On the nature of the horny scales of the pangolin

J. Linn. Soc. (Zool.), 46, 310, p. 267 With 1 figure Printed in Great Britain Spril, 1967 On the nature of the horny scales of the pangolin BY R. I. C. SPEARMAN, PH.D., F.L.S. Department of Dermatology, University College Hospital Medical School, London, W.C.l (Accepted for publication November 1966) CONTENTS PAGE Introduction.. 267 Materials and methods.. 268 Observations.. 269 The dorsal plate.. 269 The intermediate plate.. 270 The ventral plate.. 270 Discussion.. 270 Growth of the scale.. 270-4bsence of a filamentous structure... 270 The association of a granular layer with thc hair follicles.. 271 Comparison of pangolin scales with reptilian scales. 271 Comparison of pangolin scales with primate nails.. 271 Phylogenetic relationships of pangolin scales.. 272 References.. 272 SUMMARY The pangolin scale is a horny derivative of t,lie epidermis. It is complex in structure and is divisible into three distinct regions. The dorsal plate forms approximately one-sixth of the scale thickness. It is composed of flattened solid keratinized cells without basophilic nuclear remnants. This region tends to fray easily. The dorsal plate contains bound phospholipids and sulphydryl groups but is weak in disulphide bonds. The bulk of the scale is made up of the intermediate plate formed of less flattened cells without basophilic nuclei. This region is rich in disulphide bonds but contains no appreciable bound phospholipids or sulphydryl groups. The ventral plate is only a few cells thick and is rich in bound phospholipids, which also occur in the underlying scale bed epidermis. These three regions of the scale are formed from separate epidermal germinal areas which do not develop a granular layer. Keratohyalin granules are, however, formed in the epidermis between the scales. It is suggested on the basis of histological structure and dishribution of chemical constituents that pangolin scales are probably homologous with primate nails. Evidence against the views that they are homologous with reptilian scales or are derived from compressed hairs is presented. INTRODUCTION The Pholidota is a small order of eutherian mammals comprising only the scaly anteaters, or pangolins. All are inhabitants of tropical and subtropical regions (Walker, 1964; Beddard, 1902). There is only one Family, Manidae, containing the recent genus Manis. Seven living species of pangolins have been described, four of which are found

268 R. I. C. SPEARMAN, F.L.S. in Africa south of the Sahara. Three different species occur in India and South-east Asia (Bartholomew, Eagle Clarke & Grimshaw, 1911). Most pangolins are terrestrial burrowing animals but some species are arboreal. They are all nocturnal and inhabit tropical forests and thick bush or savannah country where they feed mainly on ants and termites. The Pholidota are probably not closely related to the South American anteaters, the armadillos, or the African aardvarks. Resemblances between these insectivorous mammals are now thought to be due to evolutionary convergence. Fossil pangolins are known from the Eocene period (Bartholomew et al., 1911 ; Grass&, 1955; Walker, 1964). The most characteristic feature of the Pholidota is the outer keratinized covering of the body consisting of large flat overlapping scales. These are epidermal in origin and they occur over the entire dorsal part of the animal including the head and flanks and on the outer surfaces of the limbs and over the tail. They are not found on the face or over the underside of the body, which have a sparse coat of hair. Hairs grow out from behind each scale only in the Asiatic species (Grass&, 1955). The rigid horny scales arise from a fairly narrow epidermal germinal base and they can possibly be moved to some extent by attached muscles (Walker, 1964). The scales are largest on the back, where they project backwards towards the tip of the tail. When the animal curls up, the sharp free edges of the scales are directed outwards as in a half-opened fir cone (Walker, 1964). Two main views have been proposed on the nature of pangolin scales. One view was that the scales are composed of hairs cemented together (see Flower & Lpdekker, 1891). An opposing view of Weber (1904) suggested that pangolin scales mere homologous to reptilian scales. In the present paper evidence is presented for a third view, that pangolin scales are homologous with claws or nails normally found only on the digits. The histological structure of the scales of Manis tetradactyla and of Manis pentadactyla were examind and the distribution of certain chemical constituents were compared with that in lizard scales. rodent tail scales, avian tarsal scales and human nails (Spearman, 1964, 1966). MATERIAIS AND METHODS The smaller and much less tough scales from the flank were chosen for histological examination. Formalin fixed and dried specimens were examined. The scale was removed with the surrounding epidermis and underlying dermis. Paraffin wax blocks (m.p. 58 C) were prepared after dehydration in ethanol and clearing in cedar wood oil and benzene. Sagittal sections were cut on a sledge microtome after hardening the wax block with solid carbon dioxide. The sections were stained in Ehrlich s haematoxylin and eosin for general histology. Preparations were also made for fluorescence microscopy, using an aqueous staining mixture of 2 parts of 0.02% Congo red and 1 part of0.1% Titan yellow. Thioflavine T (O*lyo aqueous solution) was used as a nuclear counterstain. The slides were washed and taken through ethanol and xylene to a non-fluorescent mountant DePeX. The sections were then examined microscopically, under ultraviolet light. This technique has been fully described by Jarrett, Spearman & Hardy (1959) and Jarrett & Spearman (1964). Keratinized cells fluoresce mainly blue or red by this method and nucleic acids present fluoresce yellow with thioflavine T. Living epidermal cells mostly fluoresce red with yellow nuclei. Disulphide bonds of cystine in keratin are not directly demonstrable by any method. We used the oxidation technique with 3% peracetic acid (Jarrett & Spearman, 1964; Pearse, 1960). Sections were afterwards washed and then stained in O.lyo methylene blue or in 0.1% thioflavine T. The methylene blue preparations were examined under ortlinary light and the thioflavine T stained sections were viewed by ordinary light and U.V. light. The staining distribution was compared with control sections previously immersed in acidified water. Since thioflavine T in addition to reacting with the product

Horny scales of the pangolin 269 of cystine oxidation also stains with nucleic acids, these normally have to be removed enzymatically prior to testing for disulphide bonds. However, the keratinized pangolin scale contained no detectable nucleic acids and so nucleases were not used. Bound sulphydryl groups of cysteine were shown by the method of Barrnett & Seligman (1952) using dihydroxy-dinaphthyl-disulphide. This technique depends on the reaction of naphthols with sulphydryl groups and the subsequent demonstration of the naphthol by azo-coupling, with fast blue B salt (Pearse, 1960; Bruce Casselman, 1962). A blue colour occurs in regions rich in sulphydryl groups. Bound phospholipids were shown by the acid haematin method modified from Baker (1944, 1946) and fully described by Jarrett et al. (1959) and Jarrett, Spearman, Riley & Cane (1965). Sections were chronied in a dichromate and calcium chloride mixture, then stained in acid haematin and differentiated in a ferricyanide mixture. A blue acid haematin lake was formed in regions rich in bound phospholipids. OBSERVATIONS The scales in both species appeared similar. The pangolin scale was found to be divisible into three distinct regions with different histological and histochemical characteristics (Fig. 1). These regions I have termed the dorsal plate, the intermediate plat,e, and the ventral plate from the three layers of the primate nail, which it closely resembles (Lewis, 1954; Jarrett & Spearman, 1966). The horny cells composing the pangolin scale are derived from the underlying germinal matrix epidermis. The free edge of the scale forms about half the total length. The dorsal surface of the scale is ridged longitudinally and tends to fray near its anterior base. A granular layer occurs in the epidermis between the scales but one was not found in the scale germinal epidermis. 5 b Fig. 1. Horny epidermal scale of Manis tetradactylrt, sagittal section, showing its relation to the germinal matrix and surrounding epidermis. Arrows represent directions of scale growth. Parallel hatching shows regions rich in bound phospholipids. The granular layer is stippled. d, Dorsal plate; i, intermediate plate; v, ventral plate; s, stratum corneum, e, free edge of scale; b, scale bed epidermis; g, intervening epidermiv with granular layer; c, corium. The dorsal plate This appears to be formed from the anterior region of the germinal epidermis and forms about one-sixth of the scale thickness. It grades into the horny layer formed between the

270 R. I. C. SPEARMAN, F.L.S. scales which, by the techniques used, was histochemically similar to the dorsal part of the scale. The epidermal cells keratinize and rapidly lose their nuclear staining as they are pushed towards the free edge of the scale. The fully keratinized cells are flattened and solid, and are arranged in layers which fray and tended to flake when sectioned on the microtome. The dorsal plate stained pink with eosin and fluoresced blue with patches of red, with Congo red. It stained strongly blue for phospholipid and there was a moderate reaction for bound sulphydryl groups. In contrast, the keratin reacted only weakly for disulphide bonds. No bound sulphydryl groups, disulphide bonds, or bound phospholipids were detected in the adjacent germinal epidermis. The intermediate plate This is the thickest region of the scale and it appears to be derived from the germinal epidermis between that of the dorsal plate matrix and the scale bed. As in the dorsal plate, the keratinized cells of the intermediate plate had lost all their basophilic nuclear staining but they appeared much less flattened and, although solid in structure, they remained practically unstained with eosin. The cell boundaries were seen when the microscope light contrast was increased but the cell membranes were hardly stained. This part of the scale fluoresced mainly blue with Congo red. No appreciable bound phospholipids or bound sulphydryl groups were found but there was a strong reaction for disulphide bonds. The germinal epidermis was similar to the dorsal plate epidermis. The ventral plate This is a thin layer of slightly flattened cells, only one or two cells in depth, which forms the base of the scale. The solid cells had lost their nuclear staining and stained pink with eosin and fluoresced blue with Congo red. The amounts of disulphide bonds and bound sulphydryl groups present appeared to be similar to the intermediate plate. However, in contrast to the intermediate plate, this layer reacted strongly for bound phospholipids, especially around the cell peripheries. The germinal nail bed epidermis contained cells which were closely palisaded together. This was in contrast to the dorsal and intermediate regions of the germinal epidermis where the cells appeared less closely packed. The uppermost cells of the germinal epidermis or scale bed, beneath the scale plate showed a strong phospholipid reaction in contrast to the epidermis in other regions. Considerable fraying of the ventral plate occurred under the free edge of the scale. DISCUSSION Growth of the scale The arrangement of the horny cell layers in the pangolin scale with their different chemical constituents suggests that the dorsal, intermediate and ventral nail plates are derived from separate areas of the germinal epidermis. Growth and kerat,inization of these epidermal germinal cells replaces the loss through wear at the free edge of the scale. Absence of a filamentous structure There was no evidence of a filamentous structure as would be expected if the scales were formed of compressed hairs as previously suggested (Flower & Lydekker, 1891). Horny filaments have been clearly shown in some keratinized structures such as the rhinoceros horn (Ryder, 1962) and in the horse s hoof (Rudall, 1957; Trautman & Fiebiger, 1960). In contrast, the arrangement of the pangolin scale horny cells in layers is quite different. The bulk of the scale is formed by the strongly disulphide bonded intermediate portion which is many cells in depth. The fraying of the exposed dorsal and ventral parts

Horny scales of the pangolin 27 1 of the scale, which are composed of layers of flattened cells, may have been mistaken for a fibrous component. The association of a granular layer with the hair follicles A granular layer has been found only in mammalian epidermis and its appearance is connected with a new form of keratinization in mammals (Spearman, 1966). Previously I put forward the hypothesis that an epidermal cell layer containing cytoplasmic keratohyalin granules was evolved in close association with the hair follicles and sweat glands (Spearman, 1964). In support of this view it was shown that in rodent tails a granular layer is confined to the necks of the follicles. Also, in the hair seals which have a sparse covering of fur, a granular layer is most prominently developed around the follicles and is often absent in intervening areas (Sokolov, 1960a; Montagna & Harrison, 1957). In the practically hairless Cetacea a granular layer is not developed (Spearman, 1966; Sokolov, 1960b). Furthermore, in foetal sheep epidermis these granules first appear in the hair follicle necks (Hardy & Lyne, 1956). Therefore, the absence of a granular layer in the germinal epidermis of the pangolin scale is further support for the view that it is not formed from agglutinated hairs. Comparison of pangolin scales with reptilian scales The horny scales of the Squamata, the tarsal scales of birds, and mammalian tail scales, such as those of rodents and marsupials, show several common features in their development and composition. None develop from a granular layer and all are composed of flattened solid horny cells which contain keratin rich in both disulphide bonds and bound sulphydryl groups. These scales also contain appreciable bound phospholipids, probably attached to the keratin molecule side chains. They all fluoresce blue when stained by Congo red. The underlying reasons why some horny cells fluoresce blue and others red is, however, not understood (Jarrett et al., 1959; Jarrett & Spearman, 1964; Jarrett et al., 1965; Spearman, 1966; Maderson, 1965a, b). The similar histological and histochemical features in these various scales suggest that they are probably homologous (Spearman, 1964). The distribution of chemical constituents in the pangolin scale is different to the reptilian type of scale. Furthermore, it is a more complex structure divided into layers, apparently derived from different epidermal germinal centres. Although the snake scale is also heterogeneous in structure it clearly grows from a single germinal epidermis and differences are explainable by cyclical changes in keratinization which occur (Spearman, 1966; Maderson, 1965a, b). The pangolin scale at different stages of growth has so far not been studied, but the possibility of a reptilian-like cycle occurring seems unlikely from its structure. Comparison of pangolin scales with primate nails The structure of the pangolin scale, although different from reptilian scales, appears similar to that of the primate nail (Lewis, 1954; Jarrett & Spearman, 1966). Thus, in both the human nail and pangolin scale there is no underlying granular layer, and there are also similarities in chemical composition of the nail and scale. The nail, like the pangolin scale, can be divided into three distinct layers. Similarly, many more disulphide bonds occur in the intermediate part of the nail than in the dorsal plate and more bound sulphydryl groups occur in the dorsal portion. Probably increased disulphide bonding helps to toughen keratin (Spearman, 1966). In both the pangolin scale and the primate nail appreciable bound phospholipid occurs only in the dorsal and ventral plates. In contrast to the reptilian type of scale, practically no bound phospholipid was found in

272 R. I. C. SPEARMAN, F.L.S. the intermediate plate of either the pangolin scale or primate nail. This intermediate region forms the bulk of both the primate nail and the pangolin scale. In both human nails and pangolin scales the germinal epithelium of the ventral nail plate contains unmasked bound phospholipid, which is unusual in living epidermal cells. The pangolin scales differ from human nails in that the latter contain a small amount of bound phospholipid around the peripheral cell membranes in the intermediate plate region, but none was demonstrable in this zone in the pangolin scale. Basophilic nuclear remnants were just weakly detectable in nail cells but were not seen in the pangolin scales. The histological and histochemical evidence, therefore, suggests that pangolin scales are not homologous with the reptilian type scale, as suggested by Weber (1904) but they may be homologous with nails. It, follows that rodent and marsupial tail scales, which closely resemble reptilian scales, are probably also not closely related to pangolin scales (Spearman, 1964). Primate nails were derived from claws which are ancient structures first developed in Amphibia and which reached their maximum development in reptiles (Noble, 1931 ; Biedermann, 1926). Normally these horny epidermal appendages are confined to the tips of the digits but there seems to be no reason why they should not have developed elsewhere. If, as seems probable, pangolin scales are homologous with nails, extra digital claws must have evolved at least once in the course of evolution and they were presumably retained in the pangolin, due to the selective advantage a scaly armour afforded against predators. The pangolin scales probably appeared quite late in evolution, when the original reptilian scales had already been replaced by a hairy epidermis which is retained on the belly of the animal. The alternative possibility, that the similarities in structure between the pangolin scales and primate nails are merely due to evolutionary convergence, cannot be entirely ruled out. However, although this might explain superficial similarities, the number of common features suggests that this is not the explanation. Moreover, the pangolin scale, claws and nails now serve quite different functions and one would not expect convergence to have occurred. Phylogenetic relationships of pangolin scales The modern concept of homology is on a genetic basis, and it is suggested that the pangolin scale and primate nail show similar inherited patterns of development, both being derived from claws normally developed on the digits. Only the Pholidota appear to have evolved this type of scale as a protective covering to the body. The horny scales of the armadillo are formed over a granular layer and are evidently quite different structures (Cooper, 1930). In the echidna a similar protective function is taken over by the quills which are thickened hairs. ACKNOWLEDGEMENTS I am grateful to Dr A. Jarrett and Dr P. A. Riley for their helpful criticism. The pangolin material was kindly supplied by the Hunterian Museum of the Royal College of Surgeons and the British Museum (Natural History). REFERENCES BAKER, J. R., 1944. The structure and chemical composition of the Golgi element. Q. J1 microsc. Sci., 85: 1. BAKER, J. R., 1946. The histochemical recognition of lipine. Q. JZ microsc. Sci., 87: 441. BARRNETT, R. J. & SELICMAN, A. M., 1952. Histochemical demonstration of protein bound sulphydryl groups. Science, N. Y., 16: 323-327. BARTEOLOMEW, J. G., EAGLE CLARKE, W., tk GRIMSHAW, P. H., 1911. Atlas of Zoogeography. Bartholomew, Edinburgh.

Horny scales of the pangolin 273 BEDDARD, F. E., 1902. Mainmalia, in The Cambridge h'atural History, 10, Macmillan, London. BIEDERMANN, W., 1926. Vergleichende Physiologie des Integuments der Wirbeltiere. Ergebn. Biob., 1 : 1-342. BRUCE CASSELMAN, W. G., 1962. Histochemical technique. Methuen, London. COOPER, Z. K., 1930. Ahistological study of tha integument of the armadillo. Am. J. Anat., 45: 1-37. FLOWER, W. H. & LYDEKKER, B. A., 1891. An introduction to the study of mammals living and extinct. p. 206. A. & C. Black, London. GRASSE, P. P. 1955. Blammifdres. Trait6 de Zoologic, 17: 1268-1284. Masson, Paris. HARDY, M. H. & LYNE, A. G., 1956. Prenatal development of wool follicles in Merino sheep. Bust. biol. Sci., 9: 423-448. JARRETT, A. & SPEARMAN, R. I. C., 1964. Histochernistry of theskin: Psoriasis. A Monograph 018 normal and parakeratotic epidermal keratinization. Eng. Univ. Press, London. JARRETT, A. & SPEARMAN, R. I. C., 1966. Histochemistry of thehuman nail. Archs. Derm. 94: 652-657. JARRETT, A,, SPEARMAN, R. I. C., & HARDY, J. A., 1959. Histochemistry of keratinization. Br. J. Derm., 71: 277-295. JARRETT, A,, SPEARMAN, R. I. C., RILEY, P. A,, & CANE, A. K., 1965. The distribution of epidermal phospholipids and their relation to alkaline phosphatase activity of the granular layer. J. invest. Derm., 44: 311-319. LEWIS, B. L., 1954. Microscopical studies of foetal and mature nail and surrounding tissue. Archs. Derm., 70: 732-747. MADERSON, P. A,, 1965a. The structure and development of the squamate epidermis. In Biology of the Skin and Hair Growth (Ed. A. G. Lyne and I3. F. Short), pp. 129-153. Angus 8: Robertson, Sydney. MADERSON, P. A., 1965b. Histological changes in the epidermis of snakes during the sloughing cycle. Proc. zool. SOC. Lond., 146: 98-113. MONTAGNA, W. & HARRISON, R. J., 1957. Specializations in the skin of the seal Phoca vitulina. Am. J. Anat., 100: 81-114. NOBLE, G. K., 1931. The Biology of the Amphibin. Dover, h'ew York. PEARSE, A. G. E., 1960. Histochemistry, theotetica1 aid applied. 2nd Ed. Churchill, London. RUDALL, K. M., 1957. The keratinization of horn. Proc. li'irst Int. Wool Textile Res. Conf. Atistralia, F. 176. RYDER, M. L., 1962. Structure of rhinoceros horn. A'oture, Lond., 193: 1199-1201. Soso~ov, W., 19FOa. The skin structure in Pinnipedia of the U.S.S.R. fauna. J. Morph., 107: 285-296. SOKOLOV, W., 1960b. Some similarities and dissimilarities in the structure of the skin aniong members of the sub-orders Odontoceli and Mystacoceti; Cetacea. Nature, LO?^., 185 : 745-747. SPEARMAN, R. I. C., 1964. The evolution of mammalian keratinized structures. In Mamnialian epidermis and its derivatives. Ed. F. J. Ebling. Symp. Zoo. SOC. Lond., 12: 67-81. SPEARMAN, R. I. C., 1966. The keratinization of epidermal scales, feathers and hairs. Biol. Rev., 41 : 59-96. TRAUTMANN, A. & FIEBIGER, J., 1960. Fundrrnic,ifnls of the histology of domestic u)timals (from 9th German edit.). Baillidre Tindall & Cox, London. WALKER, E. P., 1964. Mammals of the World. 1, 504. John Hopkins, Baltimore. WERER, M., 1904. Die Saugetiere. Gustav Fischer, Jena, 420-422. 18