- Email: [email protected]
CD44 expression in benign and malignant nevomelanocytic lesions
CD44 Expression in Benign and Malignant Nevomelanocytic Lesions CHRISTOPHER J. LEIGH, MD, PATRICIA L. PALECHEK, JENNIFER R. KNUTSON, PHD, JAMES B. McCARTHY, PHD, MICHAEL B. COHEN, MD, AND ZSOLT B. ARGENYI, MD CD44 is an integral membrane glycoprotein that is a principal receptor for hyaluronan and plays a role in cell-extraceUniar matrix interactions. Recent studies of melanomas in mouse models have suggested that increased CD44 expression by these tumors may relate to metastatic potential. Immunohistochemical expression of CD44 (standard [s] and variant [v6l) in benign and malignant nevomelanocytic lesions was assessed in formalin-fixed, paraffin-embedded tissue and was correlated with histological parameters and prognostic factors. Cases included benign nevi (three jtmctional, four compound, five intradermal, five blue, six Spitz, one deep penetrating), architecturally disordered (dysplastic) nevi (three, and primary (22) and metastatic melanomas (eight). All of the benign lesions showed diffuse and essentially uniform membrane staining of CD44s in nevomelanocyfic cells, regardless of lesion size, depth, or extent of dermal involvement. In contrast, semiquantitative analysis (0 to 3+) of the primary melanomas showed heterogeneous and decreased staining of CD44s, which inversely correlated with lesion size (-0.569) and depth of invasion (-0.622 and -0.617 for Breslow's depth and Clark's level, respectively). These results were significant at P < .05. CD44s expres-
sion in metastases paralleled that of their respective primaries. None of the benign nevomelanocytic lesions showed CD44v6 staining. In contrast, all of the malignant nevomelanocytic lesions showed cytoplasmic staining of the tumor cells. Pretreatment with chondroitinase did not alter CD44s staining. CD44s expression byimmunohistochemical determination is uniform in benign nevomelanocytic lesions. Malignant melanomas show decreased, heterogeneous staining that inversely correlates with increasing size, depth, and level of invasion. CD44 expression may be a prognostic indicator in malignant melanomas. Tumor staining with anti-chondroitin sulfate monoclonal antibodies suggests that CD44s may be expressed as a chondroitin sulfate proteoglycan in primary melanomas. HUM PATHOL 27:12881294. Copyright © 1996 by W.B. Saunders Company Key words:melanoma, nevus, cellular adhesion molecules, immunohistochemistry, CD44. Abbreviations: CSPG, cell surface chondroitin sulfate proteogiycan; CS, chondroitin sulfate; ABC, avidin-hiotin complex; VGF, vertical growth phase; DAB, diaminobenzidine; PBS, phosphate-buffered saline; Ig, immunoglobulin.
CD44 is the designation for a family of transmembrane glycoprotein molecules, a primary function of which is to serve as the principal ligand for hyaluronan, an important extracellular matrix glycosaminoglycan present in a large variety of tissues] 4 The role of CD44 in cell-extracellular matrix interactions has led some investigators to postulate that altered CD44 expression by various neoplasms may relate to their ability to metastasize. For instance, altered CD44 expression may result in different adhesive or migratory properties of neoplastic cells, thereby potentiating the ability of a tumor to infiltrate dermal or other tissues. Several different isoforms of CD44 are known to exist, and the specific isoform(s) expressed also m,a3ybe of some import with regard to metastatic potential. Cell type-specific differences in CD44 expression are also created by posttranslational addition of O- or N-linked carbohydrates. Furthermore, four serine residues potentially can serve as acceptor sites for either heparin sulfate or chondroitin sulfate glycosaminoglycans. 5-s Such posttranslational
modifications can alter the ligand binding properties of the CD44 core protein, and CD44 expressed as a cell surface chondroitin sulfate proteoglycan (CSPG) has been shown to bind to fibronectin and types I and IV collagen. 5-s Evidence for a potential role of CD44 in the dissemination of malignant melanomas comes from a variety of sources. Previous studies of malignant melanoma in mouse models have indicated that melanoma cell lines expressing high levels of CD44s (standard isoform) show a greater propensity for hematogenous dissemination than melanomas expressing low levels of CD44s. 9 Cell culture studies have shown the importance of CD44 for melanoma cell migration on hyaluronan-coated surfaces, and immunohistochemical studies have demonstrated at least some expression of CD44s in both benign and malignant nevomelanocytic lesions. 1°14 Furthermore, expression of CD44/CSPG on melanoma cells has been linked to regulation of melanoma cell migration and invasion of collagen gels or reconstituted basement membranes in vitro. "s The aim of the current study was to immunohistochemically examine CD44 expression in benign and malignant nevomelanocytic lesions in an effort to determine the relationship between expression and prognostically important histopathological characteristics of these lesions. The studies indicate that CD44s is a major isoform expressed on benign and malignant nevomelanocytic lesions. Furthermore, digestion of the tumor tissue sections with chondroitinase avidin-biotin complex (ABC) before staining with specific anti-CS (chondroitin sulfate) antibodies demonstrated positive intratumor staining for chondroitin-4 sulfate, consistent with
From the Departments of Pathology, Urology, and Dermatology, The University of Iowa and Veterans Affairs Medical Center, Iowa City, IA; and the Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN. Accepted for publication May 9, 1996. Supported by an American Cancer Society Clinical Oncology Career Development Award and by the Office of Research and Development (Medical Research Service), Department of Veterans Affairs (M.B.C.), Iowa City, IA. Address correspondence and reprint requests to Michael B. Cohen, MD, Department of Pathology, 200 Hawkins Drive, 5216 RCP, The University of Iowa, Iowa City, IA 52242-1009. Copyright © 1996 by W.B. Saunders Company
0046-8177/96/2712-002055.00/0
1288
CD44 EXPRESSION IN NEVOMELANOCYTIC LESIONS (Leigh et al)
TABLE I.
Nevomelanocytic Lesions Examined
Benign nevomelanocytic lesions Junctional nevi Compound nevi Intradermal nevi Blue nevi Spitz nevi Deep penetrating nevi Dysplastic nevi Malignant nevomelanocytic lesions Superficial spreading Lentigo maligna melanoma Nodular Acral lentiginous Metastatic melanoma
Monoclonal Antibodies and Immunohistochemistry
3 4 5 5 6 1 3 7 2 8 5 8
the in vivo expression of CD44/CSPG in nevomelanocytic lesions. MATERIALS AND METHODS Tissue Twenty-four benign nevi, three architecturally disord e r e d (dysplastic) nevi, 22 primary melanomas, and eight metastatic m e l a n o m a s were examined. Each of the lesions were from different patients, except for the metastases, which were taken from a subset of the primary m e l a n o m a patients. T h e m e l a n o m a s selected i n c l u d e d various histological types of varying Breslow d e p t h and Clark level (Tables 1 and 2). O f the 22, 14 primary m e l a n o m a s were considered to be in vertical growth phase (VGF); all VGF lesions were Clark's levels III, IV, or V. Representative 4-#m sections of formalinfixed, paraffin-embedded archival tissue were used for hematoxylin-eosin and i m m n n o h i s t o c h e m i c a l stains in each instance.
TABLE 2.
An i m m u n o p e r o x i d a s e reaction for CD44s and CD44v6 was p e r f o r m e d in each case using a labeled streptavidin-biotin method. Sections were deparaffinized in xylene, rehydrated in graded alcohols, and rinsed. Antigen retrieval was accomplished using citrate buffer for 10 minutes at high power in a microwave. E n d o g e n o u s peroxidase activity was blocked using an hydrogen p e r o x i d e / d i s t i l l e d water solution (0.3% and 1.88% for CD44s and CD44v6, respectively). Sections were covered with primary antisera incubated overnight at 4°C and rinsed. Sections were then covered with LSAB II link antibody for 10 minutes (CD44v6) or 20 minutes (CD44s) at r o o m temperature. Sections were incubated with 0.05% d i a m i n o b e n z i d i n e (DAB)/0.1 m o l / L Tris with 0.015% hydrogen p e r o x i d e to show signal of the primary antibody. A counterstain of 10% Harris hematoxylin without acid for 1 to 2 minutes was used. Negative control slides were p r e p a r e d by substituting normal mouse serum for the primary antisera. CD44s antibody was purchased from Becton Dickinson (San Jose, CA) and used at a dilution of 1:50. CD44v6 antibody was purchased f r o m B e n d e r Medical Systems (Vienna, Austria) and used at a dilution of 1:100. Biotinylated link and label were purchased from DAKO Corporation (Carpinteria, CA).
Chondroitinase Pretreatment Because CD44 may u n d e r g o posttranslational modification, including CS addition, we also attempted to d e t e r m i n e what role this might play in masking CD44 expression in nevomelanocytic lesions. Representative tissue samples of 11 tumors f r o m 10 patients were selected for chondroitinase pretreatment. Cases were selected to include a range of t u m o r types, sizes, and depths of invasion. T h e specimens were deparaffinized, rehydrated in graded alcohols, and washed in deionized water. Control slides for standard CD44 staining
Characteristics of Primary Melanomas
Case No.
Sex
Age
Diagnosis
Clark Level
Breslow Depth (mm)
Lesion Size (cm)
CD44s Staining
CD44v6 Staining
1 2 3 4 5 6 7 8 9 10 11 12 I3 14 15 16 17 18 19 20 21 22
F M M M M M F F F F F M M F F F F M F M M M
49 37 80 35 57 62 75 74 76 50 32 86 46 67 38 70 62 37 37 53 53 48
SSMIS SSMIS SSMM SSMIS SSMM SSMM LMM LMM NMM NMM NMM NMM NMM ALMMIS ALMM ALMM ALMM NMM NMM NMM SSMM ALMM
I I 1V I III II II II IV IV IV V IV I IV IV III III IV IV II V
0.00 0.00 1.49 0.00 2.10 0.19 0.54 0.32 2.80 0.85 2.86 12.40 2.31 0.00 2.20 11.00 1.15 4.50 2.50 1.85 0.22 6.00
1.0 0.8 1.1 0.6 1.3 1.3 0.8 0.3 1.0 0.7 1.5 3.6 0.4 0.4 2.l 1.5 0.7 2.0 2.2 2.5 1.9 2.0
2+ 3+ 3+ 3+ 2+ 3+ 3+ 2+ 1+ 2+ 2+ 1+ 2+ 3+ 1+ 1+ 2+ 1+ 1+ 2+ 2+ 2+
+ + + + + + + + + + + + + + + + + + + + + +
Abbreviations: MIS, melanoma-in-situ; ALMM, acral lentiginous malignant melanoma; LMM, lentigo maligna melanoma; NMM, nodular malignant melanoma; SSMM, superficial spreading malignant melanoma; CD44s, CD44, standard form; CD44v6, CD44, v6 variant (exon 10).
1289
HUMAN PATHOLOGY TABLE 3. 0 I+ 2+ 3+
Volume 27, No, 12 (December 1996)
Grading Scheme for CD44s
Tumor cells show no cytoplasmicmembrane staining < 5 0 %of tumor cells show membrane staining similar to that of basal keratinocytes 50% to 90% of tumor cells show membrane staining similar to that of basal keratinocytes >90%of tumor cells show membrane staining similar to that of basal keratinocytes
(ie, no chondroitinase pretreatment) were prepared as previously described. Slides for chondroitinase pretreatment were covered with chondroitinase-ABC (Seikagaku Kogyo Co, LTD, Tokyo, Japan) at a concentration of 0.2 U/mL. The slides were incubated in a humidity chamber at 37°C for 60 minutes and then washed twice in phosphate-buffered saline (PBS). After pretreatment with chondroitinase-avidin-biotin complex (ABC), individual slides were treated with monoclonal antibody to either CD44s or CSPG. CD44s staining was performed as previously described. Three anti-CSPG monoclonal antibodies that differed in specificity with respect to the location of the sulfate group were tested. These included antibody to unsulfated CSPG (anti-PC,-A-Di-0S; mouse immunoglobulin [Ig] G1), 4-sulfated CSPG (anti-PG-A-Di-4S; mouse IgG1), and 6-sulfated CSPG (anti-PG-A-Di-6S; mouse !gM) (all from ICN Biomedicals, Inc, Costa Mesa, CA). The anti-CSPG antibodies were used at a concentration of 1:250. Negative mouse anti-human IgG at a concentration of 1:10,000 was used as a control. The slides were incubated with antibody for 60 minutes at room temperature then washed twice with PBS. Horseradish-peroxidase goat anti-mouse IgG (Calbiochem, San Diego, CA) at a concentration of 1:250 was then added to each slide for 60 minutes at room temperature. Specimens were washed twice in 0.1 mol/L Tris and developed in 0.05% DAB in 0.03% hydrogen peroxide and 0.1 mol/L Tris for 5 minutes. The slides were then counterstained with 10% hematoxylin, dehydrated, cleared with xylene, and coverslipped.
Analysis Slides stained for CD44s were examined and graded semiquantitatively using a 0-3+ scale (Table 3). Staining of the lower (basal) epidermal layers served as an internal control for all but the metastatic lesions. CD44v6 staining of lesions was not stratified, but was considered either positive or negative. In the lesions expressing CD44v6, virtually all lesional cells were uniformly weakly positive, defying further stratification. Results were correlated with lesion size, Breslow's depth, and Clark's level using the Pearson product-moment correlation method.
RESULTS CD44s In each case, the lower epidermal layers and the epithelia of the skin appendages showed strong cytoplasmic m e m b r a n e staining for CD44s, as described in other reports 2'12'1s (Fig 1B, E). All of the benign nevomelanocytic lesions, including three architecturally disordered nevi, showed uniform m e m b r a n e staining for CD44s of an intensity similar to that of the overlying epithelium (3+) (Fig 1B). The malignant lesions, however, showed a tendency toward heterogeneous mem-
brane staining, often with large sheets and nodules of t u m o r showing very weak or no staining. Semiquantitarive analysis (0 to 3+) of CD44s staining a m o n g the primary melanomas showed an inverse correlation between lesion size (-0.569) and depth of invasion ( - 0 . 6 2 2 and - 0 . 6 1 7 for Breslow depth and Clark level, respectively) using the Pearson p r o d u c t - m o m e n t correlation. These results were all significant at P < .05. Notable exceptions to this trend were case nos. 3 and 22 (Table 2), which were sizeable and deeply invasive, but which nonetheless showed 3+ and 2+ expression, respectively. CD44s expression in metastases paralleled that of their respective primaries.
CD44v6 In each case, the lower epidermal layers and epithelia of the skin appendages showed strong cytoplasmic m e m b r a n e staining for CD44v6 (Fig 1C, F). None of the benign nevomelanocytic lesions showed CD44v6 staining. In contrast, all of the malignant nevomelanocytic lesions showed cytoplasmic staining of the tumor cells (Fig 1F). In most instances, this staining was relatively uniform, but less intense than the staining of the overlying epidermis.
Chondroitinase Treatment The degree of CD44s expression by 11 (50%) of the primary and metastatic malignant lesions after treatm e n t with chondroitinase was c o m p a r e d with pretreatm e n t expression. No pattern of change (increased or decreased) was identifiable. Specifically, CD44s staining was not altered as a result of enzymatic digestion. Results after application of the three monoclonal antibodies to unsulfated (A-0), 4-sulfated (A-4), and @sulfated (A-6) chondroitin sulfate proteoglycans were as follows. In n o n e of the 11 cases was there any significant staining with two of the three anti-proteoglycan antibodies in the malignant cells (A-0 and A-6). Only A-4 staining was weakly identified in most of the malignant melanomas. However, A-0 prominently stained vessel walls and A-4 prominently stained stroma; no difference in stroreal staining intensity was observed within or adjacent to the t u m o r or in more distantly located stroma. A-6 staining was essentially absent in both tumor cells and stromal components.
DISCUSSION CD44 is the designation for a family of integral transmembrane glycoprotein molecules e n c o d e d by a gene located on the short arm of c h r o m o s o m e 11. The CD44 gene is an approximately 60-kb segment of genomic DNA containing at least 19 exons, of which nine (exons 6 to 14) may show alternative messenger RNA splicing, leading to substantial heterogeneity in the extracellular p o r n"o n of the p ro teinmolecule. s' 15 Recognition that CD44 functions as a lymphocyte h o m i n g receptor led some investigators to further postulate that certain CD44 isoforms expressed by epithelial and other cells may play a role in the dissemination of vari-
1290
,o
-(3
0
cO O_
HUMAN PATHOLOGY
Volume 27, No. 12 (December 1996)
FIGURE I. (A) Intradermal nevus, H&E stain; (B) intradermal nevus, CD44s immunoperoxidase showing strong (3+) membrane staining of nevus cells and epidermal ceils; (C) Intradermal nevus, CD44v6 immunoperoxidase showing strong epidermal staining, but no staining of benign nevus cells; (D) Malignant melanoma, H&E stain; (E) Malignant melanoma, CD44s immunoperoxidase showing strong (3+) membrane staining of malignant ceils and epidermal cells; (F) Malignant melanoma, CD44v6 immunoperoxidase showing weak cytoplasmic staining of malignant cells and strong (3+) staining of epidermal cells (all images, original magnification ×200).
ous neoplasms. 16'~7 This idea is especially attractive because another major function of some CD44 isoforms is to serve as the principal receptor for hyaluronan, an extracellular matrix ~lyeosaminoglycan present in a large variety of tissues. ,4,18 When expressed as a cell surface proteoglycan, CD44 can also bind to other extracellular matrix components such as fibronectin and types I and IV collagen, where it has been implicated in mediating tumor cell migration and invasion in vitro. 6s The role of CD44 in cell-extracellular matrix interactions has led to several studies that attempt to determine the relationship between CD44 expression and the ability of various neoplasms to metastasize. Altered expression of CD44 molecules may facilitate dissemination by altering migratory or adhesive properties of the neoplastic cells. Additionally, the specific CD44 isoform(s) expressed by a given neoplasm may have bearing on these properties. Examination of CD44 isoform expression has yielded varying results for different neoplasms. For example, some h u m a n neoplasms show apparent upregulation of the splice variant CD44v6 (eg, colon carcinomas), whereas others show downregulation (eg, squamoeellular neoplasms). 1922
Specifically with regard to malignant melanoma, studies in mouse models have indicated that melanoma cell lines expressing high levels of CD44s (standard form) show a greater propensity for hematogenous dissemination than melanomas expressing low levels of CD44s. 9 Cell culture studies have confirmed the importance of CD44 for melanoma cell migration on hyaluronan-coated surfaces, and immunohistochemical studies have shown at least some expression of CD44s in both benign and malignant nevomelanocytic lesions. 1°14 Our intent was to immunohistochemically examine CD44 expression in nevomelanocytic lesions in an effort to determine the relationship between expression of this marker and prognostically important histopathological characteristics of these lesions. We chose to study expression of CD44s, the standard form, and CD44v6. CD44v6 was chosen because upregulation of CD44v6 has been described in rat pancreatic carcinomas, aggressive h u m a n non-Hodgkin's lymphomas, and in some studies of h u m a n colon carcinomas. 18'~1-2~ The results of our immunohistochemical study indicate that benign nevomelanocytie lesions show strong, uniform expression of the standard form of
1292
CD44 EXPRESSION IN NEVOMELANOCYTIC LESIONS (Leigh et al)
CD44 (CD44s). This finding is in keeping with that of other investigators. 12'~3 As a group, the malignant nevomelanocytic lesions examined showed heterogeneous and decreased expression o f the standard form of CD44, which inversely correlated with increasing size, depth, and level of invasion (all P < .05). Individual neoplasms followed this trend fairly closely, although occasional tumors, such as no. 3 and no. 22 (Table 2), were sizable and fairly deeply invasive, but nonetheless showed comparatively strong expression of CD44s. Interestingly, the architecturally disordered nevi examined showed strong m e m b r a n e staining like the benign nevoid lesions. Although it is compelling to consider how CD44s (and CD44v6) might be used to distinguish architecturally disordered nevi from melanoma, only three such nevi were examined. In addition, small, superficially invasive melanomas showed a similarly intense level of expression of CD44s. Therefore, we hesitate to draw any conclusions regarding the diagnostic utility of these molecules in the context of such "borderline" lesions. In a mouse model, Birch et al 9 have previously reported the propensity of melanomas that express high levels of CD44s to form larger and more numerous pulmonary metastases than melanoma cell lines expressing low levels of CD44s. 9 Based on their findings, one might expect that h u m a n melanomas that are large, deeply invasive, or have already metastasized would express high levels of CD44s by immunohistochemistry. The reason for the apparent discrepancy between that study and the current one is unclear. Si and Hershey n described a 5.7-mm-thick primary melanoma and two metastases that showed essentially no immunohistochemical expression of CD44s, and so the finding of decreased CD44s expression in advanced melanomas is not unprecedented. All of the malignant lesions examined showed some expression of CD44v6, a particular isoform that has shown altered expression in a variety of neoplasms, as discussed previously. The pattern of staining was generally weak and limited to the cytoplasm rather than the cytoplasmic m e m b r a n e - - a n unexpected finding because CD44 is a transmembrane glycoprotein. The significance of this expression is therefore unclear, and it may be spurious. However, it is interesting that, in contrast, n o n e of the benign nevoid lesions showed cytoplasmic or m e m b r a n e expression of CD44v6, making it more difficult to ascribe this observation to artifact only. Fernandez-Figueras et al ~3found no expression of CD44v6 by either benign or malignant lesions in their immunohistochemical study (also p e r f o r m e d on paraffin-embedded tissue). In that study, five Spitz nevi, two c o m m o n nevi, five primary melanomas, and seven metastatic melanomas (all cutaneous) were examined. 13 One potential explanation for the finding of decreased CD44s expression by some malignant lesions is that CS, by binding CD44s, may mask the epitope of interest for immunohistochemical detection of CD44. T r e a t m e n t of the tumors with the enzyme chondroitinase was p e r f o r m e d to unmask potentially hidden epitopes. However, subsequent assessment of CD44s expression was essentially no different from that before
treatment. Therefore, this seems an unlikely explanation for the finding of decreased CD44s expression by malignant nevoid lesions. Although the staining intensity of CD44 core protein was not dramatically altered as a result of digestion of the tumor sections with chondroitinase ABC, these studies show that tumors express low levels of CSPG containing C-4 sulfate. Melanoma cell CD44/CSPG isolated from cell lines contains predominantly C-4 sulfate. 6 Melanoma expression of CD44/CSPG has been linked to tumor cell migration of types I and 1V collagen and invasion of basement membranes in vitro. 6-s The low intensity of staining with anti-C4-CS is not surprising, given the fact that CD44/ CSPG expression is a small percentage (5% to 15%) of the total cell surface CD44 on melanoma cell lines (Knutson et al, unpublished observation, January 1996). In conclusion, benign nevomelanocytic lesions show strong, uniform expression of the standard form of CD44, whereas malignant lesions show heterogeneous and decreased expression of the standard form of CD44, which shows a statistically significant (P < .05) inverse correlation with increasing size, depth, and level of invasion. For these reasons, CD44 expression may be a useful prognostic marker in malignant melanomas. Whether CD44s expression has prospective prognostic implications i n d e p e n d e n t of these usual histopathologically important prognostic indicators has not been investigated. CD44v6 is a variant isoform of CD44 that has shown altered expression in a variety of neoplasms, but that is apparently not expressed by benign nevomelanocytic lesions and shows weak cytoplasmic expression in malignant nevomelanocytic lesions. Immunohistochemical staining with anti-CS antibodies indicates that CS does not mask the CD44s epitope in melanocytic tumors and that weak staining for A-4 CS is present. Taken together, these results indicate that CD44 may be important in the biology of malignant melanoma.
REFERENCES 1. Underhill C: CD44: The hyalnronan receptor. J Cell Sci 103:293-298, 1992 2. Wang C, Tammi M, Tammi R: Distribution of hyaluronan and its CD44 receptor in the epithelia of human skin appendages. Histochemistry 98:105-112, 1992 3. Gunthert U, Stauder R, Mayer B, et al: Are CD44 variant isoforms involved in human tumour progression? in Hart I, Hogg N (eds): Cancer Surveys. NewYork, NY, Cold Spring Harbor Laboratory Press, 1995, pp 19-42 4. Terpe HJ, Stark H, Prehm P, et al: CD44 variant isoforms are preferentially expressed in basal epithelia of non-malignant human fetal and adult tissues. Histochemistry 101:79-89, 1994 5. Jalkanen S, Jalkanen M: Lymphocyte CD44 binds the COOHterminal heparin binding domain of fibronectin. J Cell Biol 116:817825, 1992 6. Faassen AE, Schrager JA, Klein DJ, et al: A cell surface chondroitin sulfate proteoglycan immunologically related to CD44 is involved in type I collagen-mediated melanoma cell motility and invasion. J Cell Biol 116:521-531, 1992 7. Faassen AE, Mooradian DL, Tranquillo RT, et al: Cell surface CD44-related chondroitin sulfate proteoglycan is required for transforming growth factor-stimulated mouse melanoma cell motility and invasive behavior on type I collagen. J Cell Sci 105:501-511, 1993
1293
HUMAN PATHOLOGY
Volume 27, No. 12 (December 1996)
8. KnutsonJR, IidaJ, Fields G, et al: CD44/Chondroitin sulfate proteoglycan and a a2/31 integrin mediate human melanoma cell migration on type 1V collagen and invasion of basement membranes. Mol Biol Cell 7:383-396, 1996 9. Birch M, Mitchell S, Hart IR: Isolation and characterization of human melanoma cell variants expressing high and low levels of CD44. Cancer Res 51:6660-6667, 1991 10. Thomas L, Randolph Byers H, et al: CD44H regulates tumor cell migration on hyaluronate-coated substrate. J Cell Biol 118:971977, 1992 11. Si Z, Hershey P: Immunohistologieal examination of the relationship between metastatic potential and expression of adhesion molecules and "selectins" on melanoma cells. Pathology 26:6-15, 1994 12. Moretti S, Martini L, Berti E, et al: Adhesion molecule profile and malignancy of melanocytic lesions. Melanoma Res 3:235-239, 1993 13. Fernandez-Figueras MT, Ariza A, Calatrava A, et al: CD44 and melanocytic tumors: A possible role for standard CD44 in the epidelwnotropic spread of melanoma. J Cutan Pathol 22:60, 1995 14. Thomas L, Etoh T, Stamenkovic I, et al: Migration of human melanoma cells on hyaluronate is related to CD44 expression.J Invest Dermatol 100:115-120, 1993 15. Dadi HK, Greaves AM, Cox DW: Identification of a polymorphism in human CD44. Nucleic Acids Res 19:969, 1991
16, Jalkanen S, Saari S, Kalimo H, et al: Lymphocyte migration into the skin: The role of lymphocyte homing receptor (CD44) and endothelial cell antigen (HECA-452). J Invest Dermatol 94:786-792, 1990 17. Gunthert U, Hofmann M, Rudy W, et al: A new variant of glycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell 65:13-24, 1991 18. Aruffo A, Stamenkovic I, Melnick M, et al: CD44 is the principle cell surface receptor for hyaluronate. Cell 61:1303-1313, 1990 19. Fox SB, Jackson DG, Screaton GR, et al: CD44 and cancer screening. Lancet 342:548-549, 1993 20. Salmi M, Gron-Virta K, Sointu P, et al: Regulated expression of exon v6 containing isoforms of CD44 in man: Down-regulation during malignant transformation of tumors of squamocellular origin. J Cell Biol 122:431-442, 1993 21. Heider K-H, Hofmann M, Hors E, et al: A human homologue of the rat metastasis-associated variant of CD44 is expressed in colorectal carcinomas and adenomatous polyps. J Cell Biol 120:227-233, 1993 22. WielengaVJM, Heider K-H, Offerhaus GJA, et al: Expression of CD44 variant proteins in human colorectal cancer is related to tumor progression. Cancer Res 53:4754-4756, 1993 23. Koopman G, Heider K-H, Horst E, et al: Activated human lymphocytes and aggressive non-Hodgkin's lymphomas express a homologue of the rat metastasis-associated variant of CD44. J Exp Med 177:897-904, 1993
1294
sion in metastases paralleled that of their respective primaries. None of the benign nevomelanocytic lesions showed CD44v6 staining. In contrast, all of the malignant nevomelanocytic lesions showed cytoplasmic staining of the tumor cells. Pretreatment with chondroitinase did not alter CD44s staining. CD44s expression byimmunohistochemical determination is uniform in benign nevomelanocytic lesions. Malignant melanomas show decreased, heterogeneous staining that inversely correlates with increasing size, depth, and level of invasion. CD44 expression may be a prognostic indicator in malignant melanomas. Tumor staining with anti-chondroitin sulfate monoclonal antibodies suggests that CD44s may be expressed as a chondroitin sulfate proteoglycan in primary melanomas. HUM PATHOL 27:12881294. Copyright © 1996 by W.B. Saunders Company Key words:melanoma, nevus, cellular adhesion molecules, immunohistochemistry, CD44. Abbreviations: CSPG, cell surface chondroitin sulfate proteogiycan; CS, chondroitin sulfate; ABC, avidin-hiotin complex; VGF, vertical growth phase; DAB, diaminobenzidine; PBS, phosphate-buffered saline; Ig, immunoglobulin.
CD44 is the designation for a family of transmembrane glycoprotein molecules, a primary function of which is to serve as the principal ligand for hyaluronan, an important extracellular matrix glycosaminoglycan present in a large variety of tissues] 4 The role of CD44 in cell-extracellular matrix interactions has led some investigators to postulate that altered CD44 expression by various neoplasms may relate to their ability to metastasize. For instance, altered CD44 expression may result in different adhesive or migratory properties of neoplastic cells, thereby potentiating the ability of a tumor to infiltrate dermal or other tissues. Several different isoforms of CD44 are known to exist, and the specific isoform(s) expressed also m,a3ybe of some import with regard to metastatic potential. Cell type-specific differences in CD44 expression are also created by posttranslational addition of O- or N-linked carbohydrates. Furthermore, four serine residues potentially can serve as acceptor sites for either heparin sulfate or chondroitin sulfate glycosaminoglycans. 5-s Such posttranslational
modifications can alter the ligand binding properties of the CD44 core protein, and CD44 expressed as a cell surface chondroitin sulfate proteoglycan (CSPG) has been shown to bind to fibronectin and types I and IV collagen. 5-s Evidence for a potential role of CD44 in the dissemination of malignant melanomas comes from a variety of sources. Previous studies of malignant melanoma in mouse models have indicated that melanoma cell lines expressing high levels of CD44s (standard isoform) show a greater propensity for hematogenous dissemination than melanomas expressing low levels of CD44s. 9 Cell culture studies have shown the importance of CD44 for melanoma cell migration on hyaluronan-coated surfaces, and immunohistochemical studies have demonstrated at least some expression of CD44s in both benign and malignant nevomelanocytic lesions. 1°14 Furthermore, expression of CD44/CSPG on melanoma cells has been linked to regulation of melanoma cell migration and invasion of collagen gels or reconstituted basement membranes in vitro. "s The aim of the current study was to immunohistochemically examine CD44 expression in benign and malignant nevomelanocytic lesions in an effort to determine the relationship between expression and prognostically important histopathological characteristics of these lesions. The studies indicate that CD44s is a major isoform expressed on benign and malignant nevomelanocytic lesions. Furthermore, digestion of the tumor tissue sections with chondroitinase avidin-biotin complex (ABC) before staining with specific anti-CS (chondroitin sulfate) antibodies demonstrated positive intratumor staining for chondroitin-4 sulfate, consistent with
From the Departments of Pathology, Urology, and Dermatology, The University of Iowa and Veterans Affairs Medical Center, Iowa City, IA; and the Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN. Accepted for publication May 9, 1996. Supported by an American Cancer Society Clinical Oncology Career Development Award and by the Office of Research and Development (Medical Research Service), Department of Veterans Affairs (M.B.C.), Iowa City, IA. Address correspondence and reprint requests to Michael B. Cohen, MD, Department of Pathology, 200 Hawkins Drive, 5216 RCP, The University of Iowa, Iowa City, IA 52242-1009. Copyright © 1996 by W.B. Saunders Company
0046-8177/96/2712-002055.00/0
1288
CD44 EXPRESSION IN NEVOMELANOCYTIC LESIONS (Leigh et al)
TABLE I.
Nevomelanocytic Lesions Examined
Benign nevomelanocytic lesions Junctional nevi Compound nevi Intradermal nevi Blue nevi Spitz nevi Deep penetrating nevi Dysplastic nevi Malignant nevomelanocytic lesions Superficial spreading Lentigo maligna melanoma Nodular Acral lentiginous Metastatic melanoma
Monoclonal Antibodies and Immunohistochemistry
3 4 5 5 6 1 3 7 2 8 5 8
the in vivo expression of CD44/CSPG in nevomelanocytic lesions. MATERIALS AND METHODS Tissue Twenty-four benign nevi, three architecturally disord e r e d (dysplastic) nevi, 22 primary melanomas, and eight metastatic m e l a n o m a s were examined. Each of the lesions were from different patients, except for the metastases, which were taken from a subset of the primary m e l a n o m a patients. T h e m e l a n o m a s selected i n c l u d e d various histological types of varying Breslow d e p t h and Clark level (Tables 1 and 2). O f the 22, 14 primary m e l a n o m a s were considered to be in vertical growth phase (VGF); all VGF lesions were Clark's levels III, IV, or V. Representative 4-#m sections of formalinfixed, paraffin-embedded archival tissue were used for hematoxylin-eosin and i m m n n o h i s t o c h e m i c a l stains in each instance.
TABLE 2.
An i m m u n o p e r o x i d a s e reaction for CD44s and CD44v6 was p e r f o r m e d in each case using a labeled streptavidin-biotin method. Sections were deparaffinized in xylene, rehydrated in graded alcohols, and rinsed. Antigen retrieval was accomplished using citrate buffer for 10 minutes at high power in a microwave. E n d o g e n o u s peroxidase activity was blocked using an hydrogen p e r o x i d e / d i s t i l l e d water solution (0.3% and 1.88% for CD44s and CD44v6, respectively). Sections were covered with primary antisera incubated overnight at 4°C and rinsed. Sections were then covered with LSAB II link antibody for 10 minutes (CD44v6) or 20 minutes (CD44s) at r o o m temperature. Sections were incubated with 0.05% d i a m i n o b e n z i d i n e (DAB)/0.1 m o l / L Tris with 0.015% hydrogen p e r o x i d e to show signal of the primary antibody. A counterstain of 10% Harris hematoxylin without acid for 1 to 2 minutes was used. Negative control slides were p r e p a r e d by substituting normal mouse serum for the primary antisera. CD44s antibody was purchased from Becton Dickinson (San Jose, CA) and used at a dilution of 1:50. CD44v6 antibody was purchased f r o m B e n d e r Medical Systems (Vienna, Austria) and used at a dilution of 1:100. Biotinylated link and label were purchased from DAKO Corporation (Carpinteria, CA).
Chondroitinase Pretreatment Because CD44 may u n d e r g o posttranslational modification, including CS addition, we also attempted to d e t e r m i n e what role this might play in masking CD44 expression in nevomelanocytic lesions. Representative tissue samples of 11 tumors f r o m 10 patients were selected for chondroitinase pretreatment. Cases were selected to include a range of t u m o r types, sizes, and depths of invasion. T h e specimens were deparaffinized, rehydrated in graded alcohols, and washed in deionized water. Control slides for standard CD44 staining
Characteristics of Primary Melanomas
Case No.
Sex
Age
Diagnosis
Clark Level
Breslow Depth (mm)
Lesion Size (cm)
CD44s Staining
CD44v6 Staining
1 2 3 4 5 6 7 8 9 10 11 12 I3 14 15 16 17 18 19 20 21 22
F M M M M M F F F F F M M F F F F M F M M M
49 37 80 35 57 62 75 74 76 50 32 86 46 67 38 70 62 37 37 53 53 48
SSMIS SSMIS SSMM SSMIS SSMM SSMM LMM LMM NMM NMM NMM NMM NMM ALMMIS ALMM ALMM ALMM NMM NMM NMM SSMM ALMM
I I 1V I III II II II IV IV IV V IV I IV IV III III IV IV II V
0.00 0.00 1.49 0.00 2.10 0.19 0.54 0.32 2.80 0.85 2.86 12.40 2.31 0.00 2.20 11.00 1.15 4.50 2.50 1.85 0.22 6.00
1.0 0.8 1.1 0.6 1.3 1.3 0.8 0.3 1.0 0.7 1.5 3.6 0.4 0.4 2.l 1.5 0.7 2.0 2.2 2.5 1.9 2.0
2+ 3+ 3+ 3+ 2+ 3+ 3+ 2+ 1+ 2+ 2+ 1+ 2+ 3+ 1+ 1+ 2+ 1+ 1+ 2+ 2+ 2+
+ + + + + + + + + + + + + + + + + + + + + +
Abbreviations: MIS, melanoma-in-situ; ALMM, acral lentiginous malignant melanoma; LMM, lentigo maligna melanoma; NMM, nodular malignant melanoma; SSMM, superficial spreading malignant melanoma; CD44s, CD44, standard form; CD44v6, CD44, v6 variant (exon 10).
1289
HUMAN PATHOLOGY TABLE 3. 0 I+ 2+ 3+
Volume 27, No, 12 (December 1996)
Grading Scheme for CD44s
Tumor cells show no cytoplasmicmembrane staining < 5 0 %of tumor cells show membrane staining similar to that of basal keratinocytes 50% to 90% of tumor cells show membrane staining similar to that of basal keratinocytes >90%of tumor cells show membrane staining similar to that of basal keratinocytes
(ie, no chondroitinase pretreatment) were prepared as previously described. Slides for chondroitinase pretreatment were covered with chondroitinase-ABC (Seikagaku Kogyo Co, LTD, Tokyo, Japan) at a concentration of 0.2 U/mL. The slides were incubated in a humidity chamber at 37°C for 60 minutes and then washed twice in phosphate-buffered saline (PBS). After pretreatment with chondroitinase-avidin-biotin complex (ABC), individual slides were treated with monoclonal antibody to either CD44s or CSPG. CD44s staining was performed as previously described. Three anti-CSPG monoclonal antibodies that differed in specificity with respect to the location of the sulfate group were tested. These included antibody to unsulfated CSPG (anti-PC,-A-Di-0S; mouse immunoglobulin [Ig] G1), 4-sulfated CSPG (anti-PG-A-Di-4S; mouse IgG1), and 6-sulfated CSPG (anti-PG-A-Di-6S; mouse !gM) (all from ICN Biomedicals, Inc, Costa Mesa, CA). The anti-CSPG antibodies were used at a concentration of 1:250. Negative mouse anti-human IgG at a concentration of 1:10,000 was used as a control. The slides were incubated with antibody for 60 minutes at room temperature then washed twice with PBS. Horseradish-peroxidase goat anti-mouse IgG (Calbiochem, San Diego, CA) at a concentration of 1:250 was then added to each slide for 60 minutes at room temperature. Specimens were washed twice in 0.1 mol/L Tris and developed in 0.05% DAB in 0.03% hydrogen peroxide and 0.1 mol/L Tris for 5 minutes. The slides were then counterstained with 10% hematoxylin, dehydrated, cleared with xylene, and coverslipped.
Analysis Slides stained for CD44s were examined and graded semiquantitatively using a 0-3+ scale (Table 3). Staining of the lower (basal) epidermal layers served as an internal control for all but the metastatic lesions. CD44v6 staining of lesions was not stratified, but was considered either positive or negative. In the lesions expressing CD44v6, virtually all lesional cells were uniformly weakly positive, defying further stratification. Results were correlated with lesion size, Breslow's depth, and Clark's level using the Pearson product-moment correlation method.
RESULTS CD44s In each case, the lower epidermal layers and the epithelia of the skin appendages showed strong cytoplasmic m e m b r a n e staining for CD44s, as described in other reports 2'12'1s (Fig 1B, E). All of the benign nevomelanocytic lesions, including three architecturally disordered nevi, showed uniform m e m b r a n e staining for CD44s of an intensity similar to that of the overlying epithelium (3+) (Fig 1B). The malignant lesions, however, showed a tendency toward heterogeneous mem-
brane staining, often with large sheets and nodules of t u m o r showing very weak or no staining. Semiquantitarive analysis (0 to 3+) of CD44s staining a m o n g the primary melanomas showed an inverse correlation between lesion size (-0.569) and depth of invasion ( - 0 . 6 2 2 and - 0 . 6 1 7 for Breslow depth and Clark level, respectively) using the Pearson p r o d u c t - m o m e n t correlation. These results were all significant at P < .05. Notable exceptions to this trend were case nos. 3 and 22 (Table 2), which were sizeable and deeply invasive, but which nonetheless showed 3+ and 2+ expression, respectively. CD44s expression in metastases paralleled that of their respective primaries.
CD44v6 In each case, the lower epidermal layers and epithelia of the skin appendages showed strong cytoplasmic m e m b r a n e staining for CD44v6 (Fig 1C, F). None of the benign nevomelanocytic lesions showed CD44v6 staining. In contrast, all of the malignant nevomelanocytic lesions showed cytoplasmic staining of the tumor cells (Fig 1F). In most instances, this staining was relatively uniform, but less intense than the staining of the overlying epidermis.
Chondroitinase Treatment The degree of CD44s expression by 11 (50%) of the primary and metastatic malignant lesions after treatm e n t with chondroitinase was c o m p a r e d with pretreatm e n t expression. No pattern of change (increased or decreased) was identifiable. Specifically, CD44s staining was not altered as a result of enzymatic digestion. Results after application of the three monoclonal antibodies to unsulfated (A-0), 4-sulfated (A-4), and @sulfated (A-6) chondroitin sulfate proteoglycans were as follows. In n o n e of the 11 cases was there any significant staining with two of the three anti-proteoglycan antibodies in the malignant cells (A-0 and A-6). Only A-4 staining was weakly identified in most of the malignant melanomas. However, A-0 prominently stained vessel walls and A-4 prominently stained stroma; no difference in stroreal staining intensity was observed within or adjacent to the t u m o r or in more distantly located stroma. A-6 staining was essentially absent in both tumor cells and stromal components.
DISCUSSION CD44 is the designation for a family of integral transmembrane glycoprotein molecules e n c o d e d by a gene located on the short arm of c h r o m o s o m e 11. The CD44 gene is an approximately 60-kb segment of genomic DNA containing at least 19 exons, of which nine (exons 6 to 14) may show alternative messenger RNA splicing, leading to substantial heterogeneity in the extracellular p o r n"o n of the p ro teinmolecule. s' 15 Recognition that CD44 functions as a lymphocyte h o m i n g receptor led some investigators to further postulate that certain CD44 isoforms expressed by epithelial and other cells may play a role in the dissemination of vari-
1290
,o
-(3
0
cO O_
HUMAN PATHOLOGY
Volume 27, No. 12 (December 1996)
FIGURE I. (A) Intradermal nevus, H&E stain; (B) intradermal nevus, CD44s immunoperoxidase showing strong (3+) membrane staining of nevus cells and epidermal ceils; (C) Intradermal nevus, CD44v6 immunoperoxidase showing strong epidermal staining, but no staining of benign nevus cells; (D) Malignant melanoma, H&E stain; (E) Malignant melanoma, CD44s immunoperoxidase showing strong (3+) membrane staining of malignant ceils and epidermal cells; (F) Malignant melanoma, CD44v6 immunoperoxidase showing weak cytoplasmic staining of malignant cells and strong (3+) staining of epidermal cells (all images, original magnification ×200).
ous neoplasms. 16'~7 This idea is especially attractive because another major function of some CD44 isoforms is to serve as the principal receptor for hyaluronan, an extracellular matrix ~lyeosaminoglycan present in a large variety of tissues. ,4,18 When expressed as a cell surface proteoglycan, CD44 can also bind to other extracellular matrix components such as fibronectin and types I and IV collagen, where it has been implicated in mediating tumor cell migration and invasion in vitro. 6s The role of CD44 in cell-extracellular matrix interactions has led to several studies that attempt to determine the relationship between CD44 expression and the ability of various neoplasms to metastasize. Altered expression of CD44 molecules may facilitate dissemination by altering migratory or adhesive properties of the neoplastic cells. Additionally, the specific CD44 isoform(s) expressed by a given neoplasm may have bearing on these properties. Examination of CD44 isoform expression has yielded varying results for different neoplasms. For example, some h u m a n neoplasms show apparent upregulation of the splice variant CD44v6 (eg, colon carcinomas), whereas others show downregulation (eg, squamoeellular neoplasms). 1922
Specifically with regard to malignant melanoma, studies in mouse models have indicated that melanoma cell lines expressing high levels of CD44s (standard form) show a greater propensity for hematogenous dissemination than melanomas expressing low levels of CD44s. 9 Cell culture studies have confirmed the importance of CD44 for melanoma cell migration on hyaluronan-coated surfaces, and immunohistochemical studies have shown at least some expression of CD44s in both benign and malignant nevomelanocytic lesions. 1°14 Our intent was to immunohistochemically examine CD44 expression in nevomelanocytic lesions in an effort to determine the relationship between expression of this marker and prognostically important histopathological characteristics of these lesions. We chose to study expression of CD44s, the standard form, and CD44v6. CD44v6 was chosen because upregulation of CD44v6 has been described in rat pancreatic carcinomas, aggressive h u m a n non-Hodgkin's lymphomas, and in some studies of h u m a n colon carcinomas. 18'~1-2~ The results of our immunohistochemical study indicate that benign nevomelanocytie lesions show strong, uniform expression of the standard form of
1292
CD44 EXPRESSION IN NEVOMELANOCYTIC LESIONS (Leigh et al)
CD44 (CD44s). This finding is in keeping with that of other investigators. 12'~3 As a group, the malignant nevomelanocytic lesions examined showed heterogeneous and decreased expression o f the standard form of CD44, which inversely correlated with increasing size, depth, and level of invasion (all P < .05). Individual neoplasms followed this trend fairly closely, although occasional tumors, such as no. 3 and no. 22 (Table 2), were sizable and fairly deeply invasive, but nonetheless showed comparatively strong expression of CD44s. Interestingly, the architecturally disordered nevi examined showed strong m e m b r a n e staining like the benign nevoid lesions. Although it is compelling to consider how CD44s (and CD44v6) might be used to distinguish architecturally disordered nevi from melanoma, only three such nevi were examined. In addition, small, superficially invasive melanomas showed a similarly intense level of expression of CD44s. Therefore, we hesitate to draw any conclusions regarding the diagnostic utility of these molecules in the context of such "borderline" lesions. In a mouse model, Birch et al 9 have previously reported the propensity of melanomas that express high levels of CD44s to form larger and more numerous pulmonary metastases than melanoma cell lines expressing low levels of CD44s. 9 Based on their findings, one might expect that h u m a n melanomas that are large, deeply invasive, or have already metastasized would express high levels of CD44s by immunohistochemistry. The reason for the apparent discrepancy between that study and the current one is unclear. Si and Hershey n described a 5.7-mm-thick primary melanoma and two metastases that showed essentially no immunohistochemical expression of CD44s, and so the finding of decreased CD44s expression in advanced melanomas is not unprecedented. All of the malignant lesions examined showed some expression of CD44v6, a particular isoform that has shown altered expression in a variety of neoplasms, as discussed previously. The pattern of staining was generally weak and limited to the cytoplasm rather than the cytoplasmic m e m b r a n e - - a n unexpected finding because CD44 is a transmembrane glycoprotein. The significance of this expression is therefore unclear, and it may be spurious. However, it is interesting that, in contrast, n o n e of the benign nevoid lesions showed cytoplasmic or m e m b r a n e expression of CD44v6, making it more difficult to ascribe this observation to artifact only. Fernandez-Figueras et al ~3found no expression of CD44v6 by either benign or malignant lesions in their immunohistochemical study (also p e r f o r m e d on paraffin-embedded tissue). In that study, five Spitz nevi, two c o m m o n nevi, five primary melanomas, and seven metastatic melanomas (all cutaneous) were examined. 13 One potential explanation for the finding of decreased CD44s expression by some malignant lesions is that CS, by binding CD44s, may mask the epitope of interest for immunohistochemical detection of CD44. T r e a t m e n t of the tumors with the enzyme chondroitinase was p e r f o r m e d to unmask potentially hidden epitopes. However, subsequent assessment of CD44s expression was essentially no different from that before
treatment. Therefore, this seems an unlikely explanation for the finding of decreased CD44s expression by malignant nevoid lesions. Although the staining intensity of CD44 core protein was not dramatically altered as a result of digestion of the tumor sections with chondroitinase ABC, these studies show that tumors express low levels of CSPG containing C-4 sulfate. Melanoma cell CD44/CSPG isolated from cell lines contains predominantly C-4 sulfate. 6 Melanoma expression of CD44/CSPG has been linked to tumor cell migration of types I and 1V collagen and invasion of basement membranes in vitro. 6-s The low intensity of staining with anti-C4-CS is not surprising, given the fact that CD44/ CSPG expression is a small percentage (5% to 15%) of the total cell surface CD44 on melanoma cell lines (Knutson et al, unpublished observation, January 1996). In conclusion, benign nevomelanocytic lesions show strong, uniform expression of the standard form of CD44, whereas malignant lesions show heterogeneous and decreased expression of the standard form of CD44, which shows a statistically significant (P < .05) inverse correlation with increasing size, depth, and level of invasion. For these reasons, CD44 expression may be a useful prognostic marker in malignant melanomas. Whether CD44s expression has prospective prognostic implications i n d e p e n d e n t of these usual histopathologically important prognostic indicators has not been investigated. CD44v6 is a variant isoform of CD44 that has shown altered expression in a variety of neoplasms, but that is apparently not expressed by benign nevomelanocytic lesions and shows weak cytoplasmic expression in malignant nevomelanocytic lesions. Immunohistochemical staining with anti-CS antibodies indicates that CS does not mask the CD44s epitope in melanocytic tumors and that weak staining for A-4 CS is present. Taken together, these results indicate that CD44 may be important in the biology of malignant melanoma.
REFERENCES 1. Underhill C: CD44: The hyalnronan receptor. J Cell Sci 103:293-298, 1992 2. Wang C, Tammi M, Tammi R: Distribution of hyaluronan and its CD44 receptor in the epithelia of human skin appendages. Histochemistry 98:105-112, 1992 3. Gunthert U, Stauder R, Mayer B, et al: Are CD44 variant isoforms involved in human tumour progression? in Hart I, Hogg N (eds): Cancer Surveys. NewYork, NY, Cold Spring Harbor Laboratory Press, 1995, pp 19-42 4. Terpe HJ, Stark H, Prehm P, et al: CD44 variant isoforms are preferentially expressed in basal epithelia of non-malignant human fetal and adult tissues. Histochemistry 101:79-89, 1994 5. Jalkanen S, Jalkanen M: Lymphocyte CD44 binds the COOHterminal heparin binding domain of fibronectin. J Cell Biol 116:817825, 1992 6. Faassen AE, Schrager JA, Klein DJ, et al: A cell surface chondroitin sulfate proteoglycan immunologically related to CD44 is involved in type I collagen-mediated melanoma cell motility and invasion. J Cell Biol 116:521-531, 1992 7. Faassen AE, Mooradian DL, Tranquillo RT, et al: Cell surface CD44-related chondroitin sulfate proteoglycan is required for transforming growth factor-stimulated mouse melanoma cell motility and invasive behavior on type I collagen. J Cell Sci 105:501-511, 1993
1293
HUMAN PATHOLOGY
Volume 27, No. 12 (December 1996)
8. KnutsonJR, IidaJ, Fields G, et al: CD44/Chondroitin sulfate proteoglycan and a a2/31 integrin mediate human melanoma cell migration on type 1V collagen and invasion of basement membranes. Mol Biol Cell 7:383-396, 1996 9. Birch M, Mitchell S, Hart IR: Isolation and characterization of human melanoma cell variants expressing high and low levels of CD44. Cancer Res 51:6660-6667, 1991 10. Thomas L, Randolph Byers H, et al: CD44H regulates tumor cell migration on hyaluronate-coated substrate. J Cell Biol 118:971977, 1992 11. Si Z, Hershey P: Immunohistologieal examination of the relationship between metastatic potential and expression of adhesion molecules and "selectins" on melanoma cells. Pathology 26:6-15, 1994 12. Moretti S, Martini L, Berti E, et al: Adhesion molecule profile and malignancy of melanocytic lesions. Melanoma Res 3:235-239, 1993 13. Fernandez-Figueras MT, Ariza A, Calatrava A, et al: CD44 and melanocytic tumors: A possible role for standard CD44 in the epidelwnotropic spread of melanoma. J Cutan Pathol 22:60, 1995 14. Thomas L, Etoh T, Stamenkovic I, et al: Migration of human melanoma cells on hyaluronate is related to CD44 expression.J Invest Dermatol 100:115-120, 1993 15. Dadi HK, Greaves AM, Cox DW: Identification of a polymorphism in human CD44. Nucleic Acids Res 19:969, 1991
16, Jalkanen S, Saari S, Kalimo H, et al: Lymphocyte migration into the skin: The role of lymphocyte homing receptor (CD44) and endothelial cell antigen (HECA-452). J Invest Dermatol 94:786-792, 1990 17. Gunthert U, Hofmann M, Rudy W, et al: A new variant of glycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell 65:13-24, 1991 18. Aruffo A, Stamenkovic I, Melnick M, et al: CD44 is the principle cell surface receptor for hyaluronate. Cell 61:1303-1313, 1990 19. Fox SB, Jackson DG, Screaton GR, et al: CD44 and cancer screening. Lancet 342:548-549, 1993 20. Salmi M, Gron-Virta K, Sointu P, et al: Regulated expression of exon v6 containing isoforms of CD44 in man: Down-regulation during malignant transformation of tumors of squamocellular origin. J Cell Biol 122:431-442, 1993 21. Heider K-H, Hofmann M, Hors E, et al: A human homologue of the rat metastasis-associated variant of CD44 is expressed in colorectal carcinomas and adenomatous polyps. J Cell Biol 120:227-233, 1993 22. WielengaVJM, Heider K-H, Offerhaus GJA, et al: Expression of CD44 variant proteins in human colorectal cancer is related to tumor progression. Cancer Res 53:4754-4756, 1993 23. Koopman G, Heider K-H, Horst E, et al: Activated human lymphocytes and aggressive non-Hodgkin's lymphomas express a homologue of the rat metastasis-associated variant of CD44. J Exp Med 177:897-904, 1993
1294