Monoclonal T-Cell Dyscrasia of Undetermined Significance Associated With Recalcitrant Erythroderma FREE

Erythroderma is an extensive inflammatory skin reaction affecting more than 90% of the body surface. The condition is rare, with an incidence of approximately 1 per 100 000 polulation,¹ and its etiology is variable. Eczemas, psoriasis, and drug reactions are the most frequent causes of erythroderma. In 15% to30% of cases the cause cannot be found (idiopathic erythroderma),^{1 - 2} and up to 40% of these run a chronic course not manageable by local or systemic therapies (chronic idiopathic erythroderma or “red man” syndrome).^{3 - 4}

The possible association between chronic erythroderma and cutaneous T-cell lymphoma (CTCL) has long been recognized. Only a small proportion (5%-10%) of patients with chronic erythroderma develop clinically unequivocal mycosis fungoides or Sézary syndrome.^{1 ,3 - 5} However, some researchers believe that chronic erythroderma, eg, erythrodermic atopic dermatitis, is a preneoplastic condition.^{6 - 9} This concept has been promoted by Winkelmann and coworkers^{6 ,10 - 11} who coined the term pre-Sézary syndrome. This syndrome is defined as chronic erythroderma resembling that seen in Sézary syndrome, a Sézary cell count less than 10⁹/mL, and a high risk of progression into frank leukemia (Sézary syndrome). Other features are palmoplantar keratoderma, alopecia, onychodystrophy, lymphadenopathy, and an increased level of circulating IgE. However, because none of these symptoms has been found consistently in all patients, pre-Sézary syndrome is difficult to separate from pseudo–CTCL erythrodermas such as adult-onset atopic dermatitis or Ofuji papuloerythroderma.⁵ For this reason the concept of pre-Sézary syndrome has not been universally accepted.

Molecular biology and flow cytometry techniques that enable sensitive detection and characterization of clonal expansion of T cells have been used for more than a decade for the diagnosis of leukemias and lymphomas, including cutaneous lymphomas.¹² A T-cell population with monoclonally rearranged T-cell receptor (TCR) genes is readily detectable in peripheral blood in Sézary syndrome. Patients with Sézary syndrome also have an increased proportion of circulating CD4⁺CD7⁻ and CD26⁻ cells,⁵ but it is still unclear whether these represent a malignant population or merely reactive lymphocytes.^{13 - 16} CD4⁺7⁻ are also detectable in late stages of mycosis fungoides with or without erythroderma, and they might be used for monitoring response to therapy.^{17 - 18}

In view of these reports suggesting the diagnostic value of flow cytometry and TCR gene rearrangement for CTCL we have included these studies in the standard diagnostic workup of patients with erythroderma since 1999. We were able to identify a group of patients with chronic, recalcitrant erythroderma accompanied by a monoclonal expansion of CD4⁺7⁻26⁻29⁺ T lymphocytes. These patients fulfilled the criteria for pre-Sézary syndrome but none developed CTCL during the 4-year observation period. We propose that the clinical syndrome described herein represents a T-cell variant within the group of monoclonal hematologic dyscrasias with a yet undetermined risk of progression to malignancy.

A total of 30 patients, 7 women and 23 men aged between 29 and 90 years seen in our department during a 3-year period were included. Of 5 more patients with erythroderma, 3 died or were lost to follow-up before immunophenotyping and 2 did not receive peripheral blood immunophenotyping because the attending physician did not order the study. Patient distribution and diagnoses are shown in Figure 1. Six of the 10 patients with idiopathic erythroderma had chronic, recalcitrant disease defined as symptoms persisting for more than 6 months without any response to local and systemic glucocorticoid treatment. The other 4 patients experienced a single episode of erythroderma (n = 2) or relapsing disease between periods with no or minimal skin symptoms (n = 2). For control purposes (control of CD7 antibody and TCR-γ rearrangement sensitivity) blood from 6 patients with Sézary syndrome were included (the samples were provided by Mark Pittelkow, MD, Mayo Clinic, Rochester, Minn). For all patients, a minimum workup included hematologic and blood chemistry studies, 4-mm skin punch biopsy specimens for histologic studies, a chest radiograph, and abdominal ultrasonographic screening for occult tumors.

Venous blood was collected in EDTA Vacutainer tubes (Becton, Dickinson and Co, Franklin Lakes, NJ) and processed immediately. One milliliter of blood was lysed for 5 minutes at 37°C with 50 mL of lysis buffer (0.83% ammonium chloride, 0.1% kalium bicarbonate, and 0.004% EDTA) and the leukocytes were washed once in phosphate-buffered saline (PBS) solution. The cells were counted manually and resuspended in PBS solution for a concentration of 2.4 × 10⁶/mL. Antibody staining for immunophenotyping by laser scanning cytometry was performed according to the method published by Clatch and colleagues,^19-22 with slight modifications. Briefly, 3 μL of the fluorescein isothiocyanate (FITC)–, phycoerythrin (PE)-, and phycoerythrin cyanogen 5 (PECy5)–labeled antibodies (Table 1) was added to 20 μL of cell suspension and the mixture was gently pipetted onto custom-made chamber slides assembled on standard microscope glass slides. The following antibody-labeling reactions were performed: (1) CD45,CD4,CD8; (2) CD3,CD4, CLA; (3) CD4, HLA-DR, CD134; (4) CD4, CD45RO, CD45RA; (5) CD4, CD25, CD56; (6) CD4, CD7, CD26; (7) CD4, CD7, CD27; (8) CD4, CD7, CD28; (9) CD4, CD7, CD29; (10) CD4, CD7, CD30; (11) CD4, TCRαβ, TCRγδ; and (12) isotype controls. After a 30-minute incubation at 4°C the cells were washed with PBS and scanned in a laser scanning cytometer (CompuCyte, Cambridge, Mass) using the 488-nm line of argon laser as an excitation source. Integrated fluorescence in the green (FITC), orange (RPE) and far red (PC5) channels were collected on a cell-to-cell basis and presented as dot-plot diagrams following off-line fluorescence compensation with the CompuCyte software. After scanning, the chamber slides were disassembled and the cells adhering to the bottom slide were fixed briefly in methanol, air-dried, and stained with Wright-Giemsa. The slides with stained cells were repositioned in the laser scanning cytometer and the cells with the required characteristics were re-found for visual inspection.

Skin biopsy specimens (4-mm punch) and mononuclear blood cells purified on Ficoll were analyzed by means of a polymerase chain reaction with fluorescence fragment product analysis using an automated capillary electrophoresis DNA sequencer (GeneScan; Applied Biosystems, Foster City, Calif), as described elsewhere in detail.^{23 - 25} For lymphocyte depletion/enrichment the peripheral leukocytes were prepared by ammonium chloride lysis, as described above, and washed twice in PBS. CD4⁺ cells were positively isolated or selectively depleted using a Dynabeads M-450 CD4 kit (Dynal Biotech, Oslo, Norway) according to the protocol supplied by the manufacturer. For CD7 depletion, 150 μL of CD7 antibody (Dako Corp, Glostrup, Denmark) was added to 10⁷ peripheral leukocytes (total volume, 1 mL) and the cells were incubated at 4°C for 30 minutes. After washing, the cells were resuspended in 80 μL of PBS solution and 20 μL of goat anti-mouse antibodies linked to Dynabeads M-450 magnetic beads (Dynal Biotech) was added. After a 20-minute incubation at 4°C the cells were washed and resuspended in 500 μL of PBS in an Eppendorf tube and mounted on a magnetic device (Dynal). The supernatant was collected as the negative fraction. The next two 500-μL washes were discarded, and the remaining cells constituted the positive fraction. To check for the purity of the separated fractions the CD4- and CD7-depleted cells were stained for tricolor laser scanning cytometry with the following: (1) CD3-RPE + CD4-PECy5 + CD8-FITC; (2) CD3-RPE + CD4-PECy5 + CD7-FITC; and (3) CD3-RPE + CD4-PECy5 + CD26-FITC. In 2 of 4 cases it was technically feasible to deplete more than 95% of CD4⁺ or CD7⁺ cells.

Peripheral blood immunophenotypic studies were performed using a broad panel of antibodies (Table 1) to determine the CD4/CD8 cell ratio and subpopulations of CD3⁺4⁺ lymphocytes according to the expression of different surface markers. Among the 30 patients included in this study, we found 5 patients (apart from 1 patient with clinically obvious Sézary syndrome) who presented with a more than 2-fold enrichment in the ratio of CD3⁺CD4⁺CD7⁻CD26⁻/CD3⁺CD4⁺ lymphocyte subpopulations (Table 2, Figure 2, and Figure 3). Since the presence of CD4⁺CD7⁻ lymphocytes could signify the development of Sézary syndrome, we performed the TCR clonality studies. In all 5 patients we found a monoclonal population of lymphocytes by detecting a clonally rearranged TCR-γ receptor chain. In the other 4 patients only a blood sample gave this result, but in patient 5 both blood and skin samples showed identical rearrangements. None of these patients, however, showed blood or bone marrow characteristics of leukemic involvement (summary characteristics are provided in Table 2). They did not have palpable lymph nodes except patient 3, in whom findings from excisional biopsy of the left inguinal node showed reactive dermopathic changes. A slight eosinophilia and an increase in total IgE concentration were noted in most patients. Peripheral blood smears showed normal numbers of Sézary cells. Biopsy specimens from all patients showed predominantly lymphocytic, superficial perivascular infiltrates without exocytosis or epidermotropism, with slight to moderate parakeratosis and spongiosis. Their clinical course was protracted, as they were resistant to treatments that included topical and systemic glucocorticoids, systemic retinoids (acitretin), psoralen–UV-A, and methotrexate. In the case of patient 1, however, a temporary improvement was noted after he received a weekly dose of 25 mg of methotrexate, and the clinical response correlated with a decrease in the proportion of CD4⁺CD7⁻ cells from 69% to 31%. However, his erythroderma relapsed 4 months later despite the treatment, and his CD4⁺CD7⁻ cell count increased to 71%. None of these 5 patients developed unequivocal Sézary syndrome as defined by the clinical or hematologic criteria,⁵ or any other kind of malignancy during their lifetime. All died within 4 years after diagnosis of causes unrelated to erythroderma (2 of ischemic heart disease, 1 of stroke, and 1 of unknown cause).

The phenotypic characteristics of CD4⁺CD7⁻CD26⁻ cells were the following: CD3⁺, CD8⁻, CD25⁻, CD27^+/−, CD28⁺, CD29⁺, CD30⁻, CD45RO⁺, CD45RA⁻, CD56⁻, CD134⁻, HLA-DR⁻, CLA⁻, TCRαβ⁺, and TCRγδ⁻, ie, they corresponded to the phenotype of resting memory CD4⁺ cells. To investigate the possibility that the subpopulation of monoclonal T cells resided within the expanded CD4⁺CD7⁻CD26⁻ compartment, we repeated the TCRγ rearrangement studies on peripheral blood cells depleted of monoclonal CD4 and CD7 antibodies. In 2 cases (patients 1 and 3) we succeeded in obtaining pure populations containing less than 4% of CD4⁺ cells (CD4 antibody depletion) or CD7⁻ cells (CD7 antibody depletion). In both patients the monoclonal TCRγ rearrangement was absent in the CD4-depleted cells but still detectable after CD7 depletion or in positively selected CD4⁺ cells.

We were also interested in investigating the morphologic characteristics of CD4⁺CD7⁻ cells. To accomplish this task we took advantage of the laser scanning cytometry technique, which provides the possibility of finding cells with specific characteristics and observing their morphologic features under a light microscope. In each of the 4 described patients we observed 100 to 150 Wright-Giemsa–stained CD4⁺CD7⁻ cells and found that only 6 to 17 cells could be classified as having Sézary cell morphology. Otherwise, the cells were small lymphocytes. This was not different from what can be found in the peripheral blood of healthy volunteers. In conclusion, our findings indicate that some of the CD4⁺CD7⁻ cells are monoclonal but retain normal lymphocytic morphology.

Monoclonal expansions of lymphocytes occur occasionally in elderly individuals and comprise B-cell, T-cell, or plasma-cell dyscrasias. B-cell and plasma-cell dyscrasias are most common and can be detected in approximately 10% of healthy adults older than 80 years by the presence of a monoclonal immunoglobulin peak on serum electrophoresis. This condition, named monoclonal gammapathy of undetermined significance (MGUS) is considered to be preneoplastic since patients with MGUS have a severely increased risk for developing multiple myeloma and macroglobulinemia.²⁶ Another type of B−cell dyscrasia is monoclonal B lymphocytosis of undetermined significance (MLUS), which is a clinically benign variant of chronic B−cell leukemia.²⁷

Here we describe a subpopulation of elderly patients with idiopathic, chronic erythroderma with associated monoclonal dyscrasia of T cells (an abbreviation, MTUS, is proposed to be analogous with MGUS and MLUS). Monoclonal expansions of T cells are much less common that B−cell dyscrasias and their clinical relevance is poorly understood. CD8⁺ and CD4⁺ lymphocyte clones have been detected in elderly individuals^{28 - 29} and even in old mice.³⁰ These monoclonal T cells have a variable phenotype; however, the most consistent finding is CD7⁻CD45RO⁺, which resembles the phenotype seen in our patients with erythroderma.

Very little is known about the pathogenesis and clinical importance of monoclonal T−cell dyscrasias in elderly people. Some authors have suggested that they occur because of a transformational event analogous to the situation seen in MGUS, and it seems that at least in some cases (albeit rarely) clinically silent monoclonal T−cell dyscrasia progresses into malignancy.^{31 - 33} If the same pathogenic scenario was true for our patients, erythroderma with monoclonal T−cell dyscrasia could correspond to the pre−Sézary syndrome.¹⁰ Some support to this speculative notion is provided by an obvious similarity with the known surface phenotype of the cells in Sézary syndrome or advanced mycosis fungoides.

A lack of CD7 and CD26 expression is by no means specific to Sézary cells. Reactive−memory T cells in benign inflammatory skin conditions, such as atopic dermatitis, psoriasis, or infections, often have the CD7⁻CD26⁻CD45RO⁺ phenotype.^{34 - 36} According to this concept monoclonal T−cell populations represent reactive cells that were not eliminated by immunoregulatory processes. It is important to stress in this context that T−cell rearrangement studies do not allow differentiating whether the detected clone is a result of malignant tranformation or, rather, due to selection of reactive T lymphocytes. None of our patients with erythroderma and monoclonal T−cell dyscrasia fulfilled the criteria of Sézary syndrome; they all had Sézary cell counts below 5% or 1000/μL, did not develop lymphadenopathy, and their CD4/CD8 ratio was much lower than that normally seen in Sézary syndrome. The CD7 expression is much lower in Sézary syndrome than in MTUS (Figure 3B). Moreover, the CD4⁺CD7⁻ cells in our patients did not have the Sézary cell morphology but mostly resembled normal small lymphocytes. None of the patients developed lymphoma or an other malignancy during the period of observation. Taken together, we cannot eliminate the possibility that observed monoclonal cells are non−neoplastic or alternatively represent an “abortive” dead−end transformation event.

Monoclonal CD4 and CD8 cells have repeatedly been found in different autoimmune and inflammatory diseases such as rheumatoid arthritis, atherosclerosis, chronic viral infections, and multiple sclerosis.^{37 - 41} At least in patients with rheumatoid arthritis, there is ample evidence suggesting that monoclonal circulating CD4⁺CD7⁻CD45RO⁺ cells escaped from peripheral tolerance and are autoreactive.^{40 ,42} The maintenance of these clones is supported by some cytokines⁴³ and further enhanced by their resistance to apoptosis.⁴⁴ It is conceivable that the monoclonal T−cell population in patients with erythroderma results from a chronic stimulation of the immune system with a yet unidentified cutaneous autoantigen. However, in contrast to the clones detected in rheumatoid arthritis⁴⁰ and in healthy aging individuals,^{41 ,44 - 47} which are approximately enriched 5−fold in CD28⁻ cells, we did not find any differences in the proportion of CD4⁺28⁻ cells between patients with monoclonal T−cell dyscrasia and those with a known cause of erythroderma (Table 2). This seems to be an important point since it has been argued that the loss of CD28 expression occurs specifically in senescent T cells,⁴⁷ depends on the cytokine balance,⁴⁸ and, most importantly, the CD4⁺28⁻ cells may represent a functionally distinct cell type resembling the natural killer lymphocytes.^{48 - 49} If the T−cell clones detected by us in patients with erythrodermia represent inflammatory, possibly autoreactive T cells, their nature is probably different from that of CD4⁺CD7⁻CD28⁻ cells.

In summary, we describe an association between chronic idiopathic erythroderma with monoclonal T−cell dyscrasia in elderly patients (MTUS−E syndrome). It is likely that the presence of T−cell monoclonality and erythroderma are related since their association by chance is extremely improbable ( approximately 10⁻⁸, as the probability of having erythroderma is 10⁻⁵ and the probability of having T−cell dyscrasia is 10⁻³). Moreover, the case of patient 1, in whom the use of methotrexate provided a temporary alleviation of symptoms, with an associated decrease in the proportion of CD4⁺CD7⁻CD26⁻ cells, further underscores the possible causal relationships between erythroderma and T−cell dyscrasia. However, it remains unknown whether the detected monoclonal population of CD4⁺CD7⁻CD26⁻ cells represents (auto)reactive T−lymphocytes mediating the chronic inflammatory skin reaction (pseudo–CTCL erythroderma) or, rather, is a result of a transformation event (an abortive or very indolent form of erythrodermic CTCL).

Correspondence: Robert Gniadecki, MD, DSc, Department of Dermatology D, Bispebjerg Hospital, Bispebjerg bakke 23, DK−2400 Copenhagen NV, Denmark (rg01@bbh.hosp.dk).

Accepted for Publication: July 23, 2004.

Funding/Support: This work was supported by grants from Aage Bangs Fond and the Haenschs Fund (Dr Gniadecki).

Acknowledgment: We thank Ingelise Pedersen for her skillful assistance with the laser scanning cytometry.

Financial Disclosure: None.