Two new trifunctional antibodies for the therapy of human malignant melanoma

Trifunctional antibodies are able to redirect T cells and Fcgamma receptor(+) accessory immune cells to tumor targets. The simultaneous activation of these different classes of effector cells results in efficient killing of the tumor cells by different mechanisms such as phagocytosis and perforin-mediated cytotoxicity. Here, we introduce 2 new trifunctional antibodies specific for human melanoma. These trifunctional antibodies recognize with one binding arm CD3 on human T cells. The other binding arm is directed against melanoma-associated proteoglycans or melanoma-associated gangliosides (GD2 as well as GD3). They mediate specific lysis of various melanoma cell lines in correlation with the level of antigen expression in short-term cytotoxicity experiments. A combination of the 2 trifunctional antibodies was equally or even more efficient. Moreover, they induced a strong Th1 cytokine pattern with high amounts of IFN-gamma and low or no IL-4. Accordingly, CD4(+) and especially CD8(+) T cells expanded, whereas B cells, NK cells and monocytes decreased. The cytokine response was up to 16-fold higher when tumor cells were present. IFN-gamma reached cytotoxic concentrations for SK-MEL-23 melanoma cells. The induction of a T-cell-activatory and melanoma cell-inhibitory cytokine milieu together with the redirection of T-cell- and accessory cell-mediated cytotoxicity are interesting features of these trifunctional antibodies. They may be a new option for the therapy of human malignant melanoma.     

Isolation of native human monoclonal autoantibodies to breast cancer

Using a unique fusion partner cell line, MFP-2, and B-lymphocytes from breast cancer patients, we developed a set of fully human monoclonal antibodies (MAbs) that bind with high specificity and sensitivity to breast cancer cells. Immunofluorescent staining of normal tissues, primary tumors, and metastatic lymph nodes demonstrates that these antibodies are specific for breast cancer of autologous and allogeneic origin. We have also determined that many of the antibodies selected based on specific binding to breast cancer cells and tissue also bind prostate cancer cells and tissue with high specificity and sensitivity. The targets of these antibodies have been localized to the cytoplasm and membrane. Biological assays for internalization and cytotoxicity demonstrated the ability of three antibodies to rapidly internalize. Our study demonstrates that isolation of native human MAbs from the natural antibody repertoire, targeted to cancer cells, is feasible and may provide a source of tools for immunotherapy.     

Rituximab in a patient with acute renal failure due to B-cell lymphomatous infiltration of the kidneys

Renal failure is known to occur in lymphoproliferative disorders because of ureteral obstruction or parenchymal infiltration by disease. Rituximab is a genetically engineered chimeric murine/human monoclonal antibody directed against the CD20 antigen found on the surface of normal and malignant B-lymphocytes. The pharmacokinetics and metabolism of rituximab is not well established. The extent of renal clearance is not fully known, with little experience reported on the use of rituximab in patients with renal failure. We present a case where rituximab was administered to a patient with acute renal failure due to bilateral kidney infiltration by non-Hodgkin's lymphoma (NHL). The patients renal function improved on therapy, with no need for hemodialysis and there were no significant toxicities. Rituximab may be used as a treatment option for NHL patients with impaired renal function.     

A monoclonal antibody to the human c-erbB3 protein stimulates the anchorage-independent growth of breast cancer cell lines.

The c-erbB3 protein is a member of the type I growth factor receptor family. It has a widespread pattern of expression in normal tissues and is overexpressed in about 20% of breast cancers. We have raised a specific monoclonal antibody, called SGP1, against the extracellular domain of c-erbB3 which recognises the native form of the protein. The monoclonal antibody was found to modestly but significantly stimulate the anchorage-independent cloning efficiency of the breast tumour cell lines BT483 and T47D, both of which express the c-erbB3 protein. No effect was observed on 293 cells lacking expression, nor did a control isotype-matched antibody promote the growth of any of the cells tested. These results suggest that the c-erbB3 protein may normally act as a growth factor receptor.     

Aberrant expression of B203.13 antigen in acute lymphoid leukemia of B-cell origin

The B203.13 monoclonal antibody was developed by immunizing mice with the B/monocyte biphenotypic cell line B1b. During normal hematopoiesis B203.13 is expressed on a fraction of CD34+ cells, while on mature cells it is only present on B-lymphocytes. We tested this antibody as a marker of childhood B-acute lymphoblastic leukemia (B-ALL). Bone marrow aspirates from 139 cases of early B-ALL and 25 controls were studied. About 40% of the B-ALL patients expressed B203.13. In these patients, B203.13 was constantly co-expressed with CD10, but never co-expressed with CD20, contrary to the controls. The CD10(+)/B203.13(+) phenotype was specific to B-ALL, since CD10(+)/CD20(+) cells from common acute lymphoblastic leukemia (c-ALL) did not express B203.13. We concluded that the use of B203.13 in association with CD10 and CD20 provides meaningful information for distinguishing normal residual B-cells from leukemic B-lymphoblasts and that recurrence of a CD10(+)/B203.13(+) phenotype after transplantation may be a very early relapse indicator of early B-acute lymphoblastic leukemia.     

Augmented expression of programmed death-1 in both neoplastic and non-neoplastic CD4+ T-cells in adult T-cell leukemia/lymphoma

Adult T-cell leukemia/lymphoma (ATL) is a CD4(+)CD25(+) T-cell malignancy infected with human T-cell leukemia virus type-I (HTLV-I). HTLV-I infection causes the T-cell dysfunction, which contributes to the immunodeficient state of the patients. Programmed death-1 (PD-1) can negatively regulate T-cell response, when its ligand, PD-L1 or PD-L2 mainly expressed on antigen presenting cells, binds to this B7 family receptor. We investigated whether PD-1 is expressed on CD4(+) neoplastic (and/or non-neoplastic) cells or CD8(+) cytotoxic cells in peripheral blood mononuclear cells from 11 patients with ATL. By flow cytometry, we found that the levels of PD-1 expression on both CD4(+)CD25(+) and CD4(+)CD25(-) T-cell populations were increased in ATL patients compared to normal healthy volunteers, while PD-1 levels on CD8(+) T-cells were comparable between the patients and normal subjects. In stimulation with anti-CD3 antibody, the proliferation of PD-1-expressing T-cells from ATL patients was weak when compared to that of PD-1-nonexpressing normal T-cells. In addition to PD-1, PD-L1 was coexpressed on ATL cells in some patients, and PD-L1 expression was enhanced by stimulation with anti-CD3 antibody. Finally, the production of cytokines such as TNF-alpha by ATL cells was restored by blockade of PD-1/PD-L1 interaction. These findings suggest that CD4(+) T-cells are the main PD-1-expressing cells rather than CD8(+) T-cells in ATL patients, and both neoplastic and normal CD4(+) cells are exhausted as a result of PD-1 expression, and additionally PD-L1 expression on the neoplastic cell.     

Human cord blood long-term engrafting cells are CD34+ CD38-.

There have been controversies about CD34 and CD38 expression by human cord blood (CB) stem cells. Using the newborn NOD/SCID/beta2-microglobulin-null mouse assay that we recently developed, we examined the in vivo engrafting capability of human CB cells. Almost all of the 4-5 months engrafting cells were found in CD34(+) population. The capability of secondary reconstitution was found only in the CD34(+) cells. When the CD34(+) CB cells were separated into CD38(-) and CD38(+) subpopulations and tested for engraftment, the majority of the engrafting cells were detected in the CD38(-) subpopulation. These findings are consistent with the results from studies of murine stem cells and strongly indicate that the phenotype of human CB stem cells is CD34(+) CD38(-).     

The identification of irreversible rituximab-resistant lymphoma caused by CD20 gene mutations

C-terminal mutations of CD20 constitute part of the mechanisms that resist rituximab therapy. Most CD20 having a C-terminal mutation was not recognized by L26 antibody. As the exact epitope of L26 has not been determined, expression and localization of mutated CD20 have not been completely elucidated. In this study, we revealed that the binding site of L26 monoclonal antibody is located in the C-terminal cytoplasmic region of CD20 molecule, which was often lost in mutated CD20 molecules. This indicates that it is difficult to distinguish the mutation of CD20 from under expression of the CD20 protein. To detect comprehensive CD20 molecules including the resistant mutants, we developed a novel monoclonal antibody that recognizes the N-terminal cytoplasm region of CD20 molecule. We screened L26-negative cases with our antibody and found several mutations. A rituximab-binding analysis using the cryopreserved specimen that mutation was identified in CD20 molecules indicated that the C-terminal region of CD20 undertakes a critical role in presentation of the large loop in which the rituximab-binding site locates. Thus, combination of antibodies of two kinds of epitope permits the identification of C-terminal CD20 mutations associated with irreversible resistance to rituximab and may help the decision of the treatment strategy.     

Quantitative analysis of bcl-2 expression in normal and leukemic human B-cell differentiation.

Lack of apoptosis has been linked to prolonged survival of malignant B cells expressing bcl-2. The aim of the present study was to analyze the amount of bcl-2 protein expressed along normal human B-cell maturation and to establish the frequency of aberrant bcl-2 expression in B-cell malignancies. In normal bone marrow (n=11), bcl-2 expression obtained by quantitative multiparametric flow cytometry was highly variable: very low in both CD34(+) and CD34(-) B-cell precursors, high in mature B-lymphocytes and very high in plasma cells. Bcl-2 expression of mature B-lymphocytes from peripheral blood (n=10), spleen (n=8) and lymph node (n=5) was significantly higher (P<0.02) in CD23(-) as compared to CD23(+) B cells, independent of the type of tissue analyzed. Upon comparison with normal human B-cell maturation, bcl-2 expression in neoplastic B cells from 144 patients was found to be aberrant in 66% of the cases, usually corresponding to bcl-2 overexpression (63%). Follicular lymphoma (FL) carrying t(14;18) and MALT lymphoma were the only diagnostic groups constantly showing overexpression of bcl-2. Bcl-2 overexpression was also frequently found in precursor B-acute lymphoblastic leukemia (84%), typical (77%) and atypical (75%) B-cell chronic lymphocytic leukemia, prolymphocytic leukemia (two of three cases), mantle cell lymphoma (55%), but not in t(14;18)(-) FL, splenic marginal zone lymphoma, Burkitt lymphoma and multiple myeloma.     

Immunotherapy for non-Hodgkin's lymphoma: monoclonal antibodies and vaccines.

Advances in the development of monoclonal antibodies have led to new agents rapidly incorporated into standard lymphoma therapy. The characteristics of the target antigen and the properties of the antibody including interaction with the host immune system have been found to correlate with outcome. Antibodies targeting the CD20 antigen on B cells have been most widely used, led by the chimeric antibody rituximab, now used in nearly all types of B-cell non-Hodgkin's lymphoma (NHL). New antibodies targeting CD20 with augmented complement or Fc receptor binding are now being evaluated and will eventually have to be compared with rituximab. Challenges to these new antibodies include the nearly universal use of rituximab early in NHL therapy, and its increasing use as maintenance therapy. It is not clear what the activity of these antibodies will be in rituximab-refractory patients. New antibodies targeting antigens such as CD40 and CD80 are also being tested alone and in combination with rituximab. Vaccine trials using patient-specific immunization with immunoglobulin idiotype (Ig-Id present on the surface of most B-cell NHL) isolated by molecular rescue or by cell hybridization techniques are also nearing completion. These approaches attempt to actively induce specific humoral or cellular immune responses to the Ig-Id by attaching the protein to a carrier protein and the use of an immunologic adjuvant such as granulocyte macrophage colony-stimulating factor. Prior rituximab appears to delay humoral responses to the idiotype but may still allow cellular responses. The incorporation of all these approaches into optimal NHL therapy remains a challenge.     

CD22 as a target of passive immunotherapy

CD22 is a 135-kd B-cell restricted sialoglycoprotein present in the cytoplasm of virtually all B-lineage cells but expressed on the B-cell surface only at mature stages of differentiation. In humans, the vast majority of IgM(+)IgD(+) B cells express cell-surface CD22, while in lymphoid tissues CD22 expression is high in follicular mantle and marginal zone B cells and weak in germinal center B cells. In B-cell malignancies, CD22 expression ranges from 60% to 80% depending on the histological type and on the assays used. The function of the CD22 molecule is uncertain, although recent studies have suggested roles for the molecule both as a component of the B-cell activation complex and as an adhesion molecule. CD22-deficient mice have a reduced number of mature B cells in the bone marrow and circulation; the B cells have a shorter lifespan and enhanced apoptosis, thus indicating a key role of this antigen in B-cell development/survival. After binding with its natural ligand(s) or antibodies, CD22 is rapidly internalized; this provides a potent costimulatory signal in primary B-cell and proapoptotic signals in neoplastic B cells. Preclinically CD22 has been shown to be an effective target for immunotherapy of B-cell malignancies using either "naked" or toxin-labeled or radiolabeled monoclonal antibodies. Clinical trials in patients with non-Hodgkin's lymphoma (NHL) (both indolent and aggressive disease) are now ongoing with a humanized naked anti-CD22 antibody (epratuzumab, Amgen Inc, thousand Oaks, CA and Immunomedics Inc, Morris Plains, NJ) used as single agent or in combination with other monclonal antibodies (ie, rituximab) and/or chemotherapy. Preliminary data from these studies showed these approaches to be effective and well-tolerated.     

Expression of the c-erbB-2 protein in normal and transformed cells

Two synthetic peptides from the predicted sequence of the human c-erbB-2 protein were synthesized and used to raise antisera in rabbits. The sequences chosen were identical to those in the homologous rat c-neu protein. The antibodies produced reacted with the immunizing peptides in ELISA but showed little or no cross-reaction with the partly homologous peptides found in equivalent positions in the human EGF receptor. Both antipeptide antibodies, and a monoclonal antibody (MAb) specific for the rat neu protein, immunoprecipitated a 185-kDa protein from 35S-methionine-labelled lysates prepared from a rat cell line known to express high levels of the c-neu protein. The antipeptide antibodies also recognized a protein of the same size in Western blots. In addition, both antipeptide antibodies immunoprecipitated a 190-kDa protein from labelled cell lysates prepared from human and monkey cells. Antibodies to one of the peptides, which showed no detectable cross-reaction with human, rat or monkey EGF receptor, were used to examine the expression of the c-erbB-2 protein on a variety of cultured cell types. Eleven transformed, I non-established and 2 immortalized cell types were examined by immunoprecipitation for their level of expression of the c-erbB-2 protein and of the EGF receptor. The numbers of EGF receptors varied widely between different cell lines, whereas the level of the c-erbB-2 protein, which was found on all of the cell types examined, was more constant. The number of c-erbB-2 molecules present was estimated by autoradiography to be about 100,000 per cell. The antibodies were then used to examine the location and level of expression of the human c-erbB-2 and rat c-neu proteins in normal tissues. Immunohistochemical staining showed that the c-erbB-2 protein was highly expressed in rat kidney proximal tubules and loop of Henle. The c-erbB-2 protein was also present on normal human epithelial cells but in some cases with a different distribution to that of the EGF receptor.     

Cellular cytotoxicity mediated by isotype-switch variants of a monoclonal antibody to human neuroblastoma

The biological property of an antibody is determined by its antigen binding characteristics and its isotype-related effector functions. We have established monoclonal antibodies of different isotypes by stepwise selection and cloning of the hybridoma CE7. The original CE7 secretes an IgG1/kappa (CE7 gamma 1) antibody that recognises a 185 kD cell surface glycoprotein expressed on all human sympatho-adrenomedullary cells. Isotype-switch variants were isolated in the following sequence: from the original CE7 gamma 1, CE7 gamma 2b variants were isolated, and from a CE7 gamma 2b variant CE7 gamma 2a variants were isolated. The antibodies of three different isotype variant cell lines possess identical antigen binding characteristics, but display distinct effector functions as demonstrated by antibody dependent cell-mediated cytotoxicity (ADCC). ADCC was performed with the neuroblastoma line IMR-32 as the target cells, and different FcR gamma positive cells were either freshly isolated from human peripheral blood leukocytes or cultured for 6-10 days and tested as potential effector cells. Tumour lysis mediated by monocyte-derived macrophages depended on the presence of CE7 gamma 2a antibodies; antibodies from the CE7 hybridomas of gamma 2b and gamma 1 isotypes were virtually inactive in ADCC assay. Pre-exposure of macrophages to rIFN-gamma enhanced their ADCC activity, a result that is compatible with the notion that the high affinity Fc IgG receptor (FcR gamma I/CD64) is involved in the triggering of ADCC in macrophages. In contrast to macrophages, mononuclear cells, nonadherent cells and monocytes displayed considerable non-specific lytic activity, which was little influenced by the presence of antibody regardless of the isotype added.     

Peripheral T cells of patients with B cell non-Hodgkin's lymphoma show a shift in their memory status

BACKGROUND: Tumor-infiltrating T cells have a positive influence on the clinical course of B cell non-Hodgkin's lymphoma (NHL). T cells in the peripheral blood of patients with B cell NHL, however, have so far rarely been examined. METHODS: Using flow cytometry we examined lymphocyte subpopulations and numbers of na?ve/memory T cell subtypes among peripheral T cells of patients with B cell NHL (N=22), patients with metastasized solid tumors (N=27), and healthy controls (N=20). In addition, we analyzed the intracellular content of effector molecules granzyme B and perforin and expression of the T cell receptor zeta chain. RESULTS: We observed increased percentages of potentially highly cytotoxic CD8+CD56+ T cells in the peripheral blood of patients with NHL. Both, patients with NHL and patients with solid tumors showed a much higher expression of the chemokine receptors CCR4 and CCR5 on their T cells than healthy controls, suggesting a polarization of their T cells following stimulation with antigen and/or cytokines in vivo. Furthermore, patients with B cell NHL and patients with solid tumors had far lower percentages of na?ve CD45RA+CCR7+ T cells than healthy controls and, in the case of CD4+ T cells, patients with solid tumors. In contrast, patients with B cell NHL showed markedly increased levels of memory effector CD45RA-CCR7- CD4(+) T cells when compared to healthy controls and patients with metastasized solid tumors. Patients with NHL also showed elevated levels granzyme B within CD8(+) T cells, indicating that the increase in memory effector cells was of functional relevance. CONCLUSIONS: These findings indicate a marked shift in the composition of peripheral T cells of patients with B cell NHL from na?ve to memory effector-type cells.     

Chemokine C receptor 7 expression and protection of circulating CD8+ T lymphocytes from apoptosis.

Chemokine C receptor 7 (CCR7) expression is important for lymphocyte homing to tissues. We hypothesized that CCR7 also plays a role in CD8(+) T-cell protection from apoptosis. Its expression was determined on circulating T cells in patients with cancer and related to that of molecules responsible for lymphocyte susceptibility/resistance to apoptosis. Peripheral blood mononuclear cells were obtained from 36 patients with squamous cell carcinoma of the head and neck and 16 normal controls. Multicolor flow cytometry was used to evaluate CCR7, Fas, Bax, and Bcl-2 expression in CD8(+) T cells. Annexin V binding to CD8(+)CCR7(+) and CD8(+)CCR7(-) T-cell subsets was compared. Fewer CD8(+)CCR7(+) T cells bound Annexin V than CD8(+)CCR7(-) T cells in normal control and patients (P < 0.0001). CCR7 expression correlated with higher Bcl-2 but lower Bax and Fas expression levels in CD8(+) T cells in both normal control and patients (P < 0.0001). In patients, the CD8(+)CCR7(+) subset was reduced relative to normal control (P = 0.008) and replaced with an excess of apoptosis-sensitive CD8(+)CCR7(-) T cells. To study CCR7 signaling, CD8(+) T cells were stimulated with CCR7 ligands, chemokine C ligands 19 or 21. Ligand binding to CCR7 resulted in phosphorylation of Akt and increased Bcl-2 expression in CD8(+)CCR7(+) T cells, suggesting that CCR7 protects effector T cells from apoptosis through the phosphatidylinositol 3-kinase/Akt pathway. The absence of CCR7 expression on the majority of CD8(+) T cells in the peripheral circulation of patients with squamous cell carcinoma of the head and neck contributes to apoptosis and a rapid turnover of these effector cells.     

Targeting CD40 in Waldenström's Macroglobulinemia

CD40 is a member of the tumor necrosis factor (TNF) receptor family and is expressed in a majority of B-cell malignancies. In Waldenström's macroglobulinemia (WM), CD40 expression is a common feature of bone marrow infiltrating lymphoplasmacytic cells, and preclinical evidence suggests that CD40 signaling is functionally important for WM growth and survival. Two antibodies targeting CD40 (SGN-40 and HCD 122 [Chir 12.12]) are currently undergoing clinical testing in multiple myeloma (MM) and non-Hodgkin lymphoma (NHL). HCD122 is a novel, fully human, IgG1 antagonistic monoclonal antibody while SGN-40 is a humanized IgG1 partial agonistic antibody. Both agents have demonstrated activity in preclinical models and are potent mediators of antibodydependent cellular cytotoxicity (ADCC). Clinically, phase I data suggest both agents are well tolerated with no immunogenicity and have early evidence of single-agent clinical activity in relapsed and refractory NHL and MM. These observations support the testing of CD40-targeted agents in WM.     

Epratuzumab in non-hodgkin’ lymphomas

Opinion statement Non-Hodgkin’s lymphomas (NHL) are a diverse group of malignancies that result, in addition to their treatments, in significant morbidity and mortality in the population. The identification of more effective and better tolerated treatments is of vital importance. Immunotherapy using monoclonal antibodies directed against the CD20 protein has had a profound impact on the management of patients with B-cell NHL. Additional antigen targets are being aggressively investigated. The targeting of lymphoma cells with monoclonal antibodies offers a treatment with a potentially noncross-resistant mechanism of action and a more favorable toxicity profile compared to chemotherapy. Epratuzumab (humanized LL2) is a humanized immunoglobulin G1 monoclonal antibody directed against the CD22 protein in which expression is restricted to mature B cells. Because of important differences in the characteristics of CD22 compared to CD20, epratuzumab may become an important component of future therapies for NHL.     

Signals mediated by FcgammaRIIA suppress the growth of B-lineage acute lymphoblastic leukemia cells.

We examined Fc receptor expression and function in normal and leukemic human immature B cells. Fc receptor expression increased with normal B cell maturation: CD32(+) cells composed 8.1% +/- 1.2% (mean +/- s.d.) of the least mature (CD34(+)CD10(+)), 19.2% +/- 5.7% of intermediate (CD34(-)CD10(+)), and 82.4% +/- 5.0% of mature (CD34(-)CD10(-)) bone marrow CD19(+) B cells. Forty-five of 57 primary B-lineage acute lymphoblastic leukemia samples and all six cell lines studied expressed Fc receptors. By RT-PCR and antibody staining, FcgammaRIIA was the Fc receptor predominantly expressed in these cells. FcgammaRIIA ligation in RS4;11 and 380 cells induced tyrosine phosphorylation of CD32, CD19, CBL, SYK, P13-K p85 and SHIP, as well as RasGAP association with tyrosine-phosphorylated p62(dok). These signalling events resulted in a marked suppression of leukemia cell growth. After a 7-day exposure to anti-CD32, the recovery of ALL cells cocultured with stroma was reduced to 5.5% +/- 2.8% of control values in 380 cells (n = 14), 19.4% +/- 6.1% (n = 8) in RS4;11, and 4.0% +/- 1.3% (n = 6) in KOPN55bi. CD32 ligation also reduced cell recovery in five of seven CD32(+) primary leukemia samples. Thus, FcgammaRIIA mediates signals that suppress the growth of lymphoid leukemia cells.     

Identification of glioma-associated antigen MUC 2-63 as CD44.

Monoclonal antibody MUC 2-63 recognises neurogenic tumours and has been used successfully for radioimaging human malignant gliomas. We now show that the MUC 2-63 antigen has the same tissue distribution and molecular weight range as the CD44 antigen and confirm the identity of these two molecules in blocking studies using MUC 2-63 and the CD44 anti-framework antibody F10-44-2. Thus not only MUC 2-63 but also other anti-CD44 monoclonal antibodies should prove useful in imaging and, perhaps, therapy of brain tumours.     

Therapy of B-cell lymphoma with anti-CD20 antibodies can result in the loss of CD20 antigen expression.

Rituximab is a chimeric antibody with human gamma-1 and kappa constant regions and murine variable regions. It recognizes the CD20 antigen, a pan B-cell marker. Therapeutic trials in patients with B-cell non-Hodgkin's lymphoma (NHL) have shown significant efficacy with a primary response rate of 50%, and a secondary response rate of 44% after repeat treatments in prior responders. The selection for proliferating tumor cells that no longer express CD20 may compromise repeated treatment. We have identified a patient who developed a transformed NHL that lost CD20 protein expression after two courses of therapy with rituximab. In a pretreatment lymph node biopsy, 83% of B cells (as defined by CD19 and surface immunoglobulin) expressed surface CD20. A biopsy from the recurrent tumor after two courses of rituximab revealed a diffuse large cell NHL where 0% of B cells expressed CD20 with no evidence of bound rituximab. Cytoplasmic staining showed no CD20 protein. Sequencing of immunoglobulin heavy chain cDNA identified identical variable sequences in the initial and recurrent lymphomas, confirming the association between the two tumors. Literature and database review suggests that approximately 98% of diffuse large cell lymphomas express CD20, which suggests that these tumors rarely survive without CD20. This is the first identified case of loss of CD20 expression in a lymphoma that has relapsed after rituximab therapy, although several other cases have since been identified. Considering the significant number of patients treated with anti-CD20 antibodies, this may occur only rarely and is unlikely to preclude recurrent therapy with anti-CD20 antibodies in the majority of patients. However, because many patients have relapsed after anti-CD20 antibody therapy and have not been biopsied to identify clones with down-regulated CD20 antigen, we do not currently know the true frequency of this phenomenon. When possible, patients should undergo evaluation for CD20 expression before repeated courses of anti-CD20 therapy.