Supplementary Information Methods and Materials

Supplementary Information Methods and Materials:

Reagents

Birinapant (CAS 1260251-31-7) was purchased from ChemieTek (Indianapolis, IN) and carboplatin (CAS 41575-94-4) from R & D Systems (Minneapolis, MN).

Animals

Non-obese diabetic, severe combined immunodeficient, interleukin-2 receptor gamma null (NSG) female mice from Jackson Laboratories, housed according to University of California, Los Angeles (UCLA) Division of Laboratory Animal Medicine (DLAM) guidelines, were used in survival experiments. Animal experiments were approved by the UCLA Animal Research Committee and performed with DLAM oversight.

Cell Lines

Four cell lines were utilized for in vivo survival experiments. S1-GODL 1, S8-GODL, and S9-GODL were established in the G.O. Discovery Lab from primary patient HGSCs while the Ovcar-3 cell line was obtained from ATCC. S1-GODL, S8-GODL, S9-GODL, and Ovcar-3 cells were grown in RPMI/10%FBS media. A549, H226, H460, CaSki, HeLa, SiHa, 5637, J82, HT1197, DLD1, Colo205 and SW620 cells were purchased from ATCC and verified by short tandem repeat analysis of 16 alleles. All cell lines were tested for mycoplasma and were used within 10 passages.

S8-GODL and S9-GODL cell lines were derived from two independent human HGSCs clinically defined as platinum resistant. These lines were generated by plating freshly dissociated primary HGSC cells in PrEGM media (Lonza CC-3166). Cells were maintained by replacing half the media with fresh media twice per week. After 3-4 weeks, outgrown colonies were transferred to RPMI/10% FBS media without selection. Generation of the platinum sensitive cell line S1-GODL has been previously reported 1.

Survival assays

Eight-week old NSG female mice were injected intraperitoneally (IP) with 2.5x105 S1-GODL, S8-GODL, S9-GODL or Ovcar-3 cells (n=33 or 41/cohort). An IP tumor model was utilized to recapitulate intraperitoneal disease patterns observed in advanced-stage HGSC patients. After two weeks, one mouse from each cohort was sacrificed to confirm tumor establishment. Tumor-bearing mice were then randomized by simple randomization into treatment arms (n=8 or 10/arm) of vehicle, birinapant (30mg/kg IP in 12.5% Captisol 2X/week), carboplatin (50mg/kg IP in PBS 1X/week), or co-therapy for 4 weeks. A group size of n=8 per treatment arm was chosen for survival studies based on detection of statistically significant differences in progression free survival in previous work1.Immediately after completion of therapy, serum was collected from four randomly selected mice per treatment arm (S8- and S9-GODL cohorts, n=8 each) 30 min after terminal drug injections for analysis of carboplatin and birinapant levels. Serum drug levels in the mice were relatively comparable to levels observed in patients 2-4. After therapy, a researcher blinded to treatment arms observed mice until they reached NIH-defined end-point criteria 5. Confirmation of endpoint was required from two additional independent blinded researchersprior to euthanasia. Tumor, organs, cellsfrom abdominal washes and serum were harvested from euthanized mice by a blinded investigator. No individuals were excluded from the analysis. Survival comparisons were performed using Kaplan-Meier curves.

All mice under treatment were monitored for signs of toxicity. Behavior of the mice was observed daily to check for signs of neuropathy or other neurologic and organ damage as well as to monitor for signs of disease progression. Mice were weighed prior to treatment, and then weekly thereafter as part of health surveillance. In the S8- and S9-GODL survival experiments, blood was harvested from four random mice in each treatment arm (2 per cohort) immediately after the cessation of therapy and analyzed for complete blood count with differential and blood chemistry. Hematologic results were compared to historic controls from naive NSG mice (n=3). Histologic sections of organs (collected at euthanasia) from two random mice from each treatment arm were evaluated by an expert mouse pathologist. No signs of organ damage other than those caused by tumor infiltration were observed, and livers showed no signs of stress from chemotherapy.

Traf2 and cIAP2 sequencing

RNA was extracted using the QiagenAllprep kit from parental S1-GODL cells and S1-GODL cells obtained from the one mouse with tumor that had acquired resistance to carboplatin and birinapant co-therapy. Two micrograms of RNA was converted to cDNA for each sample using the Invitrogen Superscript III First Strand Synthesis Supermix kit. The cDNA was amplified using the following primer sets: TRAF2 forward: 5' -ATG GCT GCA GCT AGC GTG AC - 3' with TRAF2 reverse 5' - TTA GAG CCC TGT CAG GTC CAC AAT - 3', and cIAP2 forward 5' - AAG CGC CAA CAC GTT TGA ACT G - 3' with cIAP2 reverse 5' - CGA ACT GTA CCC TTG ATT GTA CTC CTA C - 3'. Amplification by PCR was performed using Takara ExTaq with 1.5 mM Mg2+ in the reaction buffer and the following cycling conditions: 94C for 3 min; 40 cycles of 94C for 30s, 57C for 30s, 72C 105s; 72C for 10 min. PCR products were visualized on a 1.5% agarose gel, excised and purified using the Macherry-Nagel Gel Extraction kit. Purified PCR products were sequenced in both directions by Sanger sequencing. Sequences were aligned to the NCBI database reference sequence (NM_021138 for Traf2 and NM_001165 for cIAP2) using BLAST alignment tools (

Clinical Samples

Clinical samples with pathologist confirmed HGSC were collected from consented patients in accordance with UCLA Internal Review Board (IRB) guidelines (IRB #10-000727). A detailed description of specimens is provided in Table S3.Patient samples were processed as previously described 1.

In vitro organoid bioassay

Cell lines (S1-GODL, S8-GODL, S9-GODL, Ovcar-3, A549, H226, H460, CaSki, HeLa, SiHa, 5637, J82, HT1197, DLD1, Colo205 and SW620) or dissociated tumors (primary patient HGSCs) were plated in triplicate at 20,000 cells/well in an in vitro organoid bioassay previously described 1. Cells were treated with vehicle, birinapant (300nM), carboplatin (50µM), or the combination of carboplatin and birinapant (co-therapy). Three days after treatment, organoids were dissociated. A portion of the cells were analyzed by flow cytometry for cell survival. The percentage of surviving cells was defined as the percentage of cells negative for propidium iodide and Annexin V-FITC staining at the time of the analysis. Remaining cells were re-plated in an organoid assay without drug and were qualitatively assessed for organoid growth 7 days after re-plating. Sensitivity to co-therapy was defined as ≤1% of cells surviving treatment and no growth after re-plating. The only exception to this rule was the Colo205 cell line, where approximately 3% of cells survived co-therapy based on flow cytometry but on two independent repeats no cells grew after re-plating. Therefore, the Colo205 cell line was also categorized as co-therapy sensitive.

Flow Cytometry

Populations of surviving cells were determined by AnnexinV-FITC (BD 556547)/propidium iodide (Invitrogen P3566) staining. Cells were released from organoids, immunostained as previously described1 and analyzed using a BD LSRII flow cytometer.

Western Blot

Western blot was performed as previously described 1 using the following antibodies: anti-pan-cIAP (R&D MAB3400 mouse monoclonal antibody, clone 315301 lot VXM0212081 at 0.5 µg/ µl) 1:1000, anti-cIAP1 (Abcamab108361 rabbit monoclonal antibody, clone EPR4673 lot GR140848-3) 1:1000, anti-cIAP2 (Abcam ab137393 rabbit polyclonal antibody, lot GR106772-17) 1:1000, anti-GAPDH (Abcam ab8245 mouse monoclonal antibody, clone 6C5 lot GR137268-16) 1:10000, anti-Traf2 (Cell Signaling Technologies #4724 rabbit monoclonal antibody, clone C192 Ref 10/2015 lot2) 1:1000, HRP-conjugated anti-mouse (BioRad 170-6516, lot L1706516 RevF) 1:4000, or HRP-conjugated anti-rabbit (BioRad 170-6515, Lot L005679 A) 1:4000. For visualization, Western blots were developed with Immobilon HRP substrate (Millipore) used according to manufacturer specifications.

Quantification of cIAP expression was accomplished using the following method. First, linearity of cIAP detection in this assay was tested. Three-fold dilutions of pan-cIAP (equal amounts of cIAP1 [R&D 818-P2-050 recombinant human cIAP1] and cIAP2 [R&D 818-1A-050 recombinant human cIAP2]) were runonSDS-PAGE gels (Invitrogen NuPAGE 4-12% Bis-Tris 15 well 1 mm thick gels, NP0323BOX) in 1X MOPS SDS running buffer (Invitrogen NP0001) at 140V, transferred to nitrocellulose (AmershamProtran 0.2 µm NC 10600006)at 80V for 1h at 4C, blocked in PBS+0.1% Tween-20 with 5% non-fat dried milk and probed with anti-pan-cIAP antibody diluted at 1:1000 in blocking buffer overnight. Membranes were then washed, incubated for 1h at room temperature in secondary antibody at 1:4000 in blocking buffer, washed and incubated for 5 min at RT inImmobilon HRP substrate. Signal was visualized by exposing drained membranes, wrapped in Saran plastic wrap, to autoradiography film (Santa Cruz UltraCruz film sc-201696) developed in a Konica-Minolta SRX-101A film developer. Resulting western blots were scanned as TIFF images for analysis with NIH ImageJ software v1.48 (publically available at An equal sized rectangular ROI (region of interest) was drawn around each lane. Lanes were converted to histograms and band peak intensities were determined using ImageJ tools. Positive correlation, calculated using Excel linear regression tools, was observed between signal intensity measured by western blot and amount of cIAP loaded in each lane (r2=0.9979).

To semi-quantitatively estimate cIAP levels in clinical samples, 20µg of lysate per specimen were run on an SDS-PAGE gel along with a 40ng cIAP standard (20ng each cIAP1 and cIAP2). Fractionated proteins were transferred to nitrocellulose and probed with anti-pan-cIAP and anti-GAPDH antibodies. Peak intensities for cIAP and GAPDH were determined using NIH ImageJ software as described. Cell lysates were normalized based on GAPDH levels obtained from the same exposure time western blot6. The lane with the highest GAPDH signal in each blot (GAPDHmax) was identified. GAPDH signal for each lysate (GAPDHi) was normalized to GAPDHmax using the following formula Ni=GAPDHi/GAPDHmax to yield a normalization factor (Ni) with the assumption that GAPDHmaxrepresents loading of 20 µg of tumor lysate accurately. This factor was used to normalize the cIAP signal in each lysate (Ci=cIAPpeak area per lysate/Ni). To quantify amount of cIAP per lane each lysate was then compared to the cIAP standard. The peak area of cIAP standard per blot was measured and called Cstd. Total cIAP amount of each lysate was then calculated by dividing normalized cIAP (Ci) by the cIAP standard and multiplying by 40ng (ngcIAPi = [Ci/Cstd] x 40ng)7. Analyses were performed by two independent researchers for replicate runs of each gelon at least two separate exposures of each gel. High variability of cIAP levels in individual specimens was observed between different exposures and replicate runs. To offset this variability,all results froma sample were averaged to obtain the final reported value of cIAP for each specimen.

In primary HGSC specimens and HGSC cell lines, this analysis was performed on the platinum resistant CA125 negative vs the platinum sensitive CA125 positive sorted cell fractions1 and on the bulk tumor cell population. For the platinumresistant non-HGSC cancer cell lines (A549, H226, H460, CaSki, HeLa, SiHa, 5637, J82, HT1197, DLD1, Colo205 and SW620), analysis was carried out on lysates made from cells that were treated for 72h with vehicle (bulk cells) or with 50µM carboplatin (enriched for cells surviving carboplatin treatment).

Lentiviral Plasmid Construction and Production

FU-CRW and FU-CGW lentiviral constructs, viral production, titering, and cell infection have been previously described 8,9. Knockdown constructs for cIAP1, cIAP2 and scramble controls were produced by inserting double-stranded shRNA cassettes (cIAP 1 forward: 5’TCCCTGGTTAAATCTGCCTTGGAAATCAAGAGTTTCCAAGGCAGATTTAACCA3’ and reverse: 5’AAAATGGTTAAATCTGCCTTGGAAACTCTTGATTTCCAAGGCAGATTTAACCA3’; cIAP 2 forward: 5’TCCCCAGTTCGTACATTTCTTTCATTCAAGAGATGAAAGAAATGTACGAACTG3’ and reverse: 5’AAAACAGTTCGTACATTTCTTTCATCTCTTGAATGAAAGAAATGTACGAACTG3’; scramble 1 forward: 5’TCCCGATAATTCGAGAAGACCAAGGTCAAGAGCCTTGGTCTTCTCGAATTATC3’ and reverse: 5’AAAAGATAATTCGAGAAGACCAAGGCTCTTGACCTTGGTCTTCTCGAATTATC3’. Scramble 2 forward: 5’TCCCTTTGTCACATCTTATGCTTACTCAAGAGGTAAGCATAAGATGTGACAAA3’ and reverse: 5’ AAAATTTGTCACATCTTATGCTTACCTCTTGAGTAAGCATAAGATGTGACAAA3’) into the BbsI site of psiRNA-W and then subcloningPacI cassettes containing the H1 promoter and shRNA into the FU-series vectors. Knockdown efficacy was confirmed by western blot for cIAP1 and cIAP2.

Immunohistochemistry Biomarker Assay

Immunohistochemistry (IHC) was performed on formalin-fixed paraffin-embedded sections using the pan-cIAP (Abcam ab25939) antibody at 1:250. To ensure that pan-cIAP antibodies reliably detected cIAP1 and cIAP2, lentiviral constructs (shcIAP1 and shcIAP2) were designed to selectively knockdown cIAP1 or cIAP2 respectively. HeLa cells were infected with either lentivirus alone, both lentiviruses or scramble controls (all of which were color-tagged) and isolated by fluorescence-activated cell sorting (FACS). Pellets were created from these singly infected, dually infected or scramble infected HeLa cells. Staining of these cell pellets demonstrated that the pan-cIAP antibody specificallydetected both cIAP1and cIAP2. FFPE sections of platinum resistant clinical samples were then stained using this antibody, and visualized using Alexa-594 conjugated anti-rabbit secondary antibody for immunofluorescence or biotinylated anti-rabbit secondary antibody followed by streptavidin conjugated HRP for DAB staining. Fluorescent images were taken on aKeyence BZ-X710 inverted microscope and camera system and analyzed using Keyence quantification software. Single extraction for DAPI staining was used to identify cells, which were then analyzed as positive or negative for cIAPimmunefluorescentstaining. To confirm the validity of this approach, five random high-powered fields were taken for each ofthree independent specimens. The percent cIAP positive cells (total number cIAP positive cells/total number nuclei) was quantified manually by two independent researchers and was compared to results obtained using the quantification software. For each specimen, no statistically significant differences were observed between manual and machine count quantification methods based on ANOVA (analysis of variance) analysis. Based on the agreement of quantification software and manual quantification, the quantification software was used for analysis of all remaining specimens. Positive and negative cell lines were analyzed in five independent runs to provide known standardized levels for the cIAP stain (Hela, 63.3 ± 2.3% cIAP positive; HeLa infected with shcIAP1 and shcIAP2, 1.65 ± 0.56% cIAP positive; Colo205 82.9 ± 3.1% cIAP positive; DLD1 2.4 ± 1.0% cIAP positive [mean ± SD]). These cell lines were included with each batch of immunostaining to confirm accuracy of staining and quantification.

Statistical Analysis

Survival curves were computed using the Kaplan-Meier method and adjusted p-values for comparing Kaplan-Meier curves and theirmedians were computed using the log-rank test. The p values were adjusted using the Fisher least significant difference (LSD) criterion in order to keep the overall type I error for all comparisons within the same cell line at 0.05 or less. Kaplan-Meier based median survival and the corresponding 95% confidence intervals were also computed. The Shapiro-Wiik statistic was used to assess normality of continuous variables. CBC and toxicology values were summarized by means and standard errors of the mean (SEM) and p values for their comparison were computed using ANOVA and Fisher LSD. Equality of variance for the ANOVA model was assessed using the Brown-Forsythe test. The p-values for comparing median primary HGSC expression levels of candidate markers in sensitive versus resistant samples and cancer cell lines were computed using the non-parametric Wilcoxon rank sum test. Non-parametric receiver operator characteristic (ROC) analysis was used to find the best threshold and compute accuracy statistics including: (a) the proportion of co-therapy sensitive samples correctly classified, (b) the proportion of co-therapy resistant samples correctly classified, and (c) the overall accuracy (average of a and b). The 95% confidence interval for accuracy statistics were computed using the resampling (bootstrap) method.

References

1.Janzen, D.M., et al. An apoptosis-enhancing drug overcomes platinum resistance in a tumour-initiating subpopulation of ovarian cancer. Nature communications6, 7956 (2015).

2.Oguri, S., et al. Clinical pharmacokinetics of carboplatin. Journal of clinical pharmacology28, 208-215 (1988).

3.van Hennik, M.B., et al. Comparative pharmacokinetics of cisplatin and three analogues in mice and humans. Cancer Res47, 6297-6301 (1987).

4.Amaravadi, R.K., et al. A Phase I Study of the SMAC-Mimetic Birinapant in Adults with Refractory Solid Tumors or Lymphoma. Mol Cancer Ther14, 2569-2575 (2015).

5.NIH Office of Animal Care and Use. Guidelines for endpoints in animal study proposals. (2016).

6.Laederich, M. The 4 important steps for western blot quantification. in The missing manual for bioscientists, Vol. 2016 (2015).

7.Taylor, S.C. & Posch, A. The design of a quantitative western blot experiment. BioMed research international2014, 361590 (2014).

8.Memarzadeh, S., et al. Enhanced paracrine FGF10 expression promotes formation of multifocal prostate adenocarcinoma and an increase in epithelial androgen receptor. Cancer cell12, 572-585 (2007).

9.Xin, L., Ide, H., Kim, Y., Dubey, P. & Witte, O.N. In vivo regeneration of murine prostate from dissociated cell populations of postnatal epithelia and urogenital sinus mesenchyme. Proc Natl Acad Sci U S A100 Suppl 1, 11896-11903 (2003).

1