Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`CCR FOCUS
`
`SAR3419: An Anti-CD19-Maytansinoid Immunoconjugate for
`the Treatment of B-Cell Malignancies
`
`Veronique Blanc1, Anne Bousseau1, Anne Caron1, Chantal Carrez1, Robert J. Lutz2, and John M. Lambert2
`
`Abstract
`
`SAR3419 is a novel anti-CD19 humanized monoclonal antibody conjugated to a maytansine derivate
`through a cleavable linker for the treatment of B-cell malignancies. SAR3419 combines the strengths of a
`high-potency tubulin inhibitor and the exquisite B-cell selectivity of an anti-CD19 antibody. The inter-
`nalization and processing of SAR3419, following its binding at the surface of CD19-positive human
`lymphoma cell lines and xenograft models, release active metabolites that trigger cell-cycle arrest and
`apoptosis, leading to cell death and tumor regression. SAR3419 has also been shown to be active in different
`lymphoma xenograft models, including aggressive diffuse large B-cell lymphoma, resulting in complete
`regressions and tumor-free survival. In these models, the activity of SAR3419 compared favorably with
`rituximab and lymphoma standard of care chemotherapy. Two phase I trials with 2 different schedules of
`SAR3419 as a single agent were conducted in refractory/relapsed B-cell non-Hodgkin lymphoma. Activity
`was reported in both schedules, in heavily pretreated patients of both follicular and diffuse large B-cell
`lymphoma subtypes, with a notable lack of significant hematological toxicity, validating SAR3419 as an
`effective antibody-drug conjugate and opening opportunities in the future. Numerous B-cell–specific anti-
`CD19 biologics are available to treat B-cell non-Hodgkin lymphoma, and early phase I results obtained with
`SAR3419 suggest that it is a promising candidate for further development in this disease. In addition, thanks
`to the broad expression of CD19, SAR3419 may provide treatment options for B-cell leukemias that are often
`CD20-negative. Clin Cancer Res; 17(20); 6448–58. Ó2011 AACR.
`
`Introduction
`
`Non-Hodgkin lymphoma (NHL) is the most common
`lymphoma, with an estimated 65,540 new cases and 20,210
`deaths registered in the United States in 2010 (1). NHL
`originates from the malignant development of B or T
`lymphocytes and comprises a heterogeneous group of
`malignancies (2), among which B-cell lymphomas account
`for 85% of all cases (3). Despite the significant improve-
`ment in overall survival of patients with both aggressive and
`indolent B-cell NHL that has been achieved following the
`addition of rituximab to conventional chemotherapy (4),
`not all patients who have CD20-positive tumors respond to
`the treatment (5, 6), and most patients who have disease
`response will relapse (7). More than 50% of patients with
`NHL will die of their disease (4–8). Since 2005, 5 new drugs
`have been approved for the treatment of lymphoma (6, 9),
`including 2 for B-cell NHL, namely, bortezomib, a protea-
`some inhibitor approved for the treatment of mantle cell
`
`Authors' Affiliations: 1Oncology Business Division, Sanofi, Vitry sur
`Seine, France; 2ImmunoGen, Inc., Waltham, Massachusetts
`Corresponding Author: Veronique Blanc, Sanofi, 13 quai Jules Guesde,
`Vitry sur Seine, 94403 France. E-mail:
`Veronique.Blanc2@sanofi-aventis.com
`
`doi: 10.1158/1078-0432.CCR-11-0485
`Ó2011 American Association for Cancer Research.
`
`lymphoma in 2006, and bendamustine, a DNA alkylating
`and antimetabolite agent approved for the treatment of
`indolent B-cell NHL in 2008. However, despite the large
`number of trials of novel agents, including monoclonal
`antibodies and small-molecule inhibitors (3, 10), there is
`still an important unmet medical need, particularly in
`second-line and subsequent lines of therapy. Concerning
`monoclonal antibodies, different modalities of conjugation
`have been tested in the clinic and have shown responses in
`hematological malignancies and NHL in particular, as dis-
`cussed in this CCR Focus section, dedicated to antibody
`conjugates (11). These modalities include radionuclide
`conjugation (12), toxin conjugation (13), and small-mol-
`ecule drug conjugation, also called antibody-drug conju-
`gates (ADC; refs. 14–16). ADCs represent the most active
`field today, with at least 20 such ADCs in development (11).
`Very promising clinical responses have been achieved in
`both solid tumors, as shown by T-DM1 for the treatment of
`HER2-positive breast
`tumors (14), and hematological
`malignancies, as shown by SGN35 (15) and CMC-544
`(16) for Hodgkin lymphoma and NHL, respectively.
`
`Monoclonal Antibodies beyond Rituximab for
`Treating B-Cell Malignancies
`
`Since the introduction of rituximab in 1997, only 2 more
`"naked" (nonconjugated) antibodies have been approved
`for the treatment of lymphoma: (i) alemtuzumab, an
`
`6448
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Anti-CD19-Maytansinoid Conjugate
`
`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`anti-CD52 antibody approved in 2001 for the treatment of
`CLL (17); and (ii) ofatumumab, a second-generation anti-
`CD20 antibody displaying increased complement-depen-
`dent cytotoxicity compared with rituximab,
`that was
`approved in 2009 for relapse/refractory CLL patients who
`failed fludarabine and alemtuzumab (18). In addition to
`these 2 naked antibodies, 2 murine anti-CD20 radionuclide
`conjugates were approved for consolidation treatment in
`2002 (90Y-ibritumomab) and 2003 (131I-tositumomab;
`refs. 12 and 19).
`In recent years, investigators have made a considerable
`effort to further improve NHL treatment, and several sec-
`ond- and third-generation anti-CD20 monoclonal antibo-
`dies have been examined in clinical trials (20). In addition,
`exploration of other B-cell–specific antigens,
`including
`CD19, CD22, CD37, and CD40, as potential targets for
`the development of new antibody-based treatments is grow-
`ing. Antibodies are being investigated as naked molecules,
`such as the anti-CD22 epratuzumab (21), the anti-CD40
`dacetuzumab (22), and the different anti-CD19 antibodies
`(Table 1), or in an engineered format, such as the anti-
`CD19/anti-CD3–bispecific blinatumomab (23) and the
`anti-CD37–specific small modular immunopharmaceuti-
`cal TRU-016 (24). The advent of ADC technologies has also
`led to clinical evaluation of conjugated antibodies, such as
`
`the anti-CD22-calicheamicin conjugate CMC544 (16) and
`the subject of this review, the anti-CD19-maytansinoid
`conjugate SAR3419 (25–28).
`
`CD19 Antigen
`
`CD19 is a type I transmembrane glycoprotein of the
`immunoglobulin Ig superfamily, with expression restricted
`to B cells (29). CD19 is involved in B-cell fate and differ-
`entiation through the modulation of B-cell receptor signal-
`ing at multiple stages of B-cell development (29, 30). CD19
`is ubiquitously expressed on B cells (25), as it is found
`expressed from the early pre-B stage throughout B-cell
`differentiation up to mature B cells, before it is downmo-
`dulated at the plasma cell stage (Fig. 1). Thus, CD19 has
`broader expression than CD20. The pattern of CD19 expres-
`sion is maintained in B-cell malignancies, covering all
`subtypes of B-cell lymphoma, from indolent to aggressive
`forms, as well as B-cell chronic lymphocytic leukemia and
`non-T acute lymphoblastic leukemia (31–33), and allows
`the targeting of tumor indications of early B cells, such as
`acute lymphoblastic leukemia (ALL), which cannot be
`targeted by rituximab. The quantification of the number of
`CD19 molecules at the surface of tumor cells and cell lines
`derived from malignant B cells resulted in a wide range of
`
`Table 1. Clinical trials and discovery projects targeting CD19
`
`Biologic compound
`
`Type
`
`Mechanism of action
`
`Phase (initiation date)
`
`MT-103 (blinatumomab)
`Micromet, Inc.
`SAR3419
`Sanofi-Aventis/ImmunoGen,
`Inc.
`MEDI-551 MedImmune/
`Astra-Zeneca
`MOR-208/XmAb5574
`Xencor/Morphosys
`MDX-1342 Medarex/
`Bristol-Myers Squibb
`Combotox
`University of Texas
`Southwestern/Abiogen
`
`DI-B4
`Merck KGaA/Cancer
`Research UK
`SGN-19A
`Seattle Genetics
`MDX-1206 Medarex/
`Bristol-Myers Squibb
`
`AFM-11
`Affimed Therapeutics AG
`AFM-12
`Affimed Therapeutics AG
`
`Bispecific scFv anti-CD19/anti-CD3
`BiTE
`Humanized anti-CD19 mAb
`conjugated to maytansinoid DM4
`
`Glycoengineered humanized anti-CD19
`mAb (BioWa's Potelligent)
`Fc engineered humanized
`anti-CD19 mAb
`Glycoengineered fully human
`anti-CD19 mAb (BioWa's Potelligent)
`Mixture of chimeric anti-CD19 mAb
`HD37 and anti-CD22 mAb RFB4,
`both conjugated to deglycosylated
`ricin A-chain (HD37-dgA þ RFB4-dgA)
`Chimeric anti-CD19 mAb monoclonal
`antibody
`
`Fully human anti-CD19 mAb (hBU12)
`conjugated to auristatin (vc-MMAE)
`Fully human anti-CD19 mAb (MDX1435)
`conjugated to duocarmycin
`(vc-MGBAA)
`Tetravalent tandem antibody (TandAb)
`anti-CD19/anti-CD3
`Tetravalent tandem antibody (TandAb)
`anti-CD19/anti-CD16
`
`T-cell recruitment and
`activation
`ADC (tubulin binder)
`
`I/II/III (2007) Pivotal
`þ
`ALL
`trial for MDR
`I/II (2007)
`
`Naked antibody high-affinity
`FcgRIII-enhanced ADCC
`Naked antibody high-affinity
`FcgRIII-enhanced ADCC
`Naked antibody high-affinity
`FcgRIII-enhanced ADCC
`Immunotoxin conjugate with
`deglycosylated ricin
`A-chain
`
`I (2010)
`
`I (2010)
`
`I (2008, on hold)
`
`I (2005)
`
`Naked antibody ADCC
`
`I (2010)
`
`ADC (tubulin binder)
`
`Discovery
`
`ADC (DNA alkylating agent)
`
`Discovery
`
`T-cell recruitment
`
`Discovery
`
`NK cell recruitment
`
`Discovery
`
`www.aacrjournals.org
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`6449
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`CCR FOCUS
`
`Figure 1. Pattern of expression of
`CD19 and CD20 antigens during B-
`cell development and associated
`malignancies. A simplified cartoon
`of B-cell lineage, B-cell
`malignancies, and antigen
`expression (59, 60). The positioning
`of the different B-cell malignancies
`associated with different stages of
`B-cell development is abridged and
`illustrative only; a detailed
`description is beyond the scope of
`this review.
`
`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`published values due to differences in cell line clones, the
`antibodies studied, and the methodologies used. For exam-
`ple, CD19 expression was reported to be both >100,000
`sites per cell (34) and 10,000 sites per cell (35) for the
`same cellular models. In our hands, using a calibration
`curve of beads loaded with different amounts of human
`IgG1, the quantification of CD19 antigen at the surface of
`Ramos, Raji, and Daudi lymphoma cell lines gave average
`values of 42,166, 18,315, and 14,077 antibody-binding
`sites per cell, respectively. With the same assay, the quan-
`tification of CD19 on the surface of normal B lymphocytes
`yielded values in the range of 11,000 to 16,000 antibody-
`binding sites per cell, an order of magnitude similar to that
`of the lymphoma cell lines tested.
`CD19 was shown to be internalized efficiently in lym-
`phoma tumor models with the use of different antibodies,
`such as huB4 (35) and hBU12 (36). In our laboratories, we
`quantified the internalization and processing of the anti-
`CD19 antibody huB4 using an Alexa488 antibody conju-
`
`gate and different tumor cell lines upon incubation at 37
`C.
`The kinetics of appearance of antibody-free fluorescence,
`produced by lysosomal degradation of the antibody moiety,
`showed internalization and processing of
`the huB4-
`Alexa488 in all models, with some variability between
`models (35) that may reflect differences in the capacity of
`CD19 to internalize in different cellular contexts. Indeed,
`the level of CD21 coreceptor, expressed in a subset of
`lymphomas, was described to affect the internalization and
`efficacy of an anti-CD19-maytansinoid conjugate through a
`noncleavable linker (37). However, this observation was
`not confirmed in a second study using auristatin derivatives
`conjugated to a different anti-CD19 antibody through a
`cleavable linker (36).
`
`Anti-CD19 Intervention
`
`CD19’s B-cell lineage–restricted expression and moderate
`to high homogeneous expression in most cases of lympho-
`
`ma make it an attractive target for therapies for B-cell
`malignancies, as shown by the many previous and ongoing
`therapeutic interventions that focus on CD19 (Table 1).
`Different types of molecules targeting CD19 are being
`developed that can broadly fit into 3 main classes: (i) naked
`antibodies, most of which undergo modifications of their Fc
`portion to enhance binding to FcRgIII and subsequently
`enhance antibody-dependent cell-mediated cytotoxicity
`(ADCC) activity, such as MOR-208 and MEDI-551; (ii)
`bispecific antibodies with one arm binding to CD19 and
`one arm binding to either the T-cell receptor (as exemplified
`by blinatumomab, the most advanced anti-CD19 therapeu-
`tic molecule) or NK-cell receptors (as in the molecule
`AFM12, a tetravalent tandem antibody that is in the dis-
`covery stage of development); and (iii) antibody conju-
`gates, for which anti-CD19 antibodies are conjugated to
`either a toxin (as exemplified by the ricin-based immuno-
`toxin Combotox) or a potent low-molecular-weight cyto-
`toxic molecule (as exemplified by SAR3419). As of today,
`Combotox has shown anticancer activity in NHL (38) and
`ALL (39) that may warrant further clinical testing. Studies of
`blinatumomab (23) and SAR3419 (27) have provided a
`clinical proof of concept, showing a notable rate of objective
`response in pretreated patients in phase I clinical trials.
`
`SAR3419 Structure and Mechanism of Action
`
`SAR3419 is an ADC that consists of a humanized mono-
`clonal IgG1 antibody (huB4) attached to a highly potent
`tubulin inhibitor, the maytansinoid DM4 (40), through
`reaction with an optimized cleavable linker, N-succini-
`midyl-4-(2-pyridyldithio)butyrate (SPDB linker; Fig. 2).
`The succinimidyl group of the linker reacts with amino
`groups of lysine residues of the antibody to form stable
`amide bonds, and the pyridyldithio moiety reacts with
`the sulfhydryl group of DM4 to form a hindered disulfide
`bond between the linker and DM4. The disulfide bond
`can be cleaved inside target cells by thiol-disulfide
`
`6450
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`Clinical Cancer Research
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Anti-CD19-Maytansinoid Conjugate
`
`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`Figure 2. Structure of SAR3419. Figure is adapted from Al-Katib et al. (43).
`
`exchange reactions to release fully active DM4 (28).
`SAR3419 contains an average of 3.5 DM4 molecules
`per molecule of antibody.
`The murine B4 antibody, one of the earliest antibodies to
`define the CD19 antigen (31), was humanized via a vari-
`able-domain resurfacing method (41) to yield huB4. The
`huB4 antibody displays a subnanomolar affinity for CD19
`on the surface of B cells, and this affinity is conserved
`following conjugation to create SAR3419 (35). The huB4
`antibody was shown to induce ADCC (data not shown) but
`not complement-dependent cytotoxicity as described for
`other anti-CD19 antibodies (42). The ADCC activity was
`conserved following conjugation to the maytansinoid. The
`naked antibody has no direct antiproliferative or proapop-
`totic activity. Therefore, huB4 was not found to be active in
`vivo as a single agent in severe combined immunodeficient
`(SCID) mice bearing several different lymphomas (43). The
`huB4 antibody binds only to the human CD19 antigen
`(29); it does not recognize CD19 in rodent or cynomolgus
`monkey toxicology models, and thus no B-cell depletion
`studies have been done with either the huB4 antibody or
`SAR3419 in animal models.
`Maytansine and other maytansinoids are antimitotic
`agents that bind to tubulin, inhibiting microtubule assem-
`bly and inducing G2/M arrest in the cell cycle, which
`subsequently leads to cell death (44). Maytansinoids are
`very potent, displaying cytotoxic activity in the 10 to 90
`pmol/L range across several tumor cell lines, including
`lymphoma lines (45). The SPDB linker was selected based
`on its superior activity in vivo compared with SPP (another
`cleavable linker with a more labile disulfide bond) and the
`noncleavable SMCC linker in studies analogous to those
`described by Kellogg and colleagues (46) for another anti-
`body-maytansinoid conjugate.
`SAR3419 was shown to display potent in vitro cytotoxicity
`after 5 days of exposure to different CD19-positive lym-
`phoma cell lines, with EC50 values ranging from subnano-
`molar to a few nanomolars (35). SAR3419 induced a rapid
`
`cell-cycle arrest at the G2/M phase (within the first 24 hours
`of exposure) followed by an increase in apoptotic cells from
`24 to 48 hours (35), as expected for delivery of the may-
`tansinoid payload into CD19þ cells. As has been shown in
`the case of other maytansinoid conjugates (47) and in
`experiments with SAR3419 in which the DM4 moiety was
`radiolabeled (48), the mechanism of action of SAR3419
`involves binding to the cell-surface antigen, followed by
`internalization via endocytosis and subsequent intracellu-
`lar routing to lysosomes where the huB4 antibody moiety of
`the ADC is degraded to yield the lysine-SPDB-DM4 metab-
`olite. Once it is in the reducing environment of the cyto-
`plasm, this metabolite is further subjected to cleavage by
`thiol-disulfide exchange reactions to release the free may-
`tansinoid thiol compound DM4 (47, 48). DM4 is then
`S-methylated to form S-methyl-DM4 by an endogenous
`S-methyl-transferase (47). The final metabolite of cancer
`cell metabolism of SAR3419, S-methyl-DM4, has a high
`potency as a microtubule poison that was shown to be
`similar to the potency of maytansine itself when compared
`in cytotoxicity assays in vitro on different tumor cell lines
`(40, 44, 49). The early metabolite formed upon lysosomal
`processing of the ADC, lysine-SPDB-DM4, shows a 1,000-
`fold lower activity on cell lines in vitro resulting from the
`negative effect of the charged lysine residue on its ability to
`diffuse across the plasma membrane into cells (47). How-
`ever, such lysine-linker-maytansinoid species are potent
`microtubule poisons when they are released inside cells
`following uptake and lysosomal processing of an ADC
`(28, 47).
`The mechanism of action of SAR3419 was also shown in
`vivo in immunodeficient SCID mice bearing human Ramos
`Burkitt’s lymphoma, a model that was shown to be highly
`sensitive to SAR3419 (35). In recent studies done in our
`laboratories, SAR3419 was administered at a single dose of
`20 mg/kg 14 days after tumor subcutaneous implantation.
`Tumor sampling was done at different times, and immu-
`nohistochemical
`evaluation of phospho-histone H3
`
`www.aacrjournals.org
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`6451
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`CCR FOCUS
`
`(p-histone H3, a mitosis biomarker) and cleaved caspase 3
`(a proapoptotic marker) was done. The single-dose treat-
`ment with SAR3419 induced a significant increase in stain-
`ing for p-histone H3, revealing a mitosis blockade that was
`apparent 24 hours after administration (Fig. 3). This mitosis
`blockade was then followed by tumor cell apoptosis as
`evidenced 48 hours after dosing with the ADC by a marked
`increase in cells that stained positively for cleaved caspase 3
`(Fig. 3).
`Studies were also done in the same Ramos xenograft
`model to assess whether the pattern of metabolic processing
`observed in cell lines was the same as that seen in xenograft
`models. Mice bearing the tumor xenografts were injected
`with a single administration of radiolabeled [3H]-SAR3419
`(conjugate made with tritiated DM4) at 15 mg/kg in experi-
`ments similar to previous experiments involving another
`ADC with the same linker-maytansinoid design (49).
`Tumor sampling was done at different time points, and all
`metabolites were quantified. The total amount of low-
`molecular-weight (i.e., not protein-linked) metabolites
`increased rapidly from 2 to 8 hours, with a subsequent
`linear progression from 8 to 48 hours, and an overall 6-fold
`accumulation from 2 to 48 hours (Fig. 4). At the latter time
`point, 71% of the total radioactivity in the tumor xenograft
`corresponded to low-molecular-weight metabolites, in con-
`trast to just 30% at 2 hours (data not shown). The proces-
`sing of radiolabeled SAR3419 in vivo was found to be similar
`to the in vitro findings (Fig. 4), with the sequential formation
`of lysine-SPDB-DM4 (the major species at 2 hours) fol-
`lowed by DM4 (about 10% of total nonproteic metabolites
`at 2 hours, rising to 24% at 8 hours) and then by the
`appearance of S-methyl-DM4 (first detected at 8 hours). As
`observed in vitro, similar quantities of lysine-SPDB-DM4
`and S-methyl-DM4 were found in the tumors after 48 hours
`
`(Fig. 4; ref. 49). The time course of the formation of the
`active maytansinoid metabolites in the tumor is consistent
`with the observed high induction of p-histone H3 and
`cleaved caspase 3 at 24 and 48 hours, respectively (Fig. 3).
`
`SAR3419 Preclinical Activities in Different
`Lymphoma Models
`
`SAR3419 has shown in vivo efficacy in different lympho-
`ma models, including Burkitt’s lymphomas (Namalwa,
`Ramos, and Raji) and diffuse large B-cell
`lymphoma
`[DLBCL (RL, WSU-DLCL2, and WSU-FSCCL)] implanted
`in SCID mice (35, 43). In all models, SAR3419 showed a
`high level of activity leading to complete responses, with
`mice being tumor-free at study termination at the highest
`doses and with significant tumor growth delay at the lower
`doses (both when injected as a single dose and in a multiple-
`dose schedule).
`As an illustration, in the Ramos lymphoma xenograft
`model (35), when treated by the intravenous route (2 doses
`with 4-day interval), SAR3419 was highly active at the
`dosage of 15 mg/kg (a well-tolerated dosage as evidenced
`by no loss in body weight), and 100% of the mice were
`tumor-free at the end of the study (day 124). The lower
`doses of 7.5 and 3.3 mg/kg were also highly active, with
`complete tumor regressions in 100% of the animals and
`with 5 out of 7 mice being tumor-free at the dose of 7.5 mg/
`kg on day 124. In comparison, treatment with unconjugat-
`ed DM4 at a dose equivalent to the 15 mg/kg dose of
`SAR3419 (0.35 mg/kg of DM4, 2 doses with 4-day interval)
`showed no significant antitumor activity, indicating that
`conjugation of the DM4 to the antibody was critical for
`delivering sufficient maytansinoid to the cancer cells in vivo
`to induce tumor shrinkage.
`
`pHistone H3
`C-Caspase 3
`
`48
`Time (hours)
`© 2011 American Association for Cancer Research
`
`72
`
`96
`
`6
`
`24
`
`45
`
`40
`
`35
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`Percentage of positive tumor cells area
`
`Figure 3. Kinetics of
`pharmacodynamic markers
`p-histone H3 and cleaved
`caspase 3 (C-Caspase 3) in
`Ramos xenograft model.
`Quantitative evaluation
`of p-Histone H3 and C-Caspase 3
`by image analysis quantification
`at 6, 24, 48, 72, and 96 hours after
`administration of a single
`intravenous dose of 20 mg/kg of
`SAR3419 in SCID mice bearing
`established Ramos
`subcutaneous xenograft tumors.
`
`6452
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`Clinical Cancer Research
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Anti-CD19-Maytansinoid Conjugate
`
`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`300
`
`Total nonproteic metabolites
`
`DM4-SPDB-lys
`
`DM4
`
`200
`
`S-methyl-DM4
`
`
`
`100
`
`0
`
`0
`
`8
`
`16
`
`24
`Time (hours)
`
`32
`
`40
`
`48
`
`© 2011 American Association for Cancer Research
`
`Tumor concentration (ng Eq/g)
`
`Figure 4. Kinetics of the
`production of nonproteic
`maytansinoid metabolites in the
`Ramos xenograft model.
`Quantitative evaluation of
`nonproteic maytansinoid
`metabolites (metabolites not linked
`to proteins) in nanograms per gram
`of tumor (ng Eq/g) at 2, 6, 8, 24, and
`48 hours after administration of a
`single intravenous dose of 15
`mg/kg of [3H]-SAR3419 in SCID
`mice bearing established Ramos
`xenograft tumors.
`
`The specificity of tumor targeting by SAR3419 was further
`shown by comparing the efficacy of a 10 mg/kg single dose
`of SAR3419 with the efficacy of the same dose of SAR3419
`administered together with 50 mg/kg of naked huB4 anti-
`body. The coadministration of SAR3419 with excess huB4
`antibody resulted in complete abrogation of the antitumor
`activity of SAR3419, because binding of the naked huB4
`antibody to the CD19 antigen blocks binding of the con-
`jugate to its target (Fig. 5). Coadministration of SAR3419
`with an excess of a nonspecific humanized IgG1 control
`antibody (huMy9-6) had no effect on blocking the antitu-
`mor activity of SAR3419 (Fig. 5).
`SAR3419 was tested in a systemic model of DLBCL in
`which animals that were injected intravenously with the
`WSU-FSCCL cell line developed disseminated disease with
`bone marrow, liver, spleen, lymph node, and central ner-
`vous system involvement (43). In this model, SAR3419 also
`showed a dose-response relationship (2 doses with 4-day
`interval), with all animals surviving at the end of the
`experiment (150 days postdose) at 30 mg/kg, the highest
`dose tested. In the control group, all mice were euthanized
`between day 60 and day 73 because they showed a large
`burden of disseminated tumor. At the lower dose groups of
`15 mg/kg and 7.5 mg/kg, 4 out of 7 mice and 2 out of 7 mice,
`respectively, survived until day 150 (43). In contrast, there
`was no activity of the naked huB4 injected at 30 mg/kg
`(2 doses with 4-day interval) or 0.6 mg/kg of DM4 (2 doses
`with 4-day interval), corresponding to the amount of DM4
`in the highest SAR3419 dose of 30 mg/kg.
`Taken together, the preclinical results show a high and
`specific activity of SAR3419 in a variety of different lym-
`phoma models. The results highlight the role of conjugation
`
`in directing the cytotoxic molecule to the tumor, allowing
`delivery of enough DM4 in vivo to induce tumor regression.
`It confers a large therapeutic window to a cytotoxic mole-
`cule that otherwise would be too toxic to be used alone (50),
`with free maytansine having little or no activity in most
`tumor models at
`its maximum tolerated dose (MTD;
`refs. 35, 43, and 45). In addition, the activity of SAR3419
`compared favorably with rituximab or CHOP treatments in
`the DLBCL model tested, suggesting opportunities for
`future improvements in lymphoma treatment (43).
`
`Imaging SAR3419 Activity in a Mouse Model
`
`Metabolic imaging with fluorodeoxyglucose positron
`emission tomography (FDG-PET) combined with comput-
`ed tomography is widely accepted for staging of lymphoma
`patients and assessment of response after completion of
`therapy. It is now being used to monitor response during
`treatment in lymphoma patients (51). We therefore inves-
`tigated whether FDG uptake in tumor is an appropriate
`noninvasive readout for monitoring response and predict-
`ing outcome of SAR3419 therapy in a disseminated model
`of lymphoma. Mice inoculated with human Daudi lym-
`phoma cells were followed longitudinally by FDG-PET. At
`an advanced stage of disease, involvement of kidney, ovary,
`and spinal cord sites was evident in all mice 23 days after cell
`inoculation (Fig. 6), and the PET signal was used to ran-
`domize the mice at baseline before treatment with 12.5 mg/
`kg SAR3419. In comparison with the control group, the
`metabolic activity in tumor-invaded regions was reduced by
`50% after treatment (Fig. 6). The response to SAR3419
`seemed to be homogeneous, with a decrease of signal in
`
`www.aacrjournals.org
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`6453
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`Figure 5.
`Impact of naked huB4
`antibody on the efficacy of a single
`administration of SAR3419 against
`a bulky Ramos subcutaneous
`tumor xenograft model. SCID mice
`were inoculated with Ramos tumor
`cells, and once the xenografts had
`reached about 295 mm3 on day 11,
`mice (6 per group) were treated with
`a test article given by intravenous
`injection (tail vein). The huMy9-6
`(anti-CD33 antibody) is an isotype-
`matched, nonspecific control
`humanized IgG1 antibody; huB4 is
`the anti-CD19 humanized IgG1
`antibody moiety of SAR3419;
`huB4-DM4 is a synonym for
`SAR3419.
`
`PBS
`10 mg/kg huB4-DM4
`
`10 mg/kg huB4-DM4 + 50 mg/kg huB4
`10 mg/kg huB4-DM4 + 50 mg/kg huMy9-6
`
`CCR FOCUS
`
`2500
`
`2000
`
`1500
`
`1000
`
`500
`
`Measured tumor volume (mm3)
`
`0
`
`0
`
`5
`
`10
`
`15
`
`30
`
`25
`20
`Days (post inoculation)
`© 2011 American Association for Cancer Research
`
`35
`
`40
`
`45
`
`all invaded sites, including the spinal cord. This decrease in
`PET signal translated into prolonged survival for the mice in
`the treated group, with a 40% median increase in lifespan.
`In a separate experiment, modulation of the metabolic
`signal was used to adapt the treatment schedule. The mice
`were injected twice with 15 mg/kg SAR3419 with a 5-day
`interval between both injections. Within 5 days after the
`start of therapy, the metabolic activity had decreased to
`background values. The PET signal then recurred and
`increased to reach the initial values 15 days after the end
`of the first cycle, showing relapse of the disease. A second
`cycle of treatment was then administered and was shown to
`control the disease, as evidenced by the stabilization of the
`PET signal for 15 additional days. Again, the modulation of
`the PET signal by SAR3419 therapy was predictive of out-
`come, with a 175% increase in lifespan for treated mice.
`Serial imaging of tumor uptake of FDG can therefore serve
`as a translational efficacy readout for SAR3419 therapy.
`
`SAR3419 Clinical Development
`
`Two phase I dose escalation studies exploring 2 different
`schedules of administration [every 3 weeks (Q3W; ref. 27)
`and weekly (Q1W; ref. 52)] have been conducted with
`SAR3419 in patients with refractory/relapsed B-cell NHL
`expressing CD19.
`The preliminary results from the Q3W trial, in which
`patients were eligible for up to 6 cycles of treatment, showed
`that the MTD was 160 mg/m2 (4.3 mg/kg), a dose level
`that was subsequently used to treat an expanded cohort of
`20 patients (27). The dose-limiting toxicity (DLT) at doses >
`200 mg/m2 was reversible toxicity to the cornea that did not
`preclude continued dosing (at 208 mg/m2) in patients
`
`receiving clinical benefit, albeit with dose delays of 1 to 2
`weeks (27). This ocular toxicity, which consisted mainly of
`blurred vision, was associated with microcystic epithelial
`corneal changes and was typically seen after cycle 2 or later
`cycles. The DLT observed in this Q3W phase I study of
`SAR3419 is similar to that observed in phase I studies of
`albumin-bound paclitaxel (53), as well as one other ADC
`compound from among the several ADCs in clinical eval-
`uation that have the same SPDB-DM4 linker-maytansinoid
`format (28, 54). No other clinically significant grade 3 or 4
`toxicities exceed an incidence of 10%, including peripheral
`neuropathy and gastrointestinal toxicities that are classical-
`ly seen with agents targeting tubulin (27). As with other
`antibody-maytansinoid conjugates (28), there were no
`clinically significant hematologic toxicities, and no gastro-
`intestinal toxicities were reported at the MTD, showing the
`efficacy of the linker between the huB4 antibody and the
`DM4 in the blood stream [the toxicities of free maytansine
`included myelosuppression and severe gastrointestinal
`toxicities (50)]. The half-life of SAR3419 in these patients
`was 4 to 6 days and was approximately linear across all doses
`from 20 mg/m2 to 270 mg/m2 in the phase I trial (27). This
`suggests that there was a negligible antigen sink from
`normal tissue expression, as might be anticipated for a
`population of heavily pretreated patients with NHL, all of
`whom had received prior therapy with rituximab, which
`effectively depletes normal B cells.
`Tumor shrinkage was observed in more than half of the
`35 response-evaluable patients (74%) at the time of initial
`reporting of the study, with 6 objective responses (27). Of
`note, 7 of 15 patients with rituximab-refractory disease
`showed tumor shrinkage, with 1 objective response.
`Although these are early results from a phase I trial that
`
`6454
`
`Clin Cancer Res; 17(20) October 15, 2011
`
`Clinical Cancer Research
`
`UPenn Ex. 2073
`Miltenyi v. UPenn
`IPR2022-00855
`
`

`

`Downloaded from http://aacrjournals.org/clincancerres/article-pdf/17/20/6448/1999115/6448.pdf by guest on 10 February 2023
`
`Anti-CD19-Maytansinoid Conjugate
`
`Day23
`
`Day27
`
`Day30
`
`Urinary
`bladder
`
`Adipose
`tissues
`
`Day24
`
`
`
`© 2011 American Association for Cancer Research© 2011 American Association for Cancer Research
`
`Control
`
`Tail
`
`Hind
`limb
`
`Front
`limb
`
`Head
`
`SAR3419
`12.5mg/kg
`
`Figure 6. Longitudinal monitoring of SAR3419 activity using FDG-PET in mice inoculated intravenously with Daudi lymphoma. Samples of serial FDG-PET
`images obtained in mice bearing disseminated