Supplementary Table Legends

Supplementary Table Legends

Supplementary TableS1. Surface Plasmon Resonance (SRP)-derived binding affinities and kinetic constants of HER2(T)xPRLR and its parental antibodies: HER2 antibody and H1H7672P2. HER2xPRLR bispecific antibody with HER2 trastuzumab arm (HER2(T)xPRLR) was generated as described in Materials and Methods. SPR-based Biacore technology was used to determine both on-rate and off-rate kinetic data to calculate an equilibrium dissociation constant for each of the antigen-antibody interactions.HER2(T)xPRLR was analyzed for binding to recombinant soluble monomeric extracellular domains of human PRLRor HER2 tagged with MycMycHis (mmh)(hPRLR.mmh or hHER2.mmh). HER2 antibody was analyzed for binding to hHER2.mmh. H1H7672P2 antibody was analyzed for binding to hPRLR.mmh.

Supplementary TableS2. In vitro Potency of HER2 ADC, PRLR ADC and HER2xPRLRbs ADC in a Panel of Breast Cancer Cell Lines

Breast Cancer Cell lines are arrangedbased on HER2 surface receptor numbers (determined using Quantitative Flow Cytometry). IC50 values for HER2 ADC, PRLR ADC and HER2xPRLR ADC were determined in Direct Cell Count-Based Cytotoxicity Assayfollowed by sigmoidal non-linear regression analysis with GraphPad Prism software.

Supplementary Figure Legends

Supplementary Figure S1.A, Structure of the PRLR and HER2 ADCs. PRLR antibody or HER2 antibody (blue) were conjugated via surface lysines to the maytansine derivative DM1 (red), using non-cleavable SMCC linker (green). ”n” stands for drug-to-antibody ratio (DAR)B, Upon 24h incubation, PRLR, but not HER2 antibody accumulated in low pH vesicles of T47D cells.T47D cells were grown in 96-well optical plates.PRLR or HER2 antibody conjugated to pHrodoTMdye were added to T47D cells at 10ug/ml for 24h.The pH-activated pHrodoTMbecomes fluorescent in TRITC channel upon internalization into the acidic environment of lysosomes, whereas background fluorescence of non-internalized pHrodoTM is minimal.Live cell images were taken on ImageXpressMICRO with 40x objective using Transmitted Light and TRITC channels.The same exposure setting was used to acquire wells treated with PRLR or HER antibodies. Representative images are shown. Scale = 50uM.

Supplementary Figure S2. Proteasomal degradation inhibitors, Bortezomib and Lactacystin had little or no effect on PRLR turnover. HEK293 engineered to express full length PRLRin a tetracycline-controlled fashion (Lenti-XTM Tet-On System), weregrown in 6-well plates, induced by doxycycline (0.7ug/ml) for 24h, left untreated, or pretreated with either dimethyl sulfoxide (DMSO), or bortezomib (10uM), or lactacystin (10uM), for 2h, and then treated with CHX (50ug/ml) for the indicated times, lysed and processed for Western blotting using PRLR antibodies (Life Technologies) and -tubulin antibodies as loading control.

Supplementary Figure S3. PRLR degradation was not significantly affected by PRLR antibody or Prolactin (PRL). HEK293 engineered to express full length PRLRin a tetracycline-controlled fashion (Lenti-XTM Tet-On System) weregrown in 6-well plates, induced by Doxycycline (0.7ug/ml) for 24h, left untreated (None), or treated with either PRL(100ng/ml), or PRLR antibodies (10ug/ml) (CHX/PRLR Ab), or CHX, or CHX combined with PRL (100ng/ml) (CHX/PRL), or CHX combined with PRLR antibody (10ug/ml) (CHX/PRLR Ab) for the indicated times, lysed and processed for Western blotting using PRLR antibodies (Life Technologies) and GAPDH antibodies as loading control. CHX was used at 50ug/ml.

Supplementary Figure S4. HER2xPRLR bispecific antibodieswith different HER2 and PRLR arms induced similar rates of HER2 degradation in T47D cells. HER2(T)xPRLR containing HER2 arm structurally similar to trastuzumab,and HER2xPRLR bispecific antibodiesbearing non-trastuzumab HER2 arms (HER2xPRLR bsAbs1-5) have been described (see Materials and Methods).T47D cells were grown in 6-well plates, left untreated, or incubated with HER2(T)xPRLR, or HER2xPRLR bsAb1, or HER2xPRLR bsAb2, or HER2xPRLR bsAb3, or HER2xPRLR bsAb4, or HER2xPRLR bsAb5 for the indicated times (Hours), lysed and processed for Western blotting using HER2and -actin antibodies as loading control.Quantitation of the Western blots was performed by calculating net intensity of bands with Care Stream software (Kodak).To control for loading, normalization to actin was used as follows: the sample with highest actin net intensity was used as normalization control.Actin net intensity of each sample was divided by normalization control value to get a relative value of the sample.Net intensities of HER2 bands were divided by the calculated relative actin value for the sample to get a normalized HER2 value.

Supplementary Figure S5. In T47D/HER2 cells, HER2xPRLR bsAb1 augments HER2 ADC-induced cell cycle arrest in G2M phase. T47D/HER2 cells were grown in 96-well optical plates and treated as described in the legend to Fig. 6B. Cells were processed for viability assay, then cell cycle evaluation using DNA content classification was performed as follows: for each well, DNA Integrated Intensity values of individual nuclei were measured using MetaXpressTM software, extracted and plotted as Cell Count versus DNA Integrated Intensity. Classification into G0/G1, S and G2/M populations was performed using Cell Cycle Module of MetaXpressTM software. Data for wells treated with 0.1nM, 0.5nM, 5nM, or 10nM ADC are shown. Approximately 3,000 to 18,000 cells (number in the upper right corner of each panel) were analyzed per well, depending on the treatment. Cells treated with either concentration of Non-binding control ADC combined with HER2xPRLR bsAb1 (10ug/ml) (third panel from the top) displayed a cell-cycle profile typical for untreated non-synchronized T47D/HER2 cell population (“Untreated Cells” panel). Cells treated with PRLR ADC (fourth panel from the top) displayed a cell-cycle profile typical of cells undergoing DM1-induced cell cycle arrest: diminution of the G1 peak and an approximately even distribution of cells between G0/G1 phase and G2/M phase of the cell cycle. In cells treated with HER2 ADC (second panel from the top), the size of the G0/G1 peak decreased with increasing ADC concentration, while the size of the G2/M peak remained the same, indicating that HER2 ADC induces G2/M cell cycle arrest in a concentration-dependent manner. As seen in the upper panel, HER2 ADC-induced G2/M arrest was augmented by HER2xPRLR bsAb1. This was especially evident by the shift from G0/G1 phase to G2/M phase in wells treated with 5nM and 10nM ADC.

1