11-man Ill. I

Factor XIIIa-Expressing Dermal Dendrocytes in

AIDS-Associated Cutaneous Kaposi's SarcomasKaposi's sarcoma (KS) is an angioproli-

ferative neoplasm that develops as multifocallesions, often involving the skin, accompa-nied by the accumulation of perivascular andinterstitial spindle-shaped cells (1). The na-ture of the proliferating spindle-shaped cellsin KS has been a "riddle within a puzzle" forthe past century [reviewed in (2) ]. While theendothelial cell has been linked to the spin-dle-shaped cells in KS lesions (3), twvo recentreports (4) suggest a nonvascular origin ofthe spindle-shaped cells in AIDS-associatedKS lesions. Because the cultured cells de-rived from AIDS-associated KS lesions hadmorphological features including "dendrit-ic" spindle cells (4), we asked whether thesecells could be related to the dermal dendro-cyte, a normal constituent of skin first de-scribed by Headington (5). The dermal den-drocyte is a member of the mononuclearphagocytic system (5) and shares with othermonocytes/macrophages the expression offactor XIIIa antigen (6). Furthermore, boththe dermal dendrocyte and cultured KSspindle-shaped cells possess the same enzy-matic content, including acid phophatase

and nonspecific esterase, while lacking aden-osinetriphosphatase and factor VIII-relatedantigen; they also both have pinocytoticvesicles but no Birbeck granules (4, 5). Todetermine whether the spindle-shaped cellproliferation in AIDS-associated KS lesionsincluded factor XIIIa antigen-expressingdermal dendrocytes, we analyzed 14 AIDS-associated cutaneous patch-and-plaque stageKS lesions obtained from 11 different pa-tients.The KS lesions were characterized by a

complex histological picture, including nu-merous slit-like vascular spaces, perivascularspindle-shaped cells that stained positivelyfor factor XIIIa, lymphocytes, and plasmacells (Fig. 1 and Fig. 2). In four of theAIDS-associated KS cases, the factor XIIIa-positive cells represented approximately 30to 50% of the spindle-shaped cells withinthe lesions; in eight other cases they werebetween 10 and 30% of cells; in two casesthey represented approximately 10% of thespindle-shaped cells. It appeared that inthose cases in which there was a greaterproportion of factor XIIIa-positive dermal

dendrocytes among the spindle-shaped cells,there was a greater degree of lymphocyticinflammation. The factor XIIIa staining pat-tern in all 14 KS lesions was heterogeneous;this heterogeneity is similar to the stainingpattern of another cutaneous neoplasm thatcontains a significant accumulation of der-mal dendrocytes known as the dermatofi-broma (7). The dermatofibroma, which maybe difficult to distinguish from KS lesions(8), also contains proliferating small vesselsand spindle-shaped cells embedded in a col-lagenous stroma with admixed lymphocytes.The factor XIIIa-positive cells in KS lesionsappeared as fascicles of spindle-shaped cells,or as individual dendritic-shaped cells inter-

Fig. I (top) HIV-l-associat-ed KS with collections of fac-tor XllIa--positive red-stainedspindle-shaped cells (straightarrows) in the dermis embed-ded in a collagenous stromainterspersed with numertousvascular spaces (curved ar-rows), extravasated e-y_hro-cites, and interstitial nmpho-cytes. Magnification, X200.Sections (5 ptmthick) of paraf-fin-embedded skin were cut,dewaxed in xylene, and rehy-drated. The sections to bestained for factor XIla werethen incubated with tiypsin inphosphate-buffered saline at3TC for 30 miin and endoge-nous peroxidase was blockedwith, 1% H20)2 in methanolfor 20 min. An avidin-biotinperoxidase technique (Vecta-stain ABC Kit; Vector Labs.Inc.) with the polydonal rab-bit antibody to factor XlIIa(Calbiochem Corp., diluted1:400) was used with 3-amino-9-ethylcarbazole asthe chromogen, and the sections were counterstamned with 1% hematoxylin. When either normal rab-bit serum, or rabbit antibody to the plasma-derived, extracellular form, factor Xllls was used(Calbiochem Corp.), there was no staining. Fig. 2 (bottom). HIV- 1-associated cutaneous KS.Only perivascular spindle-shaped cells dissecting between collagen bundles, and not endothelial cellslining the vessels (asterisk), are factor XIIIa-positive, as defined by red cytoplasmic staining. This site ofprominent factor XIIIa expression is accompanied by a relatively dense lymphocytic infiltrate.Magnification, X250.

Fig. 3 (top). Many dermal dendrocytes in HIV-1-associated KS lesion are LFA-1-positive withfocal formation of fascicles of spindle-shaped cells(arrow). Magnification, x25. Cryostat sections(6 pm thick) were immunohistochemically stainedwith the use of anti-LFA-1 monoclonal antibody[TS 1.18 (10)], and the avidin-biotin peroxidasetechnique described above was used. Fig. 4(bottom). HIV-1-associated psoriasis demon-strating numerous but solitary angiocentric factorXIIIa-positive dermal dendrocytes in papillarydermis immediately beneath hyperplastic epider-mis. Magnification, x 100.

SCIENCE, VOL. 243I736

on March 5, 2021

http://science.sciencem

ag.org/D

ownloaded from

spersed among dermal collagen bundles.The factor XIIIa-positive spindle-shapedcells were not S-100 positive, which rulesout a Langerhans cell or a Schwann cellorigin (9). As in normal skin, where factorXIIIa-positive dermal dendrocytes are close-ly associated with blood vessels, many factorXIIIa-positive spindle-shaped cells in theKS lesions maintained an angiocentric con-figuration. Occasional factor XIIIa-positivespindle-shaped cells also contained phago-cytic hemosiderin deposits, which would bein agreement with the previously recognizedability of dermal dendrocytes to engulf pig-ments (5). In the last two AIDS-associatedKS patients, the spindle-shaped cells wereidentified in cryostat sections as also stainingwith antibody to an antigen (LFA-l) thathas been associated with lymphocyte func-tion (Fig. 3), confirming that they werebone marrow-derived mononuclear cellsand not fibroblasts or endothelial cells (11).These results suggest that the factor

XIIIa-expressing dermal dendrocyte may bethe cell of origin for the spindle-shaped cellpopulation in AIDS-associated KS lesionsand in the KS cell line (4). Our resultsconfirm an earlier suggestion (2) that thespindle-shaped cells in KS lesions are relatedto the reticuloendothelial cell system bydemonstrating a monocyte/macrophage lin-eage marker (factor XIIIa) in the dermaldendrocytes that make up the spindle-shaped cell population. The cause of thevariation in the extent of factor XIIIa expres-sion by the spindle-shaped cells in AIDS-associated KS lesions is not known, butbecause there was a greater expression whenthere were more lymphocytes in the infil-trate, we believe that local modulation byinterferon-y may play a role (12). In one ofour AIDS patients who developed psoriasis,a skin disease with prominent blood vesselsand inflammation, we observed large num-bers of factor XIIIa-positive dermal dendro-cytes in the papillary dermis immediately

beneath the hyper epidermis (Fig. 4). Theseresults suggest that upon exposure to humanimmunodeficiency virus type 1 (HIV-l),factor XIIIa-positive dermal dendrocytesmay be activated that can give rise either toexpansile lesions such as KS or to alteredlocal immune reactions producing psoriasisin genetically susceptible individuals. In ei-ther case, it would appear that the cutaneousmanifestations of HIV-1 infection such asKS and psoriasis may represent hyperactivi-ty of the dermal dendrocytic component ofthe immune system, rather than an immuno-deficiency (13). If factor XIIIa dermal den-drocytes in vivo possess the same repertoireof cytokine production (including interleu-kin-l and basic fibroblast growth factor) asdo the cultured KS spindle-shaped cells (4,14), then it is possible to envisage themolecular basis by which HIV-l-activateddermal dendrocytes could stimulate endo-thelial, keratinocyte, and mononuclear cellproliferation in AIDS-associated KS andpsoriatic lesions (15).

BRLAN J. NICKOLOFFCHRISTOPHER E. M. GRIFFITHS

Departments ofPathology and Dermatology,University ofMichigan Medical School,

1301 Catherine Road,Ann Arbor, MI 48109-0602

REFERENCES AND NOTES1. M. Kaposi, Arch. Dernatol. Syphil. 4, 265 (1872).2. J. Rosai, in Kaposi's Sarcoma: A Text and Atlas, G.

Gottlieb and A. B. Ackerman, Eds. (Lea and Fe-biger, Philadelphia, PA, 1988), p. 3.

3. R. F. Dorfman, Hum. Pathol. 15, 1013 (1984).4. S. Nakamura et al., Science 242, 426 (1988); S. Z.

Salahuddin et al., ibid., p. 430.5. J. T. Headington, Adv. Dermatol. 1, 159 (1986).6. P. Henrikisson et al., J. Clin. Inivest. 76, 528 (1985);

Z. Nemes et al., J. Pathol. 149, 121 (1986); R.Cerio, J. J. Spaull, E. W. Jones, J. Cut. Pathol. 14,351 (1987).

7. R. Cerio, J. J. Spaull, E. Wilson-Jones, Br. J.Dermatol. 120, 197 (1989).

8. N. S. McNutt, V. Fletcher, M. A. Conant, Am. J.Pathol. 111, 62 (1983).

9. B. Recht, B. J. Nickoloff, G. S. Wood, J. Dermatol.Surg. Oncol. 12, 1192 (1986).

10. TS 1.18 was provided by C. Clayberger and A.Krensky.

11. T. A. Springer, M. L. Dustin, T. K. Kishimoto, S.D. Marlin, Ann. Rev. Immunol. 5, 223 (1987).

12. R. L. Kradin et al., Blood 69, 778 (1987).13. J. Costa and A. S. Rabson, Lancet i, 58 (1983); M. S.

Ascher and H. W. Sheppard, Clin. Exp. Immunol.73, 165 (1988).

14. B. Ensoli et al., Science 243, 223 (1989).15. J. M. Leonard et al., ibid. 242, 1665 (1988); B. J.

Nickoloff, Arch. Dermatol. 124, 1835 (1988).16. The authors thank J. T. Headington and J. J.

Voorhees for reviewing the manuscript.7 December 1988; accepted 2 March 1989

Response: Nickoloffs comment is interest-ing. It could be that the true primary neo-plastic cell of Kaposi's sarcoma (KS) is thespindle cell, as he suggests and as we andothers believe. Nickoloff also argues that thelong-sought-after origin of these cells maybe the so-called dermal dendrocytes andsuggests that the spindle cells we succeededin culturing for the first time from KSpatients may indeed be these cells. Indeed,many of the properties of our cells aresimilar to properties of macrophages. How-ever, such cells are difficult to distinguishfrom endothelial cells. For instance, in termsof uptake of low density lipoprotein, bind-ing of Ulex europeus lectin, and presence ofcytokeratin our cells are more like endotheli-al cells than macrophages. Ofcourse, dermaldendrocytes may be in the macrophage lin-eage but different from classical macro-phages in these respects. However, KS isnot limited to the skin, but occurs in numer-ous other tissues. Thus, the suggestionseems far from conclusive, but it does merittesting our cultured KS cells with the anti-bodies to factor XIIIa and LFA- l. It will becritical to confirm the specificity of theantibody to LFA- 1.

ROBERT C. GALLOS. ZAKI SALAHUDDIN

SHUJI NAKAMuRABuilding 37, Room 6A09,National Cancer Institute,

Bethesda, MD 20892

TECHNICAL COMMENTS I73731 MARCH 1989

on March 5, 2021

http://science.sciencem

ag.org/D

ownloaded from

Factor XIIIa-expressing dermal dendrocytes in AIDS-associated cutaneous Kaposi's sarcomasBJ Nickoloff and CE Griffiths

DOI: 10.1126/science.2564703 (4899), 1736-1737.243Science 

ARTICLE TOOLS http://science.sciencemag.org/content/243/4899/1736

REFERENCES

http://science.sciencemag.org/content/243/4899/1736#BIBLThis article cites 17 articles, 5 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

No claim to original U.S. Government Works.Copyright © 1989 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science.

on March 5, 2021

http://science.sciencem

ag.org/D

ownloaded from