Summary of a workshop on preclinical and translational safety assessment of CD3 bispecifics.

Currently, there is a multitude of CD3 bispecifics with different molecular designs and binding properties in preclinical and clinical development for the treatment of liquid or solid tumors. The key safety concerns with CD3 bispecifics are excessive release of cytokines, which may translate to potentially life-threating cytokine release syndrome (CRS), target organ toxicity due to redirection of T-cells to normal tissues expressing the tumor-associated antigen (TAA) (off-tumor/on-target cytotoxicity), and, in some instances, neurotoxicity. Another key challenge is to arrive at a safe clinical starting dose and an efficient escalating strategy that allows patients in early dose cohorts the potential for clinical benefit in Phase 1 trials. To expand the therapeutic index and bring more treatment options to patients, there are intense efforts to overcome these challenges through improvements in molecular design, preclinical safety assessment strategies, and clinical management practices. A recent workshop at the U.S. Food and Drug Administration (FDA) with industry, academic, and regulatory agency representation was held to discuss the challenges and explore where such improvements to the development of CD3 bispecifics can be implemented. Here, the content of the presentations and the discussion that occurred during this workshop are summarized.

Brij-58, a potential injectable protein-stabilizer used in therapeutic protein formulation.

Polysorbates (PSs) are common protein stabilizers used in biotherapeutic formulations. However, PSs are heterogeneous and unstable in liquid protein formulations [1,2]. The purpose of this work is to explore possible alternatives for polysorbate replacements that demonstrate superior protein protection, superior self-stability, low toxicity, and wide applicability. For this purpose, 8 non-ionic surfactants that have not yet been used as excipients in marketed biotherapeutic products were investigated with PS20/80 as the benchmark. Compared with PS20/80, Brij-58 showed better protein protection ability in the mAb1 formulation under forced degradation conditions when examined by visual inspection, SEC, and dynamic lighting scanning. Additionally, Brij-58 has a better inherent stability than PS20/80 in the protein formulation when detected by UPLC-CAD. Moreover, Brij-58 is an inert excipient that does not affect protein bioactivity and conformation. In addition, the LD50 and hemolysis concentration of Brij-58 were determined, which is relatively safe when used as a parenteral injection. Furthermore, Brij-58 was also an effective protein stabilizer for the other two antibody products (IgG4 subtype and bispecific antibody) in the shaking study. In summary, Brij-58 stands out as a promising PS replacement in biotherapeutic formulations with a safe, stable and effective protein-protection profile among candidate surfactants.

A Regulatory Perspective on Testing of Biological Activity of Complex Biologics.

Antibody-drug conjugates (ADCs) and bispecific antibodies are becoming increasingly popular. However, their complex structures mandate stringent regulatory guidelines to ensure their safety and efficacy. We have briefly reviewed the existing regulatory guidelines and presented our perspectives on refining them to hasten the transition of these drugs from laboratories to market.

An FDA oncology analysis of CD3 bispecific constructs and first-in-human dose selection.

We retrospectively examined the nonclinical studies conducted with 17 CD3 bispecific constructs in support of first-in-human (FIH) trials in oncology. We also collected information on the design of dose-finding clinical trials. Sponsors have used different MABEL approaches for FIH dose selection. To better assess acceptable approaches, FIH doses were computed from nonclinical studies and compared to the maximum tolerated doses (MTDs) in patients, to the highest human doses (HHDs) when an MTD was not identified, or to the recommended human dose (RHD) for blinatumomab. We concluded that approaches based on receptor occupancy, highest non-severely toxic dose, or no-observed adverse effect level are not acceptable for selecting the FIH dose as they resulted in doses close to or above the MTDs, HHDs, or the RHD. A FIH dose corresponding to 10%-30% pharmacologic activity (PA) was an acceptable approach. A FIH dose corresponding to 50% PA was acceptable for all except one construct, potentially due to its biological or structural properties. The most common toxicities in animals and patients were those related to cytokine release. Doses were better tolerated when intra-animal or intra-patient dose escalation was used. Exposing naïve patients to an MTD achieved with intra-patient dose escalation design may be unsafe.

Leverage nonclinical development of bispecifics by translational science.

Bispecific antibody constructs (Bispecifics, bsAbs) may have greater functionality compared to established monoclonal antibodies because they bind to 2 different targets or, potentially, to 2 epitopes on the same target (dual targeting). This may result in enhanced binding avidity with preferential binding to cells that express both targets or binding to targets on different cells. However, development of these next-generation biologics, including new formats, creates unique challenges due to their increased complexity. Here we review aspects of bsAbs preclinical development programs that may increase the success rates of bsAbs in clinical development.

Preclinical evaluation of a diabody-based (177)Lu-radioimmunoconjugate for CD22-directed radioimmunotherapy in a non-Hodgkin lymphoma mouse model.

Radioimmunotherapy is considered as treatment option in recurrent and/or refractory B-cell non-Hodgkin lymphoma (B-NHL). To overcome the dose limiting bone marrow toxicity of IgG-based radioimmunoconjugates (RICs), we modified a humanized diabody with 5-, 10-, or 20-kDa polyethylene glycol (PEG) for CD22-targeted radioimmunotherapy using the low-energy ß-emitter lutetium-177 ((177)Lu). A favorable pharmacokinetic profile was observed for the 10-kDa-PEG-diabody in nude mice being xenografted with subcutaneous human Burkitt lymphoma. Even at high doses of 16 MBq this diabody RIC was well tolerated by NOD Rag1(null) IL2rγ(null) (NRG) mice and did not reveal signs of organ long-term toxicity 80 days post injection. Combination therapy of the diabody RIC with unconjugated anti-CD20 Rituximab demonstrated therapeutic efficacy in established disseminated mantle cell lymphoma xenograft models. When compared with the combination of the IgG formatted (177)Lu anti-CD22 antibody and Rituximab, dual targeted therapy with the diabody RIC achieved an improved reduction of disease burden in the first nine days following treatment. The data indicate that the PEGylated anti-CD22 diabody may have potential for extending the repertoire of radiopharmaceuticals for the treatment of patients with B-NHL.

Coupling Gd­DTPA with a bispecific, recombinant protein anti­EGFR­iRGD complex improves tumor targeting in MRI.

Recombinant anti­epidermal growth factor receptor­internalizing arginine­glycine­aspartic acid (anti­EGFR single­domain antibody fused with iRGD peptide) protein efficiently targets the EGFR extracellular domain and integrin αvß/ß5, and shows a high penetration into cells. Thus, this protein may improve penetration of conjugated drugs into the deep zone of gastric cancer multicellular 3D spheroids. In the present study, a novel tumor­targeting contrast agent for magnetic resonance imaging (MRI) was developed, by coupling gadolinium­diethylene triamine pentaacetate (Gd­DTPA) with the bispecific recombinant anti­EGFR­iRGD protein. The anti­EGFR­iRGD protein was extracted from Escherichia coli and Gd was loaded onto the recombinant protein by chelation using DTPA anhydride. Single­targeting agent anti­EGFR­DTPA­Gd, which served as the control, was also prepared. The results of the present study showed that anti­EGFR­iRGD­DTPA­Gd exhibited no significant cyto-toxicity to human gastric carcinoma cells (BGC­823) under the experimental conditions used. Compared with a conventional contrast agent (Magnevist), anti­EGFR­iRGD­DTPA­Gd showed higher T1 relaxivity (10.157/mM/sec at 3T) and better tumor­targeting ability. In addition, the signal intensity and the area under curve for the enhanced signal time in tumor, in vivo, were stronger than Gd­DTPA alone or the anti­EGFR­Gd control. Thus, Gd­labelled anti­EGFR­iRGD has potential as a tumor­targeting contrast agent for improved MRI.

The double-edged sword: Neurotoxicity of chemotherapy.

The number of available therapies for hematologic malignancies continues to grow at a rapid pace. Unfortunately, many of these treatments carry both central and peripheral nervous system toxicities, potentially limiting a patient's ability to tolerate a full course of treatment. Neurotoxicity with chemotherapy is common and second only to myelosuppression as a reason to limit dosing. This review addresses the neurotoxicity of newly available therapeutic agents including brentuximab vedotin and blinatumomab as well as classic ones such as methotrexate, vinca alkaloids and platinums. Although peripheral neuropathy is common with many drugs, other complications such as seizures and encephalopathy may require more immediate attention. Rapid recognition of adverse neurologic effects may lead to earlier treatment and appropriate adjustment of dosing regimens. In addition, knowledge of common toxicities may help differentiate chemotherapy-related symptoms from actual progression of cancer into the CNS.

A two-in-one antibody against HER3 and EGFR has superior inhibitory activity compared with monospecific antibodies.

Extensive crosstalk among ErbB/HER receptors suggests that blocking signaling from more than one family member may be essential to effectively treat cancer and limit drug resistance. We generated a conventional IgG molecule MEHD7945A with dual HER3/EGFR specificity by phage display engineering and used structural and mutational studies to understand how a single antigen recognition surface binds two epitopes with high affinity. As a human IgG1, MEHD7945A exhibited dual action by inhibiting EGFR- and HER3-mediated signaling in vitro and in vivo and the ability to engage immune effector functions. Compared with monospecific anti-HER antibodies, MEHD7945A was more broadly efficacious in multiple tumor models, showing that combined inhibition of EGFR and HER3 with a single antibody is beneficial.

Highly enhanced cytotoxicity of a dimeric bispecific diabody, the hEx3 tetrabody.

We previously reported the utility for cancer immunotherapy of a humanized bispecific diabody (hEx3) that targets epidermal growth factor receptor and CD3. Here, we used dynamic and static light scattering measurements to show that the multimer fraction observed in hEx3 in solution is a monodisperse tetramer. The multimerization into tetramers increased the inhibition of cancer cell growth by the hEx3 diabody. Furthermore, 1:2 stoichiometric binding for both antigens was observed in a thermodynamic analysis, indicating that the tetramer has bivalent binding activity for each target, and the structure may be in a circular configuration, as is the case for the single-chain Fv tetrabody. In addition to enhanced cytotoxicity, the functional affinity and stability of the hEx3 tetrabody were superior to those of the hEx3 diabody. The increase in molecular weight is also expected to improve the pharmacokinetics of the bispecific diabody, making the hEx3 tetrabody attractive as a therapeutic antibody fragment for cancer immunotherapy.

The trifunctional antibody ertumaxomab destroys tumor cells that express low levels of human epidermal growth factor receptor 2.

Human epidermal growth factor receptor 2 (HER2/neu) is an important target for the treatment of the breast cancers in which it is overexpressed. However, no approved anti-HER2/neu therapy is available for the majority of breast cancer patients, who express HER2/neu at low levels (with scores of 1+ or 2+/fluorescence in situ hybridization-negative). The trifunctional antibody ertumaxomab targets HER2/neu, CD3, and activating Fcgamma receptors. In presence of ertumaxomab, tri-cell complexes consisting of tumor cells, T cells, and accessory cells form to cause tumor cell lysis. In a phase I trial with metastatic breast cancer patients, ertumaxomab could be applied safely and resulted in radiographically confirmed clinical responses. In this study, we compare ertumaxomab- and trastuzumab-mediated killing of cancer cell lines that express HER2/neu at low and high levels. Under optimal conditions for trastuzumab-mediated destruction of HER2/neu-overexpressing cells, only ertumaxomab was able to mediate the elimination of tumor cell lines that express HER2/neu at low levels (1+). Ertumaxomab-mediated activity was accompanied by a Th1-based cytokine release, a unique mode of action of trifunctional antibodies. Competitive binding studies with trastuzumab and 520C9 mapped the binding site of ertumaxomab to the extracellular regions II and III of the HER2/neu ectodomain. This site is distinct from the binding site of trastuzumab, so that HER2/neu-expressing tumor cells can be eliminated by ertumaxomab in the presence of high amounts of trastuzumab. The ability of ertumaxomab to induce cytotoxicity against various tumor cell lines, including those with low HER2/neu antigen density, may provide a novel therapeutic option for breast cancer patients who are not eligible for trastuzumab treatment.

Bispecific anti-CD20/22 antibodies inhibit B-cell lymphoma proliferation by a unique mechanism of action.

Combination immunotherapy with anti-CD20 and anti-CD22 mAbs shows promising activity in non-Hodgkin lymphoma. Therefore, bispecific mAbs (bsAbs) were recombinantly constructed from veltuzumab (humanized anti-CD20) and epratuzumab (humanized anti-CD22) and evaluated in vitro and in vivo. While none of the parental mAbs alone or mixed had notable antiproliferative activity against Burkitt lymphoma cells when not cross-linked, the bsAbs [eg, anti-CD20 IgG-anti-CD22 (scFv)(2)] were inhibitory without cross-linking and synergistic with B-cell antigen (BCR)-mediated inhibition. The bsAbs demonstrated higher antibody-dependent cellulary cytoxicity (ADCC) activity than the parental mAbs, but not complement-dependent cytoxicity (CDC) of the parental CD20 mAb. Cross-linking both CD20 and CD22 with the bsAbs resulted in the prominent redistribution of not only CD20 but also CD22 and BCR into lipid rafts. Surprisingly, appreciable translocation of CD22 into lipid rafts was also observed after treatment with epratuzumab. Finally, the bsAbs inhibited Daudi lymphoma transplant growth, but showed a significant advantage over the parental anti-CD20 mAb only at the highest dose tested. These results suggest that recombinantly fused, complementary, bispecific, anti-CD20/22 antibodies exhibit functional features distinct from their parental antibodies, perhaps representing new candidate therapeutic molecules.

Use of anti-CD3 x anti-HER2/neu bispecific antibody for redirecting cytotoxicity of activated T cells toward HER2/neu+ tumors.

Relapse after adjuvant chemotherapy or high-dose chemotherapy with stem cell transplant for high-risk breast cancer remains high and new strategies that provide additional antitumor effects are needed. This report describes methods to generate highly effective HER2/neu-specific cytotoxic T cells by arming activated T cells with anti-CD3 x anti-HER2/neu bispecific antibody (BsAb). OKT3 and 9184 (anti-HER2) monoclonal antibodies (mAb) were conjugated and used to arm T cells that were subsequently tested in binding, cytotoxicity, and cytokine secretion assays. Armed T cells aggregated and specifically killed HER2/neu(+) breast cancer cells. Cytotoxicity emerged after 6 days of culture, was higher in armed T cells than unarmed T cells at all effector to target ratios (E/T) tested, and increased as the arming dose was increased. At an E/T of 20:1, the mean cytotoxicity of armed activated T cells (ATC) from 10 normal subjects increased by 59 +/- 11% (+/-SD) over that seen in unarmed ATC (p < 0.001) and the mean cytotoxicity of armed ATC from 6 cancer patients increased by 32 +/- 9% above that seen for unarmed ATC (p < 0.0004). After arming, the BsAb persisted on ATC up to 72 h and armed ATC continued to be cytotoxic up to 54 h. The amount of interferon-gamma (IFN-gamma), tumor necrosis factor-alpha (TNF-alpha), and granulocyte-macrophage colony-stimulating factor (GM-CSF) secreted was 1699, 922, and 3092 pg/ml/10(6) cells per 24 h, respectively, when armed T cells were exposed to a HER2/neu(+) breast carcinoma cell line. These studies show the feasibility and clinical adaptability of this approach for generating large numbers of anti-HER2-specific, cytotoxic T cells for clinical trials.

Increasing the affinity for tumor antigen enhances bispecific antibody cytotoxicity.

We tested the hypothesis that bispecific Abs (Bsab) with increased binding affinity for tumor Ags augment retargeted antitumor cytotoxicity. We report that an increase in the affinity of Bsab for the HER2/neu Ag correlates with an increase in the ability of the Bsab to promote retargeted cytotoxicity against HER2/neu-positive cell lines. A series of anti-HER2/neu extracellular domain-directed single-chain Fv fragments (scFv), ranging in affinity for HER2/neu from 10(-7) to 10(-11) M, were fused to the phage display-derived NM3E2 human scFV: NM3E2 associates with the extracellular domain of human FcgammaRIII (CD16). The resulting series of Bsab promoted cytotoxicity of SKOV3 human ovarian carcinoma cells overexpressing HER2/neu by human PBMC preparations containing CD16-positive NK cells. The affinity for HER2/neu clearly influenced the ability of the Bsab to promote cytotoxicity of (51)Cr-labeled SKOV3 cells. Lysis was 6.5% with an anti-HER2/neu K(D) = 1.7 x 10(-7) M, 14.5% with K(D) = 5.7 x 10(-9) M, and 21.3% with K(D) = 1.7 x 10(-10) M at 50:1 E:T ratios. These scFv-based Bsab did not cross-link receptors and induce leukocyte calcium mobilization in the absence of tumor cell engagement. Thus, these novel Bsab structures should not induce the dose-limiting cytokine release syndromes that have been observed in clinical trials with intact IgG BSAB: Additional manipulations in Bsab structure that improve selective tumor retention or facilitate the ability of Bsab to selectively cross-link tumor and effector cells at tumor sites should further improve the utility of this therapeutic strategy.

Depletion of CCR5-expressing cells with bispecific antibodies and chemokine toxins: a new strategy in the treatment of chronic inflammatory diseases and HIV.

The chemokine receptor CCR5 is expressed on the majority of T cells and monocytes in the inflammatory infiltrate of diseases such as rheumatoid arthritis, renal diseases, and multiple sclerosis. In contrast, little expression of CCR5 is found on peripheral blood leukocytes. A specific depletion of CCR5(+) cells could therefore be a useful strategy to reduce the cellular infiltrate in chronic inflammations. Moreover, CCR5 is the major coreceptor for M-tropic HIV-1 strains. Depletion of CCR5(+) leukocytes may help to eliminate cells latently infected with HIV-1. We designed two constructs that specifically destroy chemokine receptor-positive cells. The first construct, a bispecific Ab, binds simultaneously to CCR5 and CD3. Thereby it redirects CD3(+) T cells against CCR5(+) target cells. The Ab specifically depletes CCR5(+) T cells and monocytes, but is inactive against cells that do not express CCR5. Furthermore, ex vivo the bispecific Ab eliminated >95% of CCR5(+) monocytes and T cells from the synovial fluid of patients with arthritis. Also, we designed a fusion protein of the chemokine RANTES and a truncated version of Pseudomonas. exotoxin A. The fusion protein binds to CCR5 and down-modulates the receptor from the cell surface. The chemokine toxin completely destroyed CCR5(+) Chinese hamster ovary cells at a concentration of 10 nM, whereas no cytotoxic effect was detectable against CCR5(-) Chinese hamster ovary cells. Both constructs efficiently deplete CCR5-positive cells, appear as useful agents in the treatment of chronic inflammatory diseases, and may help to eradicate HIV-1 by increasing the turnover of latently infected cells.

Development and characterization of a bispecific single-chain antibody directed against T cells and ovarian carcinoma.

Bispecific antibodies with specificity for tumor antigen and CD3 have been shown to redirect the cytotoxicity of T cells against relevant tumor. Our objective was to generate single-chain bispecific antibodies (bsSCA) that could retarget mouse cytotoxic T lymphocytes (CTL) to destroy human ovarian carcinoma in a xenogeneic setting. A bsSCA, 2C11 x B43.13, was constructed by genetic engineering and expressed in mammalian cells. Molecular characteristics, binding properties, and ability to retarget CTL were studied. Western blot analysis showed that the product is a 65-kDa protein. Purification of antibodies could be done by single-step affinity chromatography using protein L-agarose with an unoptimized yield of 200 microg/L. BsSCA 2C11 x B43.13 was capable of binding to mouse CD3 and human CA125 as detected by FACS analysis of EL4 and OVCAR Nu3H2 cells, respectively. It could also bridge activated splenic T cells and human ovarian carcinoma as demonstrated by a bridge FACS assay. Redirected mouse CTL could mediate human target cell lysis in a 20-h 51Cr release assay despite that they are xenogeneic. Prolonged incubation of redirected CTL and tumor targets resulted in a dramatic reduction in tumor cell number. CD28 co-stimulation enhanced redirected CTL function in both types of assays. BsSCA 2C11 x B43.13 thus can be used as a preclinical immunotherapeutic model for human ovarian cancer in a xenogeneic setting.

Bispecific antibodies in cancer therapy.

Based upon in vitro and animal studies, a number of Phase I and II clinical trials have been initiated to test whether bispecific antibodies could redirect immune effectors against tumor cells in cancer patients. Recently, results from those trials showed beneficial effects in some patients but it is clear many problems remain to be solved. In addition, molecular engineering approaches are providing new and improved sources of clinically relevant bispecific antibodies.

Bispecific antibody-mediated lysis of primary cultures of ovarian carcinoma cells using multiple target antigens.

We have shown previously that a bispecific antibody (BsAb) directed against both germ-cell alkaline phosphatase (GCAP) and the CD3 complex on mouse T cells could effectively eliminate GCAP-positive tumor cells in vivo using an immunocompetent mouse model. However, some GCAP-negative tumor cells were still able to grow, suggesting that BsAb therapy, when used in a clinical setting, could benefit from targeting several tumor markers to prevent outgrowth of tumor cells lacking a targeted marker. To test this hypothesis, we developed an in vitro model based on primary human ovarian carcinoma (OC) cultures and BsAbs directed against human T cells and several tumor markers [placental alkaline phosphatase (PLAP), GCAP, folate-binding protein (FBP) and CA19.9]. OC cells, isolated from primary tumors, were co-cultured with human peripheral blood mononuclear cells in the presence or absence of various concentrations of BsAbs against PLAP/GCAP, FBP and CA19.9 administered separately or in combination. Results derived from 18 primary OC samples showed that the combination treatment was better than or equally effective as the best single BsAB treatment in 60% of cases. Sometimes targeting FBP, PLAP/GCAP or CA19.9 alone was superior to targeting all simultaneously. Combining each BsAb with a low dose of IL-2 was always beneficial. These results indicate that before using a specific BsAb in the clinic, it is important to determine the optimal BsAb for each patient using this in vitro assay on cells from the removed tumor mass.

The roles of Fas and perforin in LAK T-cell/bispecific antibody-mediated killing of the murine B-lymphoma cells, BCL1.

The aim of this study was to construct bispecific F(ab')2 [anti-CD3 x anti-BCL1 idiotype (Id)] Abs (BsAbs) which would enable lymphokine-activated killer (LAK) T cells to kill Id+ mouse BCL1 lymphoma cells, and to determine the mechanism(s) underlying cell death. Using 4-day activated LAK T cells from either perforin-knockout mice or FasL-deficient gld mice, we show that the Fas pathway, but not perforin, is required for BsAb-mediated LAK T-cell-induced killing of BCL1 cells.

Tumor-infiltrating lymphocytes can be activated in situ by using in vivo activants plus F(ab')2 bispecific antibodies.

In vitro-activated T lymphocytes can be retargeted with anti-CD3 x anti-tumor bispecific antibodies (BsAb) to kill tumor cells in vitro and in vivo. The purpose of the present study was to examine the systemic and intra-tumor effects of in vivo T-cell activation with BsAb, staphylococcal enterotoxin B (SEB), and beta-glucan in combination with BsAb as a retargeting agent. CL-62 melanoma cells were injected subcutaneously into C3H/ HeN mice. Mice were subsequently treated with BsAb alone or with SEB or beta-glucan plus BsAb. Fresh splenocytes, lymph-node cells and tumor-infiltrating lymphocytes (TIL) were tested for their proliferative response using incorporation of 3H-thymidine, and for their ability to lyse CL-62 cells in the presence or absence of BsAb in 4-hr 51Chromium release assays. Toxicity of treatments was assessed in a D-galactosamine model. BsAb, alone or combined with beta-glucan, had essentially no effect on systemic T-cell cytotoxicity, and essentially no effect on systemic proliferation, unless exogenous IL-2 was provided. At the tumor site, BsAb alone, BsAb plus beta-glucan, and SEB plus BsAb all significantly increased BsAb-mediated TIL cytotoxicity. In contrast to the TIL-limited effects of BsAb and of BsAb plus beta-glucan, SEB plus BsAb markedly increased both systemic and intra-tumor T-lymphocyte activation. Toxicity correlated with measures of systemic activation, with limited effects from BsAb alone and from beta-glucan plus BsAb, and with marked lethality from SEB plus BsAb. Overall, these results suggest moderate intra-tumor activation of TIL, but no measurable systemic activation after in vivo treatment with BsAb or beta-glucan plus BsAb. SEB plus BsAb results in complete T-cell activation in both systemic and intra-tumor compartments, but at the expense of increased systemic toxicity. In conclusion, TIL can be activated in situ with different combinations of in vivo activants. In vivo-activated TIL can be retargeted with bispecific antibodies to lyse tumor cells, and may be an alternative to ex vivo T-cell activation and adoptive transfer therapy.