Abstract

To the Editor: Peripheral T-cell lymphoma, not otherwise specified (PTCL-NOS) is a broad category of biologically and clinically heterogeneous diseases that cannot be further classified into any other existing entity of lymphoma as defined by the World Health Organization classification.1,2 It more commonly affects men with an average age of presentation of 60 years.1 It can involve the bone marrow, liver, spleen, lymph nodes, and even the skin, where it can have variety of cutaneous presentations including eczematous dermatitis, erythroderma or morbilliform drug-like eruption, and generalized folliculitis-like eruption.2 CD52 antigen is expressed in approximately 40% of PTCL-NOS,3 and alemtuzumab, a monoclonal antibody targeting CD52, has been shown to be helpful in relapsed or chemotherapy refractory PTCL-NOS.4 However, the use of monoclonal antibody therapy is not without its complications. Loss of antigen expression after the use of targeted monoclonal antibody therapy has been previously reported in the literature, and a well-reported example is loss of CD20 expression after treatment with rituximab, a chimeric anti-CD20 monoclonal antibody most widely used in the treatment of non-Hodgkin B-cell lymphomas.5–7 Review of the literature reveals only rare reports of patients with leukemia having phenotypic change after alemtuzumab therapy with loss of CD52 assessed by flow cytometry.8,9 We present an instructive case in which we identify the immunohistochemical loss of CD52 expression in a biopsy of a skin lesion in a patient with PTCL-NOS after therapy with alemtuzumab. A 68-year-old white man with medical history of diabetes mellitus, pulmonary hypertension, atrial fibrillation, hypertension, arthritis, Crohn disease, ankylosing spondylitis, Agent Orange exposure, and porphyria cutanea tarda was referred to our dermatology clinic with generalized erythroderma of the face, trunk, proximal upper extremities, and red scaly papules on all extremities. He had supraclavicular and inguinal lymphadenopathy. Previous skin biopsy by an outside dermatologist showed mild acanthosis, spongiosis with only very few epidermotropic lymphocytes along with a striking band-like infiltrate of large lymphocytes. Immunohistochemistry showed strong diffuse CD3 reactivity and an abnormal CD4:CD8 ratio in excess of 10:1. CD20 highlighted dispersed aggregates of B cells, whereas CD30 highlighted scattered reactive lymphocytes. The histologic differential diagnosis for the above microscopic findings included a primary cutaneous T-cell lymphoma, in particular, plaque stage mycosis fungoides, and secondary cutaneous involvement by a systemic T-cell lymphoma. But the clinical presentation along with lack of significant epidermotropism argued against the former diagnosis somewhat. Repeat skin biopsies showed superficial to mid-dermal large lymphocytic infiltrate with only very focal epidermotropism (Fig. 1). Immunohistochemistry showed CD3- and CD7-positive T cells with a CD4–CD8 ratio of approximately 20:1. There were scattered CD30-positive lymphocytes. The infiltrate was negative for TDT, CD34, and CD56. There were only rare TCL-1 positive cells. CD68 highlighted histiocytes and CD43, myeloperoxidase and lysozyme stain showed scattered granulocytes. These findings were consistent with a T-cell lymphoma, not otherwise specified. Positron emission tomography (PET) showed extensive lymphadenopathy and subcutaneous nodules. Subsequent lymph node biopsy showed paracortical diffuse infiltrates with effacement of the normal lymph node architecture. Neoplastic cells showed CD3, CD4, CD5, CD52, and CD25 positivity and negative for CD8. CD30 highlighted scattered immunoblasts. PAX-5, CD10, and Bcl-6 highlighted B cells and the few residual germinal centers. The tumor cells were nonreactive with CD34, ALK-1, TCL-1, TdT, and CD56. The mitotic index assessed by MIB-1 (Ki-67) was approximately 30%. Epstein–Barr virus in situ hybridization (Epstein–Barr encoding region) was negative. In addition, HTLV-I and II antibodies were negative by enzyme-linked immunosorbent assay performed on peripheral blood.FIGURE 1.: Sections show a superficial to mid dermal lymphocytic infiltrate with only very focal epidermotropism (A, H&E ×40 magnification). The lymphocytes are large. Immunohistochemistry reveals CD3-reactive T cells with a CD4:CD8 ratio of approximately 20:1. There are scattered CD30 reactive lymphocytes. The infiltrate is nonreactive for TdT, CD34, and CD56. There are only rare TCL-1 reactive cells. CD68 highlights scattered histiocytes and CD43, myeloperoxidase and lysozyme stain scattered granulocytes (B, immunohistochemistry ×10 magnification).Flow cytometric analysis of the peripheral blood demonstrated a white blood cell count of 38,200/µL with 76% of cells within the small lymphocyte region, 1% within the large mononuclear region, and 23% within the granulocyte region. Within the small lymphocyte region, there was an abnormal population of immature T-lymphocytes that coexpressed CD2, CD3, CD4, CD5, CD7, and CD45. This population represented approximately 75% of all events analyzed by CD3. The neoplastic cells did not express CD8, CD1a, or CD34. There were no B cells in the specimen, and 1% of cells express the natural killer cell antigen, CD56. Based on the combined flow cytometric and cytomorphologic data, there is an abnormal, immature population of T-lymphocytes (75% of all cells) identified. This was consistent with peripheral blood involvement by T-lymphoblastic leukemia/lymphoma. Negative TdT detected in the bone marrow flow cytometry specimen and lymph node flow cytometry specimen. TdT was first positive for the peripheral blood flow cytometry but given the findings from the bone marrow and lymph node, was repeated for the peripheral blood and was negative. The immunophenotype was most consistent with a T-cell leukemia/lymphoma, CD4-positive. Molecular cytogenetic studies of a lymph node and bone marrow included conventional cytogenetic analysis, chromosome microarray analysis, and fluorescence in situ hybridization (FISH) using T-cell lymphoma panel. Results demonstrated a very complex karyotype that was resolved by chromosome microarray analysis showing interstitial deletions of 1p36, 2p22, 9p21 (twice), 10p, terminal deletions of 10q and 17p, duplications of 10p, 10q, and 17q plus gain of chromosome 8. The deletions of 9p and 9q encompassed 6 breakpoints indicating chromothripsis or genomic instability. One of the deletions in 9p included the CDKN2A tumor suppressor gene. The terminal deletion of 17p included the P53 gene. Therefore, FISH of these 2 genes using CDKN2A and P53 probes were performed to determine minimal residual disease in follow-up studies. To rule out any translocation involving T-cell receptors, FISH of T alpha- and delta-cell receptors were performed and found negative. Because FISH of T-cell rearrangements were negative and the a-CGH revealed a very complex karyotype, FISH and chromosome microarray analyzes are highly suggestive of PTCL-NOS rather than T-PLL as reported by Gesk et al.10 Taken together, the findings present in the lymph node, bone marrow, and peripheral blood, when correlated with the clinical presentation, were most compatible with a diagnosis of PTCL-NOS. Additional testing showed lymph node reactivity for CD52 (Fig. 2), and his initial skin biopsy specimen was retrospectively stained and showed similar CD52-reactivity in a subset of the cutaneous infiltrate (Fig. 3A). Given the expression of CD52 and aforementioned history of atrial fibrillation, anthracycline-based therapy was relatively contraindicated; therefore, the patient was started on alemtuzumab monotherapy. He received 15 doses of alemtuzumab over 6 weeks. He had a complete therapeutic response in the skin and lymph nodes, as confirmed with PET. His posttreatment bone marrow biopsy showed no evidence of residual/recurrent lymphoma. Unfortunately, 3 weeks after treatment, the patient developed a morbilliform eruption on the trunk and upper arms. Repeat skin biopsy showed an atypical dermal lymphocytic infiltrate, consistent with persistence/recurrence of the malignant T-cell population, and his PET/computed tomography showed disease progression. FISH of CDKN2A and P53 probes followed by cytogenetic analysis confirmed the continued presence of the original clone along with the presence of an evolved abnormal clone consisting of CDKN2A being inserted to either the short arm of chromosome 2 or translocated onto 7p and 7q indicating again a genomic instability of this region.FIGURE 2.: Sections show lymph node involvement by PTCL-NOS (A, H&E ×600 magnification) with immunoreactivity for CD52 (B, CD52 immunostain ×600 magnification).FIGURE 3.: The initial skin biopsy was retrospectively stained and showed CD52-reactivity in a subset of the malignant cutaneous lymphocytic infiltrate (A, CD52 immunostain ×400 magnification). The posttreatment biopsy showed a diffuse lymphocytic infiltrate in the superficial and mid dermis with associated dermal hemorrhage. The lymphocytes were again large with hyperchromatic nuclei and were highlighted diffusely with CD3, CD4, and CD7. CD8 stain revealed only rare positive cells. The lymphocytes were negative for CD20 and CD56. In contrast to the initial skin biopsy, the CD52 staining is now absent (B, CD52 immunostain ×400 magnification).Consequently, the patient was started on alemtuzumab for a second course because of failure to tolerate bendamustine. However, his facial eruption worsened, and a subsequent skin biopsy revealed persistent T-cell lymphoma. But, unlike his original tumor, CD52 was now negative in the neoplastic cells (Fig. 3B). After 5 weeks, alemtuzumab was discontinued because of peripheral edema and therapeutic response was unclear. He was not a candidate for bone marrow transplant given his chronic kidney disease and pulmonary hypertension. Ultimately, he opted for comfort care and died of respiratory distress. CD52 is normally expressed on all B- and T-lymphocytes as well as monocytes.11 Alemtuzumab, a monoclonal antibody against CD52, functions by attaching to CD52 on T and B cells leading to their rapid clearance. Thus, immediately after treatment, CD52-negative cells may emerge,9,12–14 yet it is interesting that the level of CD52 antigen expression need not correlate with therapeutic response,15 and that, furthermore, resistance to alemtuzumab need not have negative CD52 expression.15 Given that the CD52 gene that encodes the CD52 glycoprotein is located in the 1p36 chromosome region, an area that was found deleted in 1 copy in the bone marrow and lymph node in our patient, it is plausible that the loss of CD52 glycoprotein during the course of the disease could have been due to a second hit involving the second gene copy. An alternative hypothesis could be the acquisition of additional genetic alterations leading to genomic instability and potential downstream changes. Another theory could be that the CD52-negative cells were present but in lesser proportion at the time of initial diagnosis, and because of selective pressure on the tumor lymphocytes, resulted in resistance to alemtuzumab.8 This latter theory could be supported by the presence of both CD52-positive lymphocytes and CD52-negative lymphocytes in our patient's initial skin biopsy. To date, the exact mechanism of disease resistance and the change in phenotype remains unknown.