- Email: [email protected]
The giant melanosome: A model of deranged melanosome-morphogenesis
Printed in Sweden Copyright © 1974 by Academic Press, Inc. All rights of reproduction in any form reserved
I02
J. ULTRASTRUCTURE RESEARCH 48, 102--123
(1974)
T h e G i a n t M e l a n o s o m e : A Model of D e r a n g e d M e l a n o s o m e . M o r p h o g e n e s i s ~,2 KLAUS KONRAD, KLAUS WOLFF and HERBERT HONIGSMANN
Division of Experimental Dermatology, Department of Dermatology (I.), University of Vienna, Alserstrasse 4, A-I090 Vienna, Austria Received November 20, 1973 A unique type of melanin granule, occurring in a nevoid pigmented lesion in man, is described. It grows to extraordinary size, it melanizes asynchronously and, in contrast to all other known melanosomes of mammals, it is essentially nonfilamentous in nature. Its basic structural components are 400 ~ microvesicles, which, during melanization, are gradually embedded in the melanin matrix. Evidence is presented that identical vesciles also occur in normal filamentous melanosomes of man. The giant melanosome is considered to result from a derangement of morphogenesis at a distinct stage of melanosome ontogeny. Melanosomes are both sites of production and carriers of melanin pigment in mammals. They are round or ellipsoidal and are composed of a membrane-delimited coil of filaments upon which melanin is deposited during the process of melanization (7, 27). In principle, this basic structure applies to melanosomes of all mammalian species, but minor variations of size, structure, and shape occur. These are genetically determined, and in some species it has been possible to delineate a number of the responsible loci (5, 12, 22, 23, 26). The melanosomes of man exhibit an identical filamentous structural skeleton and the differences to the melanosomes of other species are again mainly confined to shape (ellipsoidal vs round), size, and degree of melanization (7, lO). In this paper w e describe a unique type of melanosome, observed in humans, which exhibits an heretofore unrecognized nonfilamentous substructure; it melanizes asynchronously and possesses extraordinary potentials of growth. Hence, the designation giant melanosome. 1 This investigation was supported by Fonds zur F/Srderung der wissenschaftlichen Forschung, Vienna and Schering AG, Berlin. 2 Presented, in part, at the Annual Meeting of the European Society for Dermatological Research, Amsterdam, April 25, 1973.
THE GIANT MELANOSOME
103
M A T E R I A L AND METHODS
Material Giant melanosomes constitute a consistent feature of a nevoid pigmented lesion in man, designated nevus spilus (8, 15). These lesions are not uncommon and represent caf6 au lait spot-like macular hypergpigmentations that are dotted by small macules of darker color and thus exhibit a mottled apearance. Histologically, they show features of lentigo simplex, but they also contain small nests of junctional nevus cells. Five such nevi from 5 caucasoid patients, aged 19 to 40 years, were chosen as source of giant melanosomes to be studied in this investigation.
Methods Shave biopsies were performed under local anesthesia and the tissue was processed both for light microscopic histology and histochemistry, employing routine procedures, and for electron microscopy and ultrastructural cytochemisty, as described below. Electron microscopy. Tissue specimens were minced and fixed in a 1:1 dilution of Karnovsky's paraformaldehyde-glutaraldehyde fixative (14) for 5 hours at room temperature and were rinsed and stored overnight in 0.1 M cacodylate buffer pH 7.4 at 4°C. Postfixation was carried out with 3 % osmie acid in distilled water for 1.5 hours at 0-4°C followed by staining en bloc with 0.5 % uranyl acetate in 0.l M Veronal acetate buffer, pH 7.2, for 2 hours at room temperature. Some specimens were embedded subsequent to the aldehyde-fixation and were thus not subjected to osmification or treatment with uranyl acetate. All specimens were rapidly dehydrated in a graded series of ethanol and were embedded in Epon 812 according to a procedure modified from Luft (16). Semithick sections were cut with glass knives and screened for the presence of giant melanosomes. Ultrathin sections were then cut from blocks containing giant melanosomes using Dupont and Dehmer diamond knives and Reichert OM U2 or LKB Ultrotome III ultramicrotomes. The sections were contrasted with lead citrate and examined with a Zeiss EM 9S electron microscope operating at 60 kV. Approximately 200 giant melanosomes were studied. Ultrastructural cytoehemistry. Thin strips of skin, measuring 1 x 1 × 5 ram, were fixed in the 1 : 1 dilution of Karnovsky's paraformaldehyde-glutaraldehyde fixative (14) for 3 hours at room temperature and, after an overnight rinse in 0.1 M cacodylate buffer (pH 7.4) at 4°C, they were chopped on a Smith and Farquhar (29) tissue chopper (Ivan Sorvall, Inc.). The 40-#m sections thus obtained were incubated for tyrosinase and acid phosphatase activities, as detailed below. After incubation they were rinsed and processed for electron microscopy as described for A. Incubation for tyrosinase. 0.025 g of L-DOPA (Sigma Chemical Company) was dissolved in 50 ml of 0.1 M phosphate buffer (pH 7.4) (2). Incubation was performed at 37°C and was terminated at 2 hours, 4 hours, or 8 hours. The medium was freshly prepared and changed every hour. Ineubation for acid phosphatase. The sections were rinsed in 0.2 M Tris maleate buffer (pH 7.2) and were then incubated according to Barka and Anderson (1), using sodium /3-glycerophosphate as substrate, for 1 hour at 37°C. As controls for the enzyme reactions, sections were incubated in complete media from which the substrates had been omitted.
104
KONRAD, WOLFF AND HONIGSMANN
FrG. 1. Giant melanosomes at the light microscope level. (a) Several giant melanosomes are present in a melanocyte (M) and are, in part, aligned along a dendritic process (arrow). A neighboring basal keratinocyte (K) also contains a giant melanosome. H & E, x 1 000. (b) Several giant melanosomes within the perinuclear cytoplasm of a melanocyte. Notice that the large organelles are sharply defined, strictly spherical, and of homogeneous density. H & E, × 1 300. (e) In DOPA-incubated cryostat sections the cytoplasm of melanocytes (M) is stained dark due to the formation of DOPAmelanin; the giant melanosomes (arrow) appear intensely black. DOPA. x 1 300.
Comparative studies A large quantity of material from our files on normal and experimental pigmentation in humans, guinea pigs, and mice, studied in this laboratory over the past 8 years, was screened by electron microscopy for the presence of the microvesicles described in the Results section of this paper. RESULTS
Light microscopy I n sections cut f r o m tissue e m b e d d e d in paraffin :and eosin, the giant m e l a n o s o m e s a p p e a r as s h a r p l y t o golden granules of h o m o g e n e o u s density (Fig. a p p r o a c h t h a t of a red b l o o d cell they can be seen at
a n d stained with h e m a t o x y l i n defined, spherical, d a r k b r o w n 1 a, b). Since their size m a y low magnifications. T h e y occur
THE GIANT MELANOSOME
105
F~6. 2. At the ultrastructural level the giant melanosomes are easily recognized by their size, their spherical shape a~nd their high electron density. Three giant melanosomes (1, 2, 3) one of which has been cut tangentially (1) occur in the cytoplasm of this melanocyte, which appears larger than normal but shows no other abnormalities. Notice the simultaneous occurrence of normal melanosomes of all stages of maturation (arrows). × 10 800.
106
KONRAD, WOLFFAND HONIGSMANN
in melanocytes (Fig. 1 a-c) and nevus cells, and less frequently within keratinocytes (Fig. 1 a) and macrophages. Their occurrence is focal, and it is only a smaller proportion of the total population of pigment cells that exhibit these granules in their cytoplasm. In silver-stained sections the granules acquire a pitch-black color, as does melanin; in cryostat sections incubated in the DOPA-medium they also appear intensely black, which contrasts with their natural brown to golden color (Fig. 1 c).
Electron microscopy Occurrence and distribution. As observed at the light microscope level the giant melanosomes occur within the cytoplasm of melanocytes and junctional nevus cells and, less frequently, in keratinocytes of the epidermis and melanophages within the connective tissue. These melanocytes and nevus cells appear larger than normal but show no other abnormalities (Fig. 2). In particular, they exhibit all structural prerequisities for active melanogenesis and, apart from the giant melanosomes, contain normal melanosomes of all stages (34) and of normal shapes and dimensions (Fig. 2). The melanocytes often protrude deeply into the dermis; not infrequently, multiple giant melanosomes can be seen in a single pigment cell (Figs. 2 and 4). The keratinocytes in the vicinity of giant melanosome-bearing melanocytes contain occasional giant melanosomes and a large number of normal melanosomes which are distributed either singly or as membrane-delimited aggregates (Fig. 3). The giant melanosomes are usually singly dispersed, but occasionally the membranes of two adjacent giant melanosomes may fuse so that two giant granules are enclosed within one common envelope (Fig. 3). Rarely do such fusions occur also with normal melanosomes. Dermal melanophages very rarely contain giant melanosomes but when they do the giant melanosomes are enclosed within large phagocytic vacuoles which may also contain a variable number of normal melanosomes or other cellular constituents. Morphology. The most striking ultrastructural feature of fully developed giant melanosomes is their tremendous size (Fig. 4) as they may measure up to 5 #m or more in diameter, the ranges observed being 1-6/~m. They are almost always strictly spherical and exhibit a highly ordered structure: a central core is surrounded by several concentric shell-like layers; these, in turn, are delimited by a trilaminar membrane, 70 A in width (Figs. 4 and 5). In the fully developed giant melanosomes the central core is highly melanized and uniformly electron dense, and the more Fro. 3. Giant melanosomes in the cytoplasm of two basal keratinocytes. Usually, these granules are singly dispersed but occasionllythe membranes of two giant melanosomesmay fuse as indicated by the arrow. Normal melanosomes are also present and are distributed either as singles or as aggregates. × 12 900.
THE G I A N T M E L A N O S O M E
107
108
KONRAD, WOLFF AND HONIGSMANN
peripheral layers have a more granular appearance (Figs. 4 and 10-13) whereas the outermost cortical shell is composed of densely set microvesicles which measure 300-400 A in mean diameter (Fig. 5). Closer inspection reveals that these microvesMes are also present in the subcortical zones where they and irregular melanin deposits account for the granularity observed at low magnifications (Figs. 5, 7, and 8); and that they can be even observed in the melanized core of the giant melanosomes where they appear as discrete electron lucencies within the homogeneous melanin matrix (Fig. 4, 9, 10, and 11). The microvesicles appear to represent the structural skeleton of the giant melanosomes. No other substructures and, in particular, no melanofilaments which are so typical for all other melanosomes (6, 7, 9, 22, 23) have been observed in these organelles (Figs. 5, 7, 8, and 15). These vesicles are bound by a delicate trilaminar membrane approximately 70 ~ in width which encloses a moderately electron dense matrix (Figs. 5 and 6). The latter exhibits a substructure that gives the impression of cross-sectioned filaments of extremely fine caliber (Fig. 5). Although the large majority of these microvesicles are spherical, some appear oval and, very rarely, they may be present also as short tubular profiles (Figs. 6-8). The question therefore arises whether they do not represent cross-sectioned tubules. A tightly wound coil of tubules could, if cross-sectioned, appear vesicular in an ultrathin section, but tubules should become apparent at least in other planes of section or in sections cut at different angles. A large number of the giant melanosomes studied in this investigation were sectioned serially and subserially, but we always encountered mainly vesicular profiles. Also, if the vesicles in the cortical zone of a giant melanosome represented cross-sectioned tubules one would expect that their tubular nature should become evident in tangential sections of giant melanosomes. Such sections also revealed vesicles of identical structure and dimensions (Fig. 6.) The degree of melanization is never quite uniform in a given giant melanosome. It is most pronounced in the core and decreases toward the peripheral zone of the organelle. Figures 10-13 are shown to illustrate what we consider different stages during the progressive melanization of giant melanosomes. As these structures mature, their dense homogeneous core expands toward their periphery until, in the fully melanized giant melanosome, it reaches the delimiting membrane (Fig. 14a, b). Melanin is initially deposited around and between the microvesicles (Figs. 5, 7, and 15) which are thus gradually embedded in a dense homogeneous melanin matrix FIo. 4. Multiple giant melanosomes occurring ina melanocyte. Note their large size, their spherical shape, and ordered structure: a fully melanized central core (C) which exhibits discrete electronlucencies is surrounded by several concentric shell-like layers that have a more granular appearance; the outermost, cortical, shell is composed of densely set microvesicles. The arrows denote the delimiting membrane, x 26 400.
110
KONRAD, WOLFF AND HONIGSMANN
in which they may remain discernible as membrane-delimited, electron-lucent globular bodies (Figs. 5, 9, and 14a, b). These are slightly smaller than the microvesicles proper. In the subcortical zones, however, some microvesicles may often appear as if they contained the deposited pigment but, in these instances, their delimiting membrane cannot be discerned and since the vesicles are so small as to fit in toto into a silver to silver-gray section, they probably represent microvesicles coated by melanin (Fig. 5). As melanization progresses, the number of the electron lucent vesicular profiles gradually decreases in the core of the giant melanosomes, and in fully melanized granules all structure is obliterated by the dense homogeneous melanin deposits (Figs. 13 and 14a, b). In such giant melanosomes it is only in the cortical layers that microvesicles c a n still be discerned (Fig. 14a) although, very rarely, they may disappear also from these cortical regions. In the aldehyde-fixed specimens not subjected to osmification the melanized portions of giant melanosomes exhibit the same electron density as normal, fully melanized stage IV (34) melanosomes (Fig. 16). Again, the microvesicles inside the melanin matrix are revealed as small round electron lucencies (Fig. 16). Giant melanosomes with little or hardly any melanization are rarely observed. These are small, measuring approximately 1 # m in diameter, and exhibit densely set microvesicles within their center (Fig. 15) as do mature giant melanosomes in their cortical zones. The melanocytes and nevus cells that produce giant melanosomes also regularly exhibit small membrane-delimited vacuoles, 0,5 # m or less in diameter, which contain typical microvesicles but no melanin. They are often seen in the immediate vicinity of giant melanosomes (Figs. 5, 6, and 8).
Cytochemistry Tyrosinase. Melanocytes and nevus cells that contain giant melanosomes exhibit tyrosinase activity as does any other active melanogenic pigment cell. Reaction product is found in some G o l g i cisternae, in the Golgi-associated endoplasmic reticulum, in small coated vesicles, and in n o r m a l stage I to I I I melanosomes
FIG. 5. Higher magnification of the more peripheral layers of a giant melanosome. Immediately below the delimiting membrane (double arrows) there is a cortical shell (1) composed of densely set microvesicles (arrows). These vesicles are bounded by a delicate membrane and exhibit a matrix that gives the impression of cross-sectioned filaments of extremely fine caliber. The subcortical shell (2) also contains microvesicles (arrow) and irregular melanin deposits. The more central portions (3) of the giant melanosomes are composed of a dense melanin matrix in which individual vesicles remain discernible as electronlucent globular bodies, still delimited by a delicate membrane (arrow). In the immediate vicinity of the giant melanosome there is a small vacuole (V) which contains microvesicles but no melanin. × 83 600. FIG. 6. Tangential section of a giant melanosome revealing the cortical shell and the spherical profiles of microvesicles (arrow). There is one short tubular structure (T). A small vacuole (V) containing microvesicles is present in the vicinity of the giant melanosome, x 83 600.
THE GIANT MELANOSOME
8 - 741821 J. Ultrastructure Research
111
112
KONRAD, WOLFF AND HONIGSMANN
(Fig. 17). After 2 and 4 hours of incubation, no unequivocal reaction product is detected within giant melanosomes. However, when the incubation time is increased to 8 hours or more, reaction product is definitely revealed in the subcortical and cortical shells of giant melanosomes, which, owing to the deposition of D O P A melanin, become heavily melanized up to the delimiting membrane, which also obliterates the vesicular profiles (Fig. 19). In this investigation it was not possible to determine unequivocally the exact site of the enzyme, but in the less active giant melanosomes most of the finely granular reaction product was deposited outside, on the surface of and between the microvesicles (Fig. 18). Acid phosphatase. Acid phosphatase activity was not associated with giant melanosomes. As expected, it was present within melanosome complexes and lysosomes of keratinocytes and melanophages.
Comparative studies A careful search for microvesicles within normal melanosomes revealed that these structures do occur in stage II and I I I melanosomes of normal human melanocytes and of nevus cells (Fig. 20a-e). However, they were encountered only occasionally and their numbers were small. Fully melanized stage IV melanosomes from black h u m a n hair bulbs exhibited spherical electronlucent bodies, delimited by a delicate membrane, which were morphologically identical with the electronlucent vesicles within the melanized core of giant melanosomes (Fig. 20f). They were similar to the electronlucent bodies described some time ago in the iris epithelium of M, rhesus (36), and in highly pigmented hair bulbs of humans (13, 21).
DISCUSSION Abnormally large melanin granules are known to occur in the Chediak-Higashi syndrome, a rare genetic disease in man (3, 37), in Aleutian minks (17, 25), and in a beige strain of mice (12, 18). These granules are all variants of normal melanosomes since they exhibit the classical melanofilament structure and are thus basically different from the giant melanosomes described in this paper. Similar giant melano-
FIG. 7. Part of the cortical and subcortical shell of another giant melanosome. Among the spherical microvesicles there is a short tubular profile (T). In the subcortical shell melanin is deposited around and between the microvesicles (arrows). x 83 600. FIG. 8. Another example of a tubular profile (T) in the cortical shell which, otherwise, exhibits only microvesicles. A small vacuole (V) with microvesicles is seen in the vicinity of the giant melanosome. x 83 600. FIG. 9. This electron micrograph shows at a high magnification the melanized core in which the microvesicles remain discernible as membrane-delimited electron-lucent globular bodies (arrows). × 83 600.
T H E G I A N T MELANOSOME
1 13
114
KONRAD, WOLFF AND HI~NIGSMANN
THE GIANT MELANOSOME
l 15
F~G. 14. Fully melanized giant melanosomes. (a) Nearly all structure is obliterated by the dense homogeneous melanin deposits, and only in the cortical shell there are some electronlucent vesicular profiles (arrow). (b) In this giant melanosome, melanization has not quite reached the delimiting membrane. Again, a few microvesicles (arrow) can be noticed in the cortical zone. a, b: × 83 600.
FIGS. 10-13 are shown to illustrate the progressive melanization of giant melanosomes. The degree of melanization is most pronounced in the core (C) of the organelle and decreases toward the peripheral zones; as the giant rnelanosome matures, its dense homogeneous core expands toward the periphery (Fig. •3). The concentric zonal decrements of the deposited melanin indicate different bursts of melanogenic activity that appear to proceed in a centrifugal fashion (Figs. 10, 11). × 22 200.
116
KONRAD, WOLFF AND HONIGSMANN
Fie. 15. A small giant melanosome exhibiting densely set 400 A microvesicles. Melanization has already commenced in the center (C). × 83 600. FIG. 16. Specimen was fixed with glutaraldehyde only. The melanized portions of the giant melanosome exhibit the same electron density as normal, fully melanized, stage IV melanosomes. Note the electronlucent vesicular profiles (arrow) within the dense melanin. × 83 600.
FIG. 17. Specimen incubated in DOPA, for 4 hours at 37°C, to show tyrosinase activity. Reaction product is present in Golgi-associated endoplasmic reticulum (arrow) and in normal stage I and II melanosomes (M) of a junctional nevus cell. GM: giant melanosome, x 22 200. F~G. 18. Specimen incubated in the DOPA-medium for 8 hours. Finely granular reaction product is present in the cortical and subcortical shell of the giant melanosome. It is not possible to determine the exact site of the enzyme but most of the electrondense granular deposits are outside, on the surface of, and between the microvesicles (arrows). DV marks DOPA-positive vesicles in the vicinity of the giant melanosome. × 83 600. F ~ . 19. Specimen incubated in the DOPA-medium for 8 hours. Tyrosinase activity is revealed in the peripheral shell of the giant melanosome which, due to the deposition of DOPA-melanin, is heavily melanized up to the delimiting membrane (arrows), The reaction product obscures all structural details, x 83 600.
T H E G I A N T MELANOSOME
117
118
KONRAD, WOLFF AND H(}NIGSMANN
somes have been described, at the light microscope level, in one other genetic disease in man, in neurofibromatosis (4, 32). The large granules that form the subject of this paper are generated in melanocytes and nevus cells, cells that are programmed for melanin synthesis. The fact that they contain melanin and exhibit tyrosinase activity shows that they are involved in melanogenesis, and this may justify their inclusion into the group of specialized organelles in which melanogenesis normally occurs, the melanosomes (28). Their morphologic potentials make it unlikely that they represent autophagosomes in which melanosomes are degraded, and this is further supported by their highly ordered structure and the fact that they lack acid phosphatase activity. Giant melanosomes are occasionaly found within keratinocytes, and this suggests that they are transferred to these cells as are normal melanosomes. Apparently, melanocytes and keratinocytes handle them in a similar manner as the normal secretion products of pigment cells. Finally, since within a given melanocyte the giant melanosomes constitute only a small proportion of an otherwise normal melanosome population and since these cells show no other abnormalities, the giant melanosomes cannot be ascribed to a general disturbance of the code for melanosome formation; rather they appear to reflect the deranged morphogenesis of individual pigment organelles. Giant melanosomes differ from normal melanosomes in many respects. In contrast to the ellipsoidal shape of the latter the giant melanosomes are spherical; whereas normal melanosomes have a predetermined size which, in a given individual or racial group, is maintained within rather constant limits (33, 35) the giant melanosomes appear to possess the capacity for almost unrestricted growth. The largest normal melanosomes of human skin measure up to 1.3 #m (35) along their longitudinal axis and this is still less than a third of the diameter of a giant melanosome. In the present material, the dimensions of giant melanosomes were more than ten times those of the normal melanosomes produced by the same cells. Differences also exist with regard to the mode of their melanization. In normal melanosomes the melanization is synchronous, i.e., within a given melanosome the melanin is deposited uniformly throughout the organelle, and it is thus the entire melanosome that proceeds from one stage and degree of melanization to another (7, 9, 23, 34). By contrast, in the giant melanosome all stages of melanization occur simultaneously, though subject to a rather definite direction pattern. Melanization is always most advanced in the center of the organelle and the concentric, zonal, decrements of the deposited melanin toward the periphery seem to suggest a sequence of different bursts of melanogenic activity that proceed in a wavelike fashion from the core to the cortical shell (Figs. 4 and 10-13). Synchrony of melanization thus exists only for a given concentric zone of the organelle whereas asyn-
T H E G I A N T MELANOSOME
119
FIG. 20. (a-e) Stage II to IV melanosomes. The arrows denote the microvesicles which are associated with the melanofilament sheets upon which melanin is deposited. (f) A fully melanized (stage IV) melanosome from a black human hair bulb (mongoloid race). Spherical, electron lucent, membranedelimited vesicles (arrow) are present within the melanosome. They are identical with the electron lucencies observed in giant melanosomes, a-f: x 83 600.
chrony is representative for the giant melanosome as a whole. These structures thus represent a good model of how melanin is progressively laid down on its matrix. The most remarkable difference between the giant melanosomes and normal melanosomes, however, concerns their structural skeleton. The normal melanosomes, not only of man but also of other mammalian species, are composed of a system of coiled helical filaments and of folded sheets upon which melanin is deposited (6, 7, 9, 22, 23, 30, 34). These are absent from giant melanosomes which, instead, contain microvesicles as their only structural component. Such vesicles have not been observed in normal mammalian melanosomes before but similar structures have been noted in premelanosomes of embryonal chick retinal epithelium (31) and in fowl feather melanocytes (20). In these, however, melanofilaments were also present. The origin and mode of formation of the microvesicles eludes us. They bear no resemblance to coated Golgi vesicles, and they do not occur freely in the cytoplasm. Their similarity to the vesicles of rnultivesicular bodies is only a superficial one, for
120
KONRAD, WOLFF AND H{}NIGSMANN
both their size and the thickness of their membrane are less than in the latter (11). Their occurrence in the small nonmelanized vacuoles shown in Figs. 5, 6, and 8 stimulates speculations as to whether these structures represent the earliest stages of giant melanosomes and thus raises the question of how giant melanosomes are formed and how they progressively enlarge. One could envisage a situation in which the continuity which is known to exist between endoplasmic reticulum and early premelanosomes (19, 20, 24, 31) is not disrupted as the melanosome matures (19) but remains intact. This could secure a continuous supply of precursors for membrane material and, provided the microvesicles are synthesized within the giant melanosome proper, maintain the production of microvesMes within the organelle. Our observations provide no evidence that such a continuity persists but they do suggest that giant melanosomes enlarge by the continuous addition of new vesicles to the cortical zone. Since melanization proceeds centrifugically from the core and since the vesicles are trapped within the deposited melanin, it is unlikely that they are formed in the center to be subsequently shifted to the cortical zones. The vesicles could arise in the cortex by budding off from the delimiting membrane, as is suggesed by the fact that the widths of the vesicular and giant melanosome-membranes are within the same range, but such phenomena were not observed. On the other hand, the microvesicles could be added to the giant melanosomes from the outside by the interaction with other cell components. The nonmelanized vacuoles containing vesicles shown in Figs. 5, 6, and 8 were frequently observed in the immediate vicinity of giant melanosomes and it does not appear unreasonable to assume that they could fuse with the latter. They could therefore represent not only giant melanosome precursors but also the source of new microvesicles to be added to existing giant melanosomes. Unequivocal fusions, however, were not observed. At present, it is difficult to answer the question whether the microvesicles are involved in melanogenesis or merely represent a structural component of giant melanosomes. Similar vesicles of fowl feather melanocytes contain tyrosinase activity but their significance and function are unresolved (20). In contrast, the microvesicles described in this paper do not contain tyrosinase unless the enzyme is inactive, inhibited, or located in their outer surface. DOPA-melanin was deposited outside the vesicles in the intervesicular space and, since the spontaneous melanization of giant melanosomes also spares the vesicles, we find it difficult to believe that tyrosinase is present within them, even in an inactive form. The complete obliteration of vesicular structures in maximally melanized giant melanosomes does not argue against this conclusion as the vesicles are small enough to fit completely into one section; the melanin coating them may become sufficiently electron dense to account for their apparent disappearance.
THE GIANT MELANOSOME
121
Melanogenesis is not disturbed in giant melanosomes; in fact, their melanogenic properties and their transfer to keratinocytes appear to be the only markers that link them to normal melanosomes. On the other hand, if giant melanosomes result from a derangement of melanosome morphogenesis one should expect to encounter, at least at some stage of either normal or giant melanosomes, a common structural denominator that could be taken as evidence for a relationship between the normal and abnormal granules. The microvesicles are good candidates for such a role and, indeed, a careful search for them in a large normal material has revealed that they occur in normal stage II and I I I melanosomes of epidermal melanocytes and nevus cells (Fig. 20a-e) and even in fully melanized stage IV melanosomes of hair bulbs (Fig. 20f). The fact that they are encountered only inconsistently and that their number is small may explain why, so far, they have escaped our attention as well as that of other investigators. Nonetheless, their presence in normal pigment organelles indicates that, at some stage of melanosome ontogeny, they may be normal constituents of melanosomes which may even persist, to a small degree, in mature granules. Though speculative, it is tempting to assume that they may be involved in the assembly of the melanosomal filaments. A disturbance of these assembly mechanisms could then explain both the absence of filaments from giant melanosomes and the extraordinary accumulation of vesicles within the confines of their membrane. As a final remark we should like to mention that there has been some debate whether, in normal melanosomes, the melanofilaments are a structural matrix to which tyrosinase is attached and on which melanin is eventually deposited or whether they represent tyrosinase that serves both as sole structural protein and as active enzyme. Although our findings do not disprove the latter alternative they do show that the production and deposition of melanin, by tyrosinase, can occur in the absence of melanofilaments. Thus, melanofilaments are not required for a functioning melanogenic system in mammals in vivo.
Addendum During the preparation of this manuscript an abstract entitled "Macromelanosomes in card au lait spots of neurofibromatosis" was published by J. Jimbow, G. Szab6 and T. B. Fitzpatrick (J. Invest. Dermatol. 60, 241, 1973). The "macromelanosomes" described by these authors seem to be identical with the giant melanosomes of the present study (T. B. Fitzpatrick, personal communication), and it conseqently appears that this disturbance of melanosome ontogeny may occur in more than one condition. This is further supported by a most recent observation made in our own laboratory: giant melanosomes were observed in the skin of a patient with a generalized melanosis due to metastatic melanoma.
122
KONRAD, WOLFFAND HONIGSMANN
The technical assistance of Mrs Susan Csegezi and Mrs Lotte Polasek is gratefully acknowledged. REFERENCES 1. BARKA,T. and ANDERSON, P. J., J. Histochem. Cytochem. 10, 741 (1962). 2. BECKER, S. W., PRAYER, L. L. and THATCHER, H., Arch. DermatoI. Syphilol. 31, 190 (1935). 3. BEDOYA,V., Brit. J. Dermatol. 85, 336 (1971). 4. BENEDICT,P. H., SZABO, G., FITZPATRICK,T. B. and SINESI,S. J., J. Amer. Med. Ass. 205, 618 (1968). 5. BILLINGHAM,R. E. and SILVERS,W. K., Quart. Rev. Biol. 35, 1 (1960). 6. BIRBECK, M. S. C., Ann. N.Y. Acad. Sci. 100, 540 (1963). 7. BREATHNACH,A. S., in WOLMAN,M. (Ed.), Pigments in Pathology, p. 353. Academic Press, New York, 1969. 8. COHEN, H. J., MINKIN, W. and FRANK, S. B., Arch. Dermatol. 102, 433 (1970). 9. DROCHMANS,P., in DELLAPORTA,G. and M/]IHLBOCK,O. (Eds.), Structure and Control of the Melanocyte, p. 90. Springer-Verlag, Berlin, Heidelberg, New York, 1966. 10. FITZPATRICK,T. B., QUEVEDO,W. C., SZABO,G. and SEIJI, M., in FITZPATRICK,T. B., ARNDT, K. A., CLARK,W. H., EISEN,A. Z., VAN SCOTT,E. J. and VAUGHAN,J. H. (Eds.), Dermatology in General Medicine, p. 117. McGraw-Hill, New York, 971. 11. FRIEND, D. S. and FARQUHAR,M. G., J. Cell Biol. 35, 357 (1967). 12. HEARINO,V. J., PHILLIPS,P. and LUTZNER, M. A., J. Ultrastruct. Res. 43, 88 (1973). 13. JIMBOW,K. and KUKITA,A., in KAWANARA,T., FITZPATRICK,T. B. and SEIJI, M. (Eds.), Biology of Normal and Abnormal Melanocytes, p. 171. University Park Press, Baltimore, Maryland, 1971. 14. KARNOVSKY,M. J., J. Cell Biol. 27, 137A (1965). 15. KONRAD, K., HONIGSMANN,H. and WOLFF, K., Hautarzt in press. 16. LUFT, J. H., J. Biophys. Biochem. Cytol. 9, 409 (1961). 17. LUTZNER, M. A., THIERNEV,J. H. and BENDITT,E. P., Lab. Invest. 14, 2036 (1966). 18. LUTZNER, M. A., LOWRIE, C. T. and JORDAN, H. W., at. Hered. 58, 14 (1967). 19. MAUL, G., J. Ultrastruct. Res. 26, 163 (1969). 20. MAUL, G. and BRUMBAU6H, J. A., J. Cell Biol. 48, 41 (1971). 21. MOTTAZ,J. H. and ZELICKSON,A. S., in MONTAGNA,W. and DOBSON, R. L. (Eds.), Hair Growth, p. 471. Pergamon, Oxford, 1969. 22. MOVER, F. H., Ann. N.Y. Acad. Sci. 100, 586 (1963). 23. - Amer. Zool. 6, 43 (1963). 24. NOVlKOEF, A. B., ALBALA, A. and BIEMPICA, L., J. Histochem. Cytochem. 16, 299 (1968). 25. PADGETT, G. A., LEDER, R. W., GORHAM, J. R. and O'MARY, C. C., Genetics 49, 505 (1964). 26. RITTENHOIJSE,E., Develop. Biol. 17, 351 (1968). 27. SEIJI, M., FITZPATRICK,T. B. and BIRBECK,M. S. C., J. lnvest. Dermatol. 36, 243 (1961). 28. SEIJI,M., SHIMAO,K., BIRBECK,M. S. C. and FITZPATRICK,T. B., Ann. N. I1. Acad. Sei. 100, 497 (1963). 29. SMITH, R. E. and FARQUHAR, M. G., RCA Sci. lnstr. News 10, 13 (1965). 30. SCE1ROEDER,H. E., J. Periodontal Res. 4, 1 (1969).
THE GIANT MELANOSOME
123
31. STANKA, P., Z. Zellforsch. Mikrosk. Anat. 1~2, 120 (1971). 32. SZABO, G., in GORDON, M. (Ed.), Pigment Cell Biology, p. 99. Academic Press, New York, 1959. 33. SZABO, G., GERALD, A. B., PATHAK, M. A. and FITZPATRICK,T. B., Nature (London) 222, 1081 (1969). 34. TODA, K., HORI, Y. and FITZPATRICK,T. B., Fed. Proc., Fed. Amer. Soc. Exp. Biol. 27, 722 (1968). 35. TODA, K., PATHAK, M. A., PARRISH, J. A., FITZPATRICK,T. B. and QUEVEDO, W. C., Nature (London) New Biol. 236, 143 (1972). 36. TouslMIS, A. J., Ann. N.Y. Acad. Sci. Ii00, 447 (1963). 37. ZELICKSON, A. S., WINDHORST, D. B., WHITE, J. G. and GOOD, R. A., J. Invest. Dermatol. 49, 575 (1967).
I02
J. ULTRASTRUCTURE RESEARCH 48, 102--123
(1974)
T h e G i a n t M e l a n o s o m e : A Model of D e r a n g e d M e l a n o s o m e . M o r p h o g e n e s i s ~,2 KLAUS KONRAD, KLAUS WOLFF and HERBERT HONIGSMANN
Division of Experimental Dermatology, Department of Dermatology (I.), University of Vienna, Alserstrasse 4, A-I090 Vienna, Austria Received November 20, 1973 A unique type of melanin granule, occurring in a nevoid pigmented lesion in man, is described. It grows to extraordinary size, it melanizes asynchronously and, in contrast to all other known melanosomes of mammals, it is essentially nonfilamentous in nature. Its basic structural components are 400 ~ microvesicles, which, during melanization, are gradually embedded in the melanin matrix. Evidence is presented that identical vesciles also occur in normal filamentous melanosomes of man. The giant melanosome is considered to result from a derangement of morphogenesis at a distinct stage of melanosome ontogeny. Melanosomes are both sites of production and carriers of melanin pigment in mammals. They are round or ellipsoidal and are composed of a membrane-delimited coil of filaments upon which melanin is deposited during the process of melanization (7, 27). In principle, this basic structure applies to melanosomes of all mammalian species, but minor variations of size, structure, and shape occur. These are genetically determined, and in some species it has been possible to delineate a number of the responsible loci (5, 12, 22, 23, 26). The melanosomes of man exhibit an identical filamentous structural skeleton and the differences to the melanosomes of other species are again mainly confined to shape (ellipsoidal vs round), size, and degree of melanization (7, lO). In this paper w e describe a unique type of melanosome, observed in humans, which exhibits an heretofore unrecognized nonfilamentous substructure; it melanizes asynchronously and possesses extraordinary potentials of growth. Hence, the designation giant melanosome. 1 This investigation was supported by Fonds zur F/Srderung der wissenschaftlichen Forschung, Vienna and Schering AG, Berlin. 2 Presented, in part, at the Annual Meeting of the European Society for Dermatological Research, Amsterdam, April 25, 1973.
THE GIANT MELANOSOME
103
M A T E R I A L AND METHODS
Material Giant melanosomes constitute a consistent feature of a nevoid pigmented lesion in man, designated nevus spilus (8, 15). These lesions are not uncommon and represent caf6 au lait spot-like macular hypergpigmentations that are dotted by small macules of darker color and thus exhibit a mottled apearance. Histologically, they show features of lentigo simplex, but they also contain small nests of junctional nevus cells. Five such nevi from 5 caucasoid patients, aged 19 to 40 years, were chosen as source of giant melanosomes to be studied in this investigation.
Methods Shave biopsies were performed under local anesthesia and the tissue was processed both for light microscopic histology and histochemistry, employing routine procedures, and for electron microscopy and ultrastructural cytochemisty, as described below. Electron microscopy. Tissue specimens were minced and fixed in a 1:1 dilution of Karnovsky's paraformaldehyde-glutaraldehyde fixative (14) for 5 hours at room temperature and were rinsed and stored overnight in 0.1 M cacodylate buffer pH 7.4 at 4°C. Postfixation was carried out with 3 % osmie acid in distilled water for 1.5 hours at 0-4°C followed by staining en bloc with 0.5 % uranyl acetate in 0.l M Veronal acetate buffer, pH 7.2, for 2 hours at room temperature. Some specimens were embedded subsequent to the aldehyde-fixation and were thus not subjected to osmification or treatment with uranyl acetate. All specimens were rapidly dehydrated in a graded series of ethanol and were embedded in Epon 812 according to a procedure modified from Luft (16). Semithick sections were cut with glass knives and screened for the presence of giant melanosomes. Ultrathin sections were then cut from blocks containing giant melanosomes using Dupont and Dehmer diamond knives and Reichert OM U2 or LKB Ultrotome III ultramicrotomes. The sections were contrasted with lead citrate and examined with a Zeiss EM 9S electron microscope operating at 60 kV. Approximately 200 giant melanosomes were studied. Ultrastructural cytoehemistry. Thin strips of skin, measuring 1 x 1 × 5 ram, were fixed in the 1 : 1 dilution of Karnovsky's paraformaldehyde-glutaraldehyde fixative (14) for 3 hours at room temperature and, after an overnight rinse in 0.1 M cacodylate buffer (pH 7.4) at 4°C, they were chopped on a Smith and Farquhar (29) tissue chopper (Ivan Sorvall, Inc.). The 40-#m sections thus obtained were incubated for tyrosinase and acid phosphatase activities, as detailed below. After incubation they were rinsed and processed for electron microscopy as described for A. Incubation for tyrosinase. 0.025 g of L-DOPA (Sigma Chemical Company) was dissolved in 50 ml of 0.1 M phosphate buffer (pH 7.4) (2). Incubation was performed at 37°C and was terminated at 2 hours, 4 hours, or 8 hours. The medium was freshly prepared and changed every hour. Ineubation for acid phosphatase. The sections were rinsed in 0.2 M Tris maleate buffer (pH 7.2) and were then incubated according to Barka and Anderson (1), using sodium /3-glycerophosphate as substrate, for 1 hour at 37°C. As controls for the enzyme reactions, sections were incubated in complete media from which the substrates had been omitted.
104
KONRAD, WOLFF AND HONIGSMANN
FrG. 1. Giant melanosomes at the light microscope level. (a) Several giant melanosomes are present in a melanocyte (M) and are, in part, aligned along a dendritic process (arrow). A neighboring basal keratinocyte (K) also contains a giant melanosome. H & E, x 1 000. (b) Several giant melanosomes within the perinuclear cytoplasm of a melanocyte. Notice that the large organelles are sharply defined, strictly spherical, and of homogeneous density. H & E, × 1 300. (e) In DOPA-incubated cryostat sections the cytoplasm of melanocytes (M) is stained dark due to the formation of DOPAmelanin; the giant melanosomes (arrow) appear intensely black. DOPA. x 1 300.
Comparative studies A large quantity of material from our files on normal and experimental pigmentation in humans, guinea pigs, and mice, studied in this laboratory over the past 8 years, was screened by electron microscopy for the presence of the microvesicles described in the Results section of this paper. RESULTS
Light microscopy I n sections cut f r o m tissue e m b e d d e d in paraffin :and eosin, the giant m e l a n o s o m e s a p p e a r as s h a r p l y t o golden granules of h o m o g e n e o u s density (Fig. a p p r o a c h t h a t of a red b l o o d cell they can be seen at
a n d stained with h e m a t o x y l i n defined, spherical, d a r k b r o w n 1 a, b). Since their size m a y low magnifications. T h e y occur
THE GIANT MELANOSOME
105
F~6. 2. At the ultrastructural level the giant melanosomes are easily recognized by their size, their spherical shape a~nd their high electron density. Three giant melanosomes (1, 2, 3) one of which has been cut tangentially (1) occur in the cytoplasm of this melanocyte, which appears larger than normal but shows no other abnormalities. Notice the simultaneous occurrence of normal melanosomes of all stages of maturation (arrows). × 10 800.
106
KONRAD, WOLFFAND HONIGSMANN
in melanocytes (Fig. 1 a-c) and nevus cells, and less frequently within keratinocytes (Fig. 1 a) and macrophages. Their occurrence is focal, and it is only a smaller proportion of the total population of pigment cells that exhibit these granules in their cytoplasm. In silver-stained sections the granules acquire a pitch-black color, as does melanin; in cryostat sections incubated in the DOPA-medium they also appear intensely black, which contrasts with their natural brown to golden color (Fig. 1 c).
Electron microscopy Occurrence and distribution. As observed at the light microscope level the giant melanosomes occur within the cytoplasm of melanocytes and junctional nevus cells and, less frequently, in keratinocytes of the epidermis and melanophages within the connective tissue. These melanocytes and nevus cells appear larger than normal but show no other abnormalities (Fig. 2). In particular, they exhibit all structural prerequisities for active melanogenesis and, apart from the giant melanosomes, contain normal melanosomes of all stages (34) and of normal shapes and dimensions (Fig. 2). The melanocytes often protrude deeply into the dermis; not infrequently, multiple giant melanosomes can be seen in a single pigment cell (Figs. 2 and 4). The keratinocytes in the vicinity of giant melanosome-bearing melanocytes contain occasional giant melanosomes and a large number of normal melanosomes which are distributed either singly or as membrane-delimited aggregates (Fig. 3). The giant melanosomes are usually singly dispersed, but occasionally the membranes of two adjacent giant melanosomes may fuse so that two giant granules are enclosed within one common envelope (Fig. 3). Rarely do such fusions occur also with normal melanosomes. Dermal melanophages very rarely contain giant melanosomes but when they do the giant melanosomes are enclosed within large phagocytic vacuoles which may also contain a variable number of normal melanosomes or other cellular constituents. Morphology. The most striking ultrastructural feature of fully developed giant melanosomes is their tremendous size (Fig. 4) as they may measure up to 5 #m or more in diameter, the ranges observed being 1-6/~m. They are almost always strictly spherical and exhibit a highly ordered structure: a central core is surrounded by several concentric shell-like layers; these, in turn, are delimited by a trilaminar membrane, 70 A in width (Figs. 4 and 5). In the fully developed giant melanosomes the central core is highly melanized and uniformly electron dense, and the more Fro. 3. Giant melanosomes in the cytoplasm of two basal keratinocytes. Usually, these granules are singly dispersed but occasionllythe membranes of two giant melanosomesmay fuse as indicated by the arrow. Normal melanosomes are also present and are distributed either as singles or as aggregates. × 12 900.
THE G I A N T M E L A N O S O M E
107
108
KONRAD, WOLFF AND HONIGSMANN
peripheral layers have a more granular appearance (Figs. 4 and 10-13) whereas the outermost cortical shell is composed of densely set microvesicles which measure 300-400 A in mean diameter (Fig. 5). Closer inspection reveals that these microvesMes are also present in the subcortical zones where they and irregular melanin deposits account for the granularity observed at low magnifications (Figs. 5, 7, and 8); and that they can be even observed in the melanized core of the giant melanosomes where they appear as discrete electron lucencies within the homogeneous melanin matrix (Fig. 4, 9, 10, and 11). The microvesicles appear to represent the structural skeleton of the giant melanosomes. No other substructures and, in particular, no melanofilaments which are so typical for all other melanosomes (6, 7, 9, 22, 23) have been observed in these organelles (Figs. 5, 7, 8, and 15). These vesicles are bound by a delicate trilaminar membrane approximately 70 ~ in width which encloses a moderately electron dense matrix (Figs. 5 and 6). The latter exhibits a substructure that gives the impression of cross-sectioned filaments of extremely fine caliber (Fig. 5). Although the large majority of these microvesicles are spherical, some appear oval and, very rarely, they may be present also as short tubular profiles (Figs. 6-8). The question therefore arises whether they do not represent cross-sectioned tubules. A tightly wound coil of tubules could, if cross-sectioned, appear vesicular in an ultrathin section, but tubules should become apparent at least in other planes of section or in sections cut at different angles. A large number of the giant melanosomes studied in this investigation were sectioned serially and subserially, but we always encountered mainly vesicular profiles. Also, if the vesicles in the cortical zone of a giant melanosome represented cross-sectioned tubules one would expect that their tubular nature should become evident in tangential sections of giant melanosomes. Such sections also revealed vesicles of identical structure and dimensions (Fig. 6.) The degree of melanization is never quite uniform in a given giant melanosome. It is most pronounced in the core and decreases toward the peripheral zone of the organelle. Figures 10-13 are shown to illustrate what we consider different stages during the progressive melanization of giant melanosomes. As these structures mature, their dense homogeneous core expands toward their periphery until, in the fully melanized giant melanosome, it reaches the delimiting membrane (Fig. 14a, b). Melanin is initially deposited around and between the microvesicles (Figs. 5, 7, and 15) which are thus gradually embedded in a dense homogeneous melanin matrix FIo. 4. Multiple giant melanosomes occurring ina melanocyte. Note their large size, their spherical shape, and ordered structure: a fully melanized central core (C) which exhibits discrete electronlucencies is surrounded by several concentric shell-like layers that have a more granular appearance; the outermost, cortical, shell is composed of densely set microvesicles. The arrows denote the delimiting membrane, x 26 400.
110
KONRAD, WOLFF AND HONIGSMANN
in which they may remain discernible as membrane-delimited, electron-lucent globular bodies (Figs. 5, 9, and 14a, b). These are slightly smaller than the microvesicles proper. In the subcortical zones, however, some microvesicles may often appear as if they contained the deposited pigment but, in these instances, their delimiting membrane cannot be discerned and since the vesicles are so small as to fit in toto into a silver to silver-gray section, they probably represent microvesicles coated by melanin (Fig. 5). As melanization progresses, the number of the electron lucent vesicular profiles gradually decreases in the core of the giant melanosomes, and in fully melanized granules all structure is obliterated by the dense homogeneous melanin deposits (Figs. 13 and 14a, b). In such giant melanosomes it is only in the cortical layers that microvesicles c a n still be discerned (Fig. 14a) although, very rarely, they may disappear also from these cortical regions. In the aldehyde-fixed specimens not subjected to osmification the melanized portions of giant melanosomes exhibit the same electron density as normal, fully melanized stage IV (34) melanosomes (Fig. 16). Again, the microvesicles inside the melanin matrix are revealed as small round electron lucencies (Fig. 16). Giant melanosomes with little or hardly any melanization are rarely observed. These are small, measuring approximately 1 # m in diameter, and exhibit densely set microvesicles within their center (Fig. 15) as do mature giant melanosomes in their cortical zones. The melanocytes and nevus cells that produce giant melanosomes also regularly exhibit small membrane-delimited vacuoles, 0,5 # m or less in diameter, which contain typical microvesicles but no melanin. They are often seen in the immediate vicinity of giant melanosomes (Figs. 5, 6, and 8).
Cytochemistry Tyrosinase. Melanocytes and nevus cells that contain giant melanosomes exhibit tyrosinase activity as does any other active melanogenic pigment cell. Reaction product is found in some G o l g i cisternae, in the Golgi-associated endoplasmic reticulum, in small coated vesicles, and in n o r m a l stage I to I I I melanosomes
FIG. 5. Higher magnification of the more peripheral layers of a giant melanosome. Immediately below the delimiting membrane (double arrows) there is a cortical shell (1) composed of densely set microvesicles (arrows). These vesicles are bounded by a delicate membrane and exhibit a matrix that gives the impression of cross-sectioned filaments of extremely fine caliber. The subcortical shell (2) also contains microvesicles (arrow) and irregular melanin deposits. The more central portions (3) of the giant melanosomes are composed of a dense melanin matrix in which individual vesicles remain discernible as electronlucent globular bodies, still delimited by a delicate membrane (arrow). In the immediate vicinity of the giant melanosome there is a small vacuole (V) which contains microvesicles but no melanin. × 83 600. FIG. 6. Tangential section of a giant melanosome revealing the cortical shell and the spherical profiles of microvesicles (arrow). There is one short tubular structure (T). A small vacuole (V) containing microvesicles is present in the vicinity of the giant melanosome, x 83 600.
THE GIANT MELANOSOME
8 - 741821 J. Ultrastructure Research
111
112
KONRAD, WOLFF AND HONIGSMANN
(Fig. 17). After 2 and 4 hours of incubation, no unequivocal reaction product is detected within giant melanosomes. However, when the incubation time is increased to 8 hours or more, reaction product is definitely revealed in the subcortical and cortical shells of giant melanosomes, which, owing to the deposition of D O P A melanin, become heavily melanized up to the delimiting membrane, which also obliterates the vesicular profiles (Fig. 19). In this investigation it was not possible to determine unequivocally the exact site of the enzyme, but in the less active giant melanosomes most of the finely granular reaction product was deposited outside, on the surface of and between the microvesicles (Fig. 18). Acid phosphatase. Acid phosphatase activity was not associated with giant melanosomes. As expected, it was present within melanosome complexes and lysosomes of keratinocytes and melanophages.
Comparative studies A careful search for microvesicles within normal melanosomes revealed that these structures do occur in stage II and I I I melanosomes of normal human melanocytes and of nevus cells (Fig. 20a-e). However, they were encountered only occasionally and their numbers were small. Fully melanized stage IV melanosomes from black h u m a n hair bulbs exhibited spherical electronlucent bodies, delimited by a delicate membrane, which were morphologically identical with the electronlucent vesicles within the melanized core of giant melanosomes (Fig. 20f). They were similar to the electronlucent bodies described some time ago in the iris epithelium of M, rhesus (36), and in highly pigmented hair bulbs of humans (13, 21).
DISCUSSION Abnormally large melanin granules are known to occur in the Chediak-Higashi syndrome, a rare genetic disease in man (3, 37), in Aleutian minks (17, 25), and in a beige strain of mice (12, 18). These granules are all variants of normal melanosomes since they exhibit the classical melanofilament structure and are thus basically different from the giant melanosomes described in this paper. Similar giant melano-
FIG. 7. Part of the cortical and subcortical shell of another giant melanosome. Among the spherical microvesicles there is a short tubular profile (T). In the subcortical shell melanin is deposited around and between the microvesicles (arrows). x 83 600. FIG. 8. Another example of a tubular profile (T) in the cortical shell which, otherwise, exhibits only microvesicles. A small vacuole (V) with microvesicles is seen in the vicinity of the giant melanosome. x 83 600. FIG. 9. This electron micrograph shows at a high magnification the melanized core in which the microvesicles remain discernible as membrane-delimited electron-lucent globular bodies (arrows). × 83 600.
T H E G I A N T MELANOSOME
1 13
114
KONRAD, WOLFF AND HI~NIGSMANN
THE GIANT MELANOSOME
l 15
F~G. 14. Fully melanized giant melanosomes. (a) Nearly all structure is obliterated by the dense homogeneous melanin deposits, and only in the cortical shell there are some electronlucent vesicular profiles (arrow). (b) In this giant melanosome, melanization has not quite reached the delimiting membrane. Again, a few microvesicles (arrow) can be noticed in the cortical zone. a, b: × 83 600.
FIGS. 10-13 are shown to illustrate the progressive melanization of giant melanosomes. The degree of melanization is most pronounced in the core (C) of the organelle and decreases toward the peripheral zones; as the giant rnelanosome matures, its dense homogeneous core expands toward the periphery (Fig. •3). The concentric zonal decrements of the deposited melanin indicate different bursts of melanogenic activity that appear to proceed in a centrifugal fashion (Figs. 10, 11). × 22 200.
116
KONRAD, WOLFF AND HONIGSMANN
Fie. 15. A small giant melanosome exhibiting densely set 400 A microvesicles. Melanization has already commenced in the center (C). × 83 600. FIG. 16. Specimen was fixed with glutaraldehyde only. The melanized portions of the giant melanosome exhibit the same electron density as normal, fully melanized, stage IV melanosomes. Note the electronlucent vesicular profiles (arrow) within the dense melanin. × 83 600.
FIG. 17. Specimen incubated in DOPA, for 4 hours at 37°C, to show tyrosinase activity. Reaction product is present in Golgi-associated endoplasmic reticulum (arrow) and in normal stage I and II melanosomes (M) of a junctional nevus cell. GM: giant melanosome, x 22 200. F~G. 18. Specimen incubated in the DOPA-medium for 8 hours. Finely granular reaction product is present in the cortical and subcortical shell of the giant melanosome. It is not possible to determine the exact site of the enzyme but most of the electrondense granular deposits are outside, on the surface of, and between the microvesicles (arrows). DV marks DOPA-positive vesicles in the vicinity of the giant melanosome. × 83 600. F ~ . 19. Specimen incubated in the DOPA-medium for 8 hours. Tyrosinase activity is revealed in the peripheral shell of the giant melanosome which, due to the deposition of DOPA-melanin, is heavily melanized up to the delimiting membrane (arrows), The reaction product obscures all structural details, x 83 600.
T H E G I A N T MELANOSOME
117
118
KONRAD, WOLFF AND H(}NIGSMANN
somes have been described, at the light microscope level, in one other genetic disease in man, in neurofibromatosis (4, 32). The large granules that form the subject of this paper are generated in melanocytes and nevus cells, cells that are programmed for melanin synthesis. The fact that they contain melanin and exhibit tyrosinase activity shows that they are involved in melanogenesis, and this may justify their inclusion into the group of specialized organelles in which melanogenesis normally occurs, the melanosomes (28). Their morphologic potentials make it unlikely that they represent autophagosomes in which melanosomes are degraded, and this is further supported by their highly ordered structure and the fact that they lack acid phosphatase activity. Giant melanosomes are occasionaly found within keratinocytes, and this suggests that they are transferred to these cells as are normal melanosomes. Apparently, melanocytes and keratinocytes handle them in a similar manner as the normal secretion products of pigment cells. Finally, since within a given melanocyte the giant melanosomes constitute only a small proportion of an otherwise normal melanosome population and since these cells show no other abnormalities, the giant melanosomes cannot be ascribed to a general disturbance of the code for melanosome formation; rather they appear to reflect the deranged morphogenesis of individual pigment organelles. Giant melanosomes differ from normal melanosomes in many respects. In contrast to the ellipsoidal shape of the latter the giant melanosomes are spherical; whereas normal melanosomes have a predetermined size which, in a given individual or racial group, is maintained within rather constant limits (33, 35) the giant melanosomes appear to possess the capacity for almost unrestricted growth. The largest normal melanosomes of human skin measure up to 1.3 #m (35) along their longitudinal axis and this is still less than a third of the diameter of a giant melanosome. In the present material, the dimensions of giant melanosomes were more than ten times those of the normal melanosomes produced by the same cells. Differences also exist with regard to the mode of their melanization. In normal melanosomes the melanization is synchronous, i.e., within a given melanosome the melanin is deposited uniformly throughout the organelle, and it is thus the entire melanosome that proceeds from one stage and degree of melanization to another (7, 9, 23, 34). By contrast, in the giant melanosome all stages of melanization occur simultaneously, though subject to a rather definite direction pattern. Melanization is always most advanced in the center of the organelle and the concentric, zonal, decrements of the deposited melanin toward the periphery seem to suggest a sequence of different bursts of melanogenic activity that proceed in a wavelike fashion from the core to the cortical shell (Figs. 4 and 10-13). Synchrony of melanization thus exists only for a given concentric zone of the organelle whereas asyn-
T H E G I A N T MELANOSOME
119
FIG. 20. (a-e) Stage II to IV melanosomes. The arrows denote the microvesicles which are associated with the melanofilament sheets upon which melanin is deposited. (f) A fully melanized (stage IV) melanosome from a black human hair bulb (mongoloid race). Spherical, electron lucent, membranedelimited vesicles (arrow) are present within the melanosome. They are identical with the electron lucencies observed in giant melanosomes, a-f: x 83 600.
chrony is representative for the giant melanosome as a whole. These structures thus represent a good model of how melanin is progressively laid down on its matrix. The most remarkable difference between the giant melanosomes and normal melanosomes, however, concerns their structural skeleton. The normal melanosomes, not only of man but also of other mammalian species, are composed of a system of coiled helical filaments and of folded sheets upon which melanin is deposited (6, 7, 9, 22, 23, 30, 34). These are absent from giant melanosomes which, instead, contain microvesicles as their only structural component. Such vesicles have not been observed in normal mammalian melanosomes before but similar structures have been noted in premelanosomes of embryonal chick retinal epithelium (31) and in fowl feather melanocytes (20). In these, however, melanofilaments were also present. The origin and mode of formation of the microvesicles eludes us. They bear no resemblance to coated Golgi vesicles, and they do not occur freely in the cytoplasm. Their similarity to the vesicles of rnultivesicular bodies is only a superficial one, for
120
KONRAD, WOLFF AND H{}NIGSMANN
both their size and the thickness of their membrane are less than in the latter (11). Their occurrence in the small nonmelanized vacuoles shown in Figs. 5, 6, and 8 stimulates speculations as to whether these structures represent the earliest stages of giant melanosomes and thus raises the question of how giant melanosomes are formed and how they progressively enlarge. One could envisage a situation in which the continuity which is known to exist between endoplasmic reticulum and early premelanosomes (19, 20, 24, 31) is not disrupted as the melanosome matures (19) but remains intact. This could secure a continuous supply of precursors for membrane material and, provided the microvesicles are synthesized within the giant melanosome proper, maintain the production of microvesMes within the organelle. Our observations provide no evidence that such a continuity persists but they do suggest that giant melanosomes enlarge by the continuous addition of new vesicles to the cortical zone. Since melanization proceeds centrifugically from the core and since the vesicles are trapped within the deposited melanin, it is unlikely that they are formed in the center to be subsequently shifted to the cortical zones. The vesicles could arise in the cortex by budding off from the delimiting membrane, as is suggesed by the fact that the widths of the vesicular and giant melanosome-membranes are within the same range, but such phenomena were not observed. On the other hand, the microvesicles could be added to the giant melanosomes from the outside by the interaction with other cell components. The nonmelanized vacuoles containing vesicles shown in Figs. 5, 6, and 8 were frequently observed in the immediate vicinity of giant melanosomes and it does not appear unreasonable to assume that they could fuse with the latter. They could therefore represent not only giant melanosome precursors but also the source of new microvesicles to be added to existing giant melanosomes. Unequivocal fusions, however, were not observed. At present, it is difficult to answer the question whether the microvesicles are involved in melanogenesis or merely represent a structural component of giant melanosomes. Similar vesicles of fowl feather melanocytes contain tyrosinase activity but their significance and function are unresolved (20). In contrast, the microvesicles described in this paper do not contain tyrosinase unless the enzyme is inactive, inhibited, or located in their outer surface. DOPA-melanin was deposited outside the vesicles in the intervesicular space and, since the spontaneous melanization of giant melanosomes also spares the vesicles, we find it difficult to believe that tyrosinase is present within them, even in an inactive form. The complete obliteration of vesicular structures in maximally melanized giant melanosomes does not argue against this conclusion as the vesicles are small enough to fit completely into one section; the melanin coating them may become sufficiently electron dense to account for their apparent disappearance.
THE GIANT MELANOSOME
121
Melanogenesis is not disturbed in giant melanosomes; in fact, their melanogenic properties and their transfer to keratinocytes appear to be the only markers that link them to normal melanosomes. On the other hand, if giant melanosomes result from a derangement of melanosome morphogenesis one should expect to encounter, at least at some stage of either normal or giant melanosomes, a common structural denominator that could be taken as evidence for a relationship between the normal and abnormal granules. The microvesicles are good candidates for such a role and, indeed, a careful search for them in a large normal material has revealed that they occur in normal stage II and I I I melanosomes of epidermal melanocytes and nevus cells (Fig. 20a-e) and even in fully melanized stage IV melanosomes of hair bulbs (Fig. 20f). The fact that they are encountered only inconsistently and that their number is small may explain why, so far, they have escaped our attention as well as that of other investigators. Nonetheless, their presence in normal pigment organelles indicates that, at some stage of melanosome ontogeny, they may be normal constituents of melanosomes which may even persist, to a small degree, in mature granules. Though speculative, it is tempting to assume that they may be involved in the assembly of the melanosomal filaments. A disturbance of these assembly mechanisms could then explain both the absence of filaments from giant melanosomes and the extraordinary accumulation of vesicles within the confines of their membrane. As a final remark we should like to mention that there has been some debate whether, in normal melanosomes, the melanofilaments are a structural matrix to which tyrosinase is attached and on which melanin is eventually deposited or whether they represent tyrosinase that serves both as sole structural protein and as active enzyme. Although our findings do not disprove the latter alternative they do show that the production and deposition of melanin, by tyrosinase, can occur in the absence of melanofilaments. Thus, melanofilaments are not required for a functioning melanogenic system in mammals in vivo.
Addendum During the preparation of this manuscript an abstract entitled "Macromelanosomes in card au lait spots of neurofibromatosis" was published by J. Jimbow, G. Szab6 and T. B. Fitzpatrick (J. Invest. Dermatol. 60, 241, 1973). The "macromelanosomes" described by these authors seem to be identical with the giant melanosomes of the present study (T. B. Fitzpatrick, personal communication), and it conseqently appears that this disturbance of melanosome ontogeny may occur in more than one condition. This is further supported by a most recent observation made in our own laboratory: giant melanosomes were observed in the skin of a patient with a generalized melanosis due to metastatic melanoma.
122
KONRAD, WOLFFAND HONIGSMANN
The technical assistance of Mrs Susan Csegezi and Mrs Lotte Polasek is gratefully acknowledged. REFERENCES 1. BARKA,T. and ANDERSON, P. J., J. Histochem. Cytochem. 10, 741 (1962). 2. BECKER, S. W., PRAYER, L. L. and THATCHER, H., Arch. DermatoI. Syphilol. 31, 190 (1935). 3. BEDOYA,V., Brit. J. Dermatol. 85, 336 (1971). 4. BENEDICT,P. H., SZABO, G., FITZPATRICK,T. B. and SINESI,S. J., J. Amer. Med. Ass. 205, 618 (1968). 5. BILLINGHAM,R. E. and SILVERS,W. K., Quart. Rev. Biol. 35, 1 (1960). 6. BIRBECK, M. S. C., Ann. N.Y. Acad. Sci. 100, 540 (1963). 7. BREATHNACH,A. S., in WOLMAN,M. (Ed.), Pigments in Pathology, p. 353. Academic Press, New York, 1969. 8. COHEN, H. J., MINKIN, W. and FRANK, S. B., Arch. Dermatol. 102, 433 (1970). 9. DROCHMANS,P., in DELLAPORTA,G. and M/]IHLBOCK,O. (Eds.), Structure and Control of the Melanocyte, p. 90. Springer-Verlag, Berlin, Heidelberg, New York, 1966. 10. FITZPATRICK,T. B., QUEVEDO,W. C., SZABO,G. and SEIJI, M., in FITZPATRICK,T. B., ARNDT, K. A., CLARK,W. H., EISEN,A. Z., VAN SCOTT,E. J. and VAUGHAN,J. H. (Eds.), Dermatology in General Medicine, p. 117. McGraw-Hill, New York, 971. 11. FRIEND, D. S. and FARQUHAR,M. G., J. Cell Biol. 35, 357 (1967). 12. HEARINO,V. J., PHILLIPS,P. and LUTZNER, M. A., J. Ultrastruct. Res. 43, 88 (1973). 13. JIMBOW,K. and KUKITA,A., in KAWANARA,T., FITZPATRICK,T. B. and SEIJI, M. (Eds.), Biology of Normal and Abnormal Melanocytes, p. 171. University Park Press, Baltimore, Maryland, 1971. 14. KARNOVSKY,M. J., J. Cell Biol. 27, 137A (1965). 15. KONRAD, K., HONIGSMANN,H. and WOLFF, K., Hautarzt in press. 16. LUFT, J. H., J. Biophys. Biochem. Cytol. 9, 409 (1961). 17. LUTZNER, M. A., THIERNEV,J. H. and BENDITT,E. P., Lab. Invest. 14, 2036 (1966). 18. LUTZNER, M. A., LOWRIE, C. T. and JORDAN, H. W., at. Hered. 58, 14 (1967). 19. MAUL, G., J. Ultrastruct. Res. 26, 163 (1969). 20. MAUL, G. and BRUMBAU6H, J. A., J. Cell Biol. 48, 41 (1971). 21. MOTTAZ,J. H. and ZELICKSON,A. S., in MONTAGNA,W. and DOBSON, R. L. (Eds.), Hair Growth, p. 471. Pergamon, Oxford, 1969. 22. MOVER, F. H., Ann. N.Y. Acad. Sci. 100, 586 (1963). 23. - Amer. Zool. 6, 43 (1963). 24. NOVlKOEF, A. B., ALBALA, A. and BIEMPICA, L., J. Histochem. Cytochem. 16, 299 (1968). 25. PADGETT, G. A., LEDER, R. W., GORHAM, J. R. and O'MARY, C. C., Genetics 49, 505 (1964). 26. RITTENHOIJSE,E., Develop. Biol. 17, 351 (1968). 27. SEIJI, M., FITZPATRICK,T. B. and BIRBECK,M. S. C., J. lnvest. Dermatol. 36, 243 (1961). 28. SEIJI,M., SHIMAO,K., BIRBECK,M. S. C. and FITZPATRICK,T. B., Ann. N. I1. Acad. Sei. 100, 497 (1963). 29. SMITH, R. E. and FARQUHAR, M. G., RCA Sci. lnstr. News 10, 13 (1965). 30. SCE1ROEDER,H. E., J. Periodontal Res. 4, 1 (1969).
THE GIANT MELANOSOME
123
31. STANKA, P., Z. Zellforsch. Mikrosk. Anat. 1~2, 120 (1971). 32. SZABO, G., in GORDON, M. (Ed.), Pigment Cell Biology, p. 99. Academic Press, New York, 1959. 33. SZABO, G., GERALD, A. B., PATHAK, M. A. and FITZPATRICK,T. B., Nature (London) 222, 1081 (1969). 34. TODA, K., HORI, Y. and FITZPATRICK,T. B., Fed. Proc., Fed. Amer. Soc. Exp. Biol. 27, 722 (1968). 35. TODA, K., PATHAK, M. A., PARRISH, J. A., FITZPATRICK,T. B. and QUEVEDO, W. C., Nature (London) New Biol. 236, 143 (1972). 36. TouslMIS, A. J., Ann. N.Y. Acad. Sci. Ii00, 447 (1963). 37. ZELICKSON, A. S., WINDHORST, D. B., WHITE, J. G. and GOOD, R. A., J. Invest. Dermatol. 49, 575 (1967).