Primary Immunoprevention of Epithelial Ovarian Carcinoma by Vaccination against the Extracellular Domain of Anti-Müllerian Hormone Receptor II

Epithelial ovarian carcinoma (EOC) is the most lethal gynecologic malignancy and occurs predominantly in postmenopausal women who account for >75% of all cases. EOC is typically diagnosed at late stages of the disease, resulting in a high rate of recurrence following the current standard of care and a 5-year overall survival rate under 47% (1, 2). The high rate of EOC recurrence and the relatively low 5-year overall survival indicate a great need for more effective ways to control this disease.

We have advocated that optimized control of adult onset cancers like ovarian cancer requires vaccine induction of preemptive immunity that provides safe and effective primary prevention. However, immune-mediated cancer prophylaxis is currently limited to vaccination against disease-causing pathogens, including the hepatitis B virus (HBV) linked to the development of liver cancer (3) and the human papilloma virus (HPV) linked to cervical carcinoma as well as anal, penile, and oropharyngeal cancers (4). As no such confirmed pathogen is implicated in EOC tumorigenesis, we have proposed that tissue-specific self-proteins that are “retired” from autoimmune inducing levels of expression as normal tissues age but are expressed in emerging tumors may substitute for unavailable disease-associated pathogens as vaccine targets for inducing safe and effective primary immunoprevention of adult onset cancers (5).

We found that the extracellular domain of anti-Müllerian hormone receptor, type II (AMHR2-ED) has the required features of a “retired” protein for targeting safe and effective primary immunoprevention of EOC. AMHR2 is a serine/threonine kinase receptor homologous to type II receptors of the TGFβ family (6). The human AMHR2 gene contains 11 exons with seven known alternatively spliced variants producing three known coded proteins, one additional variant with protein coding features, and three noncoding transcripts with no open reading frames (6–8). In adult women, the longest AMHR2 transcript codes for a 568 amino acid protein containing a 125 amino acid extracellular ligand-binding domain (7, 8). AMHR2 signaling causes regression of the Müllerian ducts during male fetal development and regulates oocyte development and follicle production in adult females, thereby providing substantial control of ovarian reserve and fertility (6, 9–11).

Here, we show that in normal human tissues, AMHR2-ED is expressed exclusively in the ovary, that AMHR2-ED expression declines dramatically in the postmenopausal human ovary and in the related aged murine ovary, and that AMHR2-ED is expressed in approximately 90% of human EOC. Moreover, our preclinical studies in mice show that AMHR2-ED vaccination induces safe and effective prevention and treatment of murine EOC and that this induced tumor immunity is accompanied by a surprisingly mild and transient oophoritis in young mice vaccinated at 8 weeks of age and by no detectable inflammatory ovarian changes in old mice vaccinated at 9 months of age. We found that active vaccination and specific IgG targeting of AMHR2-ED induces proinflammatory immune responses, resulting in growth arrest and apoptotic cell death of murine EOC in vivo and in vitro. Our data support the view that AMHR2-ED vaccination not only provides effective immunotherapy against EOC, but also has the potential to provide safe and effective primary immunoprevention of this disease in women at high genetic or familial risk and in cancer-free postmenopausal women who account for the majority of EOC cases.

The DNA-coding sequence of the entire 125 amino acids of the mature native AMHR2 extracellular domain lacking the signal peptide was used to generate the expression construct (NCBI Reference Sequence: NM_144547.2; Uniprot Q8K592; ref. 12). To optimize protein folding and enhance overall yield, substitutions for native codon sequences were made (Dapcel), and the optimized DNA was synthesized de novo to include an N-terminal methionine immediately followed by a FLAG tag, and a C-terminal 6xHis tag. The construct was inserted into the NdeI-BamHI sites of pET3a (GeneArt), thereby providing a recombinant dual-tagged AMHR2-ED protein. Plasmids containing these inserts were transformed in BL21 Star E. coli (Thermo Fisher Scientific), and after induction with isopropyl-β-D thiogalactopyranoside (IPTG; Amresco), high-level expression colonies were selected and sequenced to confirm proper gene orientation and alignment. The 6xHis-tagged AMHR2-ED was purified under denaturing conditions using nickel-nitrilo triacetic acid (Ni-NTA) affinity chromatography (Qiagen). The purified protein was electrophoresed on denaturing SDS-PAGE gels (Bio-Rad) and blotted onto immunblot PVDF membrane (Bio-Rad). Immune detection of AMHR2-ED was performed using the enhanced chemiluminescence system (Amersham Biosciences) with HRP-conjugated Penta-His antibody (Qiagen). Prior to use in vitro, the 6xHis-tagged AMHR2-ED was further purified by reverse-phase HPLC or treated with endotoxin removal beads (Miltenyi Biotec) to yield endotoxin-free protein (13). Levels of endotoxin were determined to be <0.05 endotoxin units (<5 pg) per mg recombinant protein.

TgMlSIIR-Tag (DR26) and TgMlSIIR-Tag (low) transgenic mice were generously provided by Dr. Denise C. Connolly, Developmental Therapeutics Program, Fox Chase Cancer Center (Philadelphia, PA). Female TgMISIIR-Tag (DR26) transgenic mice develop bilateral autochthonous EOC due to expression of the large T antigen (Tag) of SV40 under control of the AMHR2 promoter (14). TgMISIIR-Tag (low) transgenic female mice do not develop autochthonous EOC due to errant transgene insertion, but these mice are immunologically tolerant to SV40-TAg and effectively serve as histocompatible recipients of transplantable tumors derived from TgMlSIIR-Tag (DR26) mice (15). Colonies of transgenic mice were established and maintained by breeding male transgenic mice with wild-type syngeneic C57BL/6 females (The Jackson Laboratory). BALB/c, FVB/NJ, and A/J female mice were obtained commercially (The Jackson Laboratory). Mice were immunized at 8 weeks of age by a single subcutaneous injection in the abdominal flanks with 100 μg recombinant mouse AMHR2-ED or grade 7 ovalbumin (Sigma-Aldrich) in 200 μL of an emulsion of equal volumes of water and complete Freund's adjuvant (CFA; Difco) containing 200 μg of Mycobacterium tuberculosis H37Ra. For therapeutic intervention, TgMISIIR-Tag (DR26) mice were vaccinated with AMHR2-ED when tumors emerged and became palpable at 3 months of age.

MOVCAR cells were generated from TgMISIIR-Tag (DR26) ascites fluid as described previously (15). Briefly, ascites fluid from EOC-bearing mice was washed with sterile cold PBS to lyse red blood cells, and the pelleted cells were cultured in T75 flasks (BD Biosciences) in DMEM (Cell-Services Media, Cleveland Clinic, Cleveland, OH) supplemented with 5% FBS (HyClone), 0.005 mg/mL bovine insulin (Sigma-Aldrich), 5% HEPES buffer (Sigma-Aldrich), 2 mmol/L l-glutamine (Thermo Fisher Scientific), and 1% penicillin/streptomycin (Invitrogen). After overnight incubation at 37°C in 5% CO₂, adherent cells were trypsinized, washed thoroughly in PBS, and recultured with fresh supplemented media for several days until adherent monolayers formed. Adherent cells were trypsinized again, washed in PBS, and seeded with fresh supplemented media in T75 flasks for repeated passage at 3–4 day intervals until the cells became morphologically consistent and ready for inoculation into TgMISIIR-Tag (low) recipients. MOVCAR cells were periodically tested to be Mycoplasma free and their passage number did not exceed 15.

MOVCAR cells were suspended in PBS, and 4 × 10⁶ cells were injected subcutaneously in a total volume of 100 μL in the left dorsal flank of TgMISIIR-Tag (low) female mice. Tumor growth was assessed regularly using a Vernier caliper and the endpoint for all experiments involving transplantable tumors was defined by a tumor measurement of 17 mm in any direction.

Bilateral ovarian tumor growth in TgMISIIR-Tag (DR26) female mice was measured monthly using the Vevo 770 high-resolution in vivo micro-imaging system for small animal imaging (VisualSonics) as described previously (16). Measurements and calculations of tumor areas were performed using the Vevo software B-Mode measurement tool allowing for a 2D assessment of ovarian tumor size in vivo with the polygon region of interest setting (VisualSonics). Measurements of solid tumor size by B-mode sonography has been shown to correlate well with histologic measurements (17).

Tissues were excised and stored frozen in RNA-Later (AM7021, Life Technologies). RNA was either extracted from each tissue by homogenization in TRIzol reagent (Invitrogen) or purchased commercially (OriGene Technologies; and ILS Biotech). Gene expression was measured by qRT-PCR using SYBR Green PCR Mix (43098155, Applied Biosystems) with gene-specific primer pairs (Invitrogen; Table 1). Relative gene expression for each test gene was determined by normalization to β-actin expression in each tissue sample.

Transgene expression in offspring of transgenic mice was determined by PCR amplification as described previously (Table 1; ref. 16). Western blot analysis of human ovarian tissues was performed as described previously (16) using a rabbit polyclonal primary antibody specific for AMHR2-ED (Santa Cruz Biotechnology) or a mouse mAb specific for β-actin (Sigma-Aldrich). Detection was facilitated using secondary antibodies specific for rabbit and mouse IgG (Southern Biotech).

Four weeks after vaccination, splenocytes, CD4⁺ T cells, and CD8⁺ T cells were cultured in triplicate at 2 × 10⁵ cells per well in 96-well multi-screen HTS plates (EMD Millipore), precoated overnight with 4 μg/mL purified capture antibodies specific for mouse IFNγ, IL17A (eBioscience), and IL5 (BD Biosciences). Cells were cultured with 50 μg/mL of test and control antigens in a total volume of 200 μL DMEM supplemented with 10% FBS (Thermo Fisher Scientific), 5% HEPES buffer (Sigma-Aldrich), 2 mmol/L l-glutamine (Thermo Fisher Scientific), and 1% penicillin/streptomycin (Invitrogen). Positive control wells contained 2 μg/mL mouse CD3 antibody (BD Biosciences), and negative control wells contained 50 μg/mL recombinant mouse β-casein, a recombinant protein antigen generated in E. coli in a manner similar to AMHR2-ED. CD4⁺ and CD8⁺ T cells were positively enriched (>90% purity) from splenocytes by magnetic bead separation using a MidiMACS cell separator (Miltenyi Biotec). Cultures with purified CD4⁺ or CD8⁺ T cells were supplemented with 1 × 10⁵ γ-irradiated (20 Gy) syngeneic splenocyte feeder cells to facilitate antigen presentation. ELISPOTs were processed as reported previously (16), and spots were counted using an ImmunoSpot S6 analyzer with proprietary ImmunoSpot 5.1 software (Cellular Technologies Limited). T-cell frequencies are expressed as spot-forming units per 1 × 10⁶ cells.

Total serum IgG titers to AMHR2-ED were determined by serial dilutions of sera in 96-well plates as described previously (16). Isotype-specific serum antibody titers to AMHR2-ED were determined according to the manufacturer's instructions using the Mouse Typer Isotyping Panel (Bio-Rad).

Immunostaining for tissue detection of T cells was performed as described previously (16). Briefly, CD3⁺ T cells were immunostained in 5-μm sections of formalin-fixed paraffin-embedded ovarian tissues using a 1/500 dilution of rat anti-mouse CD3 (AbD Serotec; clone CD3-12), followed by a 1/1,000 dilution of biotinylated goat anti-rat IgG (Vector Laboratories). CD4⁺ and CD8⁺ T cells were immunostained in 5-μm frozen sections of ovarian tumors by incubation with a 1/50 dilution of mouse CD4 or CD8 primary antibodies (BD Biosciences), followed by treatment with secondary detection antibodies using the mouse HRP-Polymer Kit (Biocare Medical). Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining was performed using the TUNEL Kit (EMD Millipore Corporation), and apoptotic bodies were detected with Vina Green staining for caspase-3 using rabbit polyclonal caspase-3 antibody (R&D Systems). Visualization was achieved using the VECTASTAIN Elite ABC Kit, the DAB Substrate Kit (Vector Laboratories), hematoxylin and eosin counterstaining, and mounting of sections in Cytoseal 60 (Stephens Scientific) for examination by light microscopy. TUNEL-positive and caspase-3–positive cells were quantified using NIS-ELEMENTS-BR-version 4 software (Nikon).

Three weeks after vaccination of C57BL/6 females (The Jackson Laboratory) with either AMHR2-ED or grade 7 ovalbumin (Sigma-Aldrich), CD4⁺ T cells and B220⁺ B cells were enriched >90% from splenocytes by magnetic bead separation (Miltenyi Biotec). Without any prior reactivation, 2 × 10⁷ B220⁺ B cells were transferred by intraperitoneal injection into naïve 8-week-old syngeneic female TgMlSIIR-Tag (low) recipient mice. Prior to transfer, CD4⁺ T cells were activated in vitro for 3 days with 20 μg/mL immunogen in 24-well flat-bottom Falcon plates (BD Biosciences) in a total volume of 2.0 mL/well in supplemented DMEM. A total of 2 × 10⁷ activated CD4⁺ T cells were transferred into naïve recipient mice. Cell-free sera were collected 4 weeks after vaccination with either AMHR2-ED or ovalbumin, and the IgG was affinity purified from immune sera using protein G-sepharose (BioVision). TgMlSIIR-Tag (low) naïve recipients were injected intraperitoneally twice one week apart with 200 μg purified IgG in 200 μL PBS. Twenty-four hours after the final injection of cells or IgG, each TgMlSIIR-Tag (low) recipient was inoculated with 4 × 10⁶ MOVCAR cells, and tumor growth was evaluated as described above. Prior to transfer, purities of enriched lymphocytes were determined by flow cytometry using a FACSAria II flow cytometer (BD Biosciences), and results were analyzed using FlowJo software (FlowJo).

Autochthonous EOCs from TgMISIIR-Tag (DR26) mice vaccinated with CFA or AMHR2-ED were digested for 30 minutes at 37°C in DMEM containing 50 KU of DNase I (Sigma-Aldrich) and 0.2 mg/mL collagenase II (Thermo Fisher Scientific). Prior to use in ELISPOT assays, tumor-infiltrating lymphocytes (TIL) were negatively enriched by passage of the isolated leukocytes through nylon wool fiber.

Affinity-purified IgGs from sera of AMHR2-ED vaccinated and CFA vaccinated control mice were tested in three different in vitro assays for the ability to arrest growth and induce apoptosis of murine EOC cells. MOVCAR cells at 1 × 10⁴ cells/microtiter well in 50 μL media were incubated for 72 hours with varying dilutions (1/200, 1/500, and 1/1,000) of IgG. The MTS assay was performed using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega) according to the manufacturer's instructions. The bioreduction of the MTS tetrazolium compound by live cells into a soluble colored formazan product measured by absorbance at 490 nm is directly proportional to the number of living cells in culture (18). To confirm the results obtained from MTS assays, MOVCAR cell viability was determined by trypan blue exclusion by counting 200 cells per field for each sample of IgG-treated cells. In addition, apoptotic MOVCAR cells from the IgG-treated cultures were quantified by flow cytometry using the FITC Annexin V Apoptosis Detection Kit (BD Biosciences).

To determine fertility phenotypes as an assessment of ovarian function, age-matched AMHR2-ED and CFA control vaccinated C57BL/6 female mice were mated serially with the same C57BL/6 male partners over several mating cycles. Fertility efficiency of the vaccinated female mice was assessed by examining mean number of pups per litter and mean pup birth weights.

Differences between mRNA expression levels were compared using the Student t test and two-way ANOVA. Differences in ELISPOT frequencies and tumor size were compared using two-way ANOVA. Differences in fertility, growth arrest, and frequencies of trypan blue–positive and Annexin V–positive apoptotic cells were compared using the unpaired Student t test. Differences in mouse survival were determined by log-rank and χ² using correlated samples. Differences in TUNEL staining and caspase-3 positive scoring were assessed using the two sample Student t test. All experiments were repeated 3 times independently.

All protocols for animal research met with the prior approval of the Institutional Animal Care and Use Committee of the Cleveland Clinic (IACUC) in compliance with the Public Health Service policy on humane care and use of laboratory animals (IACUC #2015-1390 and IACUC #2016-1585). All human tissues examined were obtained and examined with prior approval from the Internal Review Board (IRB) of the Cleveland Clinic (IRB 12–1404) as exempt research involving use of existing data without recording subject identifiers.

Although ovarian expression of AMHR2 is well documented, extraovarian AMHR2 transcripts are known to be expressed in other adult human tissues most notably adrenal gland, spleen, pancreas, and skeletal muscle (Fig. 1A; refs. 4–6). However, none of these nonovarian AMHR2 transcripts code for exons 1–3 representing the entire extracellular ligand-binding domain for anti-Müllerian hormone (AMH; Fig. 1B; refs. 6–8). We confirmed these published observations by showing that the adrenal gland and pancreas express the cytoplasmic domain of AMHR2 (AMHR2-CD) but that AMHR2-ED expression is confined to the ovary (Fig. 1C). In addition, we also found that human postmenopausal ovaries (mean, 64 years; range, 52–95 years) and comparable old mouse ovaries (9 months of age) have significantly lower AMHR2-ED expression levels when compared respectively with human premenopausal ovaries (mean, 31 years; range, 27–45 years) and young mouse ovaries (6 weeks of age; Fig. 1D; P < 0.0001 for both human and mouse ovaries). Moreover, all human EOCs examined showed high-level expression of AMHR2-ED transcripts assessed by qRT-PCR (Fig. 1E) and substantial AMHR2-ED protein levels assessed by Western blot analysis (Fig. 1F).

To determine whether AMHR2-ED has the potential to target safe and effective EOC immunoprevention, we generated a recombinant protein spanning the entire 125 amino acids of the mature mouse AMHR2 extracellular domain (Fig. 2A and B). We found that our recombinant AMHR2-ED protein migrated as a single band on SDS-PAGE when detected by Coomassie blue staining (Fig. 2C, left) or anti-His Western blot (Fig. 2C, right). One month after AMHR2-ED vaccination of young 8-week-old C57BL/6 female mice, ELISPOT analysis showed elevated splenocyte frequencies of IFNγ-producing antigen-specific T cells at a mean of 1 per 3,831 splenocytes, and IL17 at a mean of 1 per 21,739 splenocytes, but not IL5 (Fig. 2D). The proinflammatory T cells responding to AMHR2-ED were predominantly CD4⁺ T cells but also included a minor CD8⁺ T-cell population producing IFNγ (Fig. 2E) or IL17 (Fig. 2F). Four months after AMHR2-ED vaccination, serum IgG antibody responses against AMHR2-ED were detectable at dilutions up to 1/50,000 (Fig. 2G) and consisted predominantly of IgG1 and IgG2b isotypes (Fig. 2H). AMHR2-ED was highly immunogenic in several mouse strains with completely divergent MHC H-2 haplotypes, including C57BL/6 (H-2^b), BALB/c (H-2^d), A/J (H-2^a), and FVBN/J (H-2^q) in which IFNγ-secreting splenocytes reached mean frequencies of 1 per 4,310, 1 per 4,425, 1 per 8,772, and 1 per 10,000, respectively, at one month after vaccination (Fig. 2I).

We next determined whether vaccination with AMHR2-ED could inhibit the growth of murine EOC. We found that prophylactic AMHR2-ED vaccination 15 days (Fig. 3A; P < 0.001) or 7 days (Fig. 3B; P < 0.001) prior to MOVCAR inoculation of TgMISIIR-TAg (low) transgenic mice resulted in significant inhibition in the growth of transplantable EOC when compared with control mice vaccinated with CFA alone. Similarly, prophylactic AMHR2-ED vaccination of 6–7-week-old TgMISIIR-Tag (DR26) transgenic mice delayed the appearance and significantly inhibited the growth of autochthonous EOC compared with control mice vaccinated with CFA alone (Fig. 3C; P < 0.0001). This inhibition in the emergence and growth of autochthonous EOC resulted in a highly significant 42% increase in mean overall survival (P < 0.0001) compared with control mice vaccinated with CFA alone (mean 194 ± 35 days vs. mean 135 ± 14 days, respectively; Fig. 3D). AMHR2-ED vaccination was also effective in providing highly significant immunotherapy against EOC in TgMISIIR-Tag (DR26) transgenic mice with established growing autochthonous EOC (Fig. 3E; P < 0.001). IHC analysis of autochthonous EOC from TgMISIIR-Tag (DR26) mice vaccinated with AMHR2 consistently showed prominent infiltrates (shown by arrows) of CD3⁺ T cells (Fig. 3F, top left) and CD4⁺ T cells (Fig. 3F, middle left) with occasional CD8⁺ T cells (Fig. 3F, bottom left). Corresponding immunostained EOC from control mice vaccinated with CFA alone consistently failed to show detectable T-cell infiltrates (Fig. 3F, right). TILs from EOC taken from AMHR2-ED vaccinated TgMISIIR-Tag (DR26) mice showed significantly higher ELISPOT frequencies of IFNγ (Fig. 3G; P = 0.0001) and IL17 (Fig. 3H; P = 0.0001) producing T cells in recall responses to AMHR2-ED compared with TILs from control mice vaccinated with CFA alone.

We next determined which immune population accounted for inhibition of EOC growth. We found that transfer of AMHR2-ED primed CD4⁺ T cells into TgMISIIR-Tag (low) female mice one day prior to MOVCAR inoculation resulted in highly significant inhibition of tumor growth (Fig. 4A; P < 0.0001) and enhanced overall survival (Fig. 4B; P < 0.006) compared with mice receiving ovalbumin primed CD4⁺ T cells. Moreover, transfer of AMHR2-ED primed B220⁺ B cells into TgMISIIR-Tag (low) female mice one day prior to MOVCAR inoculation also mediated significant inhibition of tumor growth (Fig. 4C; P < 0.0001) and enhanced overall survival (Fig. 4D; P < 0.009) compared with mice receiving B220⁺ B cells from ovalbumin immunized mice. In addition, transfer of affinity-purified IgG from AMHR2-ED immunized mice into TgMISIIR-Tag (low) female mice one day prior to MOVCAR inoculation resulted in significant inhibition of tumor growth (Fig. 4E; P < 0.0001) and enhanced overall survival (Fig. 4F; P < 0.002) compared with mice receiving affinity-purified IgG from ovalbumin immunized mice.

IHC analysis of autochthonous EOC from TgMISIIR-Tag (DR26) mice vaccinated with AMHR2-ED showed a diffuse distribution of TUNEL-positive cells (Fig. 5A, top left) and caspase-3–positive cells (Fig. 5A, bottom left) compared with EOC from control vaccinated mice (Fig. 5A, middle). The differences in frequencies of these apoptotic tumor cells between both vaccinated groups were significant (Fig. 5A, right; P = 0.005 for TUNEL and P = 0.0007 for caspase-3). In addition, qRT-PCR analysis showed that the increased frequencies of apoptotic tumor cells in AMHR2-ED vaccinated mice occurred in the presence of significantly increased expression of the proapoptotic Bax gene compared with EOC from control vaccinated mice (Fig. 5B; P < 0.0001). No differences were evident in gene expression of prosurvival Bcl2 or in expression of other proapoptotic genes, including Bak and Bim (Fig. 5B). We next determined whether the apoptosis occurring in EOC from AMHR2-ED vaccinated mice may be mediated by IgG. To this end, we found that treatment of MOVCAR cells for 72 hours with IgG purified from sera taken from AMHR2-ED vaccinated TgMISIIR-Tag (DR26) mice resulted in pronounced metabolic arrest as measured by the MTS assay (Fig. 5C; P = 0.02) and in corresponding apoptotic cell death as measured by trypan blue exclusion (Fig. 5D; P = 0.01) and by cell surface expression of Annexin V (Fig. 5E; P = 0.06).

We next assessed the autoimmune phenotype resulting from AMHR2-ED vaccination. Four weeks after AMHR2-ED vaccination of young 8-week-old C57BL/6 female mice, qRT-PCR analysis of ovarian tissues showed significantly elevated gene expression for IL1β but not for IFNγ when compared with control mice vaccinated with CFA alone (Fig. 6A; P = 0.0004). However, this elevated ovarian expression of inflammatory factors was not apparent by 8 weeks after vaccination (Fig. 6B). In contrast, AMHR2-ED vaccination of 9-month-old C57BL/6 mice did not result in any elevated ovarian expression of either IL1β or IFNγ at both 4 weeks (Fig. 6C; P > 0.40) and 8 weeks (Fig. 6D; P > 0.40) after vaccination. It is important to note that the transient increase in ovarian expression of IL1β after AMHR2-ED vaccination of young mice had rather benign consequences as it was not associated with any detectable changes in fertility over four mating cycles as measured by either mean number of pups/litter (Fig. 6E; P > 0.40) or mean pup birth weight (Fig. 6F; P > 0.40). It is also worth noting that splenocyte frequencies of AMHR2-ED–specific IFNγ-producing and IL17-producing T cells were actually higher in 9-month-old mice 2 months after AMHR2-ED vaccination (1 per 2,762 and 1 per 14,085 splenocytes, respectively) compared with frequencies of 8-week-old mice 1 month after vaccination (1 per 3,831 and 1 per 21,739 splenocytes, respectively; data not shown). This substantial immune response elicited in 9-month-old mice precludes aged-based decreased immunity as an explanation for the qualitative differences observed in gene expression of ovarian inflammatory factors in young versus old ovaries and most likely reflects the age-related dramatic decline in expression of AMHR2-ED that occurs in both murine and human ovaries (Fig. 1D).

Our data indicate that AMHR2-ED vaccination significantly inhibits EOC growth and does so whether vaccination occurs in prophylactic or therapeutic settings. In addition, when used prophylactically in transgenic mice engineered to develop EOC, AMHR2-ED vaccination provides a dramatic 42% enhanced overall survival (Fig. 3D; P < 0.0001). Control of EOC in young fertile mice vaccinated against AMHR2-ED occurs in the presence of mild and transient ovarian inflammatory changes that produce no detectable longlasting adverse effects on ovarian function as determined by stable fertility over several mating cycles. As normal tissue expression of AMHR2-ED is confined exclusively to the ovary, and as approximately 90% of human EOC express AMHR2-ED, vaccination against AMHR2-ED has the appeal of providing immunotherapy against EOC with substantially reduced risk of systemic autoimmune complications. This reduced risk provides a marked safety advantage over more ubiquitously expressed self-proteins that have been proposed as vaccine targets for EOC immunotherapy, including HER2, folate receptor alpha, tumor protein p53, cancer antigen 125, mucin 1 (MUC-1), and mesothelin (reviewed in ref. 19).

Reducing risk of systemic autoimmunity is certainly a desirable feature for any immunotherapeutic strategy. However, the dramatic decline in AMHR2-ED expression in older human and mouse ovaries and the lack of any detectable autoimmune oophoritis following AMHR2-ED vaccination of aged mice supports the view that AMHR2-ED vaccination of postmenopausal women may likely occur in the absence of any substantial tissue-specific ovarian autoimmunity. Thus, the decreased risk for inducing ovarian and systemic autoimmunity in aging females indicates the feasibility of achieving safe and effective primary immunoprevention of EOC in postmenopausal women who account for >75% of all EOC cases (1, 2).

In the United States, menopause occurs in women at a mean age of 51 years and indicates cessation in production of mature ovarian follicles, ovulation, and fertilization (20). This age-related event has long been associated with dramatic declines in serum levels of the AMH hormone (21). Indeed, undetectable serum levels of AMH are also associated with primary ovarian insufficiency (POI), a fertility disorder that afflicts young women in their childbearing years (22). Thus, declining levels of AMH have been proposed as a way to assess the development and progression of ovarian senescence (22). Our data indicate that a substantial decline in ovarian expression of AMHR2-ED is also associated with menopause and suggests that decreased expression of AMHR2-ED may also be involved in the etiopathogenesis of POI. This dual decline in production of both AMH ligand and its cognate receptor-binding domain appears to confirm exhaustion of the ovarian reserve and cessation in production of mature fertile follicles that are known to be regulated by AMH/AMHR2 signaling (23).

Currently, seven human AMHR2 transcripts have been identified, but only four of these transcripts code for known protein variants and only one of these four transcripts codes for the extracellular domain transcribed from exons 1–3 (Fig. 1B; ref. 6). The full-length transcript coding for the extracellular domain is found exclusively in the ovary, although transcripts coding for the human kinase domain are expressed at high levels in the adrenal gland and at somewhat lower levels in the spleen, pancreas, and skeletal muscle (Fig. 1A; 6, 24–27). It is unclear why protein variants of AMHR2 that do not contain the cognate receptor-binding domain for AMH are expressed in extraovarian tissues. Perhaps transcripts missing the extracellular domain provide transcriptional regulatory functions as has been proposed (28). Alternatively, it is feasible that the cytoplasmic serine/threonine kinase of AMHR2 may be accessed in nonovarian tissues through currently unknown alternative signaling pathways. In any event, the expression of the cytoplasmic domain in extraovarian tissues and its failure to decline significantly in expression in the aging ovary (Fig. 1D) preclude its safe use as a vaccine target for primary immunoprevention of EOC despite its substantial effectiveness as a vaccine target for EOC immunotherapy in preclinical studies (16).

It is noteworthy that immunity against EOC could be transferred into naïve recipients with either CD4⁺ T cells, B cells, or IgG specific for AMHR2-ED. This implies that both cellular and humoral immune mechanisms are involved in the ability of AMHR2-ED vaccination to mediate inhibition of EOC growth. Indeed, our data indicate that AMHR2-ED–specific IgG alone is fully capable of mediating inhibition of EOC growth in vivo (Fig. 4E and F) and cell-cycle arrest and apoptotic death of EOC in vitro (Fig. 5B–E). Thus, we propose that AMHR2-ED vaccination inhibits the growth of EOC through the helper function of CD4⁺ T cells that facilitates B cells to produce AMHR2-ED–specific IgG. The IgG agonizes a Bax/caspase-3–dependent proapoptotic signaling cascade known to inhibit tumor growth (29). The lack of increased gene expression of the proapoptotic Bim implies that the cell death signal may be mediated by binding of the proapoptotic Bid protein to Bax in the “indirect” apoptosis activation model (30). Alternatively, Bax may be activated directly by other proapoptotic proteins like p53 or Diva (31, 32). Studies are currently under way to determine the underlying apoptotic signaling elements as well as any antibody-independent roles played by B cells and CD4⁺ T cells in the inhibition of EOC growth induced by AMHR2-ED vaccination.

There are many hurdles to overcome before AMHR2-ED vaccination undergoes testing in clinical trials, not the least of which is the selection of an appropriate adjuvant. Although CFA has been most useful in preclinical proof-of-principle studies for over 70 years, its potential for generating unresolved granulomas makes it unacceptable for use in human vaccination (33). For inducing tumor prophylaxis, candidate adjuvants may best be selected on the basis of how well they mimic the features of CFA in inducing T cells from both Th1 and Th17 proinflammatory lineages each independently associated with severe autoimmune tissue damage (34). Thus, the requirements for optimized tumor immunoprevention may be substantially different from those needed for optimized tumor immunotherapy where inflammatory mediators induced by IL17 may facilitate the production of angiogenic and antiapoptotic factors that enhance the growth of established tumors (35).

Recently, primary immunoprevention has been referred to as the “holy grail” and the “great unmet need” for the control of adult onset cancers (36, 37). Indeed, several approaches to accomplish this goal have been proposed, including targeting of a variety of self-proteins, frameshift neoantigens, and viruses believed to be overexpressed in adult onset tumors (reviewed in ref. 37). Certainly, immunotherapy has already achieved breakthrough status in cancer control (38). Perhaps primary immunoprevention of adult onset cancers may soon break through as well.

S. Mazumder has ownership interest (including patents) in Shield Biotech, Inc. J.M. Johnson has ownership interest (including patents) in Shield Biotech, Inc. V.K. Tuohy reports receiving a commercial research grant from and has ownership interest (including patents) in Shield Biotech, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S. Mazumder, V.K. Tuohy

Development of methodology: S. Mazumder, J.M. Johnson, J. Ko, V.K. Tuohy

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Mazumder, E. Martelli, J. Ko

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Mazumder, E. Martelli, V.K. Tuohy

Writing, review, and/or revision of the manuscript: S. Mazumder, J.M. Johnson, V.K. Tuohy

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Mazumder, J.M. Johnson, V. Swank, N. Dvorina, J. Ko

Study supervision: V.K. Tuohy

The authors express their gratitude to Dr. Sathyamangla V. Naga Prasad, Department of Molecular Cardiology, Cleveland Clinic, Cleveland, OH, for facilitating high-resolution in vivo imaging of mouse ovarian tumors. The authors also express their thanks to Dennis Wilk, Shield Biotech, Inc., Cleveland, OH, for his assistance in purification of the recombinant proteins, and to Dr. Ritika Jaini and Mathew G. Loya, Department of Genomic Medicine, Cleveland Clinic, Cleveland, OH, for their technical assistance. Finally, the authors wish to thank Sreedevi Goparaju, Department of Immunology, Cleveland Clinic, Cleveland, OH, for her expertise in statistical analysis.

V. K. Tuohy received support for this work in the form of a sponsored research grant from Shield Biotech, Inc., Cleveland, OH, a privately owned company.

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.