CD8-positive mycosis fungoides and primary cutaneous aggressive epidermotropic CD8-positive cytotoxic T-cell lymphoma



Most cutaneous T-cell lymphomas (CTCLs) have the phenotype of T-helper lymphocytes (CD3, CD45RO and CD4 positive). CD8-positive T-cell lymphomas involving the skin are relatively rare and have been of significant interest in the recent literature. Some cases have the clinical presentation of mycosis fungoides (MF) or pagetoid reticulosis, while other have an aggressive clinical course with frequent extracutaneous dissemination.

In this review, we chart some of the earlier observations, together with current insights into this disease entity. In the process, we will see how our understanding of this entity has evolved over the years.

In 1999, Berti et al.1 studied 17 cases of CTCL that were positive for CD8. They were interested in determining whether this phenotype represented a subset of CTCL. Of the cases studied, slightly more than half behaved like other entities such as typical MF, pagetoid reticulosis and lymphomatoid papulosis. However, 8 of the 17 cases behaved in a different fashion. Clinically, they presented a spectrum ranging from patches to tumors. They generally behaved aggressively, with metastases to viscera and the central nervous system rather than lymph nodes, with a median survival of only 32 months. Histologically, these lesions showed epidermotropism of atypical T cells within an acanthotic or hyperplastic epidermis. Varying degrees of spongiosis and even keratinocytic necrosis were frequently detected, unlike the usual scenario in MF. Immunohistochemical studies showed these cells to be CD3 and CD8 positive, often with loss of CD2 and CD5 (CD7 was usually retained). βF1 and TIA-1 were also expressed. Berti et al.1 concluded from this study that a CD8-positive MF is an aggressive variant or subset of MF, with the CD8-positive cells likely derived from cytotoxic T cells (given the positivity for TIA and CD45RA). None of the lesions studied by them were γ/δ-positive. Thirteen of their 17 cases showed rearrangement of the T-cell receptor (TCR) γ-chain. Consistently, primary cutaneous epidermotropic CD8 CTCL was not associated with Epstein-Barr Virus, unlike the aggressive Natural Killer/T-cell lymphomas, nasal type. The findings of Berti et al.1 echo an even earlier report by Agnarsson et al.,2 who in 1990 reported two subsets of CD8-positive MF: those with a chronic course and those that behaved more aggressively. Histologic findings were similar in both types (epidermotropism). Interestingly, two of the five cases that behaved aggressively lost CD2 expression and were CD7 positive, similar to the CD2 loss and CD7 positivity described in the cases by Berti et al.1 (in all Agnarsson et al.2 evaluated nine cases, of which five were aggressive).

The observation that there was an aggressive CD8-positive CTCL was cemented by other reports as well. For example, Quarterman et al.3 reported an aggressive CD8-positive CTCL that involved the tongue. Urrutia et al.4 also documented a case that was resistant to therapy. Fujiwara et al.5 described a case of a 68-year-old woman who progressed in a short time from scaly macules to plaques to ulcerated lesions, leading to infiltrates in the lung and death. This case had a feature that is different from other reported cases, namely angioinvasion, and so it is possible that it is a different entity from the others being described in this review.

In contrast to these, Dummer et al.6 reported a study of three patients in the Archives of Dermatology in February 2002, where the patients fared much better. The lesions in all three patients were hyperpigmented, and the diagnosis was missed for several years. Two of the three patients had a long course of 9 and 10 years, and there was no tendency to behave like the patients described by Berti et al.1 Histologically, they showed epidermotropism, CD8 positivity and loss of CD7. All three patients exhibited TCR γ-chain rearrangement in their lesions. The authors concluded that their patients belonged to the group of patients with a chronic course as documented by Agnarsson et al.2 and can be considered a variant of CD8-positive MF. The patients in this study were successfully treated with UV-B, PUVA or steroids (with one requiring interferon-alpha and retinoid therapy as well); partial remission was achieved in two of the three patients, with complete remission in one.

Subsequently, Lu et al.,7 in 2002, reported a heterogeneous group of CD8-positive CTCL. They studied 18 patients, who were divided into four groups as follows. (a) Three had localized plaques similar to pagetoid reticulosis with histories ranging from 2 to 57 years. Two of these three patients showed TCR-γ gene rearrangement. (b) Seven patients behaved like typical (i.e. CD4-positive) MF. Both γ and β TCR gene rearrangements were detected. (c) Two patients had Sezary syndrome. (d) Six patients had nodular lesions with two of these exhibiting an aggressive disease course. Of the four patients tested, three showed TCR-β gene rearrangement and one was negative for the TCR-γ gene rearrangement. Overall, two thirds of the patients studied had epidermotropism (most prominently seen in those with pagetoid reticulosis-type clinical behavior). The immunological phenotype of the tumor cells included cutaneous lymphocyte antigen in all but three patients and granzyme B, mostly in the nodular lesions (4/5) and in 2/7 typical CD4 MF-like cases. However, granzyme B, which is a marker of cytotoxic T cells, was present in a majority of tumor cells in only 3 of 16 cases.

Into this mix of heterogeneity must be added another subtype of hypopigmented CD8-positive MF. Shabrawi-Caelen et al.,8 in a paper published in 2002, reported 15 patients with hypopigmented MF, of which 9 had a CD8-positive immunophenotype. They found the disease occurring mostly in younger patients (children and adolescents; two of their patients were 43 years at the time of diagnosis). The behavior was similar to that of typical CD4-positive MF and unlike the aggressive course described above. The histologic features were those of an epidermotropic lymphocytic infiltrate. A TCR-γ gene rearrangement was detected in most of the cases tested. Interestingly, the cytotoxic T-cell marker TIA-1 was absent in the majority of tested cases.

From the foregoing, it may be assumed that the prognosis of MF might depend on whether the T cells are cytotoxic or not. Moreover, Lu et al.7 suggested in 2002 that CD8-positive MF-like cases have a trend toward more rapid disease progression than most patients with CD4-positive MF. Massone et al.9 investigated this in a retrospective study published in 2008 evaluating 73 specimens from a total of 68 patients with early-stage disease at diagnosis (either stage Ia or Ib), therefore, including only patients with patch-stage MF. Plaque and tumor stage cases, follicular mucinosis, Sezary syndrome and small plaque parapsoriasis were excluded. This study only selected patients with a long follow up of at least 10 years (or alternatively, dead of disease). The results were classified into phenotypic categories. The majority of cases (51/68) had an α/β phenotype with CD4-positive and TIA-1-negative T cells. TCR-γ gene rearrangement was present in 27 of 50 biopsies tested from these cases. The second group comprising 10 patients had an α/β phenotype and were CD8, TIA-1 and βF1 positive. Four of the 10 cases showed TCR-γ gene rearrangement, the rest being polyclonal. The third group contained five patients. These were α/β negative, negative for CD4 and βF1, with CD8 positivity in six of seven biopsies from these five patients. TIA-1 was positive in all cases. TCR-γ gene rearrangement was present in three of five biopsies tested. The last group had only two patients, α/β positive and negative for CD4, CD8 and TIA-1, with only βF1 immunopositivity. Both cases had TCR-γ gene rearrangements.

While the numbers were small in the third and fourth categories, the finding that emerged was that in the early stage of MF, the following do not have an impact on survival: the presence or absence of a cytotoxic T cells and the presence or absence of a T-cell γ-gene rearrangement. The majority of the cases studied were of the α/β T-helper phenotype.

From the study of Massone et al.9 emerges the idea that in early stages, CD8-positive MF does not behave differently than CD4-positive MF.

This leads us to question: how does one conceptualize the entity of CD8-positive MF? Is it a separate entity? This review highlights reasons why some may not choose to think of CD8-positive MF as a single entity. It is a ‘mixed-bag’ with cases that behave like garden-variety MF and includes those found in younger patients, which are hypopigmented. There are other CD8-positive tumors that show a rather aggressive course. In fact, the aggressive variety is so distinct in its clinical behavior that the WHO-EORTC classification10 separates it in the section of Primary cutaneous peripheral T-cell lymphoma, unspecified, as a provisional entity of ‘Primary cutaneous aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma’. As the behavior of the cases described by Berti et al.1 is so different from the plodding course of other CD8-positive MF, this separation appears to have validity.

There have been recent additions to the literature further elucidating certain features of primary cutaneous aggressive epidermotropic CD8-positive cytotoxic T-cell lymphoma. Many of these are case reports that employ, in addition to immunohistochemical and gene rearrangement studies, flow cytometric (FC) analysis. Yoshizawa et al.11 describe a case of a 68-year-old man with skin lesions initially, followed by pulmonary lesions and eventual death 7 months after diagnosis. FC of skin lesions showed a CD15-positive phenotype. The authors speculated that CD15 may be expressed in aggressive CTCL, drawing on a prior publication12 from 1985. Another case report13 of a 65-year-old male patient documents the presence of CD8-positive tumor cells in the peripheral blood and cerebrospinal fluid by flow cytometry. Consequently, Introcaso et al.13 advocate using FC in these lymphomas.

Flow cytometry was utilized by Benetatos et al.,14 who found differential staining when using two different antibodies (clone SK1 and clone UCH-T4) for CD8 detection in the peripheral blood of a patient with primary cutaneous aggressive epidermotropic CD8-positive T-cell lymphoma. The more commercially available antibody, SK1, did not detect the presence of CD8 expression; however, clone UCH-T4 identified CD8-positive T cells. The patient had a short, poor clinical course resulting in death. Interestingly, the tumor cells were also CD15 positive as in the case of Yoshizawa et al.11 Given that future studies to elucidate CD8-positive lymphomas may utilize flow cytometry, the point emphasized by Benetatos et al.14 is well taken in that it may be necessary to use more than one antibody to ensure a correct diagnosis, particularly when a double-negative phenotype is encountered (CD4−/CD8−).

In summary, studies of CD8-positive T-cell lymphomas, whether early MF or an aggressive epidermotropic type, will continue to refine our concepts of these disease entities.