Junctional CD8+ Cutaneous Lymphomas With Nonaggressive Clinical Behavior: Title and subTitle BreakA CD8⁺ Variant of Mycosis Fungoides? FREE

MOST CUTANEOUS T-cell lymphomas (CTCLs) have the phenotype of T-helper memory lymphocytes (CD3⁺, CD4⁺, and CD45R0⁺).^{1 - 2} Only a minority of cases present with other phenotypes, such as CD3⁺, CD4⁻, CD8⁻, or CD3⁺CD8⁺,^{3 - 4} or with the spectrum of CD56⁺ lymphomas.^{5 - 7} The primary cutaneous lymphomas with the phenotype of cytotoxic/suppressor (CD8⁺) T cells have been of significant interest in the recent literature.^{8 - 12} The detailed analysis of histological and clinical features in several cases has led to the conclusion that CD8⁺ T-cell lymphomas represent a distinct type of CTCL with an aggressive clinical behavior.^{4 ,12}

Within the last 3 years, we have seen 3 cases that have fulfilled the criteria of an CD8⁺ epidermotropic T-cell lymphoma with a remarkable affinity of the CD8⁺ cells to the dermoepidermal junction zone. The patients in all 3 cases presented with hyperpigmented lesional skin. In 2 cases, the period between the appearance of the first skin symptoms and the diagnosis was extremely long (6 and 9 years, respectively). None of the patients developed progressive disease during the observation period.

Our study included 3 patients who were seen in the outpatient lymphoma clinic of the Department of Dermatology, University of Zurich, Zurich, Switzerland. The patients underwent a clinical examination, routine blood cell counts and chemistry studies, chest x-radiographic investigation, and ultrasound examination of the abdomen and lymph nodes. Also, serological immunoparameters were investigated, and a fluorescent activated cell sorter was used to analyze circulating peripheral blood lymphocytes.

Each patient underwent a biopsy at first presentation. Skin sections were partly fixed in paraffin and partly snap frozen for molecular biological analysis. Paraffin-embedded material was extensively studied using a panel of monoclonal antibodies (Table 1). Immunoreactivity was visualized using a standard alkaline phosphatase, anti–alkaline phosphatase technique (Dako Diagnostics AG, Zug, Switzerland).¹³

Snap-frozen material was used to extract DNA. Using previously described primers, the T-cell receptor γ chain locus was amplified, and denaturing gradient electrophoresis was used to detect clonal populations.^{14 - 17}

In situ hybridization for Epstein-Barr virus RNA was performed using a commercial kit and an Epstein-Barr virus–encoded small nuclear RNA probe (Dako Diagnostics AG).

The clinical presentations of the patients are summarized in Table 1 and shown in Figure 1). All 3 patients were in good condition and free of general symptoms. The lesions, which were confined to the trunk and the proximal aspect of the extremities, consisted of discrete, reddish brown patches, several centimeters in diameter, with minimal desquamation. The hyperpigmented aspect of the patches was prominent in all 3 patients and was not associated with a particular phototype (Table 1). In 2 patients, the lesions were asymtomatic. Patient 2 had an 8-year history of severe pruritus and urticaria factitia, both of which were refractory to therapy. Imaging did not reveal any extracutaneous involvement.

The findings of routine blood and urine laboratory tests were noncontributory. Fluorescent activated cell sorter analysis demonstrated a normal number of circulating CD4⁺ and CD8⁺ lymphocytes (CD4/CD8 ratios: patient 1, 5.0; patient 2, 2.3; and patient 3, 1.0). In cases 1 and 3, the levels of soluble interleukin 2 receptor,¹⁸ neopterin, and β₂-microglobulin¹⁹ were within normal ranges. They were not determined in case 2.

The histological findings in the 3 cases were similar. The epidermis showed a slightly compact horny layer or slight parakeratosis. The rete ridges were thin and elongated, resulting in a psoriasiform appearance. Single-cell exocytosis was evident. Beneath the epidermis, there was a moderate, bandlike lymphocytic infiltrate lining up directly opposite of the dermoepidermal junction, as well as prominent single-cell epidermotropism, but there were no Pautrier collections. Deeper in the middle dermis, there were discrete perivascular infiltrates, which were composed of small lymphocytes and histiocytes. Melanin incontinence was quite pronounced. Immunohistochemical analysis revealed that the lymphocytes at the dermoepidermal junction were CD3⁺, CD4⁻, and CD8⁺ (>80%), while the perivascular lymphocytes in the middle dermis were predominantly CD3⁺, CD4⁺, and CD8⁻ (>80%) (Figure 2), and that the CD8⁺ cells also expressed TIA-1 but not CD7. Detailed results of the immunohistochemical investigations are presented in Table 2. In all cases, additional cytotoxic molecules and natural killer cell receptors were present.²⁰

Polymerase chain reaction–denaturing gradient gel electrophoresis detected clonal T-cell populations in snap-frozen skin biopsy specimens in all 3 cases. In 2 cases, the clonal population showed a rearrangement of Vγ 1-8 and Jγ 1/2.

The other patient showed a rearrangement of Vγ 1-8 and JPγ 1/2.¹⁴ Peripheral blood mononuclear cells did not contain clonal populations according to the results of polymerase chain reaction–denaturing gradient gel electrophoresis in cases 1 and 2; however, an identical clonal T-cell rearrangement was detectable in samples of skin and blood in case 3. The results of Epstein-Barr virus RNA in situ hybridization were negative in all 3 cases.

In all 3 cases, the disease had a benign chronic course and was controlled with nontoxic and conservative therapy, including antihistamines, topical corticosteroids, and UV-B or psoralen–UV-A.The lesions in cases 1 and 2 responded to these mild therapies. The recalcitrant pruritus in case 2 posed a significant therapeutic problem. Different modalities, including a combination of psoralen–UV-A and systemic retinoids, with or without interferon alfa, were used to try to relieve the pruritis, without success. The overall duration of disease has been lengthy in patients 1 (9 years) and 2 (10 years), with no sign of progression or extracutaneous involvement. During the 10 months of follow-up in case 3, the clinical presentation of the disease has improved with local steroid treatment (betamethasone, 3-5 times a week).

Cutaneous T-cell lymphomas with a suppressor/cytotoxic (CD8⁺) phenotype are rare lymphoproliferative disorders. They are not categorized separately by recent lymphoma classifications.^{21 - 23}

We report 3 cases of CD8⁺ CTCL. All 3 cases were diagnosed as low-grade lymphomas because of the clinical features, the course of the disease, and the findings of histological and molecular biological analysis. All of them showed clonality on polymerase chain reaction–denaturing gradient gel electrophoresis and a junctional homing preference for small-cell CD8⁺ T lymphocytes. Therefore, all 3 cases fulfilled the criteria for CD8⁺ epidermotropic cytotoxic T-cell lymphomas as defined in an earlier publication,⁴ as well as for CD8⁺ patch-stage mycosis fungoides.²²

All 3 patients had more or less extensively hyperpigmented, flat skin lesions, resulting in the clinical differential diagnosis of superficial morphea, lichen sclerosus et atrophicus, or ashy dermatosis. In 2 cases, it was several years before the correct diagnosis was finally made.

In lymphoproliferative disorders, the clinical features might be helpful in making the correct diagnosis. We encourage additional studies to see whether hyperpigmentation is more frequent in CD8⁺ CTCL, as the bruiselike, contusiforme appearance, which we described previously⁵ and which has been observed by other investigators,^{6 - 7} is a common feature in CD56⁺ natural killer cell lymphomas.

After diagnosis, our patients clearly responded to mild and nonaggressive treatment, including the use of local steroids, psoralen–UV-A, retinoids, and UV-B phototherapy. One case was complicated by very severe pruritus with urticaria factitia that could not be controlled by any antipruriginous agent. This clinical behavior is in contrast to that described by Berti et al,⁴ who reported an aggressive course and median survival time of 32 months in 17 cases. In 2 of our cases, the disease course was protracted over 9 and 10 years, respectively, and in all 3 cases, there was no tendency for extracutaneous involvement, which agrees with the chronic subtype reported by Agnarsson et al in 1990.¹⁰ All 3 of our cases showed a CD7 loss, similar to Agnarsson and colleagues' cases. In view of the clinical descriptions in earlier publications, we are convinced that this variant of CD8⁺ lymphoma might also have been noticed earlier by other groups.

Because of the clinical presentation and the benign behavior in our 3 cases, we think that this type of CD8⁺ T-cell lymphoma has to be differentiated from other cytotoxic lymphoproliferations, as it appears to have a good prognosis, similar to patch-stage mycosis fungoides. It remains to be determined whether the expression of CD8 by the tumor cells has any prognostic implications at all. However, today, there is a tendency toward aggressive treatment of CD8⁺ lymphomas, which is not always justified. Therefore, all physicians involved in the care of patients with CTCL have to be aware that CD8⁺ positivity alone has no clear prognostic value.

Accepted for publication June 18, 2001.

We wish to thank Beatrix Müller for histochemical and immunohistochemical analysis, Markus Bär and Maggi Johnson for the photographic documentation, and Marion Lüthi for the preparation of the manuscript for this article.

Corresponding author and reprints: Reinhard Dummer, MD, Department of Dermatology, University Hospital, Gloriastrasse 31, CH-8091 Zürich, Switzerland (e-mail: dummer@derm.unizh.ch).