The Discovery and Preclinical Development of ASG-5ME, an Antibody–Drug Conjugate Targeting SLC44A4-Positive Epithelial Tumors Including Pancreatic and Prostate Cancer

SLC44A4, also known as CTL4, is a member of the family of solute carrier proteins known as SLC44A1-5 or choline transporter-like proteins (CTL1-5). Although this family of proteins is thought to be involved in choline transport, only SLC44A1 and SLC44A2 have been demonstrated to have choline transport activity (1–7). SLC44A4 has not been shown to be involved in choline transport, but it has been linked with acetylcholine synthesis and transport (8) as well as uptake of thiamine pyrophosphate, the phosphorylated form of vitamin B1 (9).

Through RNA expression profiling and IHC analysis, we have shown SLC44A4 to be selectively expressed in a broad range of epithelial tumors while having restricted expression in normal tissues. This represents the first thorough evaluation of SLC44A4 expression in both normal and cancer tissues in the literature. A link between SLC44A4 expression and tumorigenesis has not been reported. However, due to the expression observed in prostate and pancreatic cancers, where over 90% of cases are positive, and low expression in normal tissues SLC44A4 is a good target for an antibody–drug conjugate (ADC). Here, we describe the expression of SLC44A4 in pancreatic and prostate cancers and report on the preclinical development of ASG-5ME, an ADC composed of an SLC44A4-specific human IgG2 antibody conjugated with monomethyl auristatin E (MMAE) through a protease cleavable valine-citruline linker. Data presented here support that ASG-5ME may represent a new treatment option for patients with prostate cancer and pancreatic cancer and that SLC44A4 is a novel target in these cancers and others of epithelial origin.

For immunohistochemical (IHC) studies, a monoclonal antibody (mAb; M5-121.131) against an extracellular domain of human SLC44A4 (largest extracellular loop) was generated using standard mouse hybridoma technology. The antibody was validated for SLC44A4 specificity by correlating staining patterns on formalin-fixed, paraffin-embedded (FFPE) sections of various xenografts and cell lines with available transcript expression data. The specificity of the antibody was also confirmed by FACS analysis with recombinant cell lines expressing SLC44A4 and control cell lines not expressing the target.

SLC44A4-specific IHC was performed on tissue microarrays (TMA) of prostate cancer, pancreatic cancer, and normal tissues (US Biomax, TriStar Technologies, Accumax). After deparaffinization and rehydration, tissue sections were treated for antigen retrieval with trypsin (MP Biomedicals) for 10 minutes at 37°C and then incubated with the primary antibody M5-131.121 or isotype control antibody mouse IgG1k for 1 hour. Expression of SLC44A4 was detected with the Biogenex Super Sensitive Polymer-HRP IHC Detection Kit (Biogenex Laboratories Inc.).

Expression in patient cancer specimens was scored qualitatively based on a combination of average staining intensity of tumor cells: (0) none, (1) weak, (2) mild, (3) moderate, and (4) strong, and approximate percentage of positively staining tumor cells: (0) none; (1) >0–25%; (2) >25%–50%; (3) >50%–75%; and (4) >75%. They were then assigned to one of three different categories: samples where ≥50% of tumor cells had an average intensity score of 2, 3, or 4 were considered high/moderate and all other positives were considered low and the remainder negative. Normal tissue expression was descriptive.

The first extracellular domain of SLC44A4, amino acids 50–227, was generated using a modified version of the pAPTag5 vector system (GenHunter Corp.), in which the Alkaline phosphatase fusion gene was removed. 293T cells were transfected with the vector by calcium phosphate precipitation and selected with zeocin to derive a stable cell line secreting the recombinant protein. 293T recombinants were propagated in 293-SFM II medium (Life Technologies) in a 25-L Wave bag. Medium was harvested, filtered, and the recombinant protein was purified by metal chelate chromatography on a Ni-NTA column (Qiagen). The protein was eluted with imidazole, dialyzed into PBS pH 7.2, and quantitated by BCA assay.

Cell lines were obtained from the indicated vendors, followed by a limited number of passages after thawing (> 5 passages) to generate aliquots for freezing down to create a central master cell bank. Cell line authentication was performed on banked cell lines by short tandem repeat (STR) profiling carried out by GE Healthcare, SeqWright Genomic Services using the Promega PowerPlex 16HS kit. All studies described here utilized the master cell bank as the source of cells for respective studies. The PC-3 cell line was obtained from ATCC (catalog number CRL-1435) on March 1, 2012, and STR profiling performed on June 21, 2012. From this parental PC-3 cell line, we created the PC3-Neo and PC3-SLC44A4 recombinant lines (human and monkey) which were banked and STR profiling was performed on September 4, 2013. HT-29 and OVCAR-5 cells were obtained from NCI (catalog numbers 0507335 and 0507337), and STR profiling was performed on March 16, 2011. VCaP was obtained from ATCC on January 27, 2009 (catalog number CRL-2876) and 22Rv1 cells were obtained from ATCC on February 3, 2012, and STR profiling was performed on July 24, 2013.

Human (PC3) and rodent (3T3 clone 7, Rat1, 300.19) recombinant cell lines overexpressing SLC44A4 were generated by retroviral transduction using the replication deficient pSRalpha-MSV-TKneo amphotropic retroviral vector system (10). Cells were infected with retrovirus-containing media encoding SLC44A4. Stable-expressing cells were derived by selection for neomycin resistance using G418 (Life Technologies). Control cells were generated similarly but with retrovirus expressing only the neomycin drug resistance gene. Stable SLC44A4 expression was monitored by FACS and Western blotting.

AGS-5M2, the parental antibody used to create ASG-5ME, was selected from a large panel of SLC44A4-specific human antibodies. This panel was generated by immunizing human IgG2 producing Xenomice (11) with B300.19 cells engineered to express SLC44A4. Lymph node cells from immunized mice were fused with Sp2/0-AG14 B-cell hybridoma cells (ATCC) by electro cell fusion using an ECM2100 electro cell manipulator (BTX). SLC44A4-specific hybridomas were identified by screening hybridoma supernatants for binding to cells engineered to express loop one of the multi-transmembrane SLC44A4 protein but not to the nonexpressing parental cells.

Total RNA was extracted from the antibody-producing hybridoma cells using TRIzol Reagent (Life Technologies) according to the manufacturer's suggested protocol. cDNA synthesis was generated from total RNA using the RACE cDNA Amplification kit (Clontech) according to the manufacturer's protocol. The variable human heavy and kappa light chains were amplified using RACE primer and a reverse primer within the constant region of the human light and heavy chains, sequenced and cloned into a mammalian cell expression vector encoding human IgG2 and kappa constant domains. The plasmid was transfected into CHOK1SV host cells (Lonza) to generate a stable antibody-producing cell line according to the manufacturer's protocol.

ASG-5ME was made by conjugating AGS-5M2 with MMAE via a protease-cleavable maleimidocaproyl valine citruline p-aminobenzyl alcohol dipeptide linker (Supplementary Fig. S1). MMAE was attached to reduced cysteines in the antibody hinge region, and linked to reduced cysteines of the antibody as previously described (12).

FACS analysis was used to determine affinity of antibodies and ADCs for SLC44A4. Recombinant PC3 cells expressing either human or cynomolgus SLC44A4 or the neomycin resistance gene (negative control) were maintained at 37°C in 5% CO₂ in media containing RPMI-1640 (Gibco) supplemented with 10% fetal bovine serum (Omega Scientific). The cells were reconstituted in FACS buffer, and the AGS-5M2 and H3-1.4 isotype control antibodies as well as the ASG-5ME ADC and its isotype control H3-1.4-vcE (negative control for conjugated mAb) ADCs were mixed with cells in a dilution curve at final concentrations ranging from 0.01 to 160 nmol/L. The samples were tested in triplicate. Cells were incubated with antibodies or ADCs at 4°C on a plate shaker overnight. Following a wash, cells were incubated with the detection antibody for 1 hour, washed, and read on the flow cytometer. Detectable binding of antibodies over the range of test concentrations (MFI values of samples subtracted from their respective negative control antibodies) was analyzed using the one-site binding, nonlinear regression analysis in GraphPad Prism version 4 (Graphpad Software, Inc.). The K_D (equilibrium dissociation constant) and B_max (maximum specific binding) are reported.

Recombinant PC3 cells expressing SLC44A4 were seeded onto poly-D-lysine coated 8-well chamber slides (2.5 × 10⁴ cells per well in RPMI media) overnight and non-adherent cells were subsequently washed off. Cells were then incubated with 10 μg/mL ASG-5ME for 1 hour at 4°C (on ice) or 24 hours at 37°C. After the incubation period, unbound antibody was washed off with PBS, and cells were fixed in 4% paraformaldehyde for 20 minutes at room temperature. Cells were then permeabilized in PBS + 0.1% Triton-X-100 for 15 minutes and non-specific labeling was blocked in PBS + 10% normal goat serum. Cell surface and intracellular ASG-5ME was visualized by incubating cells with Alexa Fluor 488–labeled goat anti-human IgG (Invitrogen). Lysosomes were visualized by staining with lysosome-associated membrane protein 1, LAMP1 (mouse CD107a clone H4A3, BD Biosciences) and a secondary antibody, Alexa Fluor 568–labeled goat anti-mouse IgG (Invitrogen). Nuclei were visualized with TOPRO-3 Iodide (Invitrogen) and coverslips were mounted using Prolong Gold anti-fade reagent (Invitrogen) for imaging. High-resolution laser scanning confocal image sections were acquired using a Leica TCS SP5-II (63× oil immersion objective; NA = 1.4) and Alexa Fluor 488/568 were scanned sequentially to minimize fluorophore cross-talk and false-positive colocalization.

Exponentially growing recombinant PC3 cells expressing SLC44A4 or the neomycin drug resistance gene with a viability of 95% or greater were plated in fresh RPMI-1640 (Gibco-Invitrogen) media containing phenol red supplemented with 10% FBS. Cells were left overnight and treated with ADCs and an isotype control. After 5 days of treatment and incubation at 37°C and 5% CO₂, cell viability was measured following 1-hour incubation at 37°C with Presto Blue Reagent (Invitrogen). Samples were analyzed using a Synergy Microplate reader. Survival values were plotted using Graph Pad Prism to derive EC₅₀ values using a curve-fitting analysis model for nonlinear curve regression, sigmoidal dose response with variable slope formula.

Four- to 5-week-old ICR-SCID mice (Taconic) were maintained and used at Agensys' animal facility using Agensys' Institutional Animal Care and Use Committee (IACUC)–approved protocols. Studies were performed with either cell line–derived xenograft models or patient-derived tumor xenografts (PDX) developed at Agensys. Human tumor xenografts were established either subcutaneously or orthotopically in immunodeficient ICR SCID mice by injection of human tumor cells or fragments of PDXs. SLC44A4 expression in all xenograft models used was confirmed by IHC analysis. Tumor growth was monitored using caliper measurements every 3 to 4 days until the end of study. Tumor volume was calculated as width² × length/2, where width is the smallest dimension and length is the largest. Mean tumor volume data for each group were plotted over time with standard error bars. A statistical analysis of the tumor volume data for the last day before animal sacrifice was performed using the Kruskal–Wallis test. Pairwise comparisons were made using Tukey test procedures (two-sided) to protect the experiment-wise error rate. This implementation of the Tukey test was performed on the ranks of the data. The percent tumor growth inhibition in each treated group versus a control group was calculated as [(Control − Control baseline) − (Treated − Treated baseline)]/(Control − Control baseline) × 100%. Details on dosing regimens for individual studies are provided in respective tables or figures. Choice of doses given for drug combination studies was based on prior dose titration studies for each drug dosed individually.

To identify new targets overexpressed in cancer versus normal tissue, we performed discovery experiments using suppression subtractive hybridization. SLC44A4 was selected from a subtraction of prostate cancer minus benign prostatic hyperplasia. Gene expression of SLC44A4 was examined in a wider range of normal and cancer tissue specimens through mining of public microarray datasets contained within Oncomine Powertools cancer expression database (https://powertools.oncomine.com). Tumor types with expression observed to be higher than corresponding normal tissues include breast, gastric, ovarian, pancreatic, and possibly a subset of lung cancers (Supplementary Fig. S2). SLC44A4 gene expression in normal tissues is found to be generally low, with the exception of gastrointestinal tract and prostate.

IHC staining of FFPE confirmed SLC44A4 protein expression in patient tumors. Most notably, 85% of primary prostate cancer specimens and distant metastases examined showed a medium to high level of staining for SLC44A4 and 67% of pancreatic tumors had a medium to high level of staining. Greater than 90% of pancreatic and prostate tumors were positive for SLC44A4 expression (Table 1). We observed polarized distribution of SLC44A4 toward the apical surface in low grade, well-differentiated tumors, which is lost in high grade, poorly differentiated tumors (Fig. 1). Protein expression increased with higher disease grade in both pancreatic and prostate tumors. While the specific focus discussed here is in regard to pancreatic and prostate cancers, we did observe SLC44A4 protein expression in other indications, consistent with gene expression profiling data mentioned above.

To assess SLC44A4 expression in normal tissues, 225 samples were tested representing 36 different human tissues (Table 2). Specific positive staining for SLC44A4 was observed mainly in the epithelia of prostate, lung bronchioles, gastro-intestinal tract, a subset of tubules in the kidney cortex, fallopian tube, bladder, ureter, and uterine endometrium. SLC44A4 expression was also seen in the ductal epithelium of some samples of liver (bile ducts), breast, salivary gland, esophagus, pancreas, skin (sweat glands). The expression of SLC44A4 was generally seen with a polarized distribution toward the apical surface of the cells, with the extracellular portion of SLC44A4 facing toward the lumen and not immediately adjacent to the intravascular space (Fig. 2). SLC44A4 expression was not observed in adrenal, bladder, bone marrow, brain heart, ileum, lymph node, ovary, parathyroid, pituitary, rectum, skeletal muscle, spleen, testis, thymus, and thyroid tissues.

We observed very weak levels of expression across a large panel of cancer cell lines when looking to identify suitable in vitro cell line models. For some cell lines, SLC44A4 expression increases when implanted and grown as tumor xenografts in vivo. The level of expression is briefly maintained for several hours after harvesting and digestion of tumors and growth in cell culture conditions; however, the level of expression decreases back to initial levels after subsequent passages. One example of this phenomenon is illustrated with the OVCAR-5 ovarian cancer cell line (Supplementary Data S3). Use of an in vitro cell line model with stable high levels of expression required the generation of PC3 cells engineered to express SLC44A4 recombinantly.

AGS-5M2 was selected among a panel of anti-SLC44A4 mAbs produced in the Xenomouse screened for specific binding and high affinity for SLC44A4 recombinant protein and endogenously or recombinantly expressed in human cancer cell lines. The microtubule-disrupting agent MMAE was conjugated to the AGS-5M2 antibody as previously described to produce the ADC ASG-5ME. Results from affinity binding studies with recombinant PC3 cells engineered to express human SLC44A4 demonstrated similar affinities (K_d) for the parental antibody, AGS5-M2 (1.6 nmol/L) and the ADC ASG-5ME (1.7 nmol/L; Fig. 3A). Using PC3 cells engineered to express cynomolgus SLC44A4, we demonstrated similar affinities of AGS-5M2 (3.9 nmol/L) and ASG-5ME (1.4 nmol/L; Fig. 3B). The homology between human and cynomolgus monkey SLC44A4 is 96%. These studies demonstrated that conjugation of AGS-5M2 with vcMMAE did not significantly impact the binding affinity or B_max for human or cynomolgus monkey SLC44A4 and that affinities to human and cynomolgus monkey SLC44A4 are similar.

Laser scanning confocal fluorescence microscopy was used to visualize internalized ASG-5ME (Fig. 4). Recombinant PC3 cells expressing SLC44A4 cells incubated at 4°C showed distinct cell surface localization of ASG-5ME (Fig. 4A, green), with minimal evidence of ASG-5ME internalization (Fig. 4,C). Lysosomes visualized using an antibody to lysosome-associated membrane protein 1 (LAMP1; Fig. 4,B–C, red) appeared as discrete punctate cytosolic staining. At 37°C for 24 hours, surface membrane localization of ASG-5ME ADC was dramatically reduced and the predominant distribution of ADC localization appeared as discrete intracellular punctate staining throughout the cytosol (Fig. 4,D–F, green). The ASG-5ME was shown to colocalize with LAMP1 (Fig. 4,F, arrow, yellow), indicating that within 24 hours, ASG-5ME was trafficked to the lysosomes in the prostate cancer cells and targeted for subsequent degradation.

The potency and efficacy of ASG-5ME to induce cell death was evaluated in PC3 cells engineered to express the SLC44A4 antigen. We observed potent dose-dependent cytotoxicity in PC3-SLC44A4 cells treated with ASG-5ME with an EC₅₀ of ∼2 nmol/L (Fig. 5). ASG-5ME was not cytotoxic on mock PC3-Neo cells that do not express the SLC44A4 antigen. Moreover, the native antibody AGS-5M2, and the control ADC did not cause cytotoxicity in either PC3-SLC44A4 or PC3-Neo cells, indicating that the activity observed with the ADC was solely attributable to the released MMAE after specific binding to the SLC44A4-expressing cells.

Human cancer xenograft models were selected for in vivo testing on the basis of having favorable SLC44A4 expression. A summary of selected ASG-5ME in vivo activity studies in prostate, pancreas, colon, and ovarian xenograft models is shown in Table 3. Notably, we observed significant tumor growth inhibition in both cell line – and patient-derived tumor xenografts when implanted either orthotopically or subcutaneously in immunodeficient mice, and within androgen-dependent and castration-resistant prostate cancer models. While activity was also observed in colon and ovarian cancer models, here we focus on pancreatic and prostate cancer models due to the high percentages of positivity of SLC44A4 expression in the tumors of these two patient populations. IHC staining of SLC44A4 expression in untreated tumors is shown in the panels for the selected models (Fig. 6A–C). In all of our single-agent activity studies ASG-5ME was administered biweekly for a total of four doses. Treatment of established LAPC-9AI patient-derived prostate tumor xenografts with 1 or 3 mg/kg ASG-5ME resulted in statistically significant (P < 0.0001) tumor growth inhibition (TGI) of 88.3% and 93.3%, respectively, at the end of the study (day 27) compared with the control ADC (Fig. 6A). In the VCaP prostate xenograft model, treatment with 5 mg/kg ASG-5ME significantly inhibited tumor growth by 95% at day 38 (P < 0.0001) at the end of the study (day 52) compared with the control ADC (Fig. 6B). In the AG-Panc4 patient-derived pancreatic tumor xenograft model (Fig. 6C), treatment with 4 mg/kg ASG-5ME significantly inhibited tumor growth by 87.1% (P < 0.0001) at the end of the study (day 68) compared with the control ADC.

In vivo combination studies were performed to assess whether ASG-5ME can enhance activity of standard-of-care agents for prostate or pancreatic cancer. Agents tested included enzalutamide, gemcitabine, erlotinib, and nab-paclitaxel. Analysis of combination data revealed that ASG-5ME only significantly enhanced antitumor activity of nab-paclitaxel. Antitumor activity of ASG-5ME, nab-paclitaxel, or the combination was evaluated in AG-Panc4 pancreatic and LAPC-9AI prostate xenograft models. In the AG-Panc4 model (Fig. 6D), ASG-5ME (1 mg/kg) or nab-paclitaxel (100 mg/kg) administered every 4 days for a total of 3 doses both demonstrated modest antitumor activity compared with the control, resulting in 40.3% (P = 0.0100) and 32.6% (P = 0.0384) tumor growth inhibition, respectively, on day 29. When given in combination, a significant improvement in activity was observed resulting in 72.7% (P < 0.0001) tumor growth inhibition. The combination activity was superior to single-agent ASG-5ME (P = 0.0043) or nab-paclitaxel (P = 0.0009). In the LAPC-9AI model (day 13), neither ASG-5ME (0.5 mg/kg) nor nab-paclitaxel (60 mg/kg), both administered every four days for a total of four doses, demonstrated antitumor activity as single agents, while nab-paclitaxel at 120 mg/kg demonstrated significant antitumor activity compared with the control (40.3% TGI, p = 0.0476) (Fig. 6E). However, in combination, ASG-5ME significantly enhanced activity of nab-paclitaxel dosed at 60 mg/kg resulting in 84% tumor growth inhibition (P < 0.0001). Additionally, tumor regression (76.8%, P < 0.0001) was observed when ASG-5ME (0.5 mg/kg) was combined with the higher dose (120 mg/kg) of nab-paclitaxel. Pairwise comparisons of single-agent ASG-5ME or nab-paclitaxel at 60 or 120 mg/kg to either combination were significantly different (P < 0.0001 for all comparisons).

Our studies describe the identification of SLC44A4 as a novel differentially expressed target in pancreatic and prostate cancers. Relatively little literature has been published regarding the endogenous function or expression of SLC44A4. Based on sequence and structural similarity to other SLC44 family members, it was thought to play a role in choline transport. In functional experiments, not shown here, our results did not support a role for SLC44A4 in choline transport, nor was it involved in the transport of a number of other substrates we tested. Thiamine (vitamin B1) serves an essential role as a co-factor of enzymes involved in cellular metabolic pathways (13–14); however, it cannot be synthesized de novo in mammalian cells and must be obtained through dietary and microbiota-generated sources. A recently published report demonstrated that SLC44A4 is involved in transport of the phosphorylated form of thiamine (thiamine pyrophosphate) produced by microbiota in the intestine (9). Such a functional role would be consistent with the expression we observed within normal tissues of the gastrointestinal tract, although thiamine pyrophosphate was not one of the substrates we tested for uptake.

Additional studies performed in our lab did not identify a functional role for SLC44A4 in tumorigenesis, and there are no reports linking SLC44A4 to cancer. Furthermore, treatment with AGS-5M2 or any of our other SLC44A4-specific mAbs did not have antitumor effects. Due to this, we focused on the suitability of SLC44A4 as an ADC target.

We describe here for the first time an extensive analysis of expression of this novel target across normal and cancer specimens and identify cancer indications with overexpression of SLC44A4. IHC analysis of the expression of SLC44A4 in normal gastrointestinal tract revealed that the expression in epithelial cells is largely confined to the apical, or luminal, side of secretory epithelial cells. This apical polarity of expression is lost in advanced or poorly differentiated tumors. This may favor targeting tumor cells with an ADC and sparing normal cells that express SLC44A4. This is because tight junctions that form the epithelial barrier between systemic circulation and luminal spaces greatly reduce diffusion of large macromolecules into the lumen (15). This provides a rationale for lower exposure of apical surfaces of epithelial cells to circulating IgGs. However, several mechanisms for specifically transporting immunoglobulins across the epithelial barrier are known. The polymeric immunoglobulin receptor (pIgR) acts as a transporter of IgA and IgM molecules across epithelial barriers, but does not bind monomeric IgGs (16). FcRn, on the other hand, can transcytose IgGs across the epithelial barrier but the serum concentration of IgG1 has been reported to be 500- to 1,000-fold higher than luminal concentrations (17). This analysis suggests that the exposure of tumor cells to injected IgG is likely to be greater than the exposure at the apical side of normal epithelial cells, thus a therapeutic with anti-SLC44A4 targeting antibody conjugated to a non-selective cytotoxic molecule could be very tumor cell specific. Toxicology will be needed to validate this hypothesis.

In the process of identifying suitable in vitro and in vivo tumor models, we observed that SLC44A4 is expressed at very low levels in most tumor cell lines when grown in vitro. However, in OVCAR5 cells and additional cell lines implanted and grown as tumor xenografts in vivo, the expression of SLC44A4 increased considerably, suggesting that there might be factors or contributions from the tumor microenvironment that upregulated the expression of SLC44A4.

While the RNA levels observed in normal prostate and prostate tumors did not indicate differential expression, many patients diagnosed with prostate cancer undergo radical prostatectomy during treatment, which removes concern for on-target toxicity for this organ. Based on the high prevalence of SLC44A4 expression observed in pancreatic and prostate cancer tissue microarrays, we evaluated the activity of SLC44A4-specific ADCs in in vitro and in vivo models of these indications. From these we selected ASG-5ME as our lead candidate and we report here for the first time the development of this molecule for the treatment of pancreatic cancer, prostate cancer, and other SLC44A4-positive tumors. We have shown that this ADC has high affinity for SLC44A4 and that it is cytotoxic and able to inhibit in vivo growth of several tumor models, including patient-derived pancreatic and prostate tumor xenografts.

A number of antibody-based therapeutics are being evaluated in the clinic but none have been approved for use in prostate or pancreatic cancer. Although newer therapies have been approved for advanced pancreatic cancer (Abraxane, nab-paclitaxel) and prostate cancer (Zytiga and Xtandi), resistance to these treatments means that additional therapeutic options are needed. We evaluated whether ASG-5ME can enhance antitumor activity when combined with standard-of-care agents for pancreatic and prostate cancers in xenograft models. Our data indicated that a combination of ASG-5ME and nab-paclitaxel at sub-optimal doses showed a marked increase in antitumor activity compared with the single agents alone. Although taxane-based chemotherapy has activity in advanced prostate cancer, Abraxane has not been approved for use in prostate cancer and has not been extensively evaluated in that setting. Our combination studies provide an initial proof of concept for combining ASG-5ME with Abraxane in pancreatic and prostate cancers. Of further interest are the enhanced antitumor effects with a combination of two anti-tubulin agents. Further preclinical efficacy and safety evaluation would be necessary to justify clinical evaluation of this combination approach. Taken together, our preclinical data generated with ASG-5ME highlights its therapeutic potential in pancreatic and prostate cancer and supports its clinical development as an anti-SLC44A4 therapeutic.

No potential conflicts of interest were disclosed.

Conception and design: M. Mattie, A. Raitano, K. Morrison, Z. An, F. Doñate, I.B.J. Joseph, P. Challita-Eid, D. Benjamin, D.R. Stover

Development of methodology: M. Mattie, A. Raitano, K. Morrison, K. Morrison, Z. An, C. Guevara, Y. Li, P. Challita-Eid

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Mattie, A. Raitano, K. Morrison, L. Capo, A. Verlinsky, M. Leavitt, J. Ou, R. Nadell, H. Aviña, C. Guevara, F. Malik, R. Moser, S. Duniho, J. Coleman, Y. Li, F. Doñate

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Mattie, A. Raitano, K. Morrison, Z. An, M. Leavitt, J. Ou, H. Aviña, C. Guevara, Y. Li, F. Doñate, I.B.J. Joseph, P. Challita-Eid, D. Benjamin, D.R. Stover

Writing, review, and/or revision of the manuscript: M. Mattie, A. Raitano, K. Morrison, K. Morrison, Z. An, J. Ou, R. Nadell, Y. Li, D.S. Pereira, F. Doñate, I.B.J. Joseph, P. Challita-Eid, D.R. Stover

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Mattie, Y. Li, D.R. Stover

Study supervision: M. Mattie, A. Raitano, Z. An, I.B.J. Joseph, P. Challita-Eid

Other (cloned hybridoma heavy and light sequences into CHO cell line production plasmids, and then confirmed those constructs): F. Malik

The authors thank Jean Gudas, Aya Jakobovits, and the hardworking and thoughtful members of the Agensys research departments that did not appear on this article, but that significantly contributed to previous conference presentations or significantly contributed to historical data that ultimately led to this final article. ASG-5ME was co-developed in conjunction with Agensys' partner, Seattle Genetics, Inc.

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.