Leukemia (2002) 16, 861–864  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Phenotypic transformation of CD52pos to CD52neg leukemic T cells as a mechanism for resistance to CAMPATH-1H RE Birhiray1, G Shaw2, S Guldan3, D Rudolf3, D Delmastro4, P Santabarbara5 and L Brettman5

1Department of Hematology/Oncology,Marshfield Clinic, Marshfield, WI, USA; 2Marshfield Laboratories, Marshfield Clinic, Marshfield, WI, USA; 3Marshfield Medical Research and Education Foundation, A Division of Marshfield Clinic, Marshfield, WI, USA; 4Regional Cancer Center, Marshfield Clinic-Eau Claire, WI, USA; and 5Millennium and ILEX Partners, Cambridge, MA, USA

Immunotherapy utilizing CAMPATH-1H for patients with tory to conventional cytotoxic chemotherapy. Median survival chemotherapy-refractory chronic lymphocytic leukemia has is 7.5 months.16 Nondurable responses to the purine ana- yielded encouraging results with many reports of complete logues, particularly deoxycoformycin (DCF), have been remission. Here we report the outcome of two patients with 15,17 CD4-positive T cell prolymphocytic leukemia treated with CAM- reported. PATH-1H. Both patients responded rapidly to treatment and CAMPATH-1H therapy can result in a superior response subsequently developed CD4 lymphopenia. One patient rate compared with DCF in patients with refractory T-PLL.16 remained in complete remission after 14 weeks of treatment. A durable CR rate of 60% was reported. In three relapsed Serial peripheral blood flow cytometry revealed that the CD52 patients retreatment resulted in second CRs, two of whom antigen was present throughout treatment. The other patient who was initially CD52-positive, became CD52-negative after 6 were successfully autografted with hematopoetic stem cells weeks of treatment, and developed progressive symptoms of collected during the first CR. Patients who failed treatment T cell prolymphocytic leukemia. Immunotherapy was stopped, with CAMPATH-1H exhibited severe bone marrow aplasia chemotherapy proved futile, and the patient died. This change and pancytopenia of unknown cause. in phenotype from CD52-positive to -negative during CAM- We report the use of CAMPATH-1H as initial therapy in PATH-1H therapy points out a need to develop strategies for two patients with T-PLL. One of these patients demonstrated maintaining antigenic expression during therapy. the emergence of a CD52-negative clone resistant and, as a Leukemia (2002) 16, 861–864. DOI: 10.1038/sj/leu/2402471 consequence, was refractory to CAMPATH-1H therapy. Use Keywords: chronic lymphocytic leukemia; CAMPATH-1H; CD4; of CAMPATH-1H has resulted in CD52-negative populations CD52; monoclonal antibody therapy; T cell prolymphocytic leukemia emerging in rheumatoid arthritis and non-Hodgkin’s lym- phoma patients.18,19 This is the first reported case of pheno- typic conversion of leukemic cells from CD52-positive to Introduction CD52-negative during CAMPATH-1H therapy in T-PLL. Aspects of this case provide insight into mechanisms of CD52 antigen is a glycosylated, phosphatidylinositolglycan treatment failure. (PIG)-anchored (PIG-AP) with a sequence of 12 amino acids.1 In humans, it is expressed on normal and neoplastic , and macrophages, but not Study design erythrocytes and hematopoetic stem cells.2,3 It is present in a majority of B cell malignancies, most cases of T cell and B Peripheral blood was drawn from two patients with T-PLL and cell prolymphocytic leukemia (PLL) and some cases of acute shipped as EDTA whole blood to Cytometry Associates myelogenous and lymphocytic leukemias.4,5 (Brentwood, TN, USA) for /leukemia phenotyping. Immunotherapy with monoclonal antibodies (MoAbs) tar- Tests were performed in accordance with the recommen- geted at CD52 can be clinically useful.4,6,7 CD52 is an excel- dations of the US–Canadian Consensus Conference on the lent therapeutic target because of its abundant expression Immunophenotypic Analysis of Hematologic Neoplasia by (approximately 4.5 × 106 molecules per ), its close Flow Cytometry.20,21 Anti-CD52 MoAb was obtained from apposition to the cell membrane, and its apparent lack of Serotec (Raleigh, NC, USA). modulation.1,8 Both patients were enrolled in institutional review board CAMPATH-1H is a humanized monoclonal immunoglob- (IRB)-approved clinical studies at the Marshfield Clinic ulin (Ig) G1 MoAb directed against CD52.2,4,8,9 In vitro, CAM- Oncology Center. CAMPATH-1H was obtained from Millen- PATH-1H causes cell lysis using host effector mech- nium and ILEX Partners, Cambridge, MA, USA. anisms.2,4,10 Therapeutic activity has been shown in auto- immune disorders,7 refractory inflammatory disease,11 allo- graft T cell depletion in bone marrow transplantation,12 and Results and discussion hematolymphoid malignancies.13 Studies are ongoing with CAMPATH-1H therapy for patients with chemotherapy refrac- A 62-year-old white female presented in January 2000 with tory chronic lymphocytic leukemia (CLL), with reports of fatigue, anemia and a white blood count of 249000. Periph- complete remissions (CRs).5,14 eral blood flow cytometry on 21 January 2000 demonstrated T-PLL is a malignancy of T cells with distinctive clinical, mature (post-thymic) T cell malignancy positive for CD52, cytologic and cytogenic features.15 It is aggressive and refrac- CD2, CD3, CD7, CD4 and HLA-DR, and negative for CD5, CD8, TdT and B cell markers consistent with T-PLL (Table 1).15,21 The demonstration of CD52 expression was strong with 100% of the cells positive and staining with bright inten- Correspondence: R Birhiray, Hematology and Oncology of Indiana, 8330 Naab Road, Suite 135, Indianapolis, IN 46260–1932, USA; Fax: sity. Cytogenetics showed 11 of 21 cells with large pericentric 317–871–5518 inversion of 6, and a subclone of seven cells Received 4 May 2001; accepted 21 December 2001 also having an inverted 1p duplication. A mechanism for resistance to CAMPATH-1H RE Birhiray et al 862 Table 1 Immunophenotypic analysis of T cells from a 62-year-old white female patient

21 January 17 April 3 May 8 June 2000 2000 2000 2000

CD1 −−−− CD2 ++++ CD3 ++++ CD4 ++++ CD5 −−±− CD7 ±±±± CD8 −−−− CD16 −−−− CD25 −−−− CD52 +−−− CD56 −−−− CD57 −−−− TCR ␥/␦ −−−− HLA-DR ++++

Therapy was initiated with leukopheresis followed by CAM- PATH-1H from 11 February 2000 to 19 April 2000, with an initial test dose of 3 mg, followed by 10 mg 2 days later, and beginning on day 5, 30 mg three times a week for 9 weeks. She responded briskly, becoming neutropenic after 1 week. Figure 1 Dual staining CD52 vs CD3 or CD2 in patient 1 pre- and However, 7 weeks into therapy she developed fevers and a post-CAMPATH-1H treatment. The profile of 21 January 2000 is 3 new pre-auricular adenopathy. Over the next 2 weeks she weeks pre-treatment. CD52-positive CD3 cells predominate. Nearly developed progressive fevers, night sweats and liver function 10 weeks into treatment, on 17 April 2000, CD52-positive cells had been eliminated. Two weeks after termination of treatment, on 3 May test abnormalities. Liver biopsy showed extensive portal and 2000, two populations of CD52-negative CD3 cells had emerged. sinusoidal involvement by T-PLL. Lymph node biopsy and Seven weeks after termination of treatment, on 8 June 2000 (39 times repeat peripheral blood flow cytometry on 17 April 2000 longer than the half-life of CAMPATH-1H), very few CD52-positive revealed recurrent disease with loss of CD52-positivity (Table CD2 cells were detectable. There was no evidence of the re-emerg- 1, Figure 1). B lymphoid cells were virtually absent (0.1% of ence of significant numbers of CD52-positive cells. FACS profiles all cells in the specimen). Only 4% of mature T cells displayed were obtained on cells stained with CAMPATH-1H F(ab)2-FITCand anti CD3-PE or CD2-PE. Profiles shown were obtained after gating on an antigenic profile similar to that on 21 January 2000. The the lymphocyte population. remainder were CD52 negative. CAMPATH-1H was discontinued and chemotherapy was initiated with 2-CDA without response. On 3 May 2000, 2 content analysis of all cells showed an aneuploid pattern (60% weeks after termination of CAMPATH-1H treatment, two of all cells). The aneuploid cells displayed an increased pro- populations of CD52-negative CD3 cells had emerged (Table liferative component. This indicated the continued presence 1, Figure 1). At this point 92% of the population of T cells of a predominant population of T cells displaying an aberrant were CD52-negative. The DNA content analysis of all cells antigenic profile. The majority of these cells revealed DNA showed an aneuploid pattern (14% of all cells). There was a aneuploidy and increased cell size indicating an expanding predominant population of T cells present displaying the secondary clone compared with only 14% aneuploid cells on aberrant antigenic profile. There was a subpopulation with 3 May 2000. These cells remained negative for expression of increased cell size and aneuploidy, suggesting an evolving CD52. Subsequently the patient received cyclophosphamide, secondary clone. These cells were also negative for expression vincristine, prednisone and pentostatin without response, and of CD52. died in June 2000. Serial testing for CD52 status remained negative 7 weeks In September 1998, a second patient, an 83-year-old white (ie on 8 June 2000) after the discontinuation of CAMPATH- male presented with a white blood count of 42000, a platelet 1H (Table 1, Figure 1). More than 98% of the cells in the count of 161, and a hemoglobin of 15.2. His medical history specimen were tumor cells. The composite phenotype of the included resected melanoma, metastatic prostate carcinoma, tumor cells had been established to be CD2+CD3+CD7+CD4+. resected thyroid carcinoma and hemangiopericytoma. Periph- Because the specimen from 3 May 2000 had demonstrated a eral blood flow cytometry showed expression of CD2, CD4, bimodal expression of CD3, it was felt that CD52 should be CD5 and CD7 with aberrant loss of CD3. CD8, TdT and B looked at in the context of the other pan-T cell marker cell markers were negative. On the basis of this aberrant expressed by the tumor (ie CD2). In this way, an evaluation phenotype, a diagnosis of T-PLL was made. of CD52 expression on a positively identified tumor cell Therapy with CAMPATH-1H antibody was initiated in would be ensured. CD3 was assessed on this specimen as October 1998 and continued for 12 weeks in conjunction well, and it was uniformly brightly positive on all tumor cells with combined androgen blockade. He responded briskly to with no evidence of the bimodal distribution seen in the therapy with the development of CD4 lymphopenia for which previous specimen. he received antibiotic and anti-viral prophylaxis until Janu- Repeat cytogenetics showed evidence of clonal progression ary 2000. with a subclone having 78 to 82 . The DNA Repeat peripheral blood flow cytometry in January 1999

Leukemia A mechanism for resistance to CAMPATH-1H RE Birhiray et al 863 showed a CR. However, in January 2000, early recurrence an allograft of T cells. A CD52-negative T cell population became apparent, but clinically the patient has remained emerged that resulted in graft rejection.26 stable. His complete blood count in June 2000 showed a This case report suggests that the emergence of CD52-nega- white blood count of 13.9, hemoglobin of 13.9, and a platelet tive malignant clones can occur following CAMPATH-1H count of 245 000 with a normal differential. CD52 antigen therapy. While not a part of the current study, determination testing has remained positive. of the expression of GPI-linked antigens CD55 and CD59 (or Little is known about antigenic changes in relapsed post- lack thereof) would most definitively make the case that GPI- thymic leukemias or . This is the first reported deficient cells had emerged. Therefore, if patients progress on instance of true phenotypic transformation from positive to therapy, flow cytometric evaluation for CD52, CD55 and negative CD52 antigenicity in human leukemic T cells asso- CD59 should be considered and further treatment decisions ciated with CAMPATH-1H therapy. The consequences were made based upon the results. critical. CAMPATH-1H attaches to CD52-positive B and T cells resulting in their rapid clearance. After the CD52-positive cells Acknowledgements are cleared, a population of persistent CD52-negative B and 18,19,22 T cells sometimes emerges. Emergent cells exhibit a sev- This work was supported in part by a grant from the Marsh- ere defect in the synthesis of a glycosylphosphatidylinositol field Medical Research Foundation and funds from Millen- (GPI) precursor resulting in defective surface expression of nium and ILEX Partners, Cambridge, MA, USA. We would like 19 other GPI-anchored . The defect induced by CAM- to thank Esoterix Oncology for its continued assistance on the PATH-1H may be different in B and T cells. In B cells, the interpretation of these cases, and acknowledge support pro- defect in CD52 expression appears to be reversible in vitro. vided by the Marshfield Medical Research and Education CD52-negative T cells showed no such plasticity.22,23 Foundation through the assistance of Graig Eldred and Alice Response to CAMPATH-1H treatment is not uniform. Quan- Stargardt for manuscript preparation. titative estimates of CD52 expression on B and T cells is pre- dictive of treatment response: the greater the expression, the 5 greater the response. There may be a variation in CD52 References expression in the normal lymphocyte population, with a small proportion of cells being CD52-negative at any one time. By 1 Xia MQ, Hale G, Lifely MR, Ferguson MA, Campbell D, Packman depleting the positive cells, the negatives expand. L, Waldmann H. Structure of the CAMPATH-1 antigen, a glycosyl- Other MoAb therapies can exert selective pressure in favor phosphatidylinositol-anchored glycoprotein which is an excep- of small populations of clonal tumor cells that lack specific tionally good target for complement lysis. Biochem J 1993; 293: surface antigens resulting in tumor escape and relapse. The 633–640. 2 Hale G, Xia MQ, Tighe HP, Dyer MJ, Waldmann H. The CAM- emergence of CD20-negative clones has been reported in B PATH-1 antigen (CDw52). Tiss Antigen 1990; 35: 118–127. 24 cell lymphomas treated with . DNA analysis con- 3 Gilleece MH, Dexter TM. Effect of Campath-1H antibody on firmed that the transformed lymphoma originated from the pri- human hematopoietic progenitors in vitro. Blood 1993; 82: mary malignant clone. CD20 negativity was not simply due 807–812. to competitive inhibition of the CD20 binding sites by 4 Dyer MJ, Hale G, Marcus R, Waldmann H. Remission induction bound rituximab. in patients with lymphoid malignancies using unconjugated CAM- PATH-1 monoclonal antibodies. Leuk Lymph 1990; 2: 179–193. One explanation for the detection of only CD52-negative 5 Ginaldi L, DeMartinis M, Matutes E, Farahat N, Morilla R, Dyer cells from a patient following CAMPATH-1H treatment could MJ, Catovsky D. Levels of expression of CD52 in normal and leu- be that bound antibody competes with the detecting antibody kemic B and T cells: correlation with in vivo therapeutic responses used for flow cytometry. In our first patient, CD52 negativity to Campath-1H. Leuk Res 1998; 22: 185–191. cannot be due to competitive inhibition of the CD52 binding 6 Dyer MJ, Hale G, Hayhoe FG, Waldmann H. Effects of CAM- sites by bound residual CAMPATH-1H, because the clonal PATH-1 antibodies in vivo in patients with lymphoid malig- nancies: influence of antibody isotype. Blood 1989; 73: 1431– cells remained CD52-negative more than 7 weeks after the 1439. discontinuation of CAMPATH-1H. The half-life of CAMPATH- 7 Isaacs JD, Watts RA, Hazleman BL, Hale G, Keogan MT, Cobbold 1H at 4 weeks of treatment has been determined to range from SP, Waldmann H. Humanised monoclonal antibody therapy for 34 to 60 h: at 7.5 mg/week, t. = 34 ± 17 h; at 24 mg/week, rheumatoid arthritis. Lancet 1992; 340: 748–752. 8 Bindon CI, Hale G, Waldmann H. Importance of antigen speci- t. = 57 ± 25 h; at 75 mg/week, t. = 60 ± 36 h; at 240 mg/week, ficity for complement-mediated lysis by monoclonal antibodies. t. = 53 ± 38 h. The overall mean is 53 ± 31 h. At 1 week ± Eur J Immunol 1988; 18: 1507–1514. of treatment, the mean half-life is 42 27 h (D Rodriguez, 9 Riechmann L, Clark M, Waldmann H, Winter G. Reshaping Millennium Pharmaceutical, personal communication, human antibodies for therapy. Nature 1988; 332: 323–327. August, 2000). It is expected that after 5 half-lives (480 h or 10 Crowe JS, Hall VS, Smith MA, Cooper HJ, Tite JP. Humanized 20 days), no drug would be available to bind the neoplastic monoclonal antibody CAMPATH-1H: myeloma cell expression of lymphocytes or competitively block antigenic sites and yield genomic constructs, nucleotide sequence of cDNA constructs and false CD52-negative status in flow cytometry. Additionally, at comparison of effector mechanisms of myeloma and Chinese hamster ovary cell-derived material. Clin Exp Immunol 1992; 87: this stage no CAMPATH-1H-bound cells should remain in the + + 105–110. circulation. While the half-life of CD4 and CD8 T-lympho- 11 Dick AD, Meyer P, James T, Forrester JV, Hale G, Waldmann H, cytes has been found to be 87 and 77 days, respectively,25 Isaacs JD. Campath-1H therapy in refractory ocular inflammatory cells with bound CAMPATH-1H are cleared from the body disease. Br J Ophthalmol 2000; 84: 107–109. almost immediately after treatment. 12 Curtis RE, Travis LB, Rowlings PA, Socie G, Kingma DW, Banks This is the first reported case in which CD52-negative leu- PM, Jaffe ES, Sale GE, Horowitz MM, Witherspoon RP, Shriner DA, Weisdorf DJ, Kolb HJ, Sullivan KM, Sobocinski KA, Gale RP, kemic T cells emerged following CAMPATH-1H treatment. Hoover RN, Fraumeni JF Jr, Deeg HJ. Risk of lymphoproliferative Failure of a similar nature has been reported using CAMPATH- disorders after bone marrow transplantation: a multi-institutional 1G and CAMPATH-1M in an in vivo/ex vivo attempt to purge study. Blood 1999; 94: 2208–2216.

Leukemia A mechanism for resistance to CAMPATH-1H RE Birhiray et al 864 13 Dyer MJ. The role of CAMPATH-1 antibodies in the treatment of analysis of hematologic neoplasia by flow cytometry: medical lymphoid malignancies. Semin Oncol 1999; 26: 52–57. indications. Cytometry 1997; 30: 249–263. 14 Cull GM, Haynes AP, Byrne JL, Carter GI, Miflin G, Rebello P, 21 Harris NL, Jaffe ES, Stein H, Banks PM, Chan JK, Cleary ML, Delsol Hale G, Waldmann H, Russell NH. Preliminary experience of allo- G, De Wolf-Peeters C, Falini B, Gatter KC. A revised European– geneic transplantation for lymphoproliferative disorders American classification of lymphoid neoplasms: a proposal from using BEAM-CAMPATH conditioning: an effective regimen with the International Lymphoma Study Group. Blood 1994; 84: low procedure-related toxicity. Br J Haematol 2000; 108: 754– 1361–1392. 760. 22 Taylor VC, Sims M, Brett S, Field MC. Antibody selection against 15 Matutes E. T-cell prolymphocytic leukemia. Cancer Control 1998; CD52 produces a paroxysmal nocturnal haemoglobinuria pheno- 5: 19–24. type in human lymphocytes by a novel mechanism. Biochem J 16 Pawson R, Dyer MJ, Barge R, Matutes E, Thornton PD, Emmett E, 1997; 322: 919–925. Kluin-Nelemans JC, Fibbe WE, Willemze R, Catovsky D. Treat- 23 Rowan W, Tite J, Topley P, Brett SJ. Cross-linking of the CAM- ment of T-cell prolymphocytic leukemia with human CD52 anti- PATH-1 antigen (CD52) mediates growth inhibition in human B- body. J Clin Oncol 1997; 15: 2667–2672. and T-lymphoma cell lines, and subsequent emergence of CD52- 17 Catovsky D. Clinical experience with 2Ј-deoxycoformycin. Hema- deficient cells. Immunology 1998; 95: 427–436. tol Cell Ther 1996; 38 (Suppl. 2): S103–S107. 24 Davis TA, Czerwinski DK, Levy R. Therapy of B-cell lymphoma 18 Hertenstein B, Wagner B, Bunjes D, Duncker C, Raghavachar A, with anti-CD20 antibodies can result in the loss of CD20 antigen Arnold R, Heimpel H, Schrezenmeier H. Emergence of CD52-, expression. Clin Cancer Res 1999; 5: 611–615. phosphatidylinositolglycan-anchor-deficient T lymphocytes after 25 Hellerstein M, Hanley MB, Cesar D, Siler S, Papageorgopoulos C, in vivo application of Campath-1H for refractory B-cell non-Hodg- Wieder E, Schmidt D, Hoh R, Neese R, Macallan D, Deeks S, kin lymphoma. Blood 1995; 86: 1487–1492. McCune JM. Directly measured kinetics of circulating T lympho- 19 Brett SJ, Baxter G, Cooper H, Rowan W, Regan T, Tite J, Rapson cytes in normal and HIV-1-infected humans. Nature Med 1999; N. Emergence of CD52-, glycosylphosphatidylinositol-anchor- 5: 83–89. deficient lymphocytes in rheumatoid arthritis patients following 26 Bunjes D, Theobald M, Wiesneth M, Schrezenmeier H, Hoffmann Campath-1H treatment. Int Immunol 1996; 8: 325–334. T, Hertenstein B, Arnold R, Heimpel H. Graft rejection by a popu- 20 Davis BH, Foucar K, Szczarkowski W, Ball E, Witzig T, Foon KA, lation of primed CDw52-host T cells after in vivo/ex vivo T- Wells D, Kotylo P, Johnson R, Hanson C, Bessman D. US–Canad- depleted bone marrow transplantation. Bone Marrow Transplant ian Consensus recommendations on the immunophenotypic 1993; 12: 209–215.

Leukemia