Superior Antitumor Activity of SAR3419 to Rituximab in Xenograft Models for Non-Hodgkin's Lymphoma Ayad M

Cancer Therapy: Preclinical

Superior Antitumor Activity of SAR3419 to Rituximab in Xenograft Models for Non-Hodgkin's Lymphoma Ayad M. Al-Katib,1 Amro Aboukameel,1 Ramzi Mohammad,1 Marie-Christine Bissery,2 and Claudia Zuany-Amorim2

Abstract Purpose: To investigate the activity of SAR3419, a novel humanized anti-CD19 antibody (huB4), conjugated to a cytotoxic maytansine derivative N2'-deacetyl-N2'-(4-mercapto- 4-methyl-1-oxopentyl) maytansine, in preclinical xenograft models for non–Hodgkin's lymphoma. Experimental Design: Antitumor activity of SAR3419 was assessed as a single agent and in comparison with conventional therapies usinga subcutaneous model for diffuse large B-cell lymphoma (WSU-DLCL2) and a systemic model for follicular small cleaved cell lymphoma (WSU-FSCCL) in mice with severe combined immune deficiency. Results: Our results showed that in these chemotherapy-resistant models, SAR3419 was more effective than CHOP (cyclophosphamide-Adriamycin-vincristine-prednisone) regimen or rituximab. Only treatment with SAR3419 led to survival of the whole group of animals to the end of the experiment (150-155 days) in both models. Higher doses of SAR3419 (15 and 30 mg/kg) were more effective than lower dose of 7.5 mg/kg. The immunoconjugation was necessary because neither huB4 nor DM4 alone had significant activity. Treatment with rituximab resulted in antitumor activity in both models comparable with the low dose of SAR3419. Cyclophosphamide-Adriamycin-vincristine- prednisone alone showed modest activity in both models. Necropsy and tissue staining in the WSU-FSCCL systemic model revealed that all deaths featured leptomeningeal lymphoma in the control and treated groups. Interestingly, some of the animals that survived to the end of the experiment and seemed healthy at time of euthanasia did show microscopic evidence of lymphoma. Conclusions: Overall, SAR3419 is a very active immunotoxin in preclinical models for human B-cell lymphoma and holds promise as a novel and well-tolerated therapy in B-cell non–Hodgkin's lymphoma.

Non–Hodgkin's lymphoma (NHL) is the fifth most common 1970s and 1980s and 1% to 2% in the 1990s, resulting in al- cancer in the United States where >66,000 cases are expected to most doubling the NHL incidence (2). Although there are geo- be diagnosed in 2008 (1). An increasing incidence of NHL at a graphic variations in the distribution of different histologic rate of 3% to 4% per year was observed world-wide for the subtypes of NHL, the most common ones are the diffuse large B-cell and the follicular lymphomas (3). The addition of anti- CD20 antibody (rituximab) to cytotoxic chemotherapy has been the most notable improvement of NHL treatment since Authors' Affiliations: 1Lymphoma Research Laboratory, Wayne State University School of Medicine, Detroit, Michigan and 2Oncology the introduction of Adriamycin in the 1970s (4). Overall, the Therapeutic Department, sanofi-aventis, 94403 Vitry-sur-Seine, France 5- and 10-year relative survival of NHL has improved from Received 10/28/08; revised 3/3/09; accepted 3/22/09; published OnlineFirst 6/9/09. 50.4% and 39.4% (1990-1992) to 66.8% and 56.3% (2002- The costs of publication of this article were defrayed in part by the payment of 2004), respectively (5). Improvements were seen in all age page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. groups, in both sexes, in both nodal and extranodal disease, Part of the preclinical activity of huB4-DM4 has been reported in abstract and in both low-grade and high-grade lymphomas. There are form: Lutz RJ, Zuany-Amorim C, Vrignaud P, et al. Preclinical evaluation of a number of limitations to rituximab therapy, however. In the SAR3419 (huB4-DM4), an anti-CD19-maytansinoid immunoconjugate, for pivotal clinical trial that led to its approval for clinical use in the the treatment of B cell lymphomas. Proceedings of the 97th Annual United States, only 48% of patients with relapsed follicular Meetingof the American Association for Cancer Research; 2006 Apr 1 –5; Washington, DC. Philadelphia (PA): AACR; 2006 p. 877–878. Abstract lymphoma responded (6% complete and 42% partial re- number 3731. sponses; ref. 6). Moreover, almost 60% of initially responding Requests for reprints: Ayad M. Al-Katib, Lymphoma Research Laboratory, patients will not respond to subsequent therapy (7). Wayne State University School of Medicine, Scott Hall for Basic Medical The success (and limitations) of rituximab presents clear op- Sciences, Room #8229, 540 East Canfield, Detroit, MI 48202. Phone: – 1-313-577-0971; Fax: 1-313-577-0977; E-mail: [email protected]. portunity for exploring additional monoclonal antibody based F 2009 American Association for Cancer Research. therapy in lymphoma. To that end, different approaches are be- doi:10.1158/1078-0432.CCR-08-2808 ing explored including targeting different cell surface antigens,

Clin Cancer Res 2009;15(12) June 15, 20094038 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Anti-CD19 Immunotoxin in Lymphoma

tained in a humidified incubator at 37°C and 5% CO2. Culture medi- Translational Relevance um is changed every third day. Cells are checked periodically for immunophenotypic or karyotypic changes. The SAR 3419 reported in this article has already Xenograft models entered phase I clinical trials in B-cell non–Hodgkin's The two xenograft models used in the study were based on the WSU- lymphoma (NHL). There are such trials now open for DLCL2 and WSU-FSCCL cell lines grown in mice with severe combined accrual of patients with relapsed/refractory B-cell immune deficiency (SCID). ICR-SCID mice, age 3 to 4 wk, were pur- NHL: NCT00539682 and NCT00549185. Antibody chased from Taconic Farms, housed in special protective cages, and left therapy is proven to be the most significant advance- to adapt to the new environment for 1 wk before beginning the experiment in the treatment of NHL over the past 4 dec- ments. All in vivo studies were conducted according to Animal Investi- ades. For B-cell NHL, the only Food and Drug gation Committee–approved protocol by WSU. Administration–approved antibodies target CD20. WSU-DLCL2-SCID subcutaneous model This article describes an immunotoxin whereby a hu- 7 manized anti CD19 antibody is chemically linked to a For initial establishment of the xenograft, 10 WSU-DLCL2 cells in serum-free RPMI 1640 are injected s.c. bilaterally in each flank of 2 ICR- powerful antimicrotubule maytansinoid. SCID mice. Within 2 wk, animals developed palpable tumors at injection sites and were euthanized when tumor size reached 1,000 to 1,500 mg. Tumors were then dissected into 20 to 30 mg pieces that were re- conjugating antibodies with radioisotopes, chemotherapeutic transplanted into 56 naïve ICR-SCID mice s.c. in the flanks. Using this drugs, or toxins (reviewed in ref. 8). The most developed method, 100% of animals developed tumors. Tumor-bearing animals among these is the radio-immunotherapy where a radioisotope were randomly assigned to either control or a treatment group as is conjugated to a monoclonal antibody. Two anti-CD20 IgG- shown below. 90 radioconjugates, Y-ibritumomab tiuxetan (Zevalin) and WSU-FSCCL-SCID intravenous (systemic) model 131 I-tositumomab (Bexxar), have been approved by the U.S. 107 WSU-FSCCL cells in serum-free RPMI 1640 were injected i.v. via Food and Drug Administration for the treatment of relapsed/ tail vein into each of 56 ICR-SCID 3 to 4-wk-old female mice. Based on refractory, indolent, or transformed B-cell NHL. Several studies previous experience with this model (14), animals develop disseminat- have shown that radio-immunotherapy is well-tolerated, has ed disease involving bone marrow, liver, spleen, lymph nodes, and lep- the highest single-agent activity observed in lymphoma therapy, tomeningeal involvement of the central nervous system. There is no and can provide durable responses even in patients who had secondary propagation in this model. Animals injected with the failed previous treatments, including rituximab therapy (9). WSU-FSCCL were randomly assigned to control or a treatment group. Antibody-targeted chemotherapy is another therapeutic strat- Preclinical efficacy trial design egy that involves the use of a cytotoxic agent chemically linked to a monoclonal antibody recognizing a tumor-associated anti- Treatment of tumor-bearing animals in both models prepared as described above was started 1 wk after transplantation of tumors gen. The monoclonal antibody delivers the cytotoxic agent pref- (WSU-DLCL2) or i.v. injection of cells into mice (WSU-FSCCL). erentially to tumor cells that express the targeted antigen, hence There were 8 treatment groups, 7 animals in each treated as follows: limiting systemic toxicity. In this study, we investigated the pre- group 1, vehicle (5% glucose in water), every fourth day for 2 doses clinical efficacy of anti-CD19 humanized monoclonal antibody (q4dx2); group 2, SAR3419 at 30 mg/kg/injection, q4dx2; group 3, 2 (huB4) conjugated with a maytansine derivative [N '-deacetyl- SAR3419 at 15 mg/kg/injection, q4dx2; group 4, SAR3419 at 7.5 mg/ 2 N '-(4-mercapto-4-methyl-1-oxopentyl) maytansine, DM4] in kg/injection, q4dx2; group 5, free DM4 at 0.6 mg/kg/injection, two xenograft models representing human diffuse large B-cell q4dx2, which is comparable with the amount of DM4 detected in lymphoma and follicular lymphoma in comparison with ritux- the immunoconjugate SAR3419 at the high dose; group 6, naked imab and standard cytotoxic chemotherapy regimen (cyclo- huB4 at 30 mg/kg/injection, q4dx2; group 7, CHOP at maximum phosphamide-Adriamycin-vincristine-prednison, CHOP). tolerated doses as previously determined in our laboratory (cyclophosphamide, 40 mg/kg i.v.; doxorubicin, 3.3 mg/kg i.v.; vincristine, Maytansine is a highly cytotoxic natural product that has failed 0.5 mg/kg i.v.; and prednisone, 0.2 mg/kg orally daily for 5 d; all iv as an anticancer agent in human clinical trials because of unac- drugs were given via tail vein once on day 1; ref. 15); group 8, ritux- – ceptable systemic toxicity. The anti CD19-DM4 immunoconju- imab at 40 mg/kg/injection, q4dx2. This dose/schedule is believed to gate used in this study (SAR3419) involves a modified provide sustained blood level and saturate CD20 antigens of lym- maytansine(10)linkedtohuB4foratargeteddeliveryap- phoma cells. proach to CD19-positive B-cell lymphoma. Treatment agents CHOP components were obtained from commercial sources: Cyclo- Materials and Methods phosphamide USP from Bristol-Myers Squibb Co; Doxorubicin HCl from Bedford Laboratories; Vincristine sulfate from Mayne Pharma, Lymphoma cells Inc., and Prednisone from Schein Pharmaceutical, Inc. Rituximab was Two established cell lines were used in this study: WSU-DLCL2 and obtained from IDEC Pharmaceutical Corp. HuB4 is a humanized WSU-FSCCL. Both cell lines were established at Wayne State University mouse monoclonal antibody targeting CD19. DM4 is a cytotoxic com- (WSU), are EBV negative and were previously published (11, 12). The pound derived from maytansine (10). SAR3419 is huB4-DM4 immu- WSU-DLCL2 line is established from a patient with relapsed and resis- noconjugate (patent # US 2004/0235840, November 25, 2004). The tant diffuse large cell lymphoma. The WSU-FSCCL was established from ratio of DM4 to huB4 antibody is 3.4 moles of DM4 per mole of a patient with follicular lymphoma, grade I according to the WHO clas- huB4. The antibody is linked to DM4 by an optimized linker incorpo- sification [follicular small cleaved cell (FSCCL) by the Revised Europe- rating hindered disulfide bond [N-Succinimidyl-4-(2-pyridyldithio)bu- an American Lymphoma classification; ref. 13]. Both cell lines grow as tanoic acid (SPDB). Free drug (DM4) is <1% (by high performance suspension in RPMI 1640 supplemented with 10% fetal bovine serum liquid chromatography). The immunoconjugate is stable in plasma (FBS), 1% L-glutamine, penicillin, and streptomycin. Cultures are main- and in aqueous formulation and is cleavable inside cells. HuB4,

www.aacrjournals.org 4039 Clin Cancer Res 2009;15(12) June 15, 2009 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Cancer Therapy: Preclinical

DM4, and SAR3419 are provided by ImmunoGen, Inc., through a li- the tumor weights (LOGTUM) according to time as follows: LOGTUM cense agreement with sanofi aventis. Structure of SAR3419 is shown = slope* time + constant. Individual Td is calculated according to the in Fig. 1. following formula by dividing the base 10 logarithm of the 2 value by the slope estimated from the considered linear regression: Td = Log10 Animal monitoring (2)/slope. The control group Td is then calculated by taking the median Animals were observed for changes in weight, side effects of therapy, of the control group individual Td. measurement of subcutaneous tumor size (WSU-DLCL2 model), or T and C estimation. Defined as the time to reach 800 mg tumor signs of any sickness. They were euthanized when their tumor burden weight for the control (C) group and for the treated (T) group. Tumor ∼ ∼ reached 10% of the body weight, lost 20% of body weight, or be- weights were analyzed after Log10 transformation. T and C values were came moribund due to treatment toxicity or systemic disease in the case estimated for each animal by linear interpolation between the tumor of WSU-FSCCL model. The experiment was designed to terminate 150 d weight values surrounding the 800-mg threshold. To illustrate this for- from transplantation of tumors (WSU-DLCL2) or injection of cells mula, we use a hypothetical tumor weight target of 1,000 mg: tumor (WSU-FSCCL) because tumor-free animals at that point are considered weight of animal 1 in the control group was 905 mg (value 1) at day disease free and “cured”. This duration of experiment exceeds most stan- 31 (time 1) and 1,050 mg (value 2) at the next observation, day 34 dard protocols where experiments are usually terminated 90 to 120 d (time 2); these 2 values surround the 1,000 mg threshold value. A linear after tumor cell injections (16). Some animals in this study were eutha- interpolation between these 2 observations allows estimating the time nized after 155 d from transplantation of tumor fragments. Necropsy for the tumor to reach 1,000 mg (Time 1000) as follows: Time 1000 = was done on selected animals from each treatment group. Animals that time 2 + [([Log10(1000) − Log10 (value 2)]) × (time 2 − time 1)]/[Log10 were tumor free at that point are considered cured of their lymphoma. (value 2) − Log10 (value 1)]. So, for animal 1 of the control group, Time 1000 = 34 + [([Log10 (1000) − Log10(1050)]) × (34−31)]/[Log10 Determination of response (1050) − Log10 (905)] = 33 d. The medians of the T and C individual Different criteria were used to determine treatment efficacy in the values of the treated and control groups, respectively, are then calcu- WSU-DLCL2 model from those used for the WSU-FSCCL model. For lated taking into account censored values in animals where tumors the WSU-DLCL2-SCID model, tumor growth inhibition, tumor growth did not reach threshold value by the end of their observation period. delay, and log10kill were used, which are standard measures of tumor T − C calculation: T minus C values were calculated from T and C esti- response evaluation in subcutaneous xenograft models (16). For the mates as above. WSU-FSCCL-SCID systemic model, median day of death for each group Log10 cell kill (lck): calculated using the following formula, lck =T − C/ and number of animals alive at the end of the experiment were used as (3.32 × Td) where T is the median of individual time-to-reach 800 mg end points. This method is applicable to systemic diseases, such as leu- for the treated group; C is the median of the individual time-to-reach kemia, that disseminate in internal organs of animals but do not have 800 mg in the control group and Td is the control group doubling time. measurable subcutaneous tumors (16). Tumor growth inhibition calculation: calculated as the median of individ- Data processing ual time-to-reach 800 mg for T and control C groups, respectively, × 100. Estimation of Tumor doubling (Td) of subcutaneous (WSU-DLCL2) tumors. Td is the time for the tumor in the control group to reach twice a given tumor weight. Given that tumor weight follows a log-nor- Statistical methodology mal distribution, a base 10 logarithmic transformation is done on this Td and Log10 cell kill were calculated as descriptive parameters. Time variable. Individual Td values were estimated in the time zone between to reach 800-mg threshold for the subcutaneous tumors was deter- day 13 and day 25 posttransplantation of tumor in the control group. mined for each mouse, and then the treated groups were compared Tumor growth was fairly linear in this zone for all animals and com- with vehicle one. However, for the systemic model (WSU-FSCCL), the prised at least two measurements by animal. For each animal in the log-rank test was applied and P value for all experimental groups was control group, a linear regression is done on the base 10 logarithm of determined.

Fig. 1. Structure of SAR3419 used in this study.

Clin Cancer Res 2009;15(12) June 15, 20094040 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Anti-CD19 Immunotoxin in Lymphoma

Necropsy and histopathologic examination (7.5 mg/kg), there was significant antitumor activity (tumor All euthanized animals were dissected and examined macroscopi- growth inhibition of 0.0, tumor growth delay of 60 days, and cally for tumors or signs of gross tissue or organ abnormalities. His- >3 Log10 cell kill). There were 4 delayed deaths (days 113- topathologic examination of organs was conducted on selected 130) in the highest dose of SAR3419 (30 mg/kg) where necrop- animals in each group. Tissue sections were stained with H&E and sy and histopathologic examination did not show evidence of examined by two investigators (AMK and AA) with previous experi- tumor. We therefore believe that such deaths may represent ence in this model (14). treatment-related toxicity. In the intermediate dose level of Immunophenotyping of cells 15 mg/kg, all animals survived to the end of experiment (day Expression of CD19 and CD20 on WSU-DLCL2 and WSU-FSCCL 155) and were tumor free. Unconjugated (naked) huB4 anti- was determined using direct immunofluorescence staining with antibody showed no significant activity. DM4 alone was toxic after CD19 and anti-CD20 monoclonal antibodies (DAKO A/S) as previous- 1 or 2 injections when given at a concentration of ∼214ug/μL ly described (11). Stained cells were analyzed by flow cytometry using and resulted in severe toxicity (weight loss, dehydration, and FACScan (BD Biosciences). Background fluorescence is determined by lethargy) or death of all animals within few days of treatment. staining cells with normal mouse IgG. The proportion of positive cells The same dose (0.6 mg/kg) was tolerable when given in a more is determined by subtracting positive cells in the control from the spe- diluted form (∼93ug/μL). CHOP regimen resulted in modest cific monoclonal antibody. Channel of peak fluorescence (CPF) is visu- efficacy consistent with previous observations in this model ally determined as the peak of the flow cytometry–generated curve of (15, 17, 18). However, activity score of (+) is not considered clin- fluorescence where 20,000 cells per test are analyzed. CPF reflects the intensity of fluorescence signal and is an indirect measure of the density ically relevant. The WSU-DLCL2 is considered a chemotherapy- of antigen expression. resistantmodelsinceitwasestablishedfromapatientwith aggressive lymphoma refractory to therapy (11). Activity score in xenograft models of (+++) or (++++) corresponds with clinical Results partial and complete response, respectively (16). For rituximab, four of the seven animals survived to the end of the experiment CD19 and CD20expression by lymphoma cells. As shown in and were tumor free. In the other 3, tumors reappeared between Fig. 2, both cell lines expressed CD19 and CD20. In the WSU- days 99 and 103 after transplantation and reached ∼10% of DLCL2, 99% and 99.9% of cells were positive for CD19 and body weight between days 116 and 123. Tumor response calcu- CD20, respectively. CD20 expression was brighter in this cell lations could not be done because data were censored in more line as indicated by a higher CPF (238.2 versus 10.8, respective- than half of the animals due to activity of rituximab. ly). Similar trend was observed in the WSU-FSCCL line but with Efficacy of SAR3419 in WSU-FSCCL-SCID model. All ani- lower values; 81.4% and 93.6% of cells expressed CD19 and mals in the control group died of disseminated lymphoma be- CD20, respectively, with CPF values of 9.5 and 35.1, respective- tween days 60 and 75 (median day of death is 70; Table 2). The ly. CD19 intensity of expression was comparable in both cell pattern of lymphoma was typical of that previously described in lines, whereas CD20 expression was ∼7-fold brighter in the this model (14), i.e., lymphoma involves bone marrow, liver, WSU-DLCL2 compared with WSU-FSCCL as indicated by CPF. spleen, lymph nodes, and central nervous system (Fig. 3). There Efficacy of SAR3419 in WSU-DLCL2-SCID model. Td time in was a clear dose-response relationship of SAR3419 in this this experiment is estimated to be 5.5 days. As shown in Table 1, model. The highest dose (30 mg/kg) was most effective where SAR3419 at the higher dose levels (30 and 15 mg/kg) was effec- all animals survived to the end of the experiment (Fig. 4 and tive in eradicating the xenograft tumors. At the lower dose level Table 3). In contrast, 4 of 7 and 2 of 7 animals in the intermediate dose (15 mg/kg) and the low dose (7.5 mg/kg) survived to the end of the experiment, respectively. Unconjugated huB4 alone had very little activity in this model where the median survival for the whole group was 75 days compared with control. DM4 showed negligible activity with median day of death of 75 and no animal survived beyond 91 days after cell dose injection. When compared with SAR3419, rituximab at 40 mg/kg had antitumor activity comparable with the low dose of 7.5 mg/kg with 2 of 7 animals survived to the end of the experiment. Median day of death for the rituximab group was 110 compared with 96 days for SAR3419 low dose. CHOP chemotherapy showed modest activity in this model with median day of death of 147 days (range, 65-150). However, only three of seven animals survived to the end of the experiment. Necropsy examination was done on 37 of the 56 total animals in this experiment (66%). All animals that died before the end of the experiment (day 150) had tumors. Most common sites of involvement were lymph nodes (93%), followed by central nervous system (83%), liver (21%), spleen (10%), and bone marrow (8%). The pattern of central nervous system involve- Fig. 2. Expression of CD19 and CD20 on lymphoma cells used in this study. ment was leptomeningeal (Fig. 3), which we believe is the cause Numbers shown on top of each flow peak represent the CPF. Vertical axis, relative cell number (total 20,000 per test); horizontal axis, logfluorescence of death whenever it occurred. An unexpected finding in this intensity. study is the presence of microscopic disease in animals that

www.aacrjournals.org 4041 Clin Cancer Res 2009;15(12) June 15, 2009 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Cancer Therapy: Preclinical

Table 1. Antitumor activity of SAR 3419, CHOP, and rituximab in WSU-DLCL2–bearing SCID mice

Agent (dose) No. animals (bilateral) T/C (%) T-C^ (d) Log 10 kill TFS Drug death Activity score Gross^ Net^

Control (PBS) 7 0/7 0/7 − SAR (30 mg/kg) 7 N/a N/a N/a N/a 3/7 4/7 Cure* SAR (15 mg/kg) 7 N/a N/a N/a N/a 7/7 0/7 Cure† SAR (7.5 mg/kg) 7 0.0 60 3.29 3.12 2/7 0/7 +++ huB4 (30mg/kg) 7 57.9 5 0.27 0.11 0/7 0/7 − CHOP (MTD)‡ 7 20.4 13 0.71 0.55 0/7 0/7 + RTX (40 mg/kg)§ 7 —————— −

Abbreviations: T/C, tumor growth inhibition; T-C; tumor growth delay; N/a, not applicable. * Delayed drug toxicity (death) was noticed in 4/7 mice. †TFS as of 155 d post-WSU-DLCL2 transplantation. ‡MTD, maximum tolerated doses (see Materials and Methods section). §RTX, rituximab: 4/7 observations are censored because tumor weights never reached threshold of 800 mg. survived to the end of the experiment and seemed healthy at diarrhea, elevated liver enzymes, weakness, and lethargy pre- time of euthanasia (seen in sixof seven animals examined). cluded further development in favor of other antimicrotubule agents. The compound, therefore, is ideal for antibody conju- Discussion gation that enables selective delivery to CD19-positive cells, hence limiting toxicity. CD19 (B4) is a B-cell–restricted signal- In this article, we provide evidence that anti–CD19-DM4 ing molecule that functions as a positive regulator of the B-cell conjugate has superior antilymphoma activity compared with receptor signaling pathway. Lack of CD19 leads to arrest in B- unconjugated anti-CD19 antibody (huB4), the drug conjugate cell development and hyporesponsiveness to B-cell receptor alone (DM4), or with unconjugated anti-CD20 (rituximab). stimulation (27). Binding of anti-CD19 antibody leads to in- Findings in this study show the value of immunoconjugation ternalization of its target (28). However, the process is nega- and support their further development for clinical use. The tively influenced by CD21 expression on target cells (29). strategy can be used to deliver otherwise prohibitively toxic che- Binding of naked anti-CD19 to its target does not necessarily motherapy, maytansine derivative in this case, selectively to tu- lead to cell death or antitumor activity. In one study where mor tissue without undue host toxicity. mouse antihuman CD19 antibody was used to treat sixpa- Maytansine was originally isolated from the bark of the Af- tients with B-cell lymphoma, antibody-coated cells remained rican shrub Maytenus ovatus (19). It is the first in class of in the circulation of some patients for several days (30). The benzoansamacrolide antibiotics known as maytansinoids CD19 pathway can involve Src family kinases, lead to ampli- (20); antimitotic agents that bind to tubulin and inhibit the fication of B7-1 and B7-2 costimulatory molecules, and acti- microtubule assembly in a manner similar to but 100- to vate T cells (31). A defect in any of these steps might 1,000-fold more potent than Vinca alkaloids (21). Maytansine explain the lack of activity of the naked antibody. was evaluated by the U.S. National Cancer Institute in phase I Developing antibody-drug conjugates (ADC) that recognize and phase II clinical trials in the 1970s. The compound was tumor-associated antigens has been an active area of research found to be very effective with complete and partial responses in this decade (32) and can circumvent some of the limitations seen in a variety of tumors such as NHL, melanoma, thymo- of naked antibodies. Some of these agents have gained approv- ma, acute lymphoblastic leukemia, ovarian, and breast cancers al for clinical use as standard therapy in selected cancers. As (22-26). However, severe toxicity including nausea, vomiting, often is the case, clinical success in this area is first achieved

Table 2. Response data in WSU-FSCCL-SCID xenograft model by treatment

Treatment Group Day of death No. alive at end of exp. Median Range Mean 95% CI

Control 70 60-73 66.4 61.34-71.52 0/7 SAR (30mg/kg) 150 150-150 150 150-150 7/7 SAR (15mg/kg) 150 70-150 131.1 103.8-158.5 4/7 SAR (7.5mg/kg) 96 75-150 105.4 75.9-134.9 2/7 CHOP (MTD) 147 65-150 119 83.1-154.9 3/7 Rituxan (40mg/kg) 110 75-150 116.4 91.02-141.8 2/7 huB4 75 62-150 95.4 60.0-130.9 2/7 DM4 75 68-91 79.4 69.23-89.63 0/7

Abbreviations: 95% CI, 95% confidence interval; no.alive at end of exp.,no of alive at end of experiment.

Clin Cancer Res 2009;15(12) June 15, 20094042 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Anti-CD19 Immunotoxin in Lymphoma

Fig. 3. A, H&E-stained sections of liver in WSU-FSCCL–bearingmice. Top, shows hepatic lobules with portal veins and no evidence of involvement by lymphoma (negative control). Bottom, typical pattern of liver involvement by lymphoma (*) in the periportal areas (magnification, ×40). B, H&E sections of brains of SCID mice bearingWSU-FSCCL. Top, examples of no involvement by lymphoma. Note the thin meningeal lining of the brain (between opposing arrows). In contrast, the bottom panel shows examples of involvement of the meninges by lymphoma. Note the expanded space between the opposingarrows indicative of lymphoma cell infiltration. C, representative H&E-stained sections of SCID mice bone marrows (femurs) showing examples of normal marrow (right) and involvement by WSU-FSCCL (left). D, representative histopathologic sections of mouse lymph node showinginvolvement by WSU-FSCCL; left to right, ×200, ×400, and ×1,000 magnification, respectively.

in hematologic malignancies; calicheamicin conjugated to ed to anti-CD20 in 2002 and 2003, respectively (34, 35). Both anti-CD 33 (Mylotarg) was approved for acute myeloid leukemia Zevalin and Bexxar are proven to be more effective than in 2000 (33), followed by approval of 2 radio-immunoconju- the unlabeled anti-CD 20 antibody (rituximab). The cytotoxic gates: 90Yettrium (Zevalin) and I131 (Bexxar); isotopes conjugat- agent calicheamicin was also conjugated to anti-CD22 (CMC- 544) and has shown antilymphoma effect in preclinical animal models (36, 37). Given the encouraging results of ADCs, there are attempts now to develop rituximab-based ADCs (monomethyl auristatin E) to enhance its efficacy (38). Monomethyl auristatin E was also conjugated to anti-CD30 and found to be effective against animal models for Hodgkin lymphoma (39). Two different linkers have been used to conjugate maytansinoids to antibodies, one is the disulfide bond used in this study (SPDB-DM4; Fig. 1), which is cleavable, and the other is the thioether linker (SMCC-DM1; ref. 40). Lysosomal processing is required for the activity of antibody-maytansinoid conjugates, irrespective of the linker. SPDB-DM4 degrades in the lysosomes releasing intact maytansinoid drugs and linkers attached to lysine. The lysine-SPDB is then reduced and S- methylated to yield lipophilic and potent cytotoxic metabolite, S-methyl-DM4. DM1 conjugated to a human cancer antigen expressed on a number of solid tumors (huC242) was evaluated in preclinical animal models and in a phase I clinical trial (41, 42). The compound was also conjugated to CD79 and CD138 and evaluated in lymphoma and multiple myeloma, respectively (43, 44). Our study is the first one evaluating the DM4-anti-CD19 conjugate. It is noteworthy that ADCs with the disulfide linker (SPDB-DM4) were shown by Kovtun et al. (45), to have a “bystander” effect where antigen-positive and antigen-negative tumor cells are killed effectively. Presence of antigen-positive cells is required for such effect (Table 3). Our results presented in this study show the efficacy of the Fig. 4. Survival data of WSU-FSCCL–bearingSCID mice with different – treatments as indicated. Data in A and B are derived from the same anti CD19-DM4 conjugate against subcutaneous and system- experiment but split into two graphs for easier visualization. ic models of B-cell NHL representing diffuse large B-cell and

www.aacrjournals.org 4043 Clin Cancer Res 2009;15(12) June 15, 2009 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Cancer Therapy: Preclinical

Table 3. P values of the differences between treatment groups shown in Fig.4

0.0002 0.0011 0.0605 0.0002 0.0075 0.2208 0.0024 0.0222 0.4397 0.9223 0.0002 0.0068 0.2331 0.6071 0.9627 0.0233 0.0665 0.1435 0.3635 0.4442 0.3729 0.0233 0.0003 0.0034 0.0114 0.0802 0.0025 0.5023

follicular lymphoma. Both models are EBV negative and were mediated cytotoxicity (46). SAR3419 exhibited more signifi- established from patients with resistant disease to standard cant activity than rituximab when given at optimal doses de- chemotherapy. Standard chemotherapy regimen (CHOP) spite the fact that both models used in this study showed therefore has only modest activity in these models. The dem- brighter expression of CD20 than CD19 (Fig. 2). onstration that SAR3419 has more significant activity in these We conclude that huB4-DM4 conjugate (SAR3419) is ex- models compared with CHOP is clinically significant and tremely active in both subcutaneous and systemic preclinical illustrates the promise of ADCs as new therapeutic agents for xenograft models for diffuse large cell and follicular lympho- lymphoma. Our systemic model, the WSU-FSCCL-SCID, has mas. In addition to further development for clinical use, future predilection for central nervous system involvement, which is studies can explore its use in combination with other antilym- the major cause of death of animals (Fig. 3). SAR3419 was ef- phoma agents. fective against this model when given 7 days after a large cell 7 dose (10 ) inoculum. Similar results were reported using Disclosure of Potential Conflicts of Interest CMC-544-DM1 (anti-CD22 ADC) in the Ramos lymphoma xenograft model (37). In that study, therapy was effective when A.M. Al-Katib, commercial research grant, Sanofi-aventis. given up to 15 days after cell dose injection, although the 6 number of cells used was lower than ours (10 ). Like our Acknowledgments study, rituximab showed some activity in the xenograft models supporting earlier findings that SCID mice possess FcR-positive We thank Bertrand Monneau and Noëlle Boussac-Marlière for excellent effector cells capable of mediating antibody-dependent cell- biostatistical support.

References 1. Jemal A, Siegel R, Ward E, et al. Cancer statis- Novel concepts in radioimmunotherapy for covery of solid tumor active agents using a tics, 2008. CA Cancer J Clin 2008;58:71–96, Epub non-Hodgkin's lymphoma. Oncology (Huntingt) soft-agar-colony-formation disk-diffusion-assay. 2008 Feb 20. 2007;21:203–12, discussion 214, 217, 221. In: Cytotoxic anticancer drugs: Models and 2. Müller AMS, Ihorst G, Mertelsmann R, 10. Widdison WC, Wilhelm SD, Cavanagh EE, et al. concepts for drugdiscovery and development. Engelhardt . Epidemiology of non-Hodgkin's lym- Semisynthetic maytansine analogues for the tar- Valeriote FA, Corbett TH, Baker LH, editors. phoma (NHL): trends, geographic distribution geted treatment of cancer. J Med Chem 2006;49: Kluwer Academic Publishers; 1990. p. 35–87. and etiology. Ann Hematol 2005;84:1–12, Epub 4392–408. 17. Mohammad RM, WangS, Aboukameel A, 2004 Oct 9. Review. 11. Al-Katib AM, Smith MR, Kamanda WS, et al. et al. Preclinical studies of a nonpeptidic small- 3. Armitage JO, Weisenburger DD. New approach Bryostatin 1 down regulates MDR1 gene expres- molecule inhibitor of Bcl-2 and Bcl-X(L) [(-)- to classifyingnon-Hodgkin's lymphomas: clini- sion and potentiates vincristine cytotoxicity gossypol] against diffuse large cell lymphoma. cal features of the major histologic subtypes. in diffuse large cell lymphoma xenografts. Clin Mol Cancer Ther 2005;4:13–21. Non-Hodgkin's lymphoma classification project. Cancer Res 1998;4:1305–14. 18. Mohammad RM, Goustin AS, Aboukameel A, J Clin Oncol 1998;16:2780–95. 12. Mohammad RM, Mohamed AN, Smith MR, et al. Preclinical studies of TW-37, a new nonpep- 4. Marcus R, Hagenbeek A. The therapeutic use of Jawadi NS, Al-Katib A. A unique EBV-negative tidic small molecule inhibitor of Bcl-2, in diffuse rituximab in non-Hodgkin's lymphoma. Eur J low-grade lymphoma line (WSU-FSCCL) exhi- large cell lymphoma xenograft models reveal Haematol Suppl 2007;67:5–14. bitingboth t(14;8) and t(8;11) chromosomal drugaction on both Bcl-2 and Mcl-1. Clin Cancer – 5. Pulte D, Gondos A, Brenner H. Ongoing im- translocation. Cancer Genet Cytogenet 1993; Res 2007;13:2226 35. provement in outcomes for patients diagnosed 70:62–7. 19. Kupchan SM, Komoda Y, Court WA, et al. May- as havingnon-Hodgkinlymphoma from the 13. Harris NL, Jaffe ES, Stein H, et al. A Re- tansine, a novel antileukemic ansa macrolide 1990s to the early 21st century. Arch Intern vised European-American classification of from MaytenusoVatus. J Am Chem Soc 1972; – Med 2008;168:469–76. lymphoid neoplasms: a proposal from the 94:1354 6. 6. McLaughlin P, Grillo-Lopez AJ, Link BK, et al. International Lymphoma Study Group. Blood 20. Kupchan SM, Branfman AR, Sneden AT, Rituximab chimeric anti-CD20 monoclonal anti- 1994;84:1361–92. Verma AK, Dailey RG, Jr., Nagao KY. Novel maytansinoids. Naturally occurringand synthetic an- body therapy for relapsed indolent lymphoma: 14. Mohammad R, Abubakr Y, Dan M, et al. Bcl-2 tileukemic esters of maytansinol. J Am Chem half of patients respond to a four-dose treatment antisense oligonucleotides are effective against Soc 1975;97:5294–5. program. J Clin Oncol 1998;16:2825–33. systemic but not central nervous system disease 7. Davis TA, Grillo-Lopez AJ, White CA, et al. in severe combined immunodeficient mice bear- 21. Remillard S, Rebhun LI, Howie GA, Kupchan Rituximab anti-CD 20 monoclonal antibody ther- inghuman t(14;18) follicular lymphoma. Clin Can- SM. Antimitotic activity of the potent tumor in- apy in non-Hodgkin's lymphoma: safety and cer Res 2002;8:1277–83. hibitor maytansine. Science 1975;189:1002–5. efficacy of re-treatment. J Clin Oncol 2000;18: 15. Mohammad RM, Al-Katib A, Aboukameel A, 22. Chabner BA, Levine AS, Johnson BL, Young 3135–43. Doerge DR, Sarkar F, Kucuk O. Genistein sensi- RC. Initial clinical trials of maytansine, an antitu- 8. Bello C, Sotomayor E. Monoclonal antibodies tizes diffuse large cell lymphoma to CHOP (cyclo- mor plant alkaloid. Cancer Treat Rep 1978;62: for B-cell lymphomas: rituximab and beyond. phosphamide, doxorubicin, vincristine, 429–33. Hematology (Am Soc Hematol Educ Program) prednisone) chemotherapy. Mol Cancer Ther 23. Blum RH, Kahlert T. Maytansine. A phase I 2007;2007:233–42. 2003;2:1361–8. study of an ansa macrolide with antitumor 9. Schaefer-Cutillo J, FriedbergJW, Fisher RI. 16. Corbett TH, Valeriote FA, Polin L, et al. Dis- activity. Cancer Treat Rep 1978;62:435–8.

Clin Cancer Res 2009;15(12) June 15, 20094044 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Anti-CD19 Immunotoxin in Lymphoma

24. Cabanillas F, Rodriguez V, Hall SW, Burgess immunoconjugates. CA Cancer J Clin 2006;56: tatin-E antibody-drugconjugate(SGN-35) with MA, Bodey GP, Freireich EJ. Phase I study of 226–43. chemotherapy improves antitumour activity in maytansine usinga 3-day schedule. Cancer 33. Bross PF, Beitz J, Chen G, et al. Approval Hodgkin lymphoma. Br J Haematol 2008;142: Treat Rep 1978;62:425–8. summary: gemtuzumab ozogamicin in re- 69–73, Epub 2008 May 8. 25. Eagan RT, Ingle JN, Rubin J, Frytak S, Moertel lapsed acute myeloid leukemia. Clin Cancer 40. Erickson HK, Park PU, Widdison WC, et al. CB. Early clinical study of an intermittent sched- Res 2001;7:1490–6, Erratum in: Clin Cancer Antibody-maytansinoid conjugates are activat- ule for maytansine (NSC-153858): Brief Res 2002;8:300. ed in targeted cancer cells by lysosomal degra- communication. J Natl Cancer Inst 1978;60:93–6. 34. WitzigTE, Gordon LI, Cabanillas F, et al. Ran- dation and linker-dependent intracellular 26. Issell BF, Crooke ST. Maytansine. Cancer Treat domized controlled trial of yttrium-90-labeled processing. Cancer Res 2006;66:4426–33. Rev 1978;5:199–207. ibritumomab tiuxetan radioimmunotherapy ver- 41. Liu C, Tadayoni BM, Bourret LA, et al. Eradica- 27. Diamant E, Keren Z, Melamed D. CD19 regu- sus rituximab immunotherapy for patients with tion of large colon tumor xenografts by targeted lates positive selection and maturation in B relapsed or refractory low-grade, follicular, or delivery of maytansinoids. Proc Natl Acad Sci lymphopoiesis: lack of CD19 imposes develop- transformed B-cell non-Hodgkin's lymphoma. J U S A 1996;93:8618–23. mental arrest of immature B cells and conse- Clin Oncol 2002;20:2453–63. 42. Tolcher AW, Ochoa L, Hammond LA, et al. quential stimulation of receptor editing. Blood 35. Davis TA, Kaminski MS, Leonard JP, et al. The Cantuzumab mertansine, a maytansinoid immu- 2005;105:3247–54, Epub 2005 Jan 11. radioisotope contributes significantly to the ac- noconjugate directed to the CanAg antigen: a 28. Sapra P, Allen TM. Internalizingantibodies are tivity of radioimmunotherapy. Clin Cancer phase I, pharmacokinetic, and biologic correla- necessary for improved therapeutic efficacy of Res 2004;10:7792–8. tive study. J Clin Oncol 2003;21:211–22. antibody-targeted liposomal drugs. Cancer Res 36. DiJoseph JF, Amelino DC, Boghaert ER, 43. Poison AG, Yu S-F, Elkins K, et al. Antibody- 2002;62:7190–4. et al. Antibody-targeted chemotherapy with drugconjugatestargeted to CD79 for the treat- 29. Ingle GS, Chan P, Elliott JM, et al. High CD21 CMC-544: a CD22-targeted immunoconjugate ment of non-Hodgkin lymphoma. Blood 2007; expression inhibits internalization of anti-CD 19 of calicheamicin for the treatment of B-lym- 110:616–23, Epub 2007 Mar 20. antibodies and cytotoxicity of an anti-CD19-drug phoid malignancies. Blood 2004;103:1807–14, 44. Tassone P, Goldmacher VS, Neri P, et al. Cyto- conjugate. Br J Haematol 2008;140:46–58, Epub Epub 2003 Nov 13. toxic activity of the maytansinoid immunoconju- 2007 Nov 7. 37. DiJoseph JF, Goad ME, Dougher MM, et al. Potent gate B-B4–DM1 against CD138 + multiple 30. Hekman A, Honselaar A, Vuist WM, et al. Ini- and specific antitumor efficacy of CMC-544, a CD22- myeloma cells. Blood 2004;104:3688–96, Epub tial experience with treatment of human B cell targeted immunoconjugate of calichea micin, 2004 Aug3. lymphoma with anti-CD 19 monoclonal anti- against systemically disseminated B-cell lympho- 45. Kovtun YV, Audette CA, Ye Y, et al. Antibody- body. Cancer Immunol Immunother 1991;32: ma. Clin Cancer Res 2004;10:8620–9. drug conjugates designed to eradicate tumors 364–72. 38. Law C-L, Cerveny CG, Gordon KA, et al. Effi- with homogeneous and heterogeneous expres- 31. Yan J, Wolff MJ, Unternaehrer J, Meliman I, cient elimination of B-lineage lymphomas by sion of the target antigen. Cancer Res 2006;66: Mamula MJ. Targeting antigen to CD19 on B anti-CD20-auristatin conjugates. Clin Cancer 3214–21. cells efficiently activates T cells. Int Immunol Res 2004;10:7842–51, Erratum in: Clin Cancer 46. Dorshkind K, Pollack SB, Bosma MJ, Phillips 2005;17:869–77, Epub 2005 Jun 20. Res. 2005 May 15;11:3969. RA. Natural killer cells are present in mice 32. Sharkey RM, GoldenbergDM. Targetedthera- 39. Oflazoglu E, Kissler KM, Sievers EL, Grewal IS, with severe combined immunodeficiency py of cancer: new prospects for antibodies and Gerber H-P. Combination of the anti-CD30-auris- (SCID). J Immunol 1985;134:3798–801.

www.aacrjournals.org 4045 Clin Cancer Res 2009;15(12) June 15, 2009 Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 2009 American Association for Cancer Research. Superior Antitumor Activity of SAR3419 to Rituximab in Xenograft Models for Non-Hodgkin's Lymphoma

Ayad M. Al-Katib, Amro Aboukameel, Ramzi Mohammad, et al.

Clin Cancer Res 2009;15:4038-4045.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/15/12/4038

Cited articles This article cites 45 articles, 25 of which you can access for free at: http://clincancerres.aacrjournals.org/content/15/12/4038.full#ref-list-1

Citing articles This article has been cited by 8 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/15/12/4038.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/15/12/4038. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.