- Email: [email protected]
Benign and malignant forms of erythroderma: Cutaneous immunophenotypic characteristics
Dermatopatholoou Benign and malignant forms of erythroderma: Cutaneous immunophenotypic characteristics Elizabeth A. Abel, M.D.,* Mary L. Lindae, M.D.,* Richard T. Hoppe, M.D.,** and Gary S. Wood, M.D.*,*** Stanford and Palo Alto, CA In order to determine if imrnunohistologic features are useful in distinguishing benign from malignant types of erythroderma, we studied the immunophenotype of lesional T cells in 20 patients (8 mycosis fungoides/S6zary syndrome, 12 benign) and found them to be generally similar. In all cases, the majority of T cells were Leu-l+, Leu-4+, and Leu-5+, as is typical of mature T cells. Although in most cases a majority of Leu-3+ (helper/phenotype) T cells were present, in 2 there was a majority of the Leu-2+ (cytotoxic/suppressor) subset and in 12 others, a significant minority (20%-40%) of these ceils. Low percentages of Leu-2+ cells (<10%), resulting in high Leu-3+/Leu-2+ ratios, did not distinguish benign from malignant erythroderma. Leu-8 antigen deficiency was common in both mycosis fungoides/S6zary syndrome and benign cases (62% vs 75%, respectively). In contrast, Leu-9 antigen deficiency was present in only one patient in each group. The lack of combined Leu-8/9 antigen deficiency in our patients may be due to a heavy inflammatory T cell component, obscuring the antigen deficiencies seen in most nonerythrodermic mycosis fungoides cases. We conclude that immunophenotypic studies with the use of the current antibody panel show many similarities between benign and malignant forms of erythroderma, as well as some minor differences that may prove diagnostically useful if corroborated by future studies. (J AM ACADDEFtMArOL 1988;19:1089-95.)
Erythroderma is a cutaneous reaction pattern that can be seen in a wide variety of benign and malignant diseases, including mycosis fungoides and its leukemic variant, S+zary syndrome. The erythrodermic form of mycosis fungoides and the S6zary syndrome may be difficult to distinguish from benign forms of erythroderma. These include psoriatic erythroderma, pityriasis rubra pilaris, generalized exfoliative erythroderma secondary to drug eruption, atopic dermatitis, seborrheic dermatitis, contact dermatitis, or paraneoplastic cutaneous reaction to an internal malignancy. The most From the Departments of Dermatology,* Therapeutic Radiology,** and Pathology,*** Stanford University, Stanford, and the Veterans Administration Hospital, Palo Alto, CA.*** Supported by Merit Review funding to Dr. Wood from the Veterans Administration. Reprints not available.
important histologic features in mycosis fungoides are the presence of a bandlike lymphoid infiltrate at the dermoepidermal junction, with atypical cerebriform mononuclear cells and Pautrier intraepidermal microabscesses. Such changes were present in 7 of 11 patients with S6zary syndrome and in 2 of 4 patients with erythrodermic mycosis fungoides in one study.t In another series of 121 skin biopsy specimens from 39 patients with S6zary syndrome, only 47 (39% of specimens) showed a bandlike infiltrate with atypical lymphoid cells, of which 18 (15% of specimens) in 10 patients revealed epidermal involvement suggestive of mycosis fungoides? Features of chronic dermatitis probably related to pruritus with excoriations are present in most patients with S6zary syndrome. Furthermore, some cases reported as benign erythroderma may show 1089
1090
Journal of the American Academy of Dermatology
Abel et al.
Table I. Monoclonal antibody panel CD designation
Antibodies
Anti-Leu- 1 Anti-Leu-2 Anti-Leu-3 Anti-Leu-4 Anti-Leu-5 Anti-Leu-6 Anti-Leu-7 Anti-Leu-8
CD5 CD8 CD4 CD3 CD2 CD1
Anti-Leu-9
CD7
Anti-Tac TO 15 Anti-Leu-M3
CD25 CD22 CD14
Predominant specificity
Pan T cell (>95% T cells) Cytotoxic/suppressor T cell (15%-30% T cells) Helper T cell (70%-90% T cells) Pan T cell (>95% T cells) Pan T cell (>95% T cells) Langerhans cell/determinate cell Natural killer/killer cells Majority T cell (60% :t: 10% Leu-2+ T cells, 75% + 10% Leu-3+ T cells) Majority B cells, some granulocytes, monocytes, natural killer cells Majority T cell (100% Leu-2+ T cells, 90% Leu-3+ T cells) Interleukin 2 receptor B cell Monocyte/macrophage and myeloid cells
Table II. Leu-2 and Leu-3 antigen expression in erythroderma* Epidermis ('Leu-3+)
Dermis (Leu-3+)
Leu-2 +
__<10% 20%0-40%
-->50%
0
o
o
o
o
o
9
9
9
o
o
o
9
9
9
o
o
o
o
o
o
o
o
9
9
9
@
0
0
0
0
o
o
0
o , Benign erythroderma; 9 mycosis fungoides/S~zary syndrome erythroderma. *The Leu-3+ (helper) T cell subset generally predominated in the dermis; however, in the epidermis, Leu-2+ (eytotoxie/suppressor) T cells often equaled or exceeded Lea-3+ T cells, especially among the benign eases. Valu~ given for the dermis were virtually always the same as those for the epidermks and dermis combined. In no case did the sum of Leu-2+ and Leu-3+ cells exceed 100%.
histologic features resembling mycosis fungoides. In one report, 1 3 of 15 patients with chronic dermatitis showed dense infiltrates in the papillary dermis containing considerable numbers of large cerebriform mononuclear cells. Although there were no Pautrier microabscesses nor was there any significant epidermal infiltration, the histologic picture in these patients was indistinguishable from that in the 4 patients with erythrodermic mycosis fungoides. Additional investigations of blood and lymph nodes are usually necessary to make a definitive diagnosis of Srzary syndrome.
Immunophenotyping with conventional cell surface markers has not allowed distinction of mycosis fungoides from reactive cutaneous inflammation, as many reactive processes contain predominantly cells expressing a mature helper T cell phenotype: Leu-l+, L e u - 2 - , Leu-3+, Leu-4+, Leu-5+. However, our prior studies with the use of monoclonal antibodies to the majority T cell markers Leu-8 and Leu-9 have demonstrated major differences in cellular antigen expression between certain benign and malignant cutaneous T cell infiltrates. In a study of 41 skin biopsy specimens from 27 patients
Volume 19 Number 6 December 1988
with mycosis fungoides and 34 skin biopsy specimens from 33 controls, the Leu-8+/Leu-9+ phenotype was much more common in controls with benign disease than in patients with mycosis fungoides. Leu-8 deficiency was more sensitive but less specific than Leu-9 deficiency. 3 Similar results were found in early-stage malignant disease, such as poikilodermatous patch stage mycosis fungoides, as well as in idiopathic forms of poikiloderma, such as large-plaque parapsoriasis, suggesting the premalignant nature of this condition? This study was undertaken to determine if immunophenotyping might also help to distinguish S6zary syndrome or erythrodermic mycosis fungoides from benign forms of erythroderma. PATIENTS AND METHODS
Eight patients with cutaneous T cell lymphoma, two with erythrodermic mycosis fungoides, and six with Sdzary syndrome (mycosis fungoides/S6zary syndrome), referred to the Mycosis Fungoides Clinic at Stanford University, were included, as well as 12 patients with non-mycosis fungoides erythroderma seen in the general dermatology clinics. The diagnosis of mycosis fungoides was based either on specific skin biopsy findings, as previously described,' or on diagnostic morphologic findings in the lymph nodes or peripheral blood. As the absolute level of circulating S6zary cells necessary for the diagnosis of S~zary syndrome has been debated, we have chosen the criterion of >_10% circulating atypical cells as consistent with S6zary syndrome? The 12 patients with benign forms of erythroderma included 9 patients with psoriasis, 2 with pityriasis rubra pilaris, and 1 with a drug eruption secondary to cimetidine. Shave biopsy specimens were obtained from untreated erythrodermic skin; one haft was processed for routine histopathologic evaluation and the other half, for immunophenotyping. The latter specimen was placed in cold saline, embedded in OCT compound, frozen, sectioned, acetone-fixed, and stained with a three-stage monoclonal antibody biotin-avidin immunoperoxidase technique, as described previously.6 The panel of monoclonal antibodies employed are listed in Table I along with their CD designations and predominant specificities. Negative controls included staining with irrelevant monoclonal antibodies of similar isotype and staining with one or more stages deleted. Positive controls included staining of normal lymphoid tissues and various T cell lymphomas, some of which exhibited deficiencies of Leu-8 and Leu-9 antigens.
Erythroderma immunohistology 1091
Table III. Leu-2 antigen expression in epidermis
relative to dermis in erythroderma*
Percentage of Leu-2+ ceils
Erythroderma subgroup Benign Mycosis fungoides/ S6zary syndrome
Epidermis { Epidermis >Dermis
5
0
8
*Th~ percentage of epidermal T cells that were Leu-2+ exceeded the percentage of dermal T cells that were Leu-2+ in 7 of 12 benign cases and in no mycosis fungoides/Sgzary syndrome cases. This finding was therefore 100% specific and 58% sensitive for benign erythroderma.
As described previously, percentages of Leu-8+ and Leu-9+ mononuclear cells in semiserial sections were estimated visually by two observers and grouped into three semiquantitative categories: ~10%, 20% to 40%, and >__50%.3 Percentages of Leu-8+ and Leu-9+ cells were estimated relative to the number of Leu-4+ cells and were recorded separately for the epidermis, dermis, and both combined. RESULTS
In all cases, immunohistologic examination revealed a mononuclear cell infiltrate containing predominantly T cells, the majority of which were Leu-1 +, Leu-4+, Leu-5+, as is typical of mature T cells. T h e T helper (Leu-3+) and T cytotoxic/ suppressor (Leu-2+) antigen expression in the epidermis and dermis are shown in Table II. Although a majority of Leu-3+ T cells were present in most patients, 2 had a majority of the Leu-2+ subset, and 12 others had a significant minority (20%-40%) of these cells in the dermis and in the biopsy specimen as a whole. Multiple logistic regression applied separately to the data in Table IF failed to show that Leu-2 levels in either the dermis or the epidermis differentiate benign erythroderma from mycosis fungoides/ S6zary syndrome erythroderma. However, the sample size is small and the trend in the epidermis toward low Leu-3+/Leu-2+ ratios in benign erythroderma was noteworthy: in 6 of 7 patients with :>50% Leu-2+ cells in the epidermis, the erythroderma was benign.* *Statistical analysis of data by Dr. Jerry Halpem, Department of Biostatistics, Stanford University School of Medicine.
1092
Journal of the American Academy of Dermatology
Abel et al.
Table IV. Leu-8 and Leu-9 antigen expression in erythroderma* Epidermis (Leu-9+) "<10%
I
2~176
Dermis (Leu-9+)
I
--<10%
>_2o%
Leu-8+ <10%
o
o
o
o
o
o
o
o
o
o
9
0
9 1 4 Q 9
o
>--50% o
o
o
9
9
9
9o
o
o
o
9
.o o9
20~
20%-40%
o
o o o
>__.50%
9
9
9
o , Benign erythroderma; 9 mycosis fungoides/S6zary syndrome erythroderma. *Deficiency of Leu-8 antigen was common in both benign and mycosis fungoides/S6zary syndrome types of erythroderma. In contrast, Leu-9 antigen deficiency was rare. Leu-8 antigen deficiency was more pronounced within the epidermis than within the dermis. Valu~ given for the dermis are the same as those for epidermis and dermis combined.
Table V. TAC antigen expression in erythroderma* --<10%
Epidermis Dermis
o
20%-40%
o
o
o
o o
o
9 9
o
o
o o o
o o o
0 0
0
0
o
l
>__50% o9
0
o , Benign erythroderma; 9 mycosis fungoides/S6zary syndrome erythroderma. *The percentage of TAC+ ceils is highly variable in both patient groups but was generally 50%. No major differences were noted between epidermis and dermis. Values given for the dermis were the same as these for the epidermis and dermis combined. Benign = 11; mycosis fungoides/S6zary syndrome = 3.
Values given for the dermis were virtually always the same as those for the epidermis and dermis combined; however, the percentage of epidermal T cells that were Leu-2+ often exceeded the percentage of dermal T cells that were Leu-2+ in benign erythroderma (7/12 cases), but not in any case of mycosis fungoides/S6zary syndrome erythroderma (p=0.014, Fisher's Exact Test) (Table III). This was one of the two statistically significant immunologic differences noted between our patient groups. As shown in Table IV, deficiency of Leu-8 antigen was common in both benigh and mycosis fungoidcs/S6zary syndrome types of erythroderma, 75% and 62%, respectively. Leu-8 deficiency was more pronounced within the epidermis than within the dermis. Values given for the dermis are the same as those for the epidermis and dermis combined. In contrast, Leu-9 antigen deft-
ciency was present in only one patient in each group. TAC antigen (IL-2 receptor) expression in erythroderma in the epidermis and dermis is shown in Table V. The percentage of TAC+ cells was highly variable in both patient groups but was generally <50%. No major differences were noted between epidermis and dermis. In Table VI, Leu6 antigen expression in erythroderma is shown to be highly variable within the epidermis. Within the dermis, Leu-6+ cells were almost always increased. A significant inverse correlation was noted between Leu-6+ cells and Leu-2+ cells within the epidermis in benign disease. The rank correlation for all cases was -0.61 (p < 0.01 ). The biologic significance of this finding is unclear but points out another difference between benign and malignant forms of erythroderma (Table VII). Leu-MJ+ macrophages were studied in 5 patients (4 benign, 1 mycosis fungoides/S6zary syndrome). They were located mainly within the dermis and comprised a variable minority of the total infiltrate, except in one benign erythroderma case in which they approached 50%. TO15+ B cells were studied in 4 cases of benign erythroderma. They were essentially absent in each case. Leu-7+ natural killer/killer cells were studied in 2 cases of benign erythroderma. They were essentially absent. DISCUSSION In order to understand better the etiopathogenesis of erythroderma as a cutaneous reaction pattern, immunophenotyping studies were performed
Volume 19 Number 6 December 1988
Erythroderma immunohistotogy 1093
Table VI. Leu-6 antigen expression in erythroderma* Normal
Decreased [
Tncreased
Epidermis o o o o 0
0
0
Table VII. Correlation of intraepidermal Leu-2+ and Leu-6+ cells*
0
0
Epidermal Leu-2+ cells --<10% I 20%-40% >__50%
0
n
Epidermal Leu 6+ cells Increased o o
0 9
9
Dermis 0
0
9
9
9
9
9
0
0
0
0
0
0
9
9
9
9
9
9
9
9
0
0
o
0
o, Benign erythroderma; o, mycosis fungoidm/S6zary syndrome erythroderma. *The percentage of Leu-6+ cells was highly variable within the epidermis. Within the dermis, Leu-6+ cells were almost always increased. Benign = i 1; mycosis fungoides/S6zary syndrome = 8.
in patients with benign and malignant forms of the disease. Prior investigations have shown that immunologic abnormalities, such as combined Leu-8 and Leu-9 antigen deficiency, are common in mycosis lung9 but uncommon in benign inflammatory disease. 3,7 The previous finding that Leu-8 and Leu-9 deficiency failed to distinguish benign and idiopathic forms of poikiloderma supports the concept that atrophic large-plaque parapsoriasis is a precursor of mycosis fungoides. 4 In this study, Leu-8 and Leu-9 deficiency did not differentiate between the benign and malignant forms of erythroderma. We found that the phenotype of lesional T cells in 20 patients with erythroderma (8 mycosis fungoides/S6zary syndrome, 12 benign) was strikingly similar. In all cases, the majority of T cells was Leu-1+, Leu-4+, Leu-5+, as is typical of mature T cells. All T cells appeared to be either Leu-2+ or Leu-3+, that is, in no case of erythroderma was there a detectable lack of a major subset antigen as can occur in some cases of T cell lymphoma) Approximately two thirds of patients with both benign and malignant forms of erythroderma showed Leu-8 deficiency that, as an isolated deficiency, is not uncommon either in neoplasia or in inflammation. In contrast, Leu-9 deficiency was rare. It was present in only 2 patients (1 benign, 1 mycosis fungoides/S6zary syndrome). This differs from prior experience with mycosis fungoides skin lesions in general and leukemic S6zary cells, wherein Leu-9 deficiency is common? These discrepan-
Normal
9
9 9 *
Decreased
9
o o
9 o o o o o o
o, Benign erythroderma; o, mycosis fungoides/S~zary syfidrome erythroderma. *A statistically significant (p < 0.01) inverse correlation was seen in eases of benign erythroderma.
cies suggest that erythrodermic mycosis fungoides/ S6zary syndrome skin lesions may contain a greater proportion of cells involved with host response than other types of mycosis fungoides. Speculation concerning the roles of Leu-8 and Leu-9 T cells in cutaneous infiltrates has been discussed previously.3, 6
The dermis generaUy showed a predominance of Leu-3+ cells; however, the epidermis showed a greater number of Leu-2+ cells in seven patients. Furthermore, epidermal T cells were Leu-8deficient in every case. This nonrandom distribution of Leu-2+ T cells and Leu-8+ T cells was apparent only when the epidermis and the dermis were evaluated separately. In regard to the potential biologic and diagnostic significance of intraepidermal Leu-2+ T cells, it is interesting to note that while 7 of 12 (58%) benign erythroderma patients exhibited a greater percentage of Leu-2+ cells in the epidermis relative to the dermis, none of the mycosis fungoides/S6zary syndrome erythroderma patients showed similar findings (Table III). This difference might correlate with either an increase in host epidermal T cytotoxic/suppressor response or with a lack of expansion of clonal helper T cells in the epidermis of benign erythroderma. It will be important to expand the study of benign and mycosis fungoides/ S6zary syndrome erythroderma to determine the extent to which Leu-2+ epidermal T cell values will be helpful in distinguishing these conditions. Of interest in a prior series was the greater
1094
Journal of the American Academy of Dermatology
Abel et al.
number of cytotoxic/suppressor T cells than helper T cells in the epidermis of a patient with S6zary syndrome who had an unusually stable and protracted course. 1~Other investigators 8 also found a predominance of cytotoxic/suppressor T lymphocytes in the epidermis of 2 patients with an indolent form of mycosis fungoides. One of our 8 mycosis fungoides/S6zary syndrome patients, a S6zary syndrome patient, also 'showed ___50% Leu-2+ cells within the epidermis, but this patient has been lost to follow-up. Whether this relatively uncommon immunophenotypic feature of mycosis fungoides/ S6zary syndrome erythroderma will indicate a favorable prognosis remains to be determined. In previous studies of the blood in patients with the leukemic form of mycosis fungoides, 5 of 5 cases were deficient in 3A1, the reported equivalent of anti-Leu-9, used by Haynes et al. 9 These patients were compared to those with other forms of T cell leukemia and normal controls where Leu-9 is expressed. It would be of interest to perform comparative studies of Leu-9 expression in the blood and skin of the same patients with S6zary syndrome. If there is discordance in the phenotype within the two compartments, this finding might be explained by a paucity of tumor ceils relative to host response in the lesional skin in mycosis fungoides/S6zary syndrome. S u c h findings would correlate with the frequent lack of diagnostic histopathologic features in the skin of patients with S6zary syndrome. Staining for the T A C antigen (IL-2 receptor) was highly variable but usually less than 50%. T A C expression is associated with T ceU activation and production of interleukin-2. Although T A C expression was once thought to be a feature of adult T cell leukemia that allowed its distinction from mycosis fungoides/S6zary syndrome, it is now apparent that T A C + T cells can be present in either diseaseY Likewise, alteration of Leu-6+ cells in erythroderma was nonspecific, that is, highly variable within the epidermis but almost always increased within the dermis. There was, however, a statistically significant inverse relationship between Leu-6+ cells and Leu-2+ cells within the epidermis in benign erythroderma (Table VII). We conclude that in immunophenotypic studies with the use of the current monoclonal antibody
panel, including anti-Leu-8 and anti-Leu-9, it has been generally impossible to distinguish benign from malignant forms of erythroderma. It is known that mycosis fungoides cells in the skin are admixed with nonneoplastic cells, including reactive T cells as well as macrophages, Langerhans/ indeterminate cells, plasma ceils, and eosinophils. I~ The lack of combined Leu-8/Leu-9 antigen deficiency may be due to a heavy inflammatory T cell component, obscuring the antigen deficiencies seen in most mycosis fungoides cases. However, in 7 of 20 cases, a diagnosis of benign erythroderma could be made based on a greater percentage of Leu-2+ cells in the epidermis relative to the dermis. While this criterion was 100% specific for benign erythroderma in our series, it was only 58% senstitive, thereby limiting its general utility. It will be important to determine if this relationship is maintained in additional erythroderm a patients and whether the percentage of epidermal Leu-2+ cells has any prognostic significance in mycosis fungoides/S6zary syndrome erythroderma. Studies of combined markers such as the Leu-2, Leu-6 correlation may also prove to be useful in differential diagnosis. REFERENCES
1. Sentis H J, Willem_ze R, Scheffer E. Histopathologic studies in S6zary syndrome and erythrodermic mycosis fungoides,a comparisonwith benign forms of erythroderma. J AM Ac^D DERMATOL1986;15:1217-26. 2. Buechner SA, Winkelmann RK. S6zary syndrome: a clinicopathologic study of 39 cases. Arch Dermatol 1983;119:979-86. 3. Wood GS, Abel EA, Hoppe RT, Warnke RA. Leu-8 and Leu-9 antigen phenotypes:immunologic criteria for the distinction of mycosis fungoides from cutaneous inflammation. J AM ACADDEKMATOL1986;14:1006-13. 4. Lindae ML, Abel EA, Hoppe RT, Wood GS. Poikilodermatous mycosis fungoides and atrophic large plaque parapsoriasis exhibit similar abnormalities of T cell antigen expression. Arch Dermatol 1988;124:366-72. 5. Abel EA. Clinical and histological changes in PUVAtreated skin. In: CaUen JP, Dahl MV, Golitz LE, eds. Current issues in dermatology. Boston: G. K. Hall, 1984;2:163-91. 6. Wood GS, Warnke R. Suppressionof endogenousavidinbinding activity in tissuesand its relevanceto biotin-avidin detection systems.J Histochem Cytochem 1981;29:11961204. 7. Abel EA, Wood GS, Hoppe RT, Warnke RA. Expression of Leu-8 antigen (a majorityT-cell marker) is uncommon in mycosis fungoides. J Invest Dermatol 1985;85:199202. 8. Thomas JA, Janossy G, Graham-Brown RAC, et aL The relationship between T lymphocyte subsets and Ia-like
Volume 19
Number 6 December 1988 antigen positive nonlymphoid cells in early stages of cutaneous T cell lymphoma. J Invest Dermatol 1982; 78:169-76. 9. Haynes BF, Metzgart RS, Minna JA, Bunn PA. Phenotype characterization of cutaneous T cell lymphoma, Use of monoclonal antibodies to compare with other malignant T cells. N Engl J Med 1981;304:1319-23. 10. Piepkorn M, Marty J, Kjeldsberg CR. T cell subset
Erythroderma immunoh&tology
heterogeneity in a series of patients with mycosis fungoides and S6zary syndrome. J AM ACAD DERMATOL 1984;11:427-32. 11. Wood GS, Weiss LM, Warnke R_A, Sldar J. The immunopathology of cutaneous lymphomas: immuaophenotypic and immunogenotypic characteristics. Semin Dermatol 1986;5:334-45.
Idiopathic calcinosis of the scrotum: Histopathologic observations of fifty-one nodules Dong Hoon Song, M.D., Kwang Hoon Lee, M.D., and Won Hyoung Kang, M.D.,
Wonju, Korea A 29-year-old man had a 2-year history of multiple, asymptomatic, firm, subcutaneous nodules on the scrotal skin, which sometimes discharged a chalky material. Fifty-one nodules were observed in the histopathologic examination, which revealed, in addition to the typical findings of idiopathic calcinosis of the scrotum, various forms of intact cysts: epidermal (some calcified), pilar (calcified), hybrid (calcified), and indeterminate cysts with diffusely calcified keratinous content and attenuated walls. A mixture of calcified keratinous material and inflammatory infiltrates was detected, with or without remnants of the cyst wall. These findings suggest that idiopathic calcinosis of the scrotum derives from the dystrophic calcification of cysts. (J AM ACAD DtRMATOL 1988;19:1095-1101.)
Idiopathic calcinosis of the scrotum appears in childhood or early adulthood as multiple, asymptomatic nodules of the scrotal skin that gradually increase in size and number and sometimes exude a chalky material. Histopathologic examination shows calcific deposits of various sizes that are present in the dermis and often surrounded by foreign body granulomatous inflammation. 1 The pathogenic mechanism of this condition is not clear. Morley and Best 2 described multiple From the Department of Dermatology, Yonsel University Wonju College of Medicine. Reprint requests to: Dr. Kwang Hoon Lee, Department of Dermatology, Yonsei UniversityCollegeof Medicine, CPO Box 8044, Seoul, Korea 100-680.
calcified epidermal cysts of the scrotum, but their report lacks photomicrographs. Shapiro et al., 1in a review of pathologic findings in patients with scrotal calcinosis, observed no residual cyst or epithelial lining to the calcified mass. They suggested that true cysts might ocaur in the scrotal skin and might even be multiple but distinct from idiopathic calcinosis of the scrotum. Other authors 3-9 generally shared their opinion and regarded this condition as idiopathic. In 1979 King et al. ~~ suggested the implication of dystrophic calcification of dartos muscles but found no evidence dartos muscles in the calcified mass. Recently Swinehart and Golitz n observed epidermal cysts, some of which were calcified, and 1095
Erythroderma is a cutaneous reaction pattern that can be seen in a wide variety of benign and malignant diseases, including mycosis fungoides and its leukemic variant, S+zary syndrome. The erythrodermic form of mycosis fungoides and the S6zary syndrome may be difficult to distinguish from benign forms of erythroderma. These include psoriatic erythroderma, pityriasis rubra pilaris, generalized exfoliative erythroderma secondary to drug eruption, atopic dermatitis, seborrheic dermatitis, contact dermatitis, or paraneoplastic cutaneous reaction to an internal malignancy. The most From the Departments of Dermatology,* Therapeutic Radiology,** and Pathology,*** Stanford University, Stanford, and the Veterans Administration Hospital, Palo Alto, CA.*** Supported by Merit Review funding to Dr. Wood from the Veterans Administration. Reprints not available.
important histologic features in mycosis fungoides are the presence of a bandlike lymphoid infiltrate at the dermoepidermal junction, with atypical cerebriform mononuclear cells and Pautrier intraepidermal microabscesses. Such changes were present in 7 of 11 patients with S6zary syndrome and in 2 of 4 patients with erythrodermic mycosis fungoides in one study.t In another series of 121 skin biopsy specimens from 39 patients with S6zary syndrome, only 47 (39% of specimens) showed a bandlike infiltrate with atypical lymphoid cells, of which 18 (15% of specimens) in 10 patients revealed epidermal involvement suggestive of mycosis fungoides? Features of chronic dermatitis probably related to pruritus with excoriations are present in most patients with S6zary syndrome. Furthermore, some cases reported as benign erythroderma may show 1089
1090
Journal of the American Academy of Dermatology
Abel et al.
Table I. Monoclonal antibody panel CD designation
Antibodies
Anti-Leu- 1 Anti-Leu-2 Anti-Leu-3 Anti-Leu-4 Anti-Leu-5 Anti-Leu-6 Anti-Leu-7 Anti-Leu-8
CD5 CD8 CD4 CD3 CD2 CD1
Anti-Leu-9
CD7
Anti-Tac TO 15 Anti-Leu-M3
CD25 CD22 CD14
Predominant specificity
Pan T cell (>95% T cells) Cytotoxic/suppressor T cell (15%-30% T cells) Helper T cell (70%-90% T cells) Pan T cell (>95% T cells) Pan T cell (>95% T cells) Langerhans cell/determinate cell Natural killer/killer cells Majority T cell (60% :t: 10% Leu-2+ T cells, 75% + 10% Leu-3+ T cells) Majority B cells, some granulocytes, monocytes, natural killer cells Majority T cell (100% Leu-2+ T cells, 90% Leu-3+ T cells) Interleukin 2 receptor B cell Monocyte/macrophage and myeloid cells
Table II. Leu-2 and Leu-3 antigen expression in erythroderma* Epidermis ('Leu-3+)
Dermis (Leu-3+)
Leu-2 +
__<10% 20%0-40%
-->50%
0
o
o
o
o
o
9
9
9
o
o
o
9
9
9
o
o
o
o
o
o
o
o
9
9
9
@
0
0
0
0
o
o
0
o , Benign erythroderma; 9 mycosis fungoides/S~zary syndrome erythroderma. *The Leu-3+ (helper) T cell subset generally predominated in the dermis; however, in the epidermis, Leu-2+ (eytotoxie/suppressor) T cells often equaled or exceeded Lea-3+ T cells, especially among the benign eases. Valu~ given for the dermis were virtually always the same as those for the epidermks and dermis combined. In no case did the sum of Leu-2+ and Leu-3+ cells exceed 100%.
histologic features resembling mycosis fungoides. In one report, 1 3 of 15 patients with chronic dermatitis showed dense infiltrates in the papillary dermis containing considerable numbers of large cerebriform mononuclear cells. Although there were no Pautrier microabscesses nor was there any significant epidermal infiltration, the histologic picture in these patients was indistinguishable from that in the 4 patients with erythrodermic mycosis fungoides. Additional investigations of blood and lymph nodes are usually necessary to make a definitive diagnosis of Srzary syndrome.
Immunophenotyping with conventional cell surface markers has not allowed distinction of mycosis fungoides from reactive cutaneous inflammation, as many reactive processes contain predominantly cells expressing a mature helper T cell phenotype: Leu-l+, L e u - 2 - , Leu-3+, Leu-4+, Leu-5+. However, our prior studies with the use of monoclonal antibodies to the majority T cell markers Leu-8 and Leu-9 have demonstrated major differences in cellular antigen expression between certain benign and malignant cutaneous T cell infiltrates. In a study of 41 skin biopsy specimens from 27 patients
Volume 19 Number 6 December 1988
with mycosis fungoides and 34 skin biopsy specimens from 33 controls, the Leu-8+/Leu-9+ phenotype was much more common in controls with benign disease than in patients with mycosis fungoides. Leu-8 deficiency was more sensitive but less specific than Leu-9 deficiency. 3 Similar results were found in early-stage malignant disease, such as poikilodermatous patch stage mycosis fungoides, as well as in idiopathic forms of poikiloderma, such as large-plaque parapsoriasis, suggesting the premalignant nature of this condition? This study was undertaken to determine if immunophenotyping might also help to distinguish S6zary syndrome or erythrodermic mycosis fungoides from benign forms of erythroderma. PATIENTS AND METHODS
Eight patients with cutaneous T cell lymphoma, two with erythrodermic mycosis fungoides, and six with Sdzary syndrome (mycosis fungoides/S6zary syndrome), referred to the Mycosis Fungoides Clinic at Stanford University, were included, as well as 12 patients with non-mycosis fungoides erythroderma seen in the general dermatology clinics. The diagnosis of mycosis fungoides was based either on specific skin biopsy findings, as previously described,' or on diagnostic morphologic findings in the lymph nodes or peripheral blood. As the absolute level of circulating S6zary cells necessary for the diagnosis of S~zary syndrome has been debated, we have chosen the criterion of >_10% circulating atypical cells as consistent with S6zary syndrome? The 12 patients with benign forms of erythroderma included 9 patients with psoriasis, 2 with pityriasis rubra pilaris, and 1 with a drug eruption secondary to cimetidine. Shave biopsy specimens were obtained from untreated erythrodermic skin; one haft was processed for routine histopathologic evaluation and the other half, for immunophenotyping. The latter specimen was placed in cold saline, embedded in OCT compound, frozen, sectioned, acetone-fixed, and stained with a three-stage monoclonal antibody biotin-avidin immunoperoxidase technique, as described previously.6 The panel of monoclonal antibodies employed are listed in Table I along with their CD designations and predominant specificities. Negative controls included staining with irrelevant monoclonal antibodies of similar isotype and staining with one or more stages deleted. Positive controls included staining of normal lymphoid tissues and various T cell lymphomas, some of which exhibited deficiencies of Leu-8 and Leu-9 antigens.
Erythroderma immunohistology 1091
Table III. Leu-2 antigen expression in epidermis
relative to dermis in erythroderma*
Percentage of Leu-2+ ceils
Erythroderma subgroup Benign Mycosis fungoides/ S6zary syndrome
Epidermis { Epidermis >Dermis
5
0
8
*Th~ percentage of epidermal T cells that were Leu-2+ exceeded the percentage of dermal T cells that were Leu-2+ in 7 of 12 benign cases and in no mycosis fungoides/Sgzary syndrome cases. This finding was therefore 100% specific and 58% sensitive for benign erythroderma.
As described previously, percentages of Leu-8+ and Leu-9+ mononuclear cells in semiserial sections were estimated visually by two observers and grouped into three semiquantitative categories: ~10%, 20% to 40%, and >__50%.3 Percentages of Leu-8+ and Leu-9+ cells were estimated relative to the number of Leu-4+ cells and were recorded separately for the epidermis, dermis, and both combined. RESULTS
In all cases, immunohistologic examination revealed a mononuclear cell infiltrate containing predominantly T cells, the majority of which were Leu-1 +, Leu-4+, Leu-5+, as is typical of mature T cells. T h e T helper (Leu-3+) and T cytotoxic/ suppressor (Leu-2+) antigen expression in the epidermis and dermis are shown in Table II. Although a majority of Leu-3+ T cells were present in most patients, 2 had a majority of the Leu-2+ subset, and 12 others had a significant minority (20%-40%) of these cells in the dermis and in the biopsy specimen as a whole. Multiple logistic regression applied separately to the data in Table IF failed to show that Leu-2 levels in either the dermis or the epidermis differentiate benign erythroderma from mycosis fungoides/ S6zary syndrome erythroderma. However, the sample size is small and the trend in the epidermis toward low Leu-3+/Leu-2+ ratios in benign erythroderma was noteworthy: in 6 of 7 patients with :>50% Leu-2+ cells in the epidermis, the erythroderma was benign.* *Statistical analysis of data by Dr. Jerry Halpem, Department of Biostatistics, Stanford University School of Medicine.
1092
Journal of the American Academy of Dermatology
Abel et al.
Table IV. Leu-8 and Leu-9 antigen expression in erythroderma* Epidermis (Leu-9+) "<10%
I
2~176
Dermis (Leu-9+)
I
--<10%
>_2o%
Leu-8+ <10%
o
o
o
o
o
o
o
o
o
o
9
0
9 1 4 Q 9
o
>--50% o
o
o
9
9
9
9o
o
o
o
9
.o o9
20~
20%-40%
o
o o o
>__.50%
9
9
9
o , Benign erythroderma; 9 mycosis fungoides/S6zary syndrome erythroderma. *Deficiency of Leu-8 antigen was common in both benign and mycosis fungoides/S6zary syndrome types of erythroderma. In contrast, Leu-9 antigen deficiency was rare. Leu-8 antigen deficiency was more pronounced within the epidermis than within the dermis. Valu~ given for the dermis are the same as those for epidermis and dermis combined.
Table V. TAC antigen expression in erythroderma* --<10%
Epidermis Dermis
o
20%-40%
o
o
o
o o
o
9 9
o
o
o o o
o o o
0 0
0
0
o
l
>__50% o9
0
o , Benign erythroderma; 9 mycosis fungoides/S6zary syndrome erythroderma. *The percentage of TAC+ ceils is highly variable in both patient groups but was generally 50%. No major differences were noted between epidermis and dermis. Values given for the dermis were the same as these for the epidermis and dermis combined. Benign = 11; mycosis fungoides/S6zary syndrome = 3.
Values given for the dermis were virtually always the same as those for the epidermis and dermis combined; however, the percentage of epidermal T cells that were Leu-2+ often exceeded the percentage of dermal T cells that were Leu-2+ in benign erythroderma (7/12 cases), but not in any case of mycosis fungoides/S6zary syndrome erythroderma (p=0.014, Fisher's Exact Test) (Table III). This was one of the two statistically significant immunologic differences noted between our patient groups. As shown in Table IV, deficiency of Leu-8 antigen was common in both benigh and mycosis fungoidcs/S6zary syndrome types of erythroderma, 75% and 62%, respectively. Leu-8 deficiency was more pronounced within the epidermis than within the dermis. Values given for the dermis are the same as those for the epidermis and dermis combined. In contrast, Leu-9 antigen deft-
ciency was present in only one patient in each group. TAC antigen (IL-2 receptor) expression in erythroderma in the epidermis and dermis is shown in Table V. The percentage of TAC+ cells was highly variable in both patient groups but was generally <50%. No major differences were noted between epidermis and dermis. In Table VI, Leu6 antigen expression in erythroderma is shown to be highly variable within the epidermis. Within the dermis, Leu-6+ cells were almost always increased. A significant inverse correlation was noted between Leu-6+ cells and Leu-2+ cells within the epidermis in benign disease. The rank correlation for all cases was -0.61 (p < 0.01 ). The biologic significance of this finding is unclear but points out another difference between benign and malignant forms of erythroderma (Table VII). Leu-MJ+ macrophages were studied in 5 patients (4 benign, 1 mycosis fungoides/S6zary syndrome). They were located mainly within the dermis and comprised a variable minority of the total infiltrate, except in one benign erythroderma case in which they approached 50%. TO15+ B cells were studied in 4 cases of benign erythroderma. They were essentially absent in each case. Leu-7+ natural killer/killer cells were studied in 2 cases of benign erythroderma. They were essentially absent. DISCUSSION In order to understand better the etiopathogenesis of erythroderma as a cutaneous reaction pattern, immunophenotyping studies were performed
Volume 19 Number 6 December 1988
Erythroderma immunohistotogy 1093
Table VI. Leu-6 antigen expression in erythroderma* Normal
Decreased [
Tncreased
Epidermis o o o o 0
0
0
Table VII. Correlation of intraepidermal Leu-2+ and Leu-6+ cells*
0
0
Epidermal Leu-2+ cells --<10% I 20%-40% >__50%
0
n
Epidermal Leu 6+ cells Increased o o
0 9
9
Dermis 0
0
9
9
9
9
9
0
0
0
0
0
0
9
9
9
9
9
9
9
9
0
0
o
0
o, Benign erythroderma; o, mycosis fungoidm/S6zary syndrome erythroderma. *The percentage of Leu-6+ cells was highly variable within the epidermis. Within the dermis, Leu-6+ cells were almost always increased. Benign = i 1; mycosis fungoides/S6zary syndrome = 8.
in patients with benign and malignant forms of the disease. Prior investigations have shown that immunologic abnormalities, such as combined Leu-8 and Leu-9 antigen deficiency, are common in mycosis lung9 but uncommon in benign inflammatory disease. 3,7 The previous finding that Leu-8 and Leu-9 deficiency failed to distinguish benign and idiopathic forms of poikiloderma supports the concept that atrophic large-plaque parapsoriasis is a precursor of mycosis fungoides. 4 In this study, Leu-8 and Leu-9 deficiency did not differentiate between the benign and malignant forms of erythroderma. We found that the phenotype of lesional T cells in 20 patients with erythroderma (8 mycosis fungoides/S6zary syndrome, 12 benign) was strikingly similar. In all cases, the majority of T cells was Leu-1+, Leu-4+, Leu-5+, as is typical of mature T cells. All T cells appeared to be either Leu-2+ or Leu-3+, that is, in no case of erythroderma was there a detectable lack of a major subset antigen as can occur in some cases of T cell lymphoma) Approximately two thirds of patients with both benign and malignant forms of erythroderma showed Leu-8 deficiency that, as an isolated deficiency, is not uncommon either in neoplasia or in inflammation. In contrast, Leu-9 deficiency was rare. It was present in only 2 patients (1 benign, 1 mycosis fungoides/S6zary syndrome). This differs from prior experience with mycosis fungoides skin lesions in general and leukemic S6zary cells, wherein Leu-9 deficiency is common? These discrepan-
Normal
9
9 9 *
Decreased
9
o o
9 o o o o o o
o, Benign erythroderma; o, mycosis fungoides/S~zary syfidrome erythroderma. *A statistically significant (p < 0.01) inverse correlation was seen in eases of benign erythroderma.
cies suggest that erythrodermic mycosis fungoides/ S6zary syndrome skin lesions may contain a greater proportion of cells involved with host response than other types of mycosis fungoides. Speculation concerning the roles of Leu-8 and Leu-9 T cells in cutaneous infiltrates has been discussed previously.3, 6
The dermis generaUy showed a predominance of Leu-3+ cells; however, the epidermis showed a greater number of Leu-2+ cells in seven patients. Furthermore, epidermal T cells were Leu-8deficient in every case. This nonrandom distribution of Leu-2+ T cells and Leu-8+ T cells was apparent only when the epidermis and the dermis were evaluated separately. In regard to the potential biologic and diagnostic significance of intraepidermal Leu-2+ T cells, it is interesting to note that while 7 of 12 (58%) benign erythroderma patients exhibited a greater percentage of Leu-2+ cells in the epidermis relative to the dermis, none of the mycosis fungoides/S6zary syndrome erythroderma patients showed similar findings (Table III). This difference might correlate with either an increase in host epidermal T cytotoxic/suppressor response or with a lack of expansion of clonal helper T cells in the epidermis of benign erythroderma. It will be important to expand the study of benign and mycosis fungoides/ S6zary syndrome erythroderma to determine the extent to which Leu-2+ epidermal T cell values will be helpful in distinguishing these conditions. Of interest in a prior series was the greater
1094
Journal of the American Academy of Dermatology
Abel et al.
number of cytotoxic/suppressor T cells than helper T cells in the epidermis of a patient with S6zary syndrome who had an unusually stable and protracted course. 1~Other investigators 8 also found a predominance of cytotoxic/suppressor T lymphocytes in the epidermis of 2 patients with an indolent form of mycosis fungoides. One of our 8 mycosis fungoides/S6zary syndrome patients, a S6zary syndrome patient, also 'showed ___50% Leu-2+ cells within the epidermis, but this patient has been lost to follow-up. Whether this relatively uncommon immunophenotypic feature of mycosis fungoides/ S6zary syndrome erythroderma will indicate a favorable prognosis remains to be determined. In previous studies of the blood in patients with the leukemic form of mycosis fungoides, 5 of 5 cases were deficient in 3A1, the reported equivalent of anti-Leu-9, used by Haynes et al. 9 These patients were compared to those with other forms of T cell leukemia and normal controls where Leu-9 is expressed. It would be of interest to perform comparative studies of Leu-9 expression in the blood and skin of the same patients with S6zary syndrome. If there is discordance in the phenotype within the two compartments, this finding might be explained by a paucity of tumor ceils relative to host response in the lesional skin in mycosis fungoides/S6zary syndrome. S u c h findings would correlate with the frequent lack of diagnostic histopathologic features in the skin of patients with S6zary syndrome. Staining for the T A C antigen (IL-2 receptor) was highly variable but usually less than 50%. T A C expression is associated with T ceU activation and production of interleukin-2. Although T A C expression was once thought to be a feature of adult T cell leukemia that allowed its distinction from mycosis fungoides/S6zary syndrome, it is now apparent that T A C + T cells can be present in either diseaseY Likewise, alteration of Leu-6+ cells in erythroderma was nonspecific, that is, highly variable within the epidermis but almost always increased within the dermis. There was, however, a statistically significant inverse relationship between Leu-6+ cells and Leu-2+ cells within the epidermis in benign erythroderma (Table VII). We conclude that in immunophenotypic studies with the use of the current monoclonal antibody
panel, including anti-Leu-8 and anti-Leu-9, it has been generally impossible to distinguish benign from malignant forms of erythroderma. It is known that mycosis fungoides cells in the skin are admixed with nonneoplastic cells, including reactive T cells as well as macrophages, Langerhans/ indeterminate cells, plasma ceils, and eosinophils. I~ The lack of combined Leu-8/Leu-9 antigen deficiency may be due to a heavy inflammatory T cell component, obscuring the antigen deficiencies seen in most mycosis fungoides cases. However, in 7 of 20 cases, a diagnosis of benign erythroderma could be made based on a greater percentage of Leu-2+ cells in the epidermis relative to the dermis. While this criterion was 100% specific for benign erythroderma in our series, it was only 58% senstitive, thereby limiting its general utility. It will be important to determine if this relationship is maintained in additional erythroderm a patients and whether the percentage of epidermal Leu-2+ cells has any prognostic significance in mycosis fungoides/S6zary syndrome erythroderma. Studies of combined markers such as the Leu-2, Leu-6 correlation may also prove to be useful in differential diagnosis. REFERENCES
1. Sentis H J, Willem_ze R, Scheffer E. Histopathologic studies in S6zary syndrome and erythrodermic mycosis fungoides,a comparisonwith benign forms of erythroderma. J AM Ac^D DERMATOL1986;15:1217-26. 2. Buechner SA, Winkelmann RK. S6zary syndrome: a clinicopathologic study of 39 cases. Arch Dermatol 1983;119:979-86. 3. Wood GS, Abel EA, Hoppe RT, Warnke RA. Leu-8 and Leu-9 antigen phenotypes:immunologic criteria for the distinction of mycosis fungoides from cutaneous inflammation. J AM ACADDEKMATOL1986;14:1006-13. 4. Lindae ML, Abel EA, Hoppe RT, Wood GS. Poikilodermatous mycosis fungoides and atrophic large plaque parapsoriasis exhibit similar abnormalities of T cell antigen expression. Arch Dermatol 1988;124:366-72. 5. Abel EA. Clinical and histological changes in PUVAtreated skin. In: CaUen JP, Dahl MV, Golitz LE, eds. Current issues in dermatology. Boston: G. K. Hall, 1984;2:163-91. 6. Wood GS, Warnke R. Suppressionof endogenousavidinbinding activity in tissuesand its relevanceto biotin-avidin detection systems.J Histochem Cytochem 1981;29:11961204. 7. Abel EA, Wood GS, Hoppe RT, Warnke RA. Expression of Leu-8 antigen (a majorityT-cell marker) is uncommon in mycosis fungoides. J Invest Dermatol 1985;85:199202. 8. Thomas JA, Janossy G, Graham-Brown RAC, et aL The relationship between T lymphocyte subsets and Ia-like
Volume 19
Number 6 December 1988 antigen positive nonlymphoid cells in early stages of cutaneous T cell lymphoma. J Invest Dermatol 1982; 78:169-76. 9. Haynes BF, Metzgart RS, Minna JA, Bunn PA. Phenotype characterization of cutaneous T cell lymphoma, Use of monoclonal antibodies to compare with other malignant T cells. N Engl J Med 1981;304:1319-23. 10. Piepkorn M, Marty J, Kjeldsberg CR. T cell subset
Erythroderma immunoh&tology
heterogeneity in a series of patients with mycosis fungoides and S6zary syndrome. J AM ACAD DERMATOL 1984;11:427-32. 11. Wood GS, Weiss LM, Warnke R_A, Sldar J. The immunopathology of cutaneous lymphomas: immuaophenotypic and immunogenotypic characteristics. Semin Dermatol 1986;5:334-45.
Idiopathic calcinosis of the scrotum: Histopathologic observations of fifty-one nodules Dong Hoon Song, M.D., Kwang Hoon Lee, M.D., and Won Hyoung Kang, M.D.,
Wonju, Korea A 29-year-old man had a 2-year history of multiple, asymptomatic, firm, subcutaneous nodules on the scrotal skin, which sometimes discharged a chalky material. Fifty-one nodules were observed in the histopathologic examination, which revealed, in addition to the typical findings of idiopathic calcinosis of the scrotum, various forms of intact cysts: epidermal (some calcified), pilar (calcified), hybrid (calcified), and indeterminate cysts with diffusely calcified keratinous content and attenuated walls. A mixture of calcified keratinous material and inflammatory infiltrates was detected, with or without remnants of the cyst wall. These findings suggest that idiopathic calcinosis of the scrotum derives from the dystrophic calcification of cysts. (J AM ACAD DtRMATOL 1988;19:1095-1101.)
Idiopathic calcinosis of the scrotum appears in childhood or early adulthood as multiple, asymptomatic nodules of the scrotal skin that gradually increase in size and number and sometimes exude a chalky material. Histopathologic examination shows calcific deposits of various sizes that are present in the dermis and often surrounded by foreign body granulomatous inflammation. 1 The pathogenic mechanism of this condition is not clear. Morley and Best 2 described multiple From the Department of Dermatology, Yonsel University Wonju College of Medicine. Reprint requests to: Dr. Kwang Hoon Lee, Department of Dermatology, Yonsei UniversityCollegeof Medicine, CPO Box 8044, Seoul, Korea 100-680.
calcified epidermal cysts of the scrotum, but their report lacks photomicrographs. Shapiro et al., 1in a review of pathologic findings in patients with scrotal calcinosis, observed no residual cyst or epithelial lining to the calcified mass. They suggested that true cysts might ocaur in the scrotal skin and might even be multiple but distinct from idiopathic calcinosis of the scrotum. Other authors 3-9 generally shared their opinion and regarded this condition as idiopathic. In 1979 King et al. ~~ suggested the implication of dystrophic calcification of dartos muscles but found no evidence dartos muscles in the calcified mass. Recently Swinehart and Golitz n observed epidermal cysts, some of which were calcified, and 1095