Actionable spontaneous antibody responses antagonize malignant progression in ovarian carcinoma

Human ovarian carcinoma tissues were procured under protocols approved by the Committee for the Protection of Human Subjects at Dartmouth-Hitchcock Medical Center (#17702), by the Institutional Review Board at Christiana Care Health System (#32214), and by Advarra Institutional Review Board (#00000971) and H. Lee Moffitt Cancer Center Scientific Review Committee (MCC#18974). Human endometrioma tissues were procured under a protocol approved by the Institutional Review Board at Ponce Research Institute (#1903009574). Informed consent was obtained from all subjects. Human ovarian cancer cell lines including OVCAR3 (RRID: CVCL_0465), SKOV3 (RRID: CVCL_0532), and human endometrial stromal cells (HESC), highly invasive and immortalized with human telomerase reverse transcriptase (hTert), were obtained from ATCC. TOV21G (RRID: CVCL_3613), RMG-I (RRID: CVCL_1662), Caov3 (RRID: CVCL_0201), A2780 (RRID: CVCL_0134), OVCAR4 (RRID: CVCL_1627), OVCAR5 (RRID: CVCL_1628), OVCAR8 (RRID: CVCL_1629), Kuramochi (RRID: CVCL_1345) and BRCA OVCAR were obtained as a gift from Dr. Rugang Zhang at The Wistar Institute. Human endometriotic epithelial cells (12Z, RRID: CVCL_0Q73) were obtained as part of a collaboration with Dr. Asgerally Fazleabas and Dr. Anna Starzinski-Powitz 10 . All cell lines except RMG-I, HESC and 12Z were cultured in RPMI 1640 medium (Fisher Scientific) supplemented with 10% fetal bovine serum (FBS), penicillin (100 IU/mL), streptomycin (100IU/mL), L-glutamine (2mM), and sodium pyruvate (0.5mM). RMG-I was cultured in Ham’s F12 medium (Fisher Scientific) supplemented with 10% fetal bovine serum (FBS), penicillin (100 IU/mL), streptomycin (100 IU/mL), L-glutamine (2mM), and sodium pyruvate (0.5mM). 12Z was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM)/F12 supplemented with 10% FBS. HESC was cultured in phenol-free DMEM supplemented with charcoal-treated 10% FBS and 1% Insulin-Transferrin-Selenium (ITS). All cell lines were routinely tested for Mycoplasma by PCR. Cells were used within 20 passages from thaw for in vitro experiments and 10 passages from thaw for in vivo experiments. Cryopreserved single-cell suspensions of two stage IIIC ovarian clear cell ovarian carcinomas, two stage IIIC ovarian endometrioid carcinomas, and two endometriomas were thawed and prepared, and CD19 + B cells isolated, activated, and immortalized, as previously described 5 . The conditioned medium from each was collected and concentrated using centrifugal filter units (Millipore Sigma Amicon, UFC900324). From the concentrated medium, human IgA and IgG were purified using immunoglobulin purification kits (LigaTrap, LT-146KIT and LT-095KIT) according to the manufacturer’s protocols. To characterize the specificities of these antibodies, they were analyzed for reactivity against a proteome microarray that includes greater than 80% of the human proteome (HuProt ™ , CDI Laboratories, Mayagüez, Puerto Rico). In order to determine which peptide would be best targeted by an antibody, the amino acid sequence for SDCBP was run through two epitope prediction tools (Bepipred Linear Epitope Prediction 2.0 and ABCPred) to determine predicted epitopes. Predicted targetable extracellular domains were chosen as the peptides for tetramer analysis. We tetramerized two biotinylated peptides (GenScript) contained in the extracellular domain of syndecan binding protein (SDCBP) using PE-labelled fluorescent streptavidin (BioLegend, 405203) and used flow cytometry to determine the percent of specific antigen-reactive immortalized B cells which are secreting antibodies against the target molecule from each of the six immortalized B cell lines, using the protocol previously described. 5 We then used fluorescence-activated cell sorting with EEEIRANVAVVSGAPL peptide to sort endometrioid carcinoma-derived immortalized B cells specific for SDCBP from the pool of immortalized B cells. Single-cell V(D)J B-cell receptor sequencing was performed by the Moffitt Cancer Center Molecular Genomics Core using the 10XGenomics Chromium system. 56 cells were encapsulated and sequenced as previously described 5 . BCR reads sequenced by V(D)J assay were aligned to GRCh38 reference transcriptome using Cell Ranger VDJ (v.3.1.0, 10X Genomics). BCR heavy and light chains were assembled and annotated using Cell Ranger VDJ to determine clonotypes. Recombinant IgG4 antibodies were produced by Genscript. In brief, corresponding DNA sequences for the immunoglobulin heavy chain and light chain were synthesized and the complete sequence was subcloned into a pcDNA3.4 vector and expressed in HD 293F cells. IgG4 antibodies were eluted from cell culture supernatants. Molecular weight and purity were analyzed by SDSPAGE and high-performance liquid chromatography. RNA was extracted from ovarian clear cell carcinoma, endometrioid carcinoma, and high-grade serous carcinoma tissues and cell lines using the RNEasy Plus Mini Kit (Qiagen) to quantify SDCBP expression. Total RNA were reverse transcribed using a high-capacity cDNA reverse transcription kit with RNAse inhibitor (ThermoFisher, 4374966). Quantification of SDCBP was performed on the 7900HT Real-Time PCR system (Thermo Fisher Scientific) using SYBR Select Master Mix (Applied Biosystems, with forward primer: 5’-TCTCGAAGACTTGAAGGTAGACA-3’, and reverse primer: 5’-CGGCCACATTTGCACGTATT-3’). Expression was normalized to levels of the endogenous reference control gene GAPDH (forward primer: 5’-CCTGCACCACCAACTGCTTA-3’; and reverse primer: 5’-AGTGATGGCATGGACTGTGGT-3’). Proteins were extracted from ovarian clear cell carcinoma, endometrioid carcinoma, and high grade serous carcinoma tissues and cell lines, as well as an endometriosis cell line, and quantified as previously described 11 . Proteins were loaded onto a 10% Bis-Tris polyacrylamide gel. Membranes were incubated with recombinant anti-SDCBP IgG4 antibodies described above (Genscript) or rabbit anti-human SDCBP (Sigma). After washing with TBST, the membranes were incubated with horseradish peroxidase-conjugated rabbit anti-human IgG (1:5000, Cat. Ab6759, Abcam, RRID:AB_955434). Horseradish peroxidase-conjugated anti-β-actin antibody (1:5000, Cat. 5125S, Cell Signaling Technology, RRID: AB_1903890) was used as a loading control. Images were captured using the BioRad ChemiDoc imaging system and GE Healthcare Amersham ECL Prime Western Blotting Detection Reagents (cat. 12316992, Fisher Scientific). All in vivo protocols were approved by the University of South Florida’s Institutional Animal Care and Use Committee. Female NOD-SCID-gamma (NSG) mice, originally obtained from Jackson Laboratory, were maintained by the animal facility of H. Lee Moffitt Cancer Center and Research Institute. Mice were injected subcutaneously with 5×10 6 RMG-I, TOV21G, or OVCAR3 cells in the right flank. Once tumor uptake was demonstrated, seven to nine days post-injection, mice were randomly divided into two treatment groups of five mice each: irrelevant IgG4 control and anti-SDCBP IgG4 treatment. Twice weekly, tumor volume was measured and IgG4 control and anti-SDCBP IgG4 treatment were administered by intratumoral or peritumoral injection (100μg/100μL). Tumor volume was calculated as (L × W 2 )/2, in which L is length and W is width. Once any group became moribund, all mice were euthanized and tumor weight was recorded. Tumor specimens were formalin-fixed and paraffin-embedded or mechanically dissociated into single-cell suspensions. All experiments were repeated at least twice with similar results. The Shapiro–Wilk test was applied to determine whether data were normally distributed. Wilcoxon matched-pairs signed rank tests and unpaired t tests were applied as indicated. Two-tailed analyses were performed. Error bars represent standard error of the mean. A p value less than 0.05 was considered statistically significant. Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Prism, RRID:SCR_002798).

To identify antibodies produced by B cells recognizing possible shared targets in clear cell ovarian carcinoma, endometrioid ovarian carcinoma, and endometriosis, we analyzed viable single-cell suspensions from two cryopreserved samples of each type. B cells were isolated, activated, and immortalized using Epstein-Barr virus. These six immortalized B cell pools were found to secrete IgG and IgA at titers in the 0.7–37 mg/mL range. Using HuProt ™ proteome arrays containing >80% of the human proteome 5 , IgG and IgA tumor reactivities were decoded ( Figure 1A ). Greater than 200 targets were identified for each sample, for both IgA and IgG antibodies in independent analyses. There were nine molecules that met the following criteria: 1) They were either secreted or included an extracellular domain, and were therefore accessible to antibodies in live cells; 2) they were recognized by both IgA and IgG; and 3) reactivity was shared in every endometrioma and carcinoma sample ( Figure 1B ). Among these, we focused on SDCBP, a transmembrane molecule that links syndecan-mediated signaling to the cytoskeleton. SDCBP regulates TGF-β1-induced Smad activation and EMT by inhibiting caveolin-mediated TGF-β type I receptor internalization, and has been demonstrated to drive proliferation, migration, invasion, and angiogenesis 12 – 14 . SDCBP has been associated with unfavorable prognosis in multiple solid malignancies, including breast and colorectal cancer 15 – 16 . Furthermore, SDCBP has been reported as a therapeutic target for cancer metastases 17 , as well as cancer stemness and chemoresistance 18 . Because antibodies targeting SDCBP are being spontaneously produced in the endometriosis and ovarian cancer microenvironment, we tetramerized two different biotinylated 16–20mer peptides contained in the extracellular domain of SDCBP using fluorescent streptavidin and used flow cytometry to determine the percent of specific antigen-reactive immortalized B cells which are secreting antibodies against SDCBP from each of the six immortalized B cell lines. Based on these results, we used fluorescent activated cell sorting to sort endometrioid ovarian cancer-derived B cells specific for SDCBP and performed single-cell B cell receptor sequencing on the sorted population of B cells ( Figure 2 ). Bioinformatic analysis of the B cell receptor sequencing determined the sequence of the heavy chain and light chain of the most common B cell receptor, identified in 96% of these cells ( Figure 2 ). We then produced a recombinant antibody targeting SDCBP using these heavy chain and light chain sequences on an IgG4 backbone. IgG4 was specifically selected to avoid antibody-dependent killing of normal cells that also express SDCBP via antibody-dependent cell-mediated cytotoxicity or antibody-dependent cellular phagocytosis. The Cancer Genome Atlas RNA sequencing data were queried, demonstrating SDCBP mRNA expression in all tested cancer types, supporting its potential as a therapeutic target for patients with a diverse range of histologies ( Figure 3A , Supplementary Figure 1 ) 19 . RT-qPCR was performed to assess the baseline prevalence of SDCBP mRNA expression in ovarian cancer cell lines and tumor tissues. 26/29 (90%) of the cell lines and tissue samples express SDCBP mRNA – and all of the clear cell (5/5) and endometrioid (7/7) cell lines and tissue samples express SDCBP ( Figure 3B ). Tumor-derived, recombinant anti-SDCBP IgG4 recognizes recombinant SDCBP in Western blot analysis. Western blot demonstrated SDCBP protein expression in ovarian cancer cell lines and tumor samples tested, including clear cell, endometrioid, and high grade serous histologies ( Figure 3C , Supplementary Figure 2 ), as well as expression in endometriosis ( Figure 3D ). To investigate the antitumor effects of a recombinant anti-SDCBP IgG4 antibody in vivo , we created a high grade serous ovarian carcinoma mouse model using subcutaneous injection of OVCAR3 cells ( Figure 4A ). Mice treated with the anti-SDCBP IgG4 antibody demonstrated significantly reduced tumor growth and substantially reduced tumor weight compared to those treated with the control irrelevant IgG4 antibody (p=0.004 and p=0.05, respectively, Figure 4B – D ). To determine whether this could be extrapolated to other ovarian cancer histologies, specifically clear cell carcinoma – one of the endometriosis-associated ovarian cancer types, we next evaluated the anti-SDCBP antibody in an RMG-I tumor model. In this model, we also observed significantly reduced tumor growth and lower tumor weight in the treatment arm compared to the control arm (p<0.05 and p=0.34, respectively, Figure 5A – B ). To confirm these findings, we performed the same experiments in a second clear cell line, TOV21G, which yielded similar results ( Figure 5C – D ).

An α-SDCBP IgG4 has demonstrated anti-tumor efficacy in SDCBP + CCC and HGSOC, and SDCBP-targeted therapy for endometriosis and associated malignant conditions, as well as HGSOC, warrants further investigation.

A novel anti-SDCBP IgG4 antibody has demonstrated preclinical anti-tumor efficacy in HGSOC and CCC, with the possibility of use in EC and other tumor types given the broad expression of SDCBP among tumors. As HGSOC is the most common histologic type of ovarian cancer, and as CCC is relatively chemotherapy-resistant and associated with increased risk of poor outcomes, these two disease types represent an area of high unmet need for novel therapeutic strategies. This study establishes the potential of this technique in identifying novel therapeutic targets for CCC and verifies the utility of the technique for HGSOC, in which we had previously identified SDCBP as a target of tumor-infiltrating, IgG-producing B cells in six tumor samples 5 . This study also highlights the importance of antibody responses in different histologic subtypes of ovarian cancer and supports that intratumoral B cells, through the spontaneous production of antibodies, could exert a protective role against malignant progression. SDCBP has been previously described in small extracellular vesicles 20 ; we did not have enough serum samples from the patients analyzed to perform ELISAs, but it is theoretically possible that reactive antibodies could be detected in circulation. The concordance of antibodies being spontaneously produced in the microenvironment of endometriosis and endometriosis-related ovarian cancers raises the question of whether antibodies such as those targeting SDCBP could prevent or delay progression from endometriosis to EAOCs - in addition to delaying tumor progression in established EAOCs. In patients with ovarian cancer, high SDCBP expression is associated with a slight trend towards worse OS (p=0.37). The issue, however, is that the level of expression of SDCBP is very high in virtually all included ovarian cancers in this analysis as seen in Figure 3A , while there is more dispersion in the level of expression of SDCBP in other gynecologic and other cancer types, allowing a better comparison between high- and low-expressing tumors. Of note, high SDCBP expression is associated with statistically significant worse OS in breast, cervical, and endometrial cancers, among others ( Supplementary Figure 3 ) (Cancer Gene Prognosis Atlas, https://cgpa.moffitt.org/ ), indicating multiple avenues for further research in other cancer types - gynecologic and otherwise. The in vivo effects of the anti-SDCBP IgG4 antibody support further preclinical research investigating the underlying mechanism of action, as well as the continued in vivo study of other novel therapeutic targets identified in this study, such as OLFML2B which is the target of ongoing study. Future studies should also explore the effect of the anti-SDCBP IgG4 antibody in endometriosis models.

In 2021, there were an estimated 21,410 new cases of ovarian cancer leading to 13,770 deaths 1 . The current standard of care involves aggressive cytoreductive surgery and chemotherapy, which initially elicits a response in more than 80% of cases 2 , but the majority of patients ultimately relapse and develop chemotherapy-resistant disease. The estimated 5-year survival is 49.1% in all cases; the majority of patients are diagnosed at an advanced stage, in which case the 5-year survival drops to 30.1% 1 . There is thus an urgent need for novel therapeutics in this arena. While immunotherapies such as immune checkpoint blockade are changing the trajectory of various cancers such as melanoma, the results in ovarian cancer have been disappointing. Interestingly, tumor-infiltrating B lymphocytes are present in most ovarian cancers and have been associated with improved patient outcomes 3 – 5 . However, limited research has focused on humoral response in endometriosis-associated ovarian cancers or endometriosis. Endometriosis is a condition in which endometrial tissue is present outside of the uterine cavity, which occurs in approximately 10% of women. Endometriosis is associated with a two- to three-fold increase in a woman’s risk of developing clear cell or endometrioid ovarian cancers, and ovarian endometriosis has been associated with as high as a ten-fold increased risk of clear cell and five-fold risk of endometrioid ovarian cancer 6 , 7 . Recent studies have concluded that endometriosis is a precursor lesion to these endometriosis-associated ovarian cancers (EAOC), with corresponding somatic mutations identified in both 8 , 9 . Therefore, we hypothesized that antitumor humoral responses generated by tumor-infiltrating B lymphocytes in ovarian cancer and B lymphocytes in the microenvironment of endometriosis target specific antigens that can be used to identify novel, targetable antigen domains that can exert a protective effect on progression of established ovarian carcinomas.