Identification and Validation of a Novel Biologics Target in Triple Negative Breast Cancer

The goal of this study was to identify a novel target for antibody-drug conjugate (ADC) development in triple negative breast cancer (TNBC), which has limited treatment options, using gene expression datasets and in vitro siRNA/CRISPR and in vivo functional assays. We analyzed 4467 breast cancers and identified GABRP as top expressed gene in TNBC with low expression in most normal tissues. GABRP protein was localized to cell membrane with broad range of receptors/cell (815–53,714) and expressed by nearly half of breast cancers tissues. GABRP gene knockdown inhibited TNBC cell growth and colony formation in vitro and growth of MDA-MB-468 xenografts in nude mice. Commercially available anti-GABRP antibody (5–100 μg/ml) or de novo generated Fabs (20 μg/ml) inhibited TNBC cell growth in vitro. The same antibody conjugated to mertansine (DM1) also showed significant anticancer activity at nanomolar concentrations. Our results indicate that GABRP is a potential novel therapeutic target for ADC development.


Differential gene expression in TNBC and non-TNBC datasets
To identify genes that are highly expressed on TNBC cell surface, we used two independent data sets (MDACC and Wang cohorts) to define genes overexpressed in ER/PR-and HER2-negative cancers. In MDACC cohort, probe sets were tested for differential expression in TNBC (n=73) and non-TNBC cases (n=221) using an unequal variance t-test. A second dataset obtained from frozen tissues of surgically resected breast cancer specimens from 286 lymph-nodenegative patients including 56 TNBC and 230 receptor-positive cases (Wang et al. 2005) was used to confirm overexpression of genes in TNBC. We focused on the 1871 genes that are overexpressed in TNBC vs non-TNBC at an FDR level of <0.00001. Sixty-two percent of these genes (n=1162) were also overexpressed in TNBC cases in second (Wang) independent human breast cancer dataset (individual P-values <0.05). Out of these, we discovered 681 genes that have at least two-fold overexpression in TNBC with p<0.0001 observed in independent datasets (Supplementary Table 1). To account for multiple comparisons we performed beta uniform mixture analysis (BUM) of the p values, that showed a non-uniform distribution and was used to calculate false discovery rates (FDR) for particular p-values.  All media were supplemented with 10% fetal bovine serum (FBS; Gibco), 100 units/ml penicillin, and 100 µg/ml streptomycin.

Validation Cohorts
Immunoblotting PVDF membranes were blocked with 2% BSA in 10 mM Tris-HCl,50 mM Immunohistochemistry Briefly, slides were deparaffinized using xylene and rehydrated to distilled water. Endogenous peroxidase was quenched with hydrogen peroxide, and 1:75 dilution of primary antibody against GABRP was used. The slides were then washed with 10 mmol/L Tris-HCl, 50 mmol/L NaCl, 0.1% Tween 20, pH 7.4 (TBST) followed by incubation with horseradish peroxide-conjugated secondary antibody. Immunoreactivity was detected with the peroxidase-based Envision+ system (Dako). Diaminobenzidine (DAB) was used to detect the antibody complex (Dako). The slides were subsequently washed, counterstained with hematoxylin, dehydrated, cleared and coverslipped with resinous mounting media. Yale index breast cancer TMA (YTMA279-18) containing breast cancer tumors and cell lines including MDA-MB-468 and SK-BR3, was simultaneously stained using the same antibody.
TMAs were scanned to create bright field digital images using the ScanScope CS (Aperio, Vista, CA). All digital images were viewed in ImageScope by a breast pathologist (V.P.) that scored tumor epithelial cells as a percentage of cells with GABRP signal. GABRP positivity threshold was set at >1% with minimum number of 100 tumor cells from each case on TMA.
Areas with necrosis or inadequate quality of tissue/staining were excluded from scoring.

Flowcytometry
For this QuantiBRITE PE flow-cytometric analysis, isotype control IgG-PE antibody (R&D Systems) and ECD binding GABRP antibody conjugated to phycoerythrin (PE) were used.   that lacks identity with known gene targets was used as a scrambled control for nonsequence specific effects. In brief, HiPerFect reagent was diluted with serum-free medium in two-thirds the transfection volume for 10 minutes. It was then added to the diluted siRNA (GABRP or nonspecific siRNA pool) and incubated at ambient temperature for 20 minutes.
The culture medium was aspirated from the cells, and the cells were washed with serumfree medium. The siRNA-HiPerFect complex was added drop wise to the cells and incubated at 37°C for 4-6 hours, at which time one-third volume medium with 30% FBS was added and the cells were incubated for an additional 20 hours (for a total incubation period of 24 hours). The cells were then ready to be collected for further experiments.

Stable siRNA transfection
For stable knockdown, plasmids were packaged as virus by using phoenix-packaging cells

Soft agar assay
In a 6-well culture plate, base layer was formed by 1.6% low-gelling temperature agarose (Sigma Aldrich) mixed with 2X RPMI complete medium (20% FBS), while the top layer contained 0.6% agarose, 2X RPMI complete medium and 1X10 4 cells per well. 1X RPMI complete media was added and replenished every 3 days. Colony spheres formed were fixed and stained with 0.1% crystal violet and 2.1% citric acid.

Tumor Growth
Single-cell suspensions with >95% viability were used for mammary orthotopic injections in (Length x Width 2 ) x π/6. The animals were euthanized 10 weeks after tumor cell inoculation.
Representative data were obtained from five mice per experimental group and the entire experiment was repeated in three independent trials.

GABRP antibody-DM1 ADC Immunogen method uses succinimidyl-4-(N-maleimidomethyl)-
cyclohexane-1-carboxylate (SMCC) as linker which introduces maleimido group on the antibody to enable linkage of the DM1 via a non-reducible thioether bond. This non-cleavable linker is therefore released only intracellularly. In addition to the rabbit polyclonal GABRP antibody, rabbit polyclonal isotype control antibody (Abcam) was also conjugated simultaneously to DM1 using same method. ADCs were characterized by hydrophobic interaction, size exclusion and reversed phase chromatography, UV-vis spectrophotometry, and the concentration of free drug in ADC was limited to <5%.

GABRP Overexpression
Lentivirus carrying vector control or GABRP plasmids with CMV promoter and GFP and puromycin cassettes was directly purchased from Applied Biological Materials Inc. Briefly, MDA-MB-231 cells were infected with control or GABRP lentivirus alongwith 4mg/ml polybrene in complete RPMI medium in 6-well culture plates. After 24h, virus-containing medium was replaced with fresh complete RPMI medium containing 1mg/ml of puromycin for selection.

Supplementary Table and Figure Legends
Supplementary