Search

Abstract 1847 CD37 is a tetraspanin transmembrane family protein that is strongly expressed on the surface of mature human B cells and transformed mature B cell lymphoma and leukemia cells, including CLL cells. It is absent or minimally expressed on normal T cells, NK cells, monocytes, and granulocytes. Predominant expression of CD37 on CLL cells makes it an ideal candidate to target with potential agents for treatment of CLL. TRU-016, a small modular immunopharmaceutical protein (SMIP) that specifically binds to an extracellular region of CD37, is presently in clinical trials in CLL patients. TRU-016 includes humanized immunoglobulin variable regions (scFv) fused to a human IgG1 Fc region. We have previously reported that SMIP-016, the chimeric version of the humanized TRU-016, induced apoptosis in CLL B cells in the presence of goat anti-human Fc antibody cross-linker through a novel, caspase-independent pathway. Furthermore, SMIP-016 showed potent in vivo activity in a SCID xenograft mouse model. Aside from direct cytotoxicity, SMIP-016 mediates antibody-dependent cellular cytotoxicity (ADCC) by NK cells both in vitro and in vivo. In an attempt to enhance its ADCC function, a new variant of SMIP-016, SMIP-016GV, was designed with a modification of the glycosylation of the Fc portion of the molecule. SMIP-016GV exhibits enhanced binding to both low- and high-affinity molecular variants of human CD16 (FcγRIII) and augments ADCC potency when compared to SMIP-016. In this study, we compared SMIP-016GV and SMIP-016 in direct cytotoxicity and ADCC against CLL B cells. While SMIP-016 and SMIP-016GV mediated comparable direct cytotoxicity at 48 hrs in the presence of goat anti-human Fc crosslinker, the SMIP-016GV resulted in 2 to 4 fold increase in NK cell mediated ADCC function at all effector to target ratios tested. This increased ADCC with SMIP-016GV was observed using NK cell effectors derived from both normal as well as CLL-affected individuals. In addition, this enhanced cytotoxicity was sustained at concentrations of SMIP-016GV as low at 5E-6 μg/ml. These low concentrations of SMIP-016GV were also able to mediate superior ADCC in 697 cells expressing as few as 10,000 molecules of surface CD37 antigen. Furthermore, NK cells stimulated with the glycovariant were potently activated and released 3 to 4 fold more IFNγ compared to SMIP-016. Ongoing studies are aimed at defining other effector cells which may interact with SMIP-016GV via different Fcγ Receptors. Collectively, these results suggest potential use of the SMIP-016GV with enhanced ADCC function as an alternate for TRU-016 in B cell malignancies including CLL therapy. Disclosures: Siadak: Trubion Pharmaceuticals: Employment.

Feline leukaemia viruses (FeLVs) are classified into subgroups A, B and C by their use of different host cell receptors on feline cells, a phenotype which is determined by the viral envelope. FeLV-A is the ubiquitous, highly infectious form of FeLV, and FeLV-C isolates are rare variants which are invariably isolated along with FeLV-A. The FeLV-C isolates share the capacity to induce acute non-regenerative anaemia and the prototype, FeLV-C/Sarma, has strongly age-restricted infectivity for cats. The FeLV-C/Sarma env sequence is closely related to that of common, weakly pathogenic FeLV-A isolates. We now show by construction of chimeric viruses that the receptor specificity of FeLV-A/Glasgow-1 virus can be converted to that of FeLV-C by exchange of a single env variable domain, Vr1, which differs by a three codon deletion and nine adjacent substitutions. Attempts to dissect this region further by directed mutagenesis resulted in disabled proviruses. Sequence analysis of independent natural FeLV-C isolates showed that they have unique Vr1 sequences which are distinct from the conserved FeLV-A pattern. The chimeric viruses which acquired the host range and subgroup properties of FeLV-C retained certain FeLV-A-like properties in that they were non-cytopathogenic in 3201B feline T cells and readily induced viraemia in weanling animals. They also induced a profound anaemia in neonates which had a more prolonged course than that induced by FeLV-C/Sarma and which was macrocytic rather than non-regenerative in nature. Although receptor specificity and a major determinant of pathogenicity segregate with Vr1, it appears that sequences elsewhere in the genome influence infectivity and pathogenicity independently of the subgroup phenotype.

Abstract 1744 Poster Board I-770 CD37 is a tetraspanin transmembrane family protein that is strongly expressed on the surface of mature human B-cells and transformed mature B-cell lymphoma and leukemia cells, including CLL cells. It is expressed minimally or is absent on normal T-cells, natural killer cells, monocytes, and granulocytes. Predominant expression of CD37 on CLL cells makes it an ideal candidate to target with potential agents for treatment of CLL. TRU-016, a Small Modular ImmunoPharmaceutical protein (SMIP) targeted towards the extracellular region of CD37, is presently in clinical trials in CLL patients. TRU-016 consists of variable regions (scFv) and engineered constant regions encoding the human IgG1 domains. We have previously reported that SMIP-016, the chimeric precursor of the fully humanized TRU-016, induced apoptosis in CLL B cells in the presence of goat anti-human Fc ab cross-linker through a novel, caspase-independent pathway. Furthermore, SMIP-016 showed potent in-vivo activity in a SCID xenograft mouse model. Aside from direct cytotoxicity, SMIP-016 mediates antibody-dependent cellular cytotoxicity (ADCC) by NK cells both in vitro and in vivo. Recently, in an attempt to enhance the ADCC function, a new variant of SMIP-016, Tru-ADhanCe SMIP-016, has been created with a modification of the glycosylation of the Fc portion of the molecule. TRU-ADhanCe SMIP-016 has been shown to exhibit enhanced binding to both low- and high-affinity molecular variants of human CD16 (FcRIII) and augmented ADCC potency when compared to SMIP-016. In this study, we compared TruADhanCe SMIP-016 and SMIP-016 in direct cytotoxicity and ADCC experiments using CLL B-cells. While SMIP-016, and TruADhanCe SMIP-016 mediated comparable direct cytotoxicity at 24, 48 and 72 hrs in the presence of anti-human Fc crosslinker, the TruADhanCe SMIP-016 resulted in 2 to 4 fold increased NK cell mediated ADCC function. Consistent with the comparable direct cytotoxic effects, the early phosphorylation patterns were similar in cells treated with TruADhanCe SMIP-016 or SMIP-016 in the presence of anti-human Fc cross linker. Ongoing studies are aimed to define the mechanistic basis of the enhanced ADCC function by TruADhanCe SMIP-016 and to determine if use of soluble CD16.Fc as a cross-linker, an in vitro model of in vivo Fc receptor binding, may reveal enhanced apoptotic-signaling of TruADhanCe SMIP-016. These results suggest potential use of TruADhanCe versions of TRU-016 with enhanced ADCC function as an alternate for TRU-016 in B cell malignancies including CLL therapy. [This work was supported by D. Warren Brown Foundation, Leukemia and Lymphoma Society and National Cancer Institute.] Disclosures Thompson: Trubion Pharmaceuticals: Employment. Siadak:Trubion Pharmaceuticals: Employment. Algate:Trubion Pharmaceuticals: Employment. Cerveny:Trubion Pharmaceuticals: Employment.

Abstract 3725 Poster Board III-661 CD19 is a lineage-specific B-cell antigen, expressed at a high density on CLL cells, that contributes to B-cell receptor signaling but to date has not been effectively targeted with therapeutic monoclonal antibodies. XmAb5574 is a novel engineered anti-CD19 monoclonal antibody with a modified Fc-domain designed to enhance binding of FcγRIIIa that is predominately expressed on Natural Killer (NK)-cells. Utilizing freshly isolated chronic lymphocytic leukemia (CLL) patient B-cells we demonstrate that XmAb5574 lacks significant internalization seen with other anti-CD19 antibodies [maximum internalization for XmAb5574 was only 27.9% at 30-minutes (95%CI 14.5%, 41.4%)], thereby enhancing its ability to induce potent antibody-dependent cellular cytotoxicity (ADCC). Annexin V/PI flow cytometry analysis revealed that XmAb5574 mediates modest direct cytotoxicity not significantly different from Rituximab (0.6% increase, 95%CI -10.5%, 11.7%, p=0.91), and no complement mediated cytotoxicity (CDC) against primary CLL B-cells. Multi-color flow cytometry and monocyte derived macrophages (MDM) were used to assess XmAb5574 antibody dependent cellular phagocytosis (ADCP) against CLL cells and revealed no significant impact of the Fc-domain modification on MDM induced ADCP against CLL cells as compared to the wild type parental anti-CD19 antibody (12.37% vs. 10.51%, p=0.58). Interestingly, utilizing NK-cells and CLL cells isolated from normal donors and CLL patients, and employing autologous and allogeneic effector-target (E:T) conditions, XmAb5574 was found to mediate significantly higher ADCC when compared to the control humanized anti-CD19 non-engineered antibody (26.9% higher at E:T 25:1, p=0.0004 for allogeneic conditions, and 23.6% higher, p=0.004 for autologous conditions). ADCC mediated by XmAb5574 was also significantly higher as compared to Rituximab (33.5% higher at E:T 25:1, p Disclosures: Desjarlais: Xencor: Employment.

7094 Chemoimmunotherapy with fludarabine and rituximab has demonstrated promising clinical activity in CLL. We hypothesized that the IL-21 might enhance the efficacy of this regimen and play a role in immune mediated clearance of CLL cells. Expression of IL21R-a was present (>20% cells +) in 13 of 16 (81%) pts. IL-21 induced direct cytotoxicity of CLL cells (>20% apoptosis) in subset of patients (n=9, 40% ± 15.9, ptyr- 701 and STAT-3tyr-705 was seen in cells with apoptosis >20% whereas phosphorylation of STAT-1tyr-701 was not seen in cells failing to undergo apoptosis. In contrast to this, responder and non responder cells exhibited comparable levels of STAT-1, STAT-3 and constitutive Ser727 phosphorylated STAT-3 levels. Further, IL-21 induced upregulation of Bim was observed in the cells undergoing apoptosis. Interestingly, IL-21 enhanced additively, the direct apoptosis by either antibody (rituximab or alemtuzumab) or fludarabine therapy (Pretreatment of CLL cells with IL-21 for 18–20 hours). In contrast to the enhanced fludarabine mediated direct cytotoxic death of B cells, IL-21 failed to enhance the fludarabine mediated death of T cells (n=5, % alive 56±15 vs 58±16 with IL-21) suggesting that this immunotherapeutic strategy will not augment the immunosuppressive effect of fludarabine. IL-21 also enhanced innate immune activation as demonstrated by killer inhibitory receptor matched autologous natural killer cell mediated antibody dependent cellular cytotoxicity against rituximab coated CLL cells by 33% (E: T ratio 25:1, 42% ± 4.4 versus 28% ± 3.1, p No significant financial relationships to disclose.

CD37 is a lineage-specific B-cell antigen that represents an attractive target for immunotherapy in B-cell malignancies, especially in chronic lymphocytic leukemia (CLL) and non-Hodgkin’s lymphoma (NHL). CD37-specific small modular immuno pharmaceutical (CD37-SMIP™) drug is an engineered protein therapeutic directed to the CD37 antigen using a single chain variable region (scFv) linked to a modified human IgG1 hinge, CH2, and CH3 domains. We have previously presented that CD37-SMIP™ drug induces both ADCC and apoptosis against primary CLL cells and B-cell lymphoma cells and therapeutic efficacy was observed in a Raji cell disseminated leukemia xenograft model. Herein, we sought to determine the effector cell type(s) mediating ADCC and explore if agents that deplete NK cell inhibitory T-regulatory cells influence CD37-SMIP™ efficacy. Immunostaining of human peripheral blood mononuclear cells (PBMCs) from CLL patients demonstrated no expression of CD37 on CD3+ T cells, CD16+ or CD56+ NK cells, or CD64+ monocytes whereas CD37 was highly expressed on CD19+ B cells. Using purified human PBMCs as effector cells and Cr-51 labelled CLL B cells as targets, we found the CD37-SMIP™ dependent ADCC was predominantly mediated by NK cells, but not naïve or activated monocytes. Consistent with these in-vitro results, the in-vivo therapeutic efficacy of CD37-SMIP™ drug was significantly compromised by depletion of NK cells in the Raji cell disseminated leukemia xenograft SCID mouse model. The median survival time of CD37-SMIP™ treated mice decreased from 51 days (95% CI: 38, 78) to 27 days (95% CI: 25, 37) (p=0.017) with the depletion of NK cells. Consistent with previous studies, ADCC is not diminished by fludarabine, an agent that may deplete T-regulatory cells, suggesting that the anti-CD37 protein and fludarabine might combine for increased efficacy in-vivo. We therefore examined the effect of fludarabine on CD37-SMIP™ in-vitro apoptotic activity. These data demonstrate that direct cell death mediated by CD37-SMIP™ drug (5 ug/mL) synergizes with fludarabine-induced caspase-dependent apoptosis, as measured by both the MTT assay and annexin V/PI staining of CLL cells (combination index, CI=0.44). Overall, these data suggest that the CD37-SMIP™ is a promising therapeutic agent against CD37+ B-cell malignancies as either monotherapy or in combination with fludarabine. Further clinical development is warranted to investigate the effect of CD37-SMIP™ drug in CLL and NHL. (SMIP trademark is owned by Trubion Pharmaceuticals).

Background: The monoclonal anti-CD20 antibody rituximab induces the death of chronic lymphocytic leukemia (CLL) in part through apoptosis. Tumor necrosis factor (TNF)-α and other cytokines up-regulate bcl-2 and other anti-apoptotic proteins, thereby inhibiting apoptosis. In addition, TNF-α may play a role in the cytokine release syndrome and infusion toxicity associated with monoclonal antibody therapy. Therefore, antagonists of TNF-α represent potential therapeutic agents that may decrease infusion toxicity and enhance rituximab-induced apoptosis. Methods: We conducted a phase II trial of the TNF-α inhibitor etanercept in combination with rituximab in patients (pts) with relapsed CLL or small lymphocytic lymphoma (SLL), to determine if inhibition of TNF-α by etanercept increases the clinical activity of rituximab and reduces infusion toxicity. Pts received etanercept, 25 mg subcutaneously (SC) twice weekly during weeks 1–5, and rituximab, 375 mg/m2 intravenously (IV) three times weekly during weeks 2–5. Stepped up dosing was utilized with the first two rituximab treatments. Results: Thirty-six pts (25 male) with a median of 2 prior therapies (range 1–8) were enrolled; 23 pts had previously received rituximab. Therapy was generally well tolerated. Infusion reactions were mild and were not associated with severe adverse events. Fourteen pts experienced grade 1 (n=5) or grade 2 (n=9) infusion reactions. These infusion reactions extinguished with subsequent does of rituximab. Twelve pts developed grade 4 neutropenia, and neutropenic infections occurred in 4 pts, with 3 of these infections occurring during rapid CLL progression. Twelve pts developed grade 2 (n=10) or grade 3 (n=2) non-neutropenic infections. Nine of 32 evaluable pts (28%) achieved an NCI 96 response, with 8 partial responses (PR) and 1 complete response (CR). Median duration of response was 10 months. Eighteen pts had stable disease (SD), while 5 pts had progressive disease (PD). No responses were observed in 8 pts with del(17p13). In contrast, 9 of 24 pts (38%) without del(17p13) responded, including 3 pts with complex karyotypes. Sixteen pts with responsive or stable disease did not require additional treatment for CLL for 6 months after end of therapy, and the median treatment free survival for this group was 12 months. Eight of these 16 pts have not yet required other CLL therapy. Conclusions: This study confirms our previous finding that single agent thrice weekly rituximab is ineffective in relapsed CLL with del(17p13) (Cancer Res 63: 36, 2003). The combination of rituximab and etanercept was well tolerated and demonstrated clinical activity in relapsed CLL pts without del(17p13). However, the addition of etanercept did not improve the clinical response rate beyond that expected with single agent thrice weekly rituximab.

The CD40 antigen is involved in cell survival and differentiation of B-cells and is uniformly expressed on chronic lymphocytic leukemia (CLL) cells. The CD40/CD40L interaction stimulates B-cells, dendritic cells and monocytes to proliferate, differentiate, up regulate co-stimulatory molecules and increase antigen presentation. While activation of CD40 can protect CLL cells against early fludarabine-induced apoptosis, these cells become sensitive to delayed death by extrinsic pathway apoptosis. (Blood, 105: 3193–8, 2005). SGN-40 is a humanized anti-CD40 antibody entering clinical trials and has been reported to have weak agonistic properties following CD40 ligation. To pursue rational clinical development of SGN-40, we studied the effects of this antibody in fresh, non-cryopreserved primary CLL cells. These studies included classic antibody mediated killing mechanisms and evidence of both CLL cell activation and protection against early fludarabine-mediated apoptosis. CLL cells treated with SGN-40 (10 mcg/ml) for 2 hours (hrs) in the presence of human serum promoted no complement mediated cytoxicity (CDC) in 8 pts tested. Direct SGN-40 induced apoptosis of human CLL cells with or without anti-Fc IgG cross-linking at 24, 48 and 72 hrs was not increased over that observed with the isotype control antibody trastuzumab in 8 pts studied. In contrast, SGN-40 induced antibody dependent cellular cytotoxicity (ADCC) against CLL cells an average of 12% (±11.39 SD, range 2–32%) killing at 4 hrs (effector to target cell ratio 25:1) in 6 pts tested. The SGN-40 induced ADCC against CLL cells were similar to that observed with alemtuzumab (average 19%, SD 6.9, range10–30%) or rituximab (average 18%, SD 12.48, range 8–42.5%). SGN-40 also mediated death in Raji and 697 lymphoblastic lymphoma cell lines via ADCC. Similar to reports by others with CD40 ligand, SGN-40 mediated activation was noted with modest up-regulation of CD80 and HLA-DR at 48hrs. When administered prior to fludarabine, SGN-40 also protected against death in 5 consecutive samples, although this was less than observed with CD40 ligand transfected HeLa cells, consistent with incomplete CD40 activation. Concurrent administration of SGN-40 and fludarabine did not protect from drug-mediated apoptosis. In conclusion, these findings suggest that SGN-40 has dual property of mediating cytotoxic effect by ADCC and partial CD40 activation. Development of SGN-40 as a therapeutic agent in CLL is justified and future studies by our group are focusing on enhancing SGN-40 mediated ADCC against CLL cells and potentially designing combination studies with SGN-40 to exploit this agent’s ability to engage the CD40/CD40L network.

Tru 16.4 is a single chain protein with a modified CD37 binding Fv domain linked to a modified human IgG1 hinge, CH2 and CH3 domains. It is a member of a novel composition class called small modular immunopharmaceuticals (SMIP) construct. CD37 is a stable marker for B-cell chronic lymphocytic leukemia (BCLL) and non-Hodgkins lymphoma (NHL). Although radiolabeled murine anti-CD37 antibody MB-1 has been evaluated in NHL patients and shown to be able to target tumors, efficacy was not established. There has been no evaluation of an unconjugated anti-CD37 antibody in human clinical trials. Here we report that Tru 16.4, binds to CD37 on primary CLL cells surface, and induces caspase-independent apoptosis that is further enhanced by a secondary crosslinking with goat anti-human antibody. The percentage of dead cell population over control based upon Annexin V/PI staining of 12 BCLL patient samples (28.3%, SD 15.4%) is significantly greater than that of isotype control, trastuzumab (3.4%, SD 7.6%). The expression of CD37 on T cells is low, and the exposure of Tru 16.4 under the same conditions does not demonstrate significant binding or induce noticeable apoptosis on normal T cells at 24 hours (0.47%, SD 0.99%). This provides a selective advantage for BCLL antibody therapy. Results from a comparison study utilizing 6 additional patient samples demonstrated that Tru 16.4 induced apoptosis (31.8%, SD 12.2%) is significantly higher than that of the anti-CD20 antibody, rituximab (21.8%, SD 5.25%) and similar to that observed with the anti-CD52 antibody, alemtuzumab (36.6%, SD 19.5%). The cell death induced by Tru 16.4 is dependent upon dose and duration of treatment, indicating 5μg/mL as a saturation concentration and 24 hours as the optimal time for maximal in vitro apoptosis. The degree of CLL cell death was found to be directly proportionate to cell surface membrane expression of CD37, based on nonparametric Spearman rank correlation analysis of the mean fluoresecnce intensity (MFI) after PE-labelled anti-CD37 antibody immunostaining and percentages of Annexin V/PI positive cells, with a correlation factor rs = 0.80 ( P < 0.01). Ongoing studies suggest Tru 16.4 induces intracellular signaling following CD37 antigen ligation and current efforts are focused on identifying the specific death pathways mediated by this treatment. Overall, these findings indicate that Tru 16.4 mediates B-cell specific apoptosis through a novel caspase independent mechanism and supports effort for further clinical investigation in CLL.

Campath-1H is a humanized IgG1 kappa antibody directed against the cell surface antigen CD52, a GPI linked protein of unknown function that is present on B and T lymphocytes as well as CLL cells. Campath-1H is approved for fludarabine refractory CLL and ongoing studies will continue to define its role in the treatment of CLL. Despite its clinical use, little is known about the in vivo mechanism of its cytotoxicity. Mechanisms invoking complement, effector cells and direct antibody- induced cytotoxicity have been proposed. Here we show that campath-1H is directly toxic to human CLL cells in vitro in the absence of complement or effector cells. Furthermore, we observed that normal T-cells isolated from random blood donors are less sensitive to direct campath-1H induced cytotoxicity than CLL cells. Using annexin V-FITC/propidium iodide (PI) staining with FACS analysis, we found that treatment of CLL samples for 4 hrs with campath-1H (10μg/ml) or campath-1H with a cross-linking anti-Fc IgG (10μg/ml) leads to positive staining in 9.5 % and 29% of the cell respectively (n=11). Parallel study using the MTT assay showed a 7.9 % and 20.4 % loss of viability with campath-1H or campath-1H + anti-Fc IgG (n=5). Antibody directed cellular cytoxicity, using the chromium release assay, showed 20 % killing at 4 hrs (effector to target cell ratio 25:1, n = 4). After treating primary CLL cells with campath-1H for 4 hrs in the presence of human complement, 67 % of the cells stained positive with annexin V-FITC/PI (n=5). Direct campath-1H induced killing of normal T-cells was reduced compared to CLL cells. After a 4 hr treatment with campath-1H + anti-Fc IgG, 6.7 % of the cells stained positive with annexin V-FITC/PI (n=5). The direct campath-1H cytoxicity we observed in CLL cells is caspase-independent. Immunoblot analysis showed no evidence of caspase-3, 8, or 9 activation or PARP cleavage, nor was there a reduction in cytotoxicity with the addition of Z-VAD-fmk. Using minimal detergent solubilization of CLL cells followed by density-gradient centrifugation and immunoblot analysis, we found that CD52 colocalizes with the glycosphingolipid GM1. Fluorescence microscopy analysis revealed that upon cross-linking CD52 with campath-1H + anti-Fc IgG, large patches, and in many cases caps, enriched in CD52 and GM1 formed on the plasma membrane of primary CLL cells. Cholesterol depletion of lymphocyte membranes with methyl-β-cyclodextran diminished the formation of these CD52 and GM1 enriched patches and caps. Inhibition of actin polymerization with cytochalasin B also reduced the formation of campath-1H + anti-Fc IgG induced caps. The cytotoxic effect of cross-linked campath-1H upon CLL cells was reduced by methyl-β-cyclodextran or cytochalasin B pretreatment as measured by annexin V-FITC/PI staining and MTT assay analysis. In conclusion, we found that campath-1H treatment of CLL cells leads to rapid cell death through a caspase-independent mechanism in the absence of complement or effector cells. Experiments using methyl-β-cyclodextran or cytochalasin B treated cells, suggest that the direct cytotoxicity of campath-1H is dependent on formation of large antigen/lipid raft aggregates upon the cell membrane, with underlying cytoskeletal rearrangement. The difference in direct killing observed between CLL cells and normal T-cells suggests that death mechanisms may differ significantly and thus provide the opportunity to refine CD52 directed antibodies with the aim of reducing the suppression of cellular immunity.

The monoclonal anti-CD20 antibody rituximab (Rituxan) induces the death of chronic lymphocytic leukemia (CLL) cells in part through apoptosis. Tumor necrosis factor (TNF)-α and other cytokines up-regulate bcl-2 and other anti-apoptotic proteins, thereby inhibiting apoptosis. In addition, TNF-α may play a role in the cytokine release syndrome and infusion toxicity associated with monoclonal antibody therapy. Therefore, antagonists of TNF-α represent potential therapeutic agents that may be able to decrease infusion toxicity and enhance rituximab-induced apoptosis. We initiated a phase II trial of etanercept in combination with rituximab in patients (pts) with relapsed CLL or small lymphocytic lymphoma (SLL), with the goals of increasing clinical activity of rituximab and reducing infusion toxicity through neutralization of TNF-α. Pts received etanercept 25 mg subcutaneously (SC) twice weekly during weeks 1–5 and rituximab 375 mg/m2 intravenously (IV) three times weekly during weeks 2–5. Stepped up dosing of rituximab was used with the first two rituximab treatments. From 12/02 to 10/03, 19 pts (14 male) with a median of 2 prior therapies (range 1–8) were enrolled. Therapy was generally well tolerated, although 5 pts developed grade 4 neutropenia and 5 pts developed grade 3–4 infection or neutropenic fever. Four of 19 pts achieved an NCI 96 partial response, with no responses in 8 pts with del(17p13.1), corresponding to loss of the p53 gene. In contrast, 4 of 11 pts without del(17p13.1) achieved a PR, including 1 pt with a complex karyotype. This study was stopped for interim analysis and has reopened, with the exclusion of patients with a del(17p13.1) based upon the response outcome above and a previous preliminary report by our group demonstrating that rituximab was ineffective for del(17p13.1) CLL (Cancer Res63:36, 2003). We conclude that the combination of rituximab and etanercept has clinical activity in heavily treated, relapsed CLL pts without del(17p13.1), irrespective of prior rituximab treatment. Accrual to this study continues, and the results will be updated at the 2004 ASH meeting.

The direct effect of the feline leukaemia virus (FeLV) on erythroid colony formation in vitro was investigated. Bone marrow mononuclear cells (BMMC) from FeLV-naive, specificpathogen-free (SPF), adult cats were inoculated with FeLVs of characterized strains and biologically cloned subgroups and the subsequent development of colony forming units-erythroid (CFU E ) and burst forming units-erythroid (BFU E ) and colony forming units-granulocyte-macrophage (CFU GM ) was monitored. Exposure to the anaemia-causing Kawakami-Theilen strain of FeLV (FeLV-KT), a phenotypic mixture of subgroups A, B, and C, caused constant depression of day 2 CFU E (to 47% of sham-inoculated controls), day 4 CFU E (41% of controls), and day 10 BFU E (38% of controls). CFU GM were unaffected. The lymphoma-causing Rickard strain of FeLV (FeLV-R-TL) caused sporadic depression of CFU E and BFU E . In contrast, neither FeLV-R passaged through feline embryonic kidney fibroblasts (FeLV-R-CRFK) nor biologically cloned, subgroup-specific, FeLVs of fibroblast origin, caused decrements in CFU E or BFU E , suggesting that fibroblast passage attenuated the direct erythrosuppressive effect of FeLV. Suppression of CFU E and BFU E by lymphoma cell-origin FeLV was dependent on infectious virus and was associated with FeLV replication by the cultured myelomonocytic precursor cells. Attenuation of infectivity by heat or u.v. restored CFU E and BFU E development. Examination of the relationship between viral infectivity (VI), viral protein concentration, and CFU E suppression showed that the infectious FeLV was 20-fold more effective than u.v.-inactivated FeLV as an inhibitor of erythrogenesis in vitro .