and thyroxine 5’ - monodeiodinating activity in thyroid hormone-insensitive Amphibia. General and Comparative Endocrinology, 57, 465–471.

Galton, V. A. (1991). Thyroid hormone receptors and iodothyronine deiodinases in the developing Mexican axolotl, Ambystoma mexicanum. General and Comparative Endocrinology, 85, 62–70.

Galton, V. A. (1992). The role of thyroid hormone in amphibian metamorphosis. Trends in Endocrinology and Metabolism, 3, 96–100.

Gancedo, B., Corpas, I., Alonso-Gomez, A. L., Delgado, M. J., Morreale de Excobar, G., & Alonso-Bedate, M. (1992). Corticotropin-releasing factor stimulates metamorphosis and increases thyroid hormone concentration in prometamorphic Rana perezi larvae. General and Comparative Endocrinology, 87, 6–13.

Gelse, K., Pöschl, E., Aigner, T. (2003). Collagens–structure, function, and biosynthesis. Advanced Drug Delivery Review, 55, 1531–1546.

Gilbert, L. I., Tata, J. R., & Atkinson, B. G. (Eds.), (1996). Metamorphosis: Postembryonic reprogramming of gene expression in amphibians and insect cells. San Diego: Academic Press.

Goplen, D., Bougnaud, S., Rajcevic, U., Boe, S. O., Skaftnesmo, K. O., Voges, J., et al. (2010). αB-crystallin is elevated in highly infiltrative apoptosis-resistant glioblastoma cells. The American Journal of Pathology, 177, 1618-1628.

Gorbman, A. (1957). The thyroid gland of Typhlomolge rathbuni. Copeia, 1957, 41–43.

Gould, S. J. (1977). Ontogeny and phylogeny. Cambridge, UK: Belkap Press.

Gould, S. J. (1981). Change in developmental timing as a mechanism of macroevolution. In J. T. Bonner (Ed.), Evolution and development (pp. 333–346). Berlin, Germany: Springer- Verlag.

Grant, M. P. (1930). The release of follicular colloid from the thyroid of Necturus maculosus following heteroplastic anterior-pituitary implants. The Anatomical Record, 46, 205–221.

Graue, W. V. (1998). Estudio gene´tico y demogra´fico de la poblacion del amfibio Ambystoma mexicanum (Caudata: Ambystomaidae) del Lago de Xochimilco. Mexico: Instituto de Ciencias del Mar y Limnologia, UNAM.

Gudernatsch, J. F. (1912). Feeding experiments in tadpoles: I. The influence of specific organs given as food on growth and differentiations: A contribution to the knowledge of organs with internal secretion. Archiv fu¨r Entwicklungsmechanik der Organismen, 35, 457–483.

Guillot, L., Carre, A., Szinnai, G., Castanet, M., Tron, E., Jaubert, F., et al. (2010). NKX2-1 mutations leading to surfactant protein promoter dysregulation cause interstitial lung disease in “Brain-Lung-Thyroid Syndrome”. Human Mutations, 31, E1146-1162.

Hagemann, T. L., Boelens, W. C., Wawrousek, E. F., Messing, A. (2009). Suppression of GFAP toxicity by alphaB-crystallin in mouse models of Alexander disease. Human Molecular Genetics, 18, 1190-1199.

Hanon, E. A., Lincoln, G. A., Fustin, J.-M., Dardente, H., Masson-Pevet, M., Morgan, P. J., et al. (2008). Ancestral TSH mechanism signals summer in a photoperiodic mammal. Current Biology, 18, 1147–1152.

Harris, R. N. (1987). Density-dependent paedomorphosis in the salamander Notophthalamus viridescens dorsalis. Ecology, 68, 705–712.

Hayes, T., Chan, R., & Licht, P. (1993). Interactions of temperature and steroids on larval growth, development, and metamorphosis in a toad (Bufo boreas). The Journal of Experimental Zoology, 266, 206–215.

Heino, J. (2007). The collagen family members as cell adhesion proteins. Bioessays, 29, 1001–1010.

Hennemann, G., Docter, R., Friesema, E. C., De Jong, M., Krenning, E. P., & Visser, T. J. (2001). Plasma membrane transport of thyroid hormones and its role in thyroid hormone metabolism and bioavailability. Endocrine Reviews, 22, 451–476.

Hervant, F., Mathieu, J., Durand, J. P. (2001). Circadian rhythmicity, respiration and behavior in hypogean and epigean salamanders. Natura Croatica, 10, 141-152.

Huggins, P., Johnson, C. K., Schoergendorfer, A., Putta, S., Bathke, A. C., Stromberg, A. J., et al. (2012). Identification of differentially expressed thyroid hormone responsive genes from the brain of the Mexican Axolotl (Ambystoma mexicanum). Comparative Biochemistry and Physiology Part C: Toxicology, 155, 128–135.

Huxley, J. S., & Hogben, L. T. (1922). Experiments on amphibian metamorphosis and pigment response in relation to internal secretions. Proceedings of the Royal Society B, 93, 36–53.

Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J., Scherf, U. et al. (2003). Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics, 4, 249-264.

Jiang, C. H., Tsien, J. Z., Schultz, P. G., Hu, Y. (2001). The effects of aging on gene expression in the hypothalamus and cortex of mice. Proceedings of the National Academy of Science of the United States of America, 98, 1930-1934.

Katzav, A., Faust-Socher, A., Kvapil, F., Michaelson, D. M., Blank, M., Pick, C. G., et al. (2011). Antiphospholipid syndrome induction exacerbates a transgenic Alzheimer disease model on a female background. Neurobiology of Aging, 32, 272–279.

Kezer, J. (1952). Thyroxine-induced metamorphosis of the neotenic salamanders Eurycea tynerensis and Eurycea neotenes. Copeia, 1952, 234–237.

Krenz, J. D., & Server, D. M. (1995). Mating and ovoposition in paedomorphic Ambystoma talpoideum preceeds the arrival of terrestrial males. Herpetologica, 51, 387–393.

Kuchenbauer, F., Hopfner, U., Stalla, J., Arzt, E., Stalla, G. K., Paez-Pereda, M. (2001). Extracellular matrix components regulate ACTH production and proliferation in corticotroph tumor cells. Molecular and Cellular Endocrinology, 175, 141-148.

Kuhn, E. R., De Groef, B., Van der Geyten, S., & Darras, V. M. (2005). Corticotropin- releasing hormone-mediated metamorphosis in the neotenic axolotl Ambystoma mexicanum: Synergistic involvement of thyroxine and corticoids on brain type II deiodinase. General and Comparative Endocrinology, 143, 75–81.

Larsen, J. H. (1968). Ultrastructure of thyroid follicle cells of three salamanders (Ambystoma, Amphiuma, and Necturus) exhibiting varying degrees of neoteny. Journal of Ultrastructure Research, 24, 190–209.

Larson, A., Weisrock, D. W., & Kozak, K. H. (2003). Phylogenetic systematics of salamanders (Amphibia:Urodela), a review. In D. M. Sever (Ed.), Reproductive biology and phylogeny of Urodela (pp. 31–108). Enfield, NH: Science Publishers, Inc.

Lathe, R. (2002). Steroid and sterol 7-hydroxylation: ancient pathways. Steroids, 67, 967- 977.

Laudet, V. (2011). The origins and evolution of vertebrate metamorphosis. Current Biology, 21, R726–R737.

Ledesma, M. D., Martin, M. G., Dotti, C. G. (2012). Lipid changes in the aged brain: effect on synaptic function and neuronal survival. Progress in Lipid Research, 51, 23-25.

Light, L. E., & Lowcock, L. A. (1991). Genome size and metabolic rate in salamanders. Comparative Biochemistry and Physiology, 100B, 83-92.

Lynn, W. G., & Wachowski, H. E. (1951). The thyroid gland and its functions in cold- blooded vertebrates. The Quarterly Review of Biology, 26, 123-168.

Madsen, J., Kliem, A., Tornoe, I., Skjodt, K., Koch, C., Holmskov, U. (2000). Localization of lunch surfactant protein D on mucosal surfaces in human tissues. Journal of Immunology, 164, 5866-5870.

Manzon, R. G., & Denver, R. J. (2004). Regulation of pituitary thyrotropin gene expression during Xenopus metamorphosis: Negative feedback is functional throughout metamorphosis. The Journal of Endocrinology, 182, 273–285.

Melcangi, R. C., Magnaghi, V., Cavarretta, I., Riva, M. A., Piva, F., Martini, L. (1998). Effects of steroid hormones on gene expression of glial markers in the central and peripheral nervous system: variations induced by aging. Experimental Gerontology, 33, 827-836.

Nakao, N., Ono, H., Yamamura, T., Anraku, T., Takagi, T., Higashi, K., et al. (2008). Thyrotrophin in the pars tuberalis triggers photoperiodic response. Nature, 452, 317–322.

Norman, M. F., Carr, J. A., & Norris, D. O. (1987). Adenohypophysial-thyroid activity of the tiger salamander, Ambystoma tigrinum, as a function of metamorphosis and captivity. The Journal of Experimental Zoology, 242, 55–66.

Norris, D. O., & Gern, W. A. (1976). Thyroxine-induced activation of hypothalamo- hypophysial axis in neotenic salamander larvae. Science, 194, 525–527.

Norris, D. O., & Platt, J. E. (1973). Effects of pituitary hormones, melatonin, and thyroidal inhibitors on radioiodide uptake by the thyroid glands of larval and adult tiger salamanders, Ambystoma tigrinum (Amphibia: Caudata). General and Comparative Endocrinology, 21, 368–376.

Nugent, B. M., Tobet, S. A., Lara, H. E., Lucion, A. B., Wilson, M. E., Recabarren, S. E., et al. (2012). Hormonal programming across the lifespan. Hormone Metabolism Research, 44, 577-586.

Page, R. B., & Voss, S. R. (2009). Induction of metamorphosis in axolotls (Ambystoma mexicanum). Cold Spring Harbor Protocols.

Page, R. B., Boley, M. A., Smith, J. J., Putta, S., & Voss, S. R. (2010). Microarray analysis of a salamander hopeful monster reveals transcriptional signatures of paedomorphic brain development. BMC Evolutionary Biology, 10, 199.

Page, R. B., Monaghan, J. R., Samuels, A. K., Smith, J. J., Beachy, C. K., & Voss, S. R. (2007). Microarray analysis identifies keratin loci as sensitive biomarkers for thyroid hormone disruption in salamanders (Ambystoma). Comparative Biochemistry and Physiology Part C: Toxicology and Pharmacology, 145, 15–27.

Page, R. B., Monaghan, J. R., Walker, J. A., & Voss, S. R. (2009). A model of transcriptional and morphological changes during thyroid hormone-induced metamorphosis of the axolotl. General and Comparative Endocrinology, 162, 219–232.

Page, R. B., Voss, S. R., Samuels, A. K., Smith, J. J., Putta, S., & Beachy, C. K. (2008). Effect of thyroid hormone concentration on the transcriptional response underlying induced metamorphosis in the Mexican axolotl (Ambystoma). BMC Genomics, 9, 78.

Power, D. M., Llewellyn, L., Faustino, M., Nowell, M. A., Bjornsson, B. T., Einarsdottir, I. E., et al. (2001). Thyroid hormones in growth and development of fish. Comparative Biochemistry and Physiology. Part C: Toxicology and Pharmacology, 130, 447–459.

Prahlad, K. V. (1968). Induced metamorphosis: Rectification of a genetic disability by thyroid hormone in the Mexican axolotl, Siredon mexicanum. General and Comparative Endocrinology, 11, 21–30.

Prahlad, K. V., & DeLanney, L. E. (1965). A study of induced metamorphosis in the axolotl. The Journal of Experimental Zoology, 160, 137–145.

Riddiford, L. M. (1996). Molecular aspects of juvenile hormone action in insect metamorphosis. In Gilbert, L. I., Tata, J. R., Atkinson, B. G., (Eds). Metamorphosis: postembryonic reprogramming of gene expression in amphibian and insect cells. (pp. 223-251) San Diego: Academic Press.

Ritchie, J. W., Shi, Y. B., Hayashi, Y., Baird, F. E., Muchekehu, R. W., Christie, G. R., et al. (2003). A role for thyroid hormone transporters in transcriptional regulation by thyroid hormone receptors. Molecular Endocrinology, 17, 653–661.

Rosenkilde, P., & Ussing, A. P. (1996). What mechanisms regulate neoteny and induced metamorphosis in urodeles? The International Journal of Developmental Biology, 40, 665–673.

Ryan, T. J., & Semlitsch, R. D. (1998). Intraspecific heterochrony and life history evolution: Decoupling somatic and sexual development in a facultatively paedomorphic salamander. Proceedings of the National Academy of Sciences of the United States of America, 95, 5643–5648.

Ryan, T. J., & Semlitsch, R. D. (2003). Growth and the expression of alternative life cycles in the salamander Ambystoma talpoideum (Caudata: Ambystomatidae). Biological Journal of the Linnean Society (London), 80, 639–646.

Safi, R., Bertrand, S., Marchand, O., Duffraisse, M., de Luze, A., Vanacker, J. M., et al. (2004). The axolotl (Ambystoma mexicanum), a neotenic amphibian, expresses functional thyroid hormone receptors. Endocrinology, 145, 760–772.

Safi, R., Vlaeminck-Guillem, V., Duffraisse, M., Seugnet, I., Plateroti, M., Margotat, A., et al. (2006). Paedomorphosis revisited: Thyroid hormone receptors are functional in Necturus maculosus. Evolution and Development, 8, 284–292.

Scott, D. E. (1993). Timing of reproduction in paedomorphic and metamorphic Ambystoma talpoideum. The American Midland Naturalist, 129, 397–402.

Semlitsch, R. D. (1987). Paedomorphosis in Ambystoma talpoideum: Effects of density, food, and pond drying. Ecology, 68, 994–1002.

Semlitsch, R. D., Harris, R. D., & Wilbur, H. M. (1990). Paedomorphosis in Ambystoma talpoideum: Maintenance of population variation and alternative life history pathways. Evolution, 44, 1604–1613.

Sexton, O. J., & Bizer, J. R. (1978). Life history patterns of Ambystoma tigrinum in montane Colorado. The American Midland Naturalist, 99, 101–118.

Shaffer, H. B. (1984). Evolution in a paedomorphic lineage. I. An electrophoretic analysis of the Mexican ambystomatid salamanders. Evolution, 38, 1194–1206.

Shaffer, H. B., & Breeden, F. (1989). The relationship between allozyme variation and life history: Non-transforming salamanders are less variable. Copeia, 1989, 1016–1023.

Shaffer, H. B., & Voss, S. R. (1996). Phylogenetic and mechanistic analysis of a developmentally integrated character complex: Alternate life history modes in ambystomatid salamanders. American Zoologist, 36, 24–35.

Shi, Y. B. (2000). Amphibian metamorphosis: From morphology to molecular biology. New York: Wiley.

Smith, H. M. (1969). The Mexican axolotl: Some misconceptions and problems. BioScience, 19, 593–597.

Smith, H. M. (1989). Discovery of the axolotl and its early history in biological research. In G. M. Malacinski & J. B. Armstrong (Eds.), Developmental biology of the axolotl. New York: Oxford University Press.

Smith, J. J., Putta, S., Walker, J. A., Kump, D. K., Samuels, A. K., Monaghan, J. R., et al. (2005). Sal-Site: integrating new and existing ambystomatid salamander research and informational resources. BMC Genomics, 6.

Snyder, R. C. (1956). Comparative features of the life histories of Ambystoma gracile (Baird) from populations at low and high altitudes. Copeia, 1956, 41–50.

Song, F., Poljak, A., Crawford, J., Kochan, N. A., Wen, W., Cameron, B., et al. (2012). Plasma apolipoprotein levels are associated with cognitive status and decline in a community cohort of older individuals. PLoS One, 7, e34078.

Sprules, W. G. (1974). The adaptive significance of paedogenesis in North American species of Ambystoma (Amphibia: Caudata): An hypothesis. Canadian Journal of Zoology, 52, 393–400.

Swingle, W. W. (1922). Experiments on the metamorphosis of neotenous amphibians. The Journal of Experimental Biology, 36, 397–421.

Takahashi, M. K., & Parris, M. J. (2008). Life cycle polyphenism as a factor affecting ecological divergence within Notophthalmus viridescens. Oecologia, 158, 23–34.

Tassava, R. A. (1969). Survival and limb regeneration of hypophysectomized newts with pituitary xenografts from larval axolotls, Ambystoma mexicanum. The Journal of Experimental Zoology, 171, 451–458.

Taurog, A. C., Oliver, R., Eskay, L., Porter, J. C., & Mckenzie, J. M. (1974). The role of TRH in the neoteny of the Mexican axolotl (Ambystoma mexicanum). General and Comparative Endocrinology, 24, 267-279.

Teng, Y. W., Mehedint, M. G., Garrow, T. A., Zeisel, S. H. (2011). Deletion of betaine- homocysteine S-methyltranferase in mice perturbs choline and 1-carbon metabolism, resulting in fatty liver and hepatocellular carcinomas.

Teng, Y. W., Ellis, J. M., Coleman, R. A., Zeisel, S. H. (2012). Mouse betaine- homocysteine S-methyltransferase deficiency reduces body fat via increasing energy expenditure and impairing lipid synthesis and enhancing glucose oxidation in white adipose tissue. Joural of Biological Chemistry, 287, 16187-16198.

Thambisetty, M., Simmons, A., Velayudhan, L., Hye, A., Campbell, J., Zhang, Y., et al. (2010). Association of plasma clusterin concentration with severity, pathology, and progression in Alzheimer disease. Archives of General Psychiatry, 67, 739– 748.

Thomas, P.D., Campbell, M.J., Kejariwal, A., Mi, H., Karlak, B., Daverman, R., Diemer, K., Muruganujan, A., Narechania, A. (2003). PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 13, 2129–2141.

Tompkins, R., & Townsend, J. K. (1977). Control of metamorphic events in a neotenous urodele Ambystoma mexicanum. Journal of Experimental Zoology, 200, 191-196.

Tulla, M., Pentikäinen, O., Viitasalo, T., Käpylä, J., Impola, U., Nykvist, P., et al. (2001). Selective binding of collagen subtypes by integrin α1I, α2I, and α 10I domains. Journal of Biological Chemisty, 276, 48206–48212.

Ueno, T., Tanaka, K., Kaneko, K., Taga, Y., Sata, T., Irie, S., et al. (2010). Enhancement of procollagen biosynthesis by p180 through augmented ribosome association on the endoplasmic reticulum in response to stimulated secretion. Journal of Biological Chemisty. 285, 29941-29950.

Van der Laan, S., Lachize, S. B., Vreugdenhil, E., de Kloet, E. R., Meijer, O. C. (2008). Nuclear receptor coregulators differentially modulate induction and glucocorticoid receptor-mediated repression of the corticotropin-releasing hormone gene. Endocrinology, 149, 725-732.

Van Vliet, P. (2012). Cholesterol and late-life cognitive decline. Journal of Alzheimers Disease, 30, S147-S162.

Vlaeminck-Guillem, V., Safi, R., Guillem, P., Leteurtre, E., Duterque-Coquillaud, M., & Laudet, V. (2006). Thyroid hormone receptor expression in the obligatory paedomrophic salamander Necturus maculosus. The International Journal of Developmental Biology, 50, 553–560.

Voss, S. R., Kump, K. D., Walker, J. A., Shaffer, H. B., & Voss, G. J. (2012). Thyroid hormone responsive QTL and the evolution of paedomorphic salamanders. Heredity, 109, 293–298.

Voss, S. R., & Shaffer, H. B. (2000). Evolutionary genetics of metamorphic failure using wild-caught versus laboratory axolotls (Ambystoma mexicanum). Molecular Ecology, 9, 1401–1408.

Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58, 236-244.

Weins, J. J., & Hoverman, J. T. (2008). Digit reduction, body size, and paedomorphosis in salamanders. Evolutionary Development, 10, 449-463.

Wells, K. D. (2007). The ecology and behavior of amphibians. Chicago: University of Chicago Press.

Whiteman, H. H. (1994). Evolution of facultative paedomorphosis. The Quarterly Review of Biology, 69, 205–221.

Whiteman, H. H., Wissinger, S. A., Denoel, M., Mecklin, C. J., Gerlanc, N. M., & Gutrich, J. J. (2012). Larval growth in polyphenic salamanders: Making the best of a bad lot. Oecologia, 168, 109–118.

Wilbur, H. M. (1980). Complex life cycles. Annual Review of Ecology and Systematics, 11, 67–93.

Wilbur, H. M., & Collins, J. P. (1973). Ecological aspects of amphibian metamorphosis. Science, 182, 1305–1314.

Yamaguchi, Y., Mann, D. M., Ruoslathi, E. (1990). Negative regulation of transforming growth factor-h by the proteoglycan decorin. Nature 346, 281 – 284.

Yau, J. L., & Seckl, J. R. (2012). Local amplification of glucocorticoids in the aging brain and impaired spatial memory. Frontiers in Aging Neuroscience, 4, 24.

Zahn, J. M., Poosala, S., Owen, A. B., Ingram, D. K., Lustig, A., Carter, A., et al. (2007). AGEMAP: a gene expression database for aging in mice. PLoS Genetics, 3, 2326 – 2337.

Zambrano, L., Vega, E. E., Herrera, M., Prado, E., & Reynoso, V. H. (2007). A population matrix model and population viability analysis to predict the fate of endangered species in highly managed water systems. Animal Conservation, 10, 297–303.

VITA Carlena K. Johnson

PLACE OF BIRTH: Unadilla, New York

EDUCATION: Bachelor of Art, Biology, 2007 Hartwick College, Oneonta, NY

PROFESSIONAL POSITIONS HELD: Research Assistant, June 2005-July 2005 Dept. of Biology, Hartwick College Research Assistant, June 2006-August 2006 Dept. of Biology, Hartwick College Wetland/Turtle Intern, April 2008-December 2008 Hudsonia Ltd., Red Hook, NY Research Assistant, August 2010-July 2012 Dept. of Biology, University of Kentucky Research Assistant, August 2012-December 2012 Kentucky EPSCoR, University of Kentucky Teaching Assistant, January 2013-April 2013 Dept. of Biology, University of Kentucky

SCHOLASTIC AND PROFESSIONAL HONORS: Frank G. Brooks Award, April 2007 BBB, Hartwick College Daniel R. Reedy Quality Achievement Fellowship, August 2010-April 2012 Dept. of Biology, University of Kentucky Academic year Fellowship, August 2012-December 2012 Dept. of Biology, University of Kentucky