For treatment of patients with prostate cancer (PCa), we developed a novel T cell-engaging (BiTE) antibody designated AMG 212 or BAY2010112 that is bispecific for prostate-specific membrane antigen (PSMA) and the CD3 epsilon subunit of the T cell receptor complex. AMG 212/BAY2010112 induced target cell-dependent activation and cytokine release of T cells, and efficiently redirected T cells for lysis of target cells. In addition to Chinese hamster ovary cells stably expressing human or cynomolgus monkey PSMA, T cells redirected by AMG 212/BAY2010112 also lysed human PCa cell lines VCaP, 22Rv1, MDA PCa 2b, C4-2, PC-3-huPSMA, and LnCaP at half maximal BiTE concentrations between 0.1 and 4 ng/mL (1.8–72 pmol/L). No lysis of PSMA-negative human PCa cell lines PC-3 and DU145 was observed. The subcutaneous (s.c.) formation of tumors from PC-3-huPSMA cells in NOD/SCID mice was significantly prevented by once daily intravenous (i.v.) injection of AMG 212/BAY2010112 at a dose level as low as 0.005 mg/kg/d. Rapid tumor shrinkage with complete remissions were observed in NOD/SCID mice bearing established s.c. 22Rv1 xenografts after repeated daily treatment with AMG 212/BAY2010112 by either the i.v. or s.c. route. Of note, 22Rv1 tumors were grown in the absence of human T cells followed by intraperitoneal injection of T cells 3 days before BiTE treatment. No effects on tumor growth were observed in the absence of human T cells or AMG 212/BAY2010112. On the basis of these preclinical results, AMG 212/BAY2010112 appears as a promising new BiTE antibody for the treatment of patients with PSMA-expressing PCa. Mol Cancer Ther; 11(12); 2664–73. ©2012 AACR.

Prostate cancer (PCa) is the most common type of cancer in men in the United States with estimated 192,000 new cases and 27,400 deaths in 2009 (1), and is the second leading cause of cancer death in men after lung cancer. Relapse after primary therapy by radical prostatectomy remains a common problem in PCa patients because an estimated 30% to 40% of men experience recurrence within 10 years (2, 3). There is a critical need for new targeted agents in this PCa patient population and for those patients who become hormone refractory and require chemotherapy.

Antibodies that target PSMA are a potential addition to the armentarium of PCa therapies. Developments comprise naked humanized or human IgG1 antibodies, and antibodies of enhanced activity such as antibody drug conjugates and radioimmunoconjugates (4). PSMA is also an established marker for imaging of residual PCa and metastases using the radiolabeled diagnostic antibody ProstaScint (5). The antigen is expressed in a high percentage of PCa patients across all stages of the disease (6). A highly promising approach to enhance the anticancer activity of antibodies is the engagement of T cells by using bispecific antibodies and related constructs. One example is blinatumomab, a CD19/CD3-bispecific antibody construct of the BiTE (for bispecific T cell engager) class that has shown high response rates and an acceptable benefit/risk profile in patients with B-cell lineage lymphoma (7) and leukemia (8). The group of Elsässer–Beile was first to design and characterize PSMA/CD3-bispecific antibodies of the so-called diabody format (9–11). While such bacterially expressed diabodies showed conditional T cell activation, redirected PCa cell lysis and activity against outgrowing xenografts in mouse models, their murine nature and highly restricted species cross-reactivity of the CD3 binding moiety greatly limit nonclinical safety assessment.

Here, we characterized AMG 212/BAY 2010112, the first PSMA/CD3-bispecific single-chain antibody construct that (i) is very close in its amino acid sequence to human germline Ig-V segments, (ii) is cross-reactive to human and non-human primate PSMA and CD3 antigens, (iii) is highly active against PCa tumor formation as well as established PCa xenograft tumors in a novel mouse model, and (iv) is meeting essential requirements as are desired for clinical development of novel antibody-based therapeutics.

Expression, production, and purification of AMG 212/BAY2010112

AMG 212/BAY2010112 is a single-chain bispecific antibody construct comprising of an anti-PSMA and an anti CD3 scFv, both reacting with human and cynomolgus targets. AMG 212/BAY2010112 was constructed by recombinant DNA technology and produced in supernatants from stably transfected Chinese hamster ovary (CHO) cells. Purification of the monomeric protein was done in a 2-step process using Ni chelate chromatogaraphy followed by gel filtration as described for an EGF receptor–specific BiTE antibody in ref. 12.

Affinity determination on target positive cell lines using Scatchard evaluation

For affinity analysis, saturation-binding experiments were conducted using a monovalent detection system (anti-His Fab/Alexa 488, Micromet GmbH) to determine monovalent binding of AMG 212/BAY2010112 to PSMA or CD3 positive cell lines. The human PCa cell line LNCaP, CHO cells transfected with human PSMA (CHO-huPSMA), CHO cells transfected with cynomolgus monkey PSMA (CHO-cyPSMA), the human T cell line HPB ALL and the rhesus monkey T cell line 4919 LnPx (kindly provided by Prof. Fickenscher, University Erlangen-Nuernberg, Germany; ref. 13) were used. Two to 5 × 104 cells of the respective cell line were incubated with 50 μL from a 1:3 dilution series (starting at 800 nmol/L) of AMG 212/BAY2010112, followed by 45 minutes of incubation on ice, and 1 residual washing step. Subsequently, cells were incubated for 30 minutes with 30 μL of an anti-His Fab/Alexa488 solution (30 μg/mL). After 1 washing step, cells were resuspended in 120 μL fluorescence-activated cell sorting (FACS) buffer and analyzed using FACS flow or a FACS Canto machine (Becton Dickinson), and Cell quest software (GraphPad Software). Data are means from 3 experiments from at least 2 different sets of experiments.

Epitope determination using linear peptide mapping

The PSMA-binding scFv of AMG 212/BAY2010112 was subjected to an epitope mapping approach based on peptide scanning (Pepscan). 693 different 15-mer peptides were synthesized that span the entire extracellular amino acid sequence of human PSMA and overlap with each neighboring 15-mer peptide by 14 amino acids. These peptides were synthesized directly in 455-well peptide arrays. For this experiment, the anti-PSMA scFv of AMG 212/BAY 2010112 was produced in Escherichia coli and crude periplasmic extracts were used for the analysis. The scFv was incubated on polypropylene cards containing the covalently linked peptides and specific binding was detected and calculated as firstly described in ref. 14.

Cell lines and cell culture

LNCaP, 22Rv1, PC-3, and CHO-dhfr- (CHO) were purchased from DSMZ (Germany), VcaP, and MDA PCa 2b from ATCC, and C4-2 from ViroMed Laboratories (Minnetonka, Minnesota), respectively. The cell lines were obtained before the experiment series and analyzed for authenticity by the respective cell bank (DNA-fingerprinting techniques such as short tandem repeat profiling). No additional identification was conducted in-house. The 22Rv1 and C4-2 were cultured in RPMI 1640 and 10% fetal calf serum (FCS; Invitrogen), LNCaP in Iscove's Basal Medium (Biochrom AG) and 10% FCS, VCaP in Dulbecco's Modified Eagle's Media (Biochrom AG), and 10% FCS. MDA PCa 2b cells were cultured in F-12 K-Nutrient Mixture (Invitrogen) and 20% FCS supplemented with 5 mL Insulin-Transferrin-Selenium-Solution (Invitrogen), 10 ng/mL recombinant human EGF (MBL), and 100 pg/mL hydrocortisone (Sigma-Aldrich). Cell lines were cultured at 37°C in a 5% CO2 chamber.

PC-3 were stably transfected with the human PSMA cDNA (PC-3-huPSMA) using the eukaryotic expression vector pEF-puromycin (15). CHO cells stably expressing human (CHO-huPSMA) or cynomolgus monkey PSMA (CHO-cyPSMA) were generated by transfecting cells with plasmids containing the respective cDNAs. Selection of clones and amplification of expression was conducted in the presence puromycin (0.8 μg/mL) and methotrexate (20 nmol/L).

PK analysis in mice

Female BALB/c mice received a single intravenous (i.v.) bolus injection of 0.1, 0.3, or 1 mg/kg AMG 212/BAY2010112 or were subcutaneous (s.c.) injected with a single dose of 0.3 mg/kg AMG 212/BAY2010112. Whole blood samples were collected at the following times: 0 (predose), 0.083, 0.167, 0.5, 1, 2, 4, 7, 24, and 48 hours after administration. AMG 212/BAY2010112 serum concentrations were determined by an electrochemiluminescence-based assay. A polyclonal goat anti-AMG 212/BAY2010112 serum was coated to a high binding plate (MSD). After blocking overnight with Dulbecco's Phosphate Buffered Saline containing 5% bovine serum albumin, plates were washed with an ELISA washer. Standards, quality-control samples, and study samples were added to the carbon plate and were subsequently incubated on a rotation shaker for 1 hour at room temperature. After another wash step with the ELISA washer, the detection anti-penta His-biotin-labeled detection antibody (GenScript) was added and incubated for 1 hour on a rotation shaker at room temperature. Streptavidin SulfoTag (MSD) was added and the plates were incubated for 1 hour at room temperature, followed by another wash step. Finally, reading buffer T (MSD) was added and the plates were analyzed at the Sector Imager 2400 according to the manufacturer's instructions. All samples were determined in triplicates. The pharmacokinetic parameters were calculated from the geometric mean plasma concentrations by noncompartmental analysis (16, 17) using the KinEx program version 3.0 (Bayer Pharma AG, 2010).

Cytotoxicity, T cell activation, and cytokine release

Redirected T cell cytotoxicity was assayed by flow cytometry using human and cynomolgus monkey peripheral blood mononuclear cells (PBMC) as effector cells and PSMA-positive human PCa cell lines or PSMA-transfected CHO cells as targets. PBMC were isolated by Ficoll density gradient centrifugation and if applicable enrichment for CD3+ T cells was conducted by using the Pan T Cell Isolation Kit (Miltenyi Biotech) according to the manufacturer's instructions. Effector cells were cocultivated with target cells labeled with Vybrant DiO (Invitrogen) at effector-to-target (E:T) cell ratios of 5:1 or 10:1 with increasing concentrations of AMG 212/BAY2010112 for 48 hours and cell lysis was monitored by nuclear uptake of propidium iodide by flow cytometry (12).

Activation of CD4+ and CD8+ T cell subsets was determined by using phycoerythrin (PE)-labeled anti-CD4 and allophycocyanine (APC)-Cy7-labeled anti-CD8 antibodies. PE-Cy7-conjugated anti-CD69 and APC-conjugated anti-CD25 were used to characterize activated T cells. All fluorescently labeled antibodies were purchased from BD Biosciences (Heidelberg). For staining of cell surface proteins, cells were incubated with the appropriate antibody mixtures for 30 minutes at 4°C and samples were analyzed by flow cytometry.

Cytokine concentrations were determined in supernatants of cytotoxicity assays by use of cytometric bead arrays (CBA-kit, BD Biosciences) in accordance to the manufacturer's protocol.

Animal tumor studies

In vivo studies to evaluate antitumor activity were conducted in, NOD.CB17-Prkdcscid/J mice (Charles River). Mice were housed in individually ventilated cages with laboratory chow and water available ad libitum. All experiments were conducted according to the German Animal Protection Law with permission from the responsible local authorities.

In the tumor formation study, 5 × 106 PC-3-huPSMA cells were injected s.c. with or without PBMC obtained from a single healthy human donor at an E:T cell ratio of 1:2 into the flank of female nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice (n = 8). Mice were treated once daily from the day of tumor/PBMC injection with 5, 0.5, 0.05 and 0.005 mg/kg AMG 212/BAY2010112 BiTE antibody for 10 days by i.v. bolus injection into the lateral tail vein. Controls included 2 vehicle-treated groups (tumor cells injected with or without human PBMC). Treatment efficacy in terms of tumor formation and growth was determined by external caliper measurements, and tumor volumes were calculated using a standard hemi-ellipsoid formula: {length (mm) × width (mm)2}/2.

For the established tumor model, male NOD/SCID mice were irradiated with a single dose (2 Gy) of 137Cs γ-ray before s.c. injection of 22Rv1 (5 × 106) tumor cells dispersed in medium with 50% Matrigel (Cultrex, R&D Systems) before injection. Mice were irradiated to facilitate engraftment of the injected human T cells. Animals (n = 8) were allocated to treatment groups when tumors had reached a volume of approximately 200 mm3 and were treated with a single i.v. bolus injection of anti-asialo GM1 rabbit antibody (WAKO) into the lateral tail vein 1 day before human T cell injection to deplete murine NK cells. In vitro activated and expanded human T cells isolated from the blood of a single donor (T cell activation/expansion kit, Miltenyi Biotech) were injected into the peritoneal cavity on day 11 after tumor cell injection. One control group of animals (n = 5) did not receive human T cells. Animals received AMG 212/BAY2010112 or vehicle once daily by either s.c., or i.v. bolus injection into the lateral tail vein for a total of 28 days starting at day 14 posttumor cell inoculation. Growth of tumors was determined as described above.

Statistical analysis

Data were analyzed by ANOVA, and differences in tumor volumes of the AMG 212/BAY2010112 BiTE-treated groups to the vehicle control group injected with human effector cells were assessed with a Dunnett's posttest. P < 0.05 was considered to be statistically significant. All statistical analyses were conducted using GraphPad Prism version 5.00 for Windows (GraphPad Software).

Generation of human PSMA/CD3-bispecific BiTE antibody AMG 212/BAY2010112

The PSMA/CD3-bispecific BiTE antibody AMG 212/BAY2010112 as shown in Fig. 1A was purified in larger amounts to apparent homogeneity (Fig. 1B). It showed one major protein species (95.6%) of defined charge upon analytical high-resolution cation exchange chromatography (Fig. 1C). Only the monomeric form of AMG 212/BAY2010112 was used for further in vitro and in vivo characterization. The BiTE antibody showed a high stability in human plasma at 37°C where after incubation for 24 hours EC50 values for redirected target cell lysis were not significantly different to untreated controls.

Figure 1.

Characteristics of purified AMG 212/BAY 2010112. The domain arrangement of AMG 212/BAY2010112 is schematically shown in A. Purified AMG 212/BAY2010112 was separated by SDS-PAGE and stained with Coomasie Blue (B) and subjected to high-resolution cation exchange chromatography (C) to show purity and charge homogeneity. The elution profile of purified AMG 212/BAY2010112 from the cation exchange column was recorded by absorption at 280 nm. The calculated molecular weight of AMG 212/BAY2010112 is 55 kDa. mAU, milli absorption units.

Figure 1.

Characteristics of purified AMG 212/BAY 2010112. The domain arrangement of AMG 212/BAY2010112 is schematically shown in A. Purified AMG 212/BAY2010112 was separated by SDS-PAGE and stained with Coomasie Blue (B) and subjected to high-resolution cation exchange chromatography (C) to show purity and charge homogeneity. The elution profile of purified AMG 212/BAY2010112 from the cation exchange column was recorded by absorption at 280 nm. The calculated molecular weight of AMG 212/BAY2010112 is 55 kDa. mAU, milli absorption units.

Close modal

Bispecific binding of AMG 212/BAY2010112 to CD3 and a defined epitope of PSMA

Saturation binding curves of AMG 212/BAY2010112 to CHO cells expressing either human or cynomolgus monkey PSMA and to T cells of human or cynomolgus monkey origin were used to determine equilibrium binding constants (KD) of AMG 212/BAY2010112. As shown in Fig. 2A, AMG 212/BAY2010112 bound to CD3 and PSMA of both species. KD values for human and cynomolgus monkey CD3 were 9.4 ± 4.3 nmol/L and 16.3 ± 2.1 nmol/L, respectively, and 47.0 ± 8.1 nmol/L and 212.6 ± 29 nmol/L for human and cynomolgus monkey PSMA, respectively.

Figure 2.

Binding of AMG 212/BAY2010112 to human and cynomolgus monkey PSMA and CD3 antigens. Increasing concentrations of monomeric AMG 212/BAY2010112 were incubated with 10,000 cells of the indicated cell lines. Cell-bound BiTE antibody was detected by flow cytometry using a monovalent antihexahistidine Fab antibody fragment covalently coupled to ALEXA 488. Binding signals are given as median relative fluorescence intensity (RFI). A, hyperbolic binding curves were plotted and half-maximal binding concentrations (KD) were calculated using the algorithm by George Scatchard. Exemplary binding curves are depicted. KD values with standard deviations were determined from 3 independent experiments per cell line. B, the epitope on human PSMA recognized by AMG 212/BAY2010112 was determined by binding of the anti-PSMA single-chain antibody moiety of AMG 212/BAY2010112 to a panel of 15-mer peptides permutated by 1 amino acid that spanned the entire extracellular domain of human PSMA. Binding signals for peptides in the ELISA assay are given as relative signal units. C, the position of the most reactive peptide is highlighted in the crystal structure of human PSMA (20).

Figure 2.

Binding of AMG 212/BAY2010112 to human and cynomolgus monkey PSMA and CD3 antigens. Increasing concentrations of monomeric AMG 212/BAY2010112 were incubated with 10,000 cells of the indicated cell lines. Cell-bound BiTE antibody was detected by flow cytometry using a monovalent antihexahistidine Fab antibody fragment covalently coupled to ALEXA 488. Binding signals are given as median relative fluorescence intensity (RFI). A, hyperbolic binding curves were plotted and half-maximal binding concentrations (KD) were calculated using the algorithm by George Scatchard. Exemplary binding curves are depicted. KD values with standard deviations were determined from 3 independent experiments per cell line. B, the epitope on human PSMA recognized by AMG 212/BAY2010112 was determined by binding of the anti-PSMA single-chain antibody moiety of AMG 212/BAY2010112 to a panel of 15-mer peptides permutated by 1 amino acid that spanned the entire extracellular domain of human PSMA. Binding signals for peptides in the ELISA assay are given as relative signal units. C, the position of the most reactive peptide is highlighted in the crystal structure of human PSMA (20).

Close modal

To map the binding site of AMG 212/BAY2010112 to human PSMA, overlapping peptides of the extracellular domain of PSMA were synthesized and tested for binding of AMG 212/BAY2010112 using a PepBlot (18, 19). The highest binding signal of AMG 212/BAY2010112 was seen with a peptide corresponding to amino acid residues 325 to 348 of human PSMA (Fig. 2B). According to the published crystal structure of PSMA (20), this peptide localizes to a small loop structure on the apical side of the protein facing away from the plasma membrane (Fig. 2C). The sequence of this peptide is identical between human and cynomolgus monkey PSMA, which explains the cross-species binding of AMG 212/BAY2010112 to PSMA.

Human/cynomolgus monkey cross-reactivity of AMG 212/BAY2010112 for T cell activation, cytokine release, and redirected lysis

The pharmacologic effects of AMG 212/BAY2010112 on effector T cells and PSMA-expressing target cells were analyzed in vitro using cell coculture assays. AMG 212/BAY2010112 activity was compared in both human and cynomolgus monkey cell-based assay systems to investigate whether the cynomolgus monkey can serve as relevant species for pharmacologic and toxicological safety assessments.

The dose-dependent activation of CD4+ and CD8+ T cell populations by AMG 212/BAY2010112 was analyzed using FACS by the appearance of the immediate-early and late activation markers CD69 and CD25, respectively, after 48 hours of incubation in the presence of human LNCaP PCa cells (Fig. 3A). At the maximum, between 60% to 80% of CD4+ and CD8+ T cells of both species were activated, whereas less than 10% of T cells were found activated when freshly isolated. Values for half-maximal effective concentration (EC50) of T cell activation ranged between 3.4 and 6.7 ng/mL of AMG 212/BAY2010112 for human and between 13.7 and 21.2 ng/mL for cynomolgus monkey cell cocultures. The largest difference between human and cynomolgus monkey T cells was observed with the CD4+/CD25+ T cell population where EC50 values differed by 6.2-fold. No T cell activation was observed in the absence of PSMA-expressing target cells (data not shown).

Figure 3.

Effects of AMG 212/BAY2010112 on T cell activation, cytokine release, and redirected lysis in coculture experiments. Cocultures of LNCaP human PCa cells and isolated human or cynomolgus monkey CD3+ T cells at effector-to-target (E:T) cell ratios of 5:1 (human) or 10:1 (cynomolgus monkey) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. A, T cell activation. The percentage of activated CD4+ and CD8+ T cells was determined by flow cytometry using antigen-specific fluorochrome-conjugated monoclonal antibodies against CD69 and CD25. B, cytokine release. Concentrations of the indicated human and cynomolgus monkey cytokines in the cell culture supernatants were determined using cytometry-based bead arrays. C, redirected lysis of target cells. CHO cells transfected with either human or cynomolgus monkey PSMA or untransfected cells and isolated human or cynomolgus monkey CD3+ T cells (E:T cell ratio: 10:1) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. The percentage of target cell lysis was monitored by flow cytometric determination of propidium iodine uptake into nuclei of lysed cells. Data points in all assays represent the mean of duplicate wells. Error bars represent standard deviation of the mean. Concentrations that induced half maximal T cell activation and target cell lysis (EC50), respectively, are shown.

Figure 3.

Effects of AMG 212/BAY2010112 on T cell activation, cytokine release, and redirected lysis in coculture experiments. Cocultures of LNCaP human PCa cells and isolated human or cynomolgus monkey CD3+ T cells at effector-to-target (E:T) cell ratios of 5:1 (human) or 10:1 (cynomolgus monkey) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. A, T cell activation. The percentage of activated CD4+ and CD8+ T cells was determined by flow cytometry using antigen-specific fluorochrome-conjugated monoclonal antibodies against CD69 and CD25. B, cytokine release. Concentrations of the indicated human and cynomolgus monkey cytokines in the cell culture supernatants were determined using cytometry-based bead arrays. C, redirected lysis of target cells. CHO cells transfected with either human or cynomolgus monkey PSMA or untransfected cells and isolated human or cynomolgus monkey CD3+ T cells (E:T cell ratio: 10:1) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. The percentage of target cell lysis was monitored by flow cytometric determination of propidium iodine uptake into nuclei of lysed cells. Data points in all assays represent the mean of duplicate wells. Error bars represent standard deviation of the mean. Concentrations that induced half maximal T cell activation and target cell lysis (EC50), respectively, are shown.

Close modal

Next, the AMG 212/BAY2010112 dose-dependent release of 6 different cytokines by activated T cells in the presence of target cells was compared. The cytokine release profiles were very similar in human and cynomolgus monkey cell coculture systems with absolute amounts of cytokines reaching approximately 2-fold higher levels in the human system. Highest concentrations were observed for interferon-γ, which reached 3 to 6 ng/mL in cell culture medium after 48 hours (Fig. 3B). Intermediate increases reaching levels of 1 to 1.5 ng/mL were observed for TNF-α, IL-2 and IL-10. For technical reasons, IL-10 levels could not be determined in cynomolgus monkey cocultures, whereas IL-5 levels were not detected in human cocultures. Little or no increase was observed for IL-4 and IL-6. As observed previously, cytokine release of T cells was observed only in the presence of target cells (data not shown), which is one hallmark of BiTE antibodies (21).

The dose-dependent redirected lysis of target cells by AMG 212/BAY2010112 was investigated with CHO cells, which expressed either human or cynomolgus monkey PSMA, to have the same cellular background of target cells. Overall, potent lysis of transfected CHO cells was observed with EC50 values ranging across species between 0.37 and 6.69 ng/mL (Fig. 3C). In homologous assays where T cells and PSMA from the same species were used, an 18-fold higher concentration of AMG 212/BAY2010112 was required to reach half maximal lysis in the cynomolgus monkey system compared with the human system. Notably, target cells expressing human PSMA were more potently lysed by either human or cynomolgus monkey T cells than those expressing cynomolgus monkey PSMA, which may be attributed to the somewhat lower affinity of AMG 212/BAY2010112 to cynomolgus monkey PSMA (Fig. 2A). Human T cells showed a 5-fold higher potency of redirected lysis with AMG 212/BAY2010112 than cynomolgus monkey T cells, regardless whether human tumor cells or CHO cells expressing either human or cynomolgus monkey PSMA were used. Taken together, AMG 212/BAY2010112 showed dose-dependent activity for T cell activation, cytokine release and redirected lysis with human and cynomolgus monkey T cells, and reacted with PSMA antigen from both species.

Redirected lysis of human PCa lines by AMG 212/BAY2010112

To assess the potency of redirected lysis by AMG 212/BAY2010112 with target cells naturally expressing PSMA, the 7 human PCa cell lines VCaP, 22Rv1, MDA PCa 2b, C4-2, PC-3, DU145, and LNCaP were procured. FACS staining showed that, except for PC-3 and DU145 cells, all PCa cell lines expressed detectable levels of PSMA. Surface densities of PSMA were quantified and ranged from 1.2 × 104 (VCaP) to 5 × 105 (PC-3-huPSMA) PSMA molecules per cell (see Fig. 4B). Target cells were incubated in the presence of increasing concentrations of AMG 212/BAY2010112 with a 10-fold excess of unstimulated PBMC as effector cells and lysis was measured after 48 hours by nuclear uptake of propidium iodine via FACS. All 5 PCa cell lines, which naturally express PSMA as well as PC-3-huPSMA cells (stably expressing a human PSMA cDNA), were lysed with EC50 values ranging between 0.1 and 4 ng/mL (i.e., 1.8–64 pmol/L) AMG 212/BAY2010112 (Fig. 4A). No lysis of parental PSMAnegative PC-3 cells was observed. EC50 values for redirected lysis correlated with the number of surface expressed PSMA molecules per cell in a statistically significant manner (Fig. 4B).

Figure 4.

Redirected lysis of human PCa cell lines by AMG 212/BAY2010112 using unstimulated human effector cells. A, the human PCa cell lines VCaP, 22Rv1, MDA PCa 2b, C4-2, LNCaP, PC-3-huPSMA, and PC-3 cells and human PBMC (E:T cell ratio: 10:1) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. The percentage of tumor cell lysis was measured by uptake of propidium iodide using flow cytometry. Each data point represents the mean of duplicate wells. Error bars represent standard error of mean. B, concentrations that induced half maximal target cell lysis (EC50) were plotted against mean number of PSMA molecules on the surface of the respective human PCa cell line.

Figure 4.

Redirected lysis of human PCa cell lines by AMG 212/BAY2010112 using unstimulated human effector cells. A, the human PCa cell lines VCaP, 22Rv1, MDA PCa 2b, C4-2, LNCaP, PC-3-huPSMA, and PC-3 cells and human PBMC (E:T cell ratio: 10:1) were incubated with increasing concentrations of AMG 212/BAY2010112 for 48 hours. The percentage of tumor cell lysis was measured by uptake of propidium iodide using flow cytometry. Each data point represents the mean of duplicate wells. Error bars represent standard error of mean. B, concentrations that induced half maximal target cell lysis (EC50) were plotted against mean number of PSMA molecules on the surface of the respective human PCa cell line.

Close modal

Pharmacokinetics of AMG 212/BAY2010112 in BALB/c mice

Pharmacokinetic parameters of AMG 212/BAY2010112 were determined after i.v. bolus administration of 0.1, 0.3, and 1 mg/kg to BALB/c mice. A dose proportional increase of serum concentrations was observed. Serum clearance was low and amounted to approximately 0.3 L/(kg·h) independently of the dose. An intermediate volume of distribution at steady state of approximately 1 L/kg was observed. This indicates that the BiTE antibody did not only distribute into the extravascular space (as observed for IgGs) but also into tissues, which resulted in a considerably extended terminal half-life of approximately 8 hours.

After s.c. administration of 0.3 mg/kg to mice, maximum serum concentrations of approximately 64 μg/L were reached 2 hours after administration. Bioavailability was 18%, indicating significant metabolism during absorption via the lymphatic system. Pharmacokinetic parameters after i.v. and s.c. bolus administration to mice are summarized in Table 1.

Table 1.

Pharmacokinetic parameters of AMG 212/BAY2010112 after i.v. and s.c. administration to BALB/C mice

Administrationi.v. boluss.c. bolus
Dose(mg/kg)0.100.301.00.30
AUC (mg h/L) 0.33 0.93 3.3 0.17 
AUCnorm (kg h/L) 3.3 3.1 3.3 0.55 
CLmatrix (L/h/kg) 0.30 0.32 0.30 n.a. 
Cmax norm (kg/L) 10.4 11 9.5 0.21 
Vss (L/kg) 0.89 1.2 0.94 n.a. 
t1/2 (h) 6.5 9.7 7.9 11 
F (bioavailability) (%) 100 100 100 18 
Administrationi.v. boluss.c. bolus
Dose(mg/kg)0.100.301.00.30
AUC (mg h/L) 0.33 0.93 3.3 0.17 
AUCnorm (kg h/L) 3.3 3.1 3.3 0.55 
CLmatrix (L/h/kg) 0.30 0.32 0.30 n.a. 
Cmax norm (kg/L) 10.4 11 9.5 0.21 
Vss (L/kg) 0.89 1.2 0.94 n.a. 
t1/2 (h) 6.5 9.7 7.9 11 
F (bioavailability) (%) 100 100 100 18 

NOTE: Serum levels of AMG 212/BAY2010112 after single bolus i.v. administration of 0.1, 0.3, and 1 mg/kg and after single bolus s.c. administration of 0.3 mg/kg to female BALB/c mice were determined as described in Materials and Methods.

Abbreviations: AUC, area under the curve; AUCnorm; normalized AUC; CLmatrix, serum clearance; Cmax norm, maximum serum levels; Vss, volume of distribution at steady state; t1/2, terminal serum half-life (beta); n.a., not applicable.

Inhibition of tumor growth and tumor regression by AMG 212/BAY2010112-redirected human T cells in immunodeficient mice

Following the characterization of AMG 212/BAY2010112 in vitro, its efficacy on the inhibition of tumor growth was investigated in mouse xenograft models. AMG 212/BAY2010112 was first studied for its preventive effect in a tumor outgrowth model. PC-3-huPSMA cells were mixed with unstimulated human PBMC shortly before injection and mixtures were administered s.c. to immunodeficient NOD/SCID mice. Immediately following tumor cell inoculation, mice were treated i.v. with AMG 212/BAY2010112 at doses ranging from 5 mg/kg down to 0.005 mg/kg by once daily injection for 10 consecutive days. Efficacy was evaluated by monitoring tumor formation. A significant and dose-dependent delay of tumor formation/growth was observed with tumor growth inhibition after a monitoring period of 43 days ranging from 86% (0.005 mg/kg/d) to 99% (5 mg/kg/d) compared with the vehicle plus PBMC control (Fig. 5A). The inhibition was statistically significant even at the lowest dose of 0.005 mg/kg, which corresponded to only 0.1 ng AMG 212/BAY2010112 per mouse/d. AMG 212/BAY2010112 completely prevented tumor formation in 1 and 2 of 8 animals at doses of 0.5 or 5 mg/kg/d, respectively, whereas small nodules became visible in some animals on Days 46 and 53, respectively.

Figure 5.

Antitumor activity of AMG 212/BAY2010112 in mouse xenograft models. A, PC-3-huPSMA tumor formation model. PC-3-huPSMA, human prostate carcinoma cells stably transfected with a human PSMA-encoding cDNA, were injected s.c. with or without PBMC at, as applicable, an E:T cell ratio of 1:2 into the right dorsal flank of female NOD/SCID mice. Beginning at the time of tumor cell implantation, mice were injected once daily i.v. with the indicated doses of AMG 212/BAY2010112 BiTE or with vehicle for 10 days. Tumor growth was determined as described in Materials and Methods. Values represent mean tumor volume (cm3) +/− SEM (n = 8). Inhibition of tumor formation was statistically significant in all AMG 212/BAY2010112-treated groups from day 8 onward when compared with vehicle-treated animals that had received PBMCs (P < 0.001; 1-way ANOVA with Dunnett's posttest). B, established 22Rv1 PCa tumor model. 22Rv1 PCa xenografts were grown subcutaneously for 11 days until they reached a volume of approximately 180 to 200 mm3. Then mice were injected with activated and ex vivo expanded human T cells into the peritoneal cavity except for 5 animals that served as control. Three days later (day 14), at which tumor volumes had reached approximately 220 mm3 without significant deviations among study groups, mice were treated once daily for 28 days either i.v. or s.c. with the indicated doses of AMG 212/BAY2010112 BiTE or with vehicle. Tumor growth was determined by external caliper measurements and tumor volumes were calculated using a standard hemiellipsoid formula. Values shown represent mean tumor volume [cm3] +/− SEM (n = 8 per group). Inhibition of tumor growth was statistically significant in the AMG 212/BAY2010112 dose groups compared with the respective vehicle-treated animal that had received T cells [P < 0.01; days 15 and 16 (s.c.); day 16 (i.v.); P < 0.001; day 17 onward (s.c. and i.v.); 1-way ANOVA with Dunnett's posttest; ‡, no tumors were visible at necropsy in 3 animals of the s.c.-treated group.

Figure 5.

Antitumor activity of AMG 212/BAY2010112 in mouse xenograft models. A, PC-3-huPSMA tumor formation model. PC-3-huPSMA, human prostate carcinoma cells stably transfected with a human PSMA-encoding cDNA, were injected s.c. with or without PBMC at, as applicable, an E:T cell ratio of 1:2 into the right dorsal flank of female NOD/SCID mice. Beginning at the time of tumor cell implantation, mice were injected once daily i.v. with the indicated doses of AMG 212/BAY2010112 BiTE or with vehicle for 10 days. Tumor growth was determined as described in Materials and Methods. Values represent mean tumor volume (cm3) +/− SEM (n = 8). Inhibition of tumor formation was statistically significant in all AMG 212/BAY2010112-treated groups from day 8 onward when compared with vehicle-treated animals that had received PBMCs (P < 0.001; 1-way ANOVA with Dunnett's posttest). B, established 22Rv1 PCa tumor model. 22Rv1 PCa xenografts were grown subcutaneously for 11 days until they reached a volume of approximately 180 to 200 mm3. Then mice were injected with activated and ex vivo expanded human T cells into the peritoneal cavity except for 5 animals that served as control. Three days later (day 14), at which tumor volumes had reached approximately 220 mm3 without significant deviations among study groups, mice were treated once daily for 28 days either i.v. or s.c. with the indicated doses of AMG 212/BAY2010112 BiTE or with vehicle. Tumor growth was determined by external caliper measurements and tumor volumes were calculated using a standard hemiellipsoid formula. Values shown represent mean tumor volume [cm3] +/− SEM (n = 8 per group). Inhibition of tumor growth was statistically significant in the AMG 212/BAY2010112 dose groups compared with the respective vehicle-treated animal that had received T cells [P < 0.01; days 15 and 16 (s.c.); day 16 (i.v.); P < 0.001; day 17 onward (s.c. and i.v.); 1-way ANOVA with Dunnett's posttest; ‡, no tumors were visible at necropsy in 3 animals of the s.c.-treated group.

Close modal

We next addressed the question whether AMG 212/BAY2010112 could affect established PCa xenografts. A new study design was established in which human effector T cells were delivered shortly before initiation of BiTE treatment by intraperitoneal (i.p.) injection into tumor-bearing NOD/SCID mice. Previous experiments have shown that i.p.-delivered human T cells enter the circulation of mice and remain detectable for approximately 2 weeks (22). For the present study, human 22Rv1 PCa xenografts were grown s.c. in male NOD/SCID in the absence of human T cells until they had reached a volume of 180 to 200 mm3. Then, human T cells expanded in vitro for 18 days were injected i.p. and treatment with AMG 212/BAY2010112 was initiated 3 days after the adoptive T cell transfer by once daily i.v. or s.c. injections with doses of 0.5 or 2.5 mg/kg, respectively. To achieve similar drug exposures for i.v. and s.c. administrations a 5 times higher dose was given s.c. as based on the 18% bioavailability of AMG 212/BAY2010112 after s.c administration (see Table 1). Established tumors began to readily regress after initiation of either i.v. or s.c. treatment and tumor volumes maximally declined within the first week of treatment (Fig. 5B). Complete tumor remissions were observed in 3/8 animals of the s.c.-treated group after necropsy, whereas small nodules with sizes less than 50 mm3 were detected in the residual animals. In the i.v.-treated group small nodules of less than 50 mm3, most likely representing fibrous tissue remnants, were detected in 3/8 animals, whereas small tumors remained in the other animals. Unrestricted tumor growth occurred in animals treated with human T cells alone or a vehicle. In summary AMG 212/BAY2010112 very potently inhibited growth of PSMA expressing human PCa xenografts in mice, which eventually led to complete remission in some animals after s.c. application.

In the present study, the anti-PSMA/anti-CD3 single-chain antibody construct AMG 212/BAY2010112 was examined for its in vitro and in vivo pharmacologic properties and its suitability for nonclinical safety studies with non-human primates.

Cell binding studies revealed that CD3 antigens of human and cynomolgus monkey origin were bound by AMG 212/BAY2010112 with similar equilibrium dissociation constants in the nanomole per liter range. The differences seen can be compensated in respective toxicology studies by studying higher dose levels of the BiTE antibody in cynomolgus monkeys. Importantly, qualitative responses of human and cynomolgus monkey T cells to AMG 212/BAY2010112 were largely comparable with respect to the magnitude of target cell lysis, T cell activation, and cytokine release, supporting the use of cynomolgus monkeys as a relevant species for the nonclinical safety assessment of AMG 212/BAY2010112. Respective nonclinical safety studies are currently ongoing. The first in human study (phase 1) is planned in patients with advanced castration-resistant PCa but expect AMG 212/BAY2010112 to be beneficial also for patients in earlier stages of the disease.

Previous work has provided in vitro proof of concept that PSMA is a useful target for lysis of PSMA-expressing PCa cells by redirected T cells using bispecific diabodies (9–11). These are produced in E. coli, are of murine sequence and their CD3-binding moiety, which is related to the anti-CD3 antibody OKT-3, does not react with T cells from cynomolgus monkeys (23). Moreover, information of such diabodies concerning pharmaceutical properties such as productivity, serum stability, aggregation behavior, and charge homogeneity is scarce. Using the clinically validated BiTE format, we generated a molecule, which appears for a number of reasons better suited for development as a human therapeutic than currently existing diabodies: (i) we used anti-PSMA and anti-CD3 single chain antibodies that are very close in their amino acid sequence to human germline Ig-V segments, an established feature reducing potential immunogenicity of antibodies. (ii) We used single-chain antibodies that bind to epitopes shared by human and monkey target antigens providing the basis for nonclinical safety assessment of AMG 212/BAY2010112 in a pharmacologically relevant species. (iii) AMG 212/BAY2010112 was selected from a panel of candidate molecules as lead candidate for its superior activity and CMC related properties. Activity-related selection parameters comprised the cytotoxic activity with activated human T cells and unstimulated human PBMC as effector cells directed against human PSMA transfected CHO cells and human PCa cell lines expressing native PSMA. Cytotoxic activity in the macaque system was assessed with macaque PSMA transfected CHO cells and a macaque T cell line. CMC-related selection parameters comprised production yield, protein homogeneity in high-resolution ion exchange chromatography, stability in human plasma, thermal stability, and dimer conversion. Compared with published PSMA/CD3-bispecific diabodies, which have EC50 values for redirected lysis of approximately 0.25 nmol/L (15 ng/mL; refs. 9, 10), AMG 212/BAY2010112 is considerably more potent with EC50 values ranging from 1.8 to 72 pmol/L (0.1 and 4 ng/mL) depending on the cell line. Only recently, a newly constructed diabody was published with an EC50 value of approximately 23 pmol/L (1.4 ng/mL) using C4-2 target cells (11). (iv) Finally, we observed that AMG 212/BAY2010112 can mediate remission of 22Rv1 PCa xenografts, whereas published PSMA/CD3-bispecific diabodies were thus far only shown to inhibit tumor formation. On the basis of all of the above features, we believe that AMG 212/BAY2010112 meets all properties required for development of a potent novel human therapeutic.

This is also the first report on a novel in vivo study design for analyzing the activity of T cell-engaging antibodies against established tumors in mice. Most importantly, tumors were grown s.c. to an average size of 200 mm3 in NOD/SCID mice in the absence of human T cells. Ex vivo expanded human T cells were then injected into the peritoneal space of mice from where they had to enter the peripheral blood compartment to reach the s.c. implanted tumor. Three days following adoptive T cell transfer, dosing of AMG 212/BAY2010112 was started, which triggered a rapid onset of tumor regression. This is consistent with an immediate engagement of peripheral T cells for redirected tumor lysis. The fast onset is different from vaccines and anti-CTLA-4 antibodies, which require days to weeks for mounting a tumor-specific T cell response in man (22). Rapid tumor regression also suggests that both AMG 212/BAY2010112 and i.p.-delivered T cells readily gained access to tumor tissue. In our present study, the s.c. administration of the BiTE antibody appeared to be even more efficacious than the i.v. route.

The bioavailability of AMG 212/BAY2010112 in mice after administration of a single s.c. dose was 18%. Preliminary studies indicate that AMG 212/BAY2010112 is bioavailable and pharmacologically active after s.c. application in cynomolgus monkeys. Additional studies are in progress to better understand the activity of AMG 212/BAY2010112 after s.c. administration and evaluate its applicability for treatment of PCa patients.

M. Friedrich, T. Raum, R. Lutterbuese, M. Voelkel, P. Deegen, D. Rau, R. Kischel, P. Hoffmann, C. Brandl, P. Mueller, M. Fuergut, P.A. Baeuerle, B. Rattel, and P. Kufer are Amgen employees and have received stock. T. Raum, R. Lutterbuese, R. Kischel, P. Hoffmann, P. Kufer have ownership interest (including patents). J. Schuhmacher, R. Finnern, and D. Zopf are Bayer Pharma employees. J.W. Slootstra is employed by Pepscan Presto.

Conception and design: R. Lutterbuese, R. Finnern, P.A. Baeuerle, B. Rattel, P. Kufer

Development of methodology: C. Brandl

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Raum, M. Voelkel, P. Deegen, D. Rau, R. Kischel, C. Brandl, M. Fuergut, J.W. Slootstra

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Friedrich, T. Raum, R. Lutterbuese, P. Deegen, P. Hoffmann, J. Schuhmacher, R. Finnern, M. Fuergut, B. Rattel, P. Kufer

Writing, review, and/or revision of the manuscript: M. Friedrich, T. Raum, R. Lutterbuese, P. Deegen, P. Hoffmann, J. Schuhmacher, P. Mueller, R. Finnern, D. Zopf, P.A. Baeuerle, B. Rattel, P. Kufer

Study supervision: M. Friedrich, B. Rattel, P. Kufer

Project Management Support: P. Mueller

The authors thank Maren Voges for preparing figures.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Jemal
A
,
Siegel
R
,
Ward
E
,
Hao
Y
,
Xu
J
,
Thun
MJ
. 
Cancer statistics, 2009
.
CA Cancer J Clin
2009
;
59
:
225
49
.
2.
Simmons
MN
,
Stephenson
AJ
,
Klein
EA
. 
Natural history of biochemical recurrence after radical prostatectomy: risk assessment for secondary therapy
.
Eur Urol
2007
;
51
:
1175
84
.
3.
Ward
JF
,
Moul
JW
. 
Rising prostate-specific antigen after primary prostate cancer therapy
.
Nat Clin Pract Urol
2005
;
2
:
174
82
.
4.
Elsasser-Beile
U
,
Buhler
P
,
Wolf
P
. 
Targeted therapies for prostate cancer against the prostate specific membrane antigen
.
Curr Drug Targets
2009
;
10
:
118
25
.
5.
Mohammed
AA
,
Shergill
IS
,
Vandal
MT
,
Gujral
SS
. 
ProstaScint and its role in the diagnosis of prostate cancer
.
Expert Rev Mol Diagn
2007
;
7
:
345
9
.
6.
Joniau
S
,
Abrahamsson
PA
,
Bellmunt
J
,
Figdor
C
,
Hamdy
F
,
Verhagen
P
, et al
Current vaccination strategies for prostate cancer
.
Eur Urol
2012
;
61
:
290
306
.
7.
Bargou
R
,
Leo
E
,
Zugmaier
G
,
Klinger
M
,
Goebeler
M
,
Knop
S
, et al
Tumor regression in cancer patients by very low doses of a T cell-engaging antibody
.
Science
2008
;
321
:
974
7
.
8.
Topp
MS
,
Kufer
P
,
Gokbuget
N
,
Goebeler
M
,
Klinger
M
,
Neumann
S
, et al
Targeted therapy with the T-cell-engaging antibody blinatumomab of chemotherapy-refractory minimal residual disease in B-lineage acute lymphoblastic leukemia patients results in high response rate and prolonged leukemia-free survival
.
J Clin Oncol
2011
;
29
:
2493
8
.
9.
Buhler
P
,
Wolf
P
,
Gierschner
D
,
Schaber
I
,
Katzenwadel
A
,
Schultze-Seemann
W
, et al
A bispecific diabody directed against prostate-specific membrane antigen and CD3 induces T-cell mediated lysis of prostate cancer cells
.
Cancer Immunol Immunother
2008
;
57
:
43
52
.
10.
Buhler
P
,
Molnar
E
,
Dopfer
EP
,
Wolf
P
,
Gierschner
D
,
Wetterauer
U
, et al
Target-dependent T-cell activation by coligation with a PSMA x CD3 diabody induces lysis of prostate cancer cells
.
J Immunother
2009
;
32
:
565
73
.
11.
Fortmuller
K
,
Alt
K
,
Gierschner
D
,
Wolf
P
,
Baum
V
,
Freudenberg
N
, et al
Effective targeting of prostate cancer by lymphocytes redirected by a PSMA x CD3 bispecific single-chain diabody
.
Prostate
2011
;
71
:
588
96
.
12.
Lutterbuese
R
,
Raum
T
,
Kischel
R
,
Hoffmann
P
,
Mangold
S
,
Rattel
B
, et al
T cell-engaging BiTE antibodies specific for EGFR potently eliminate KRAS- and BRAF-mutated colorectal cancer cells
.
Proc Natl Acad Sci U S A
2010
;
107
:
12605
10
.
13.
Knappe
A
,
Feldmann
G
,
Dittmer
U
,
Meinl
E
,
Nisslein
T
,
Wittmann
S
, et al
Herpesvirus saimiri-transformed macaque T cells are tolerated and do not cause lymphoma after autologous reinfusion
.
Blood
2000
;
95
:
3256
61
.
14.
Slootstra
JW
,
Puijk
WC
,
Ligtvoet
GJ
,
Langeveld
JP
,
Meloen
RH
. 
Structural aspects of antibody-antigen interaction revealed through small random peptide libraries
.
Mol Divers
1996
;
1
:
87
96
.
15.
Raum
T
,
Gruber
R
,
Riethmuller
G
,
Kufer
P
. 
Anti-self antibodies selected from a human IgD heavy chain repertoire: a novel approach to generate therapeutic human antibodies against tumor-associated differentiation antigens
.
Cancer Immunol Immunother
2001
;
50
:
141
50
.
16.
Yamaoka
K
,
Nakagawa
T
,
Uno
T
. 
Statistical moments in pharmacokinetics
.
J Pharmacokinet Biopharm
1978
;
6
:
547
58
.
17.
Benet
LZ
,
Galeazzi
RL
. 
Noncompartmental determination of the steady-state volume of distribution
.
J Pharm Sci
1979
;
68
:
1071
4
.
18.
Bernard
J
,
Harb
C
,
Mortier
E
,
Quemener
A
,
Meloen
RH
,
Vermot-Desroches
C
, et al
Identification of an interleukin-15alpha receptor-binding site on human interleukin-15
.
J Biol Chem
2004
;
279
:
24313
22
.
19.
Teeling
JL
,
Mackus
WJ
,
Wiegman
LJ
,
van den Brakel
JH
,
Beers
SA
,
French
RR
, et al
The biological activity of human CD20 monoclonal antibodies is linked to unique epitopes on CD20
.
J Immunol
2006
;
177
:
362
71
.
20.
Davis
MI
,
Bennett
MJ
,
Thomas
LM
,
Bjorkman
PJ
. 
Crystal structure of prostate-specific membrane antigen, a tumor marker and peptidase
.
Proc Natl Acad Sci U S A
2005
;
102
:
5981
6
.
21.
Brischwein
K
,
Parr
L
,
Pflanz
S
,
Volkland
J
,
Lumsden
J
,
Klinger
M
, et al
Strictly target cell-dependent activation of T cells by bispecific single-chain antibody constructs of the BiTE class
.
J Immunother
2007
;
30
:
798
807
.
22.
Hoos
A
,
Ibrahim
R
,
Korman
A
,
Abdallah
K
,
Berman
D
,
Shahabi
V
, et al
Development of ipilimumab: contribution to a new paradigm for cancer immunotherapy
.
Semin Oncol
2010
;
37
:
533
46
.
23.
Kipriyanov
SM
,
Moldenhauer
G
,
Martin
AC
,
Kupriyanova
OA
,
Little
M
. 
Two amino acid mutations in an anti-human CD3 single chain Fv antibody fragment that affect the yield on bacterial secretion but not the affinity
.
Protein Eng
1997
;
10
:
445
53
.