The pigmentary system of developing axolotls: IV. An analysis of the axanthic phenotype

The axanthic mutation in axolotls was first described by Lyerla & Dalton in 1971. Axanthic is a simple Mendelian recessive trait. Axolotls homozygous for this mutation appear normal except for lack of visible xanthophores and iridophores (Lyerla & Dalton, 1971). Axanthic animals also lack the yellow pteridine pigments that are characteristic of normal xanthophores.

Subsequent to the initial description of the axanthic phenotype, Dalton & Hoerter (1974) examined purine synthesis in wild-type, melanoid and axanthic axolotl skin in an effort to determine why iridophores fail to develop in either melanoid or axanthic axolotls. They concluded only that the chromatographic patterns of purines observed in these three types of animals were different; however, they were unable to identify positively any of the purines. Thus, little additional information was gained which might have contributed to further defining either the melanoid or axanthic defects. Moreover, it seems unlikely that more will be learned about purine pigment synthesis and iridophore differentiation until the process is better understood in wild-type animals.

Dunson (1974) searched for structural evidence of either xanthophores or iridophores in axanthic skin using transmission electron microscopy. Iridophores were not observed anywhere in axanthic skin. Cells that were present in ‘positions occupied by xanthoblasts in wild type larvae…’ (Dunson, 1974) were identified as potential xanthoblasts in axanthic skin. These cells generally lacked distinctive organelles, especially organelles resembling pterinosomes (characteristic of xanthophores). Based on the apparent presence and characteristics of presumed xanthoblasts, Dunson speculated that the axanthic mutation might be responsible for either (1) an early block in xanthophore differentiation (prior to pteridine biosynthesis or pterinosome formation), (2) a block in pterinosome formation leading to failure of pteridines to accumulate or (3) a block in pteridine biosynthesis leading to a failure of pterinosomes to form.

The goal of this study was to define more carefully and completely the chemical and structural features of skin in developing axanthic axolotls, and to determine which, if any, of Dunson’s (1974, see above) speculative explanations for the axanthic defect might be valid. We also report two additional significant findings regarding the axanthic phenotype. (1) Defective xanthophores are clearly present in axanthic skin. They are fragile cells, identifiable on the basis of location and morphological criteria, such as the presence of pterinosome-like organelles in the cell processes. (2) Some cells in the dermis (and epidermis) of axanthic axolotls are heavily contaminated with a virus. The virus particles are most often packaged in lysosome-like vesicles within cells that appear to be macrophages. Occasionally, in older axanthic animals, virus is also observed to be packaged within vesicles (resembling pigment granules) in melanophores. Finally, we suggest that the axanthic defect is most likely to be a defect associated with pteridine/purine pigment biosynthesis.

Axolotls homozygous for the axanthic gene (ax) were obtained from the Indiana University Axolotl Colony, Bloomington, Indiana. Feeding, maintenance and the categorization of axolotls into three arbitrary age classes (larva, juvenile, adult) are described in Frost et al. (1984a).

Axolotl skin was prepared for transmission electron microscopy (TEM) as described previously (Frost, Epp & Robinson, 1984a). The fixative used was 2·5% glutaraldehyde in 0·1 M-cacodylate buffer, pH 7·4, with postfixation in 2% osmium tetroxide in the same buffer, and en bloc staining in 2% aqueous uranyl acetate.

Axanthic axolotls, by definition are pteridine deficient (Lyerla & Dalton, 1971). To confirm this, axolotl skin was extracted in 70% ethanol and analysed by thin-layer chromatography (TLC) as described in detail by Frost & Bagnara (1978) and Frost et al. (1984a). TLC plates were coated with a mixture of cellulose: silica gel G (1:1). The solvent used for pigment separations in one dimension was n-propanol: 7 % ammonia (2:1, v/v). Pigments were identified by comparing u.v.-fluorescent properties (colour) and chromatographic mobility with similar data from commercially purchased standards (see Frost et al. 1984a for a list of the pteridine pigments that have been identified in wild-type axolotls and for which we searched in axanthics).

The only well-differentiated pigment cell types observed in axanthic skin are melanophores (Fig. 1). In young larval axanthic axolotls, melanophores are localized beneath the basement membrane of the epidermis, surrounded by a loose collagen matrix (Fig. 1A,B). In older axanthics, the collagen layer is much thicker and melanophores are often observed embedded within the collagen as shown in Fig. 1C. Morphologically, melanophores from axanthic skin are similar to those described from wild-type skin (Frost et al. 1984a). ‘Mature’ melanophores (like the one shown in Fig. 1C) contain primarily pigment granules (melanosomes), occasional mitochondria, few premelanosomes and little else in the way of cell organelles (Fig. 1C). In young (larval and young juvenile) axanthics melanophores are generally ‘less differentiated’ and contain a variety of cytoplasmic organelles including numerous melanosomes, premelanosomes, large and small vesicles, extensive Golgi, some endoplasmic reticulum (ER), mitochondria, occasional microtubules and intermediate filaments (see Figs 1A,B, 2A,B). In addition to organelles that are characteristic of axolotl melanophores in general, axanthic melanophores also have been observed to have virus-containing organelles (Figs 1C, 2), some of which are nearly indistinguishable from pigment granules (melanosomes) (see the virus-containing particle (arrow) in Fig. 1C). Although clearly infected by a virus, the melanophores appear to be otherwise healthy (Fig. 2).

Evidence of xanthophores in axanthic skin is rare, but when such cells are encountered their morphology is distinctive (Fig. 3). The characteristics of the cells shown in Fig. 3 and those’of the cell processes that are parts of xanthophores (Fig. 4) include: numerous prepterinosome-like (large, empty) vesicles (the prepigment organelles of xanthophores), a cytoplasm typical of pigment cells and distinctly different from fibroblasts or blood cells (the only other cell types commonly encountered in axolotl dermis; see also Frost et al. 1984a,b, 1986), numerous small, electron-dense multivesicular bodies (MVBs), many small empty vesicles, mitochondria, smooth and rough ER (much of which is swollen), microtubules, intermediate filaments and occasional Golgi complexes that appear to be underdeveloped compared to those observed in melanophores (compare Golgi in Fig. 3B,C with that in Fig. 2A, for example). Axanthic melanophores also do not contain electron-dense MVBs, they exhibit little evidence of ER (no evidence of swollen ER) and have numerous melanosomes. (Compare Figs 1, 2 with Figs 3, 4.) Thus far, xanthophore cell bodies have only been observed to occur beneath the collagen layer of the dermis, and not within the collagen layer as is the normal case for melanophores in older axanthic animals and all pigment cells in older wild-type axolotls (Frost et al. 1984a). The positional location of axanthic xanthophores is typical of pigment cells in larval skin, wherein pigment cells are initially located beneath the dermis and subsequently migrate to just below the basement membrane of the epidermis (Frost et al. 1984a). It is possible, based on this observation, that axanthic xanthophores may be defective in their ability to migrate to their proper location.

In addition to melanophores and occasional xanthophores, a specific blood cell type is frequently encountered in the dermis of axanthic axolotl skin. Based on the morphology and contents of this cell, we suggest that it is a macrophage (Fig. 5). In axanthic skin macrophages are most often observed adjacent to or near melanophores (a position often occupied by xanthophores). Such cells contain numerous empty vesicles (presumed lysosomes) and a variety of vesicles containing electron-dense materials (Fig. 5A). Close inspection of the electron-dense vesicles reveals that they are filled with virus particles close packed in hexagonal array (Fig. 5B,C). Individual particles measure approximately 40 nm in diameter and are morphologically identical to the virus particles observed to occur occasionally in melanophores (Figs 1C, 2). Aside from the presence of numerous electron-dense virus-containing particles, the macrophage shown in Fig. 5A is similar in general features to the xanthophore shown in Fig. 3A. The macrophage, however, does not possess extensive cell processes as is typical of pigment cells.

Pigment extraction and analyses confirmed the earlier findings of Lyerla & Dalton (1971). There are no pteridine pigments detectable in axanthic skin at any stage during development. Riboflavin (a yellow compound) is, however, present in skin as shown on the thin-layer chromatogram in Fig. 6. Also, there appears to be more riboflavin in skin from older axanthic axolotls compared to that from younger animals.

From 1983 to 1985 we examined skin from 17 different axanthic axolotls. These animals represent a variety of ages and were from a number of different spawnings carried out during this two-year period. Every axanthic animal examined was found to be virus infected, some more severely so than others.

During the course of the past two years we also tried repeatedly to rear axanthic larvae to sexual maturity. From more than 50 axanthic larvae only one survived to attain adult secondary sexual characteristics (a female), and this animal was never bred successfully. The animal eventually died in apparently healthy condition and was subsequently found also to be virus infected. The Indiana University Axolotl Colony has also had similar difficulties in rearing axanthic animals to sexual maturity (F. Briggs, personal communication), and only recently have two homozygous axanthic adults been successfully bred by colony personnel.

It seems likely that the virus found in axanthics is responsible for the reduced survivability of axanthic animals. The virus does not, however, appear to be immediately lethal to these animals, because otherwise healthy-appearing axanthics have been found to be virus-infected. During our survey of the development of pigmentation in wild-type, melanoid, albino and axanthic axolotls (Frost et al. 1984a,b, 1986, and the present study), we encountered only one nonaxanthic, nonexperimental (see below) animal that was virus-infected (Frost & Robinson, unpublished data). This animal, phenotypically wild-type, was possibly ax/+, having arisen from a cross involving two ax/+ parents. Considering the large number of animals examined during these studies (>100 of all phenotypes over a 4-year period), it seems likely that the defect caused by the axanthic gene is severe enough that viral expression and this mutant phenotype are coincidental.

More recently, we have undertaken experiments designed to alter the pigment phenotype of wild-type axolotls by administering either allopurinol (Frost & Bagnara, 1979) or guanosine to young larval axolotls (see below). In a significant number of cases these drug-treated animals eventually developed virus infections (especially those animals receiving guanosine) (Frost & Robinson, unpublished data). Thus, it seems that all axolotls may carry the genetic information for virus expression, but only ax/ax animals and those animals that have been ‘stressed’ by drug treatment actually express the virus. Furthermore, it is intriguing that guanosine (a precursor to both purine and pteridine pigment biosynthesis) and allopurinol (an inhibitor of purine/pteridine synthesis; Spector, 1977) appear to allow the virus to be expressed.

The significance of the virus (which remains unidentified) to the axanthic phenotype is obscure at the present time. That it may, in fact, be linked, perhaps indirectly, to the expression of the axanthic gene is plausible in view of the fact that all axanthic animals express the virus. It is anticipated that future studies will be designed (1) to isolate and identify the virus and (2) to determine what factors control the expression of the virus.

With regard to axanthic xanthophores, the cytoplasmic contents of these cells differ distinctly from other pigment cells, including other xanthophores (see descriptions and figures of wild-type xanthophores: Frost et al. 1984a; melanoid xanthophores: Frost et al. 1984b; albino xanthophores: Frost et al. 1986), iridophores (Frost et al. 1984a, 1986) and melanophores (e.g. the present study) that we have observed. Based on these observations, it appears that pterinosomes do form in axanthic xanthophores, although no pigment is synthesized in such cells. Furthermore, axanthics appear to have all of the standard ‘cellular machinery’ found in other pigment cells although some organelles, notably the Golgi, ER and electron-dense MVBs, contribute to the ‘abnormal’ appearance of these cells.

In view of these observations, we assert that none of Dunson’s (1974) hypotheses regarding the underlying mechanism of the axanthic defect (see Introduction) are likely because these speculations were based on the presumed absence of pterinosome-like organelles in axanthic skin. This is clearly not the case, and we would thus offer the suggestion that the axanthic defect is one that blocks pigment biosynthesis (pteridines in particular) at a point very early in the biosynthetic pathway, but does not affect organelle formation. To our knowledge this is the first case in which pigment biosynthesis and organelle formation have been clearly separated from each other as independent cellular events in cells other than melanophores. [Albinism is the classic example wherein ‘unpigmented’ pigment cells occur (see Frost et al. 1986).]

Iridophores have not been observed in axanthic skin and may fail entirely to differentiate. Iridophores are also absent in melanoid axolotl skin (Frost et al. 1984b), and in neither case is there evidence to suggest why these cell types are missing. It may be that iridophores do differentiate in both mutant axolotl types but have not been identified. ‘Prereflecting platelets’ may not resemble mature reflecting platelets enough to facilitate identification of unpigmented iridophores (for an example of what a ‘prereflecting platelet’ might look like see Frost et al. 1986). Nevertheless, in both melanoid and axanthic axolotls, pteridine metabolism is abnormal (Frost et al. 1984a,b) and it is further known that both xanthophore and iridophore pigments derive from purine precursors (Frost & Malacinski, 1980). Consequently, we suggest that the synthesis of the purine precursor(s) may be the key to understanding these defects. The next obvious step towards gaining a better understanding of these defects is thus to examine iridophores more closely and attempt to characterize purine pigments precisely in terms of those purines that function as pigments and how they are synthesized.

It is well known that all pigment cells share a common embryonic origin (the neural crest) and it has been further hypothesized that the various types of pigment organelles also share a common origin, i.e. at the very least they all arise from a combination of Golgi and ER components (Bagnara et al. 1979). If this is true, then the different cell types might be suspected to arise as a result of differential synthesis of specific pigments. If the axanthic defect resulted simply from an inability to synthesize pteridines, then one would expect to observe more- or less-normal-appearing pigment cells with organelles devoid of pigment. This, in fact, seems to be the case; however, this simple explanation raises a great number of questions.

For example, given the ubiquitous nature of pteridines and their, role as essential enzymatic cofactors (Kaufman, 1967), could an animal survive if it were totally unable to synthesize these compounds? Are axanthic axolotls, in fact, totally unable to synthesize pteridines? Regarding the xanthophores, what is the significance of swollen ER, sparse Golgi and numerous electron-dense MVBs? Are these characteristics typical of a pigment cell unable to synthesize pigment? Also, why do unpigmented xanthophores occur only rarely in axanthic skin, and why do they fail to migrate up to their normal position just beneath the basement membrane of the epidermis? Finally, given the known plasticity of pigment cell differentiation (Ide, 1978, 1979), why do xanthophores (or their precursors) not simply differentiate into melanophores as we suspect may happen in melanoid animals (Frost et al. 1984b)? [This possibility is unlikely because melanophores have not been shown to be present in abnormally large numbers in axanthic skin (Lyerla & Dalton, 1971).]

At the present time very little is known about the actual mechanism of differentiation of pigment cell types. Why some chromatoblasts become melanophores while others become xanthophores is largely speculative. It is hoped that further investigations into the nature of the axanthic defect coupled with studies that are currently in progress in this laboratory whereby pigment phenotypes are being altered by specific drugs will eventually elucidate the mechanisms responsible for altering pathways of cell differentiation.

The authors wish to thank Fran Briggs (IU Axolotl Colony; recently deceased) for providing the animals used in all of these studies. We dedicate this series of four papers to her memory.

This work was supported by NSF PCM80-22599, NIH AM34478 and the KU Biomedical Research Fund. The Center for Biomedical Research (KU) provided equipment and facilities and Mrs Lorraine Hammer provided invaluable assistance with the electron microscopy.