Novel anti-Mullerian hormone receptor 2 binding peptide prevents chemotherapy-related ovarian follicle loss in a mouse model.

AMHR2BP is composed of 17 L- and D-enantiomeric amino acids (VGSNAYNQFrVSGValS), has a molecular weight of 1768.95 Da, a net charge of + 1, and exhibits high affinity for AMHR2 according to in silico testing. Epitope mapping implied AMHR2BP would have a high affinity for AMHR2, confirmed by AlphaFold3-predicted HMHR2:AMHR2BP structure. To verify this affinity, we plotted the relationship between increasing concentrations of AMHR2BP against a fixed amount of total protein (50 µg) from an OC cell line expressing AMHR2 (A2780) (Fig.  1 A). Extrapolating from the tangents of the highest and lowest points of the graph, binding affinity (Kd) was calculated as 3.75 micromoles/l. This low Kd signifies a high affinity. Fig. 1 Binding affinity and dose effect of AMHR2BP. A Plot of the relationship between increasing concentration of AMHR2BP with a fixed amount of protein (50 µg). The Kd of 3.75 micromoles/l is extrapolated from the tangents of the highest and lowest points of the graph.  B Real-time RT-PCR analysis of AMHR2BP dose effect on granulosa cell expression of miRNA181b1, SMAD1 and SMAD5, and C FSH-R, miRNA181a1, miRNA181a2, miRNA181b2, and SMAD8, expressed as Log2 fold change compared to placebo.  The effect of AMHR2BP on cells in culture. D MTT proliferation assay measuring percent proliferation in three ovarian cancer cell lines (MDAH-2774, SKOV-3, and A2780) after treatment with AMHR2BP concentrations of 5, 10, or 20 µg/ml. E Expression of AMHR2 based on RT-PCR in the same ovarian cancer cell lines, MDAH-2774, SKOV-3, and A2780, showing differential expression in the different cell lines. F Ki67 and Caspase 3 and G AMH, AMHR2, FSH-R, and inhibin B, in granulosa cells cultured for 24 h in plain medium (control, black), a negative peptide (cross hair), or 10 µg/ml AMHR2BP (blue). The asterisks indicate significance of the comparisons between the different treatments with Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005; *** p  < 0.001) Binding affinity and dose effect of AMHR2BP. A Plot of the relationship between increasing concentration of AMHR2BP with a fixed amount of protein (50 µg). The Kd of 3.75 micromoles/l is extrapolated from the tangents of the highest and lowest points of the graph.  B Real-time RT-PCR analysis of AMHR2BP dose effect on granulosa cell expression of miRNA181b1, SMAD1 and SMAD5, and C FSH-R, miRNA181a1, miRNA181a2, miRNA181b2, and SMAD8, expressed as Log2 fold change compared to placebo.  The effect of AMHR2BP on cells in culture. D MTT proliferation assay measuring percent proliferation in three ovarian cancer cell lines (MDAH-2774, SKOV-3, and A2780) after treatment with AMHR2BP concentrations of 5, 10, or 20 µg/ml. E Expression of AMHR2 based on RT-PCR in the same ovarian cancer cell lines, MDAH-2774, SKOV-3, and A2780, showing differential expression in the different cell lines. F Ki67 and Caspase 3 and G AMH, AMHR2, FSH-R, and inhibin B, in granulosa cells cultured for 24 h in plain medium (control, black), a negative peptide (cross hair), or 10 µg/ml AMHR2BP (blue). The asterisks indicate significance of the comparisons between the different treatments with Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005; *** p  < 0.001) We previously showed that applying rAMH to GC in culture leads to an increase in SMAD1, SMAD5, and miRNA181b1 as well as a decrease in FSH-R, SMAD8, and miRNAs 181a1, 181a2, and 181b2 [ 19 ]. We thus asked whether AMHR2BP would act through the same mechanism. Following application of AMHR2BP 10 µg/ml, 20 µg/ml, or placebo (control group) to a primary culture of GC for 24 h, both AMHR2BP doses caused activation of miRNA 181b1 ( p  < 0.001) as well as SMAD 1 ( p  = 0.047) and 5 ( p  < 0.005) relative to PBS (Fig.  1 B). We also observed downregulation of FSH-R ( p  < 0.005), miRNA181a1 ( p  = 0.005), miRNA181a2 ( p  < 0.009), miRNA181b2 ( p  < 0.001), and SMAD 8 ( p  = 0.023) with both AMHR2BP doses (C), consistent with our previous findings following application of rAMH [ 19 ]. ELISA measurements of protein representation also showed increased protein concentration of SMAD 5 and 8/9 protein cascade proportionate to the AMHR2BP dose (Table  3 ). Table 3 Cellular ELISA levels of SMAD 1, 5, 8/9, in GC cultured in plain medium (untreated), 10 μg/ml AMHR2BP (AMHR2BP 10), and 20 μg/ml (AMHR2BP 20) Marker Untreated (± SD) AMHR2BP 10 μg/ml (± SD) AMHR2BP 20 μg/ml (± SD) p -value* FSH-R (Pg/μg protein) 54.98 ± 1.71 39.38 ± 2.08 11.78 ± 2.73 0.015 SMAD 5 (Pg/μg protein) 27.18 ± 7.19 82.63 ± 8.74 118.34 ± 2.79 0.041 SMAD 8/9 (Pg/μg protein) 10.42 ± 6.80 53.33 ± 5.98 95.68 ± 11.29 0.030 * t -test AMHR2BP 10 vs. 20 μg/ml Cellular ELISA levels of SMAD 1, 5, 8/9, in GC cultured in plain medium (untreated), 10 μg/ml AMHR2BP (AMHR2BP 10), and 20 μg/ml (AMHR2BP 20) * t -test AMHR2BP 10 vs. 20 μg/ml To prove that AMHR2BP would prevent proliferation, we used the TACS® MTT cell viability assay in three AMHR2-expressing OC cell lines and confirmed that AMHR2BP downregulated proliferation in a dose-dependent manner when exposed to 5, 10, or 20 μg/ml (Fig.  1 D, p  < 0.001). Interestingly, the effect was smaller in MDAH-2774 cells, which also had the lowest expression of AMHR2 (Fig.  1 E). To corroborate specificity, we performed MTT in MDAH-2774 and SKOV-3 cell lines following administration of two analogs compared to AMHR2BP. AMHR2BP again significantly inhibited proliferation, reducing cell growth to 19.7 ± 2.6%, compared to untreated controls (100.0 ± 0.9% and 100.3 ± 0.5%, respectively; p  < 0.001, one-way ANOVA), while analog #1 (86.4 ± 3.7%) and analog #2 (88.1 ± 2.2%) exhibited only modest, non-significant reductions. Because GC are the primary target of AMH, we next measured Ki67 in GC 24 h after application of PBS (control group), negative peptide, or AMHR2BP 10 µg/ml (AMHR2BP group). AMHR2BP-treated cells showed > 50% less Ki67 (Fig.  1 F, p  < 0.005) suggesting a decrease in proliferation. A significantly lower Caspase 3 mRNA than control ( p  < 0.005) indicates that the decreased proliferation was due to cell quiescence, not cell death. Expression of AMH, AMHR2, FSH-R, and inhibin B was also significantly reduced in AMHR2BP compared to the control groups ( p  ≤ 0.005 for all (Fig.  1 G)). To investigate whether AMHR2BP could inhibit follicular development in vivo , we treated 18-week-old C57BL/6 J female mice continuously via intraperitoneal pumps with AMHR2BP 50 µg/day (6 mice), rAMH 1.8 µg/day (6 mice), or normal saline (placebo, 6 mice) and compared to a pre-treatment group (baseline, 6 mice). Treatments were continued for 3 weeks, equivalent to 5 estrous cycles. After euthanasia, we measured the number of ovarian follicles and corpora lutea (CL; Fig.  2 A–E). We observed physiological progression of follicular development in the placebo group (21 weeks’ age mice, in black). The number of PDF decreased while PRF increased because of PDF activation, whereas SEF, TEF, and CL counts remained relatively unchanged compared to baseline. Compared to the placebo group, both the AMHR2BP-treated (blue) and the rAMH-treated group (green) exhibited higher PDFs ( p  = 0.011) and lower PRFs ( p  = 0.016), while SEFs and TEFs were lower, but not significantly. CL counts were instead significantly higher ( p  = 0.004; Fig.  2 E), driven by a lower representation of CL in the placebo group that is physiologically difficult to explain. When we compared the AMHR2BP-treated group (blue) to baseline (white), we saw no significant change in any of the follicle types, confirming that AMHR2BP, similarly to rAMH, maintains the follicular reservoir by halting follicular development. Fig. 2 The effect of AMHR2BP on ovarian cortex. Follicle measurements in C57BL/6 J female mice at baseline (12 weeks old, white), or after treatment with AMHR2BP (blue), rAMH (green), or placebo (PBS, black) for 3 weeks. A Primordial follicles (PDF). B Primary follicles (PRF). C Secondary follicles (SEF). D Tertiary follicles (TRF). E Corpora lutea. Log2 fold change of real-time RT-PCR ovarian cortex tissue expression of F inhibin B, as well as G BMP15, Caspase 3 (Cas-3), FSH-R2, GDF9, and Ki67, in the AMHR2BP, rAMH, and placebo (PBS) groups compared to the baseline group. H Log2 fold change of ovarian tissue protein concentrations of BMP15, Caspase 3, FSH-R2, GDF9, and Ki67 in C57BL/6 J female mice at baseline (12 weeks old), or after treatment with AMHR2BP, rAMH, or placebo (PBS) for 3 weeks. Both treatment regimens exhibited a significant decrease in concentration of all parameters, showing inhibition of cellular function. The asterisks indicate significance of the comparisons between the different treatments using Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005) The effect of AMHR2BP on ovarian cortex. Follicle measurements in C57BL/6 J female mice at baseline (12 weeks old, white), or after treatment with AMHR2BP (blue), rAMH (green), or placebo (PBS, black) for 3 weeks. A Primordial follicles (PDF). B Primary follicles (PRF). C Secondary follicles (SEF). D Tertiary follicles (TRF). E Corpora lutea. Log2 fold change of real-time RT-PCR ovarian cortex tissue expression of F inhibin B, as well as G BMP15, Caspase 3 (Cas-3), FSH-R2, GDF9, and Ki67, in the AMHR2BP, rAMH, and placebo (PBS) groups compared to the baseline group. H Log2 fold change of ovarian tissue protein concentrations of BMP15, Caspase 3, FSH-R2, GDF9, and Ki67 in C57BL/6 J female mice at baseline (12 weeks old), or after treatment with AMHR2BP, rAMH, or placebo (PBS) for 3 weeks. Both treatment regimens exhibited a significant decrease in concentration of all parameters, showing inhibition of cellular function. The asterisks indicate significance of the comparisons between the different treatments using Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005) Ovarian tissue expression of key regulators of follicle development, such as BMP15 and GDF9, in addition to Ki67, Caspase 3, and inhibin B, using both real-time RT-PCR (Fig.  2 F, G) and ELISA (H), showed both rAMH and AMHR2BP producing a significant decrease in concentration of all the other markers ( p  < 0.005), showing inhibition of cell differentiation and function. To assess whether AMHR2BP could prevent CX-induced follicle loss in mice, we counted follicle stages at day 3 and day 14 (Fig.  3 A) in 12-week-old female mice that received CX on day 2, with or without AMHR2BP pre-treatment on day 1. Despite the euthanasia of 3 mice per group on day 3 having the purpose to ensure that subsequent results would not be misinterpreted, the results were sufficiently compelling and were therefore included in the analyses. On day 3, the number of PDF/mm 3 was significantly higher in AMHR2BP than in CX ( p  < 0.0001) and control ( p  = 0.0027) groups (B), while PRF, SEF, and TEF/mm 3 were lower, but non-significantly (Fig.  3 C–E). On day 14, PDF and PRF/mm 3 were higher in AMHR2BP than in CX ( p  < 0.0001 for PDF and p  = 0.037 for PRF, respectively) and control group ( p  = 0.003 for PDF and p  = 0.005 for PRF, respectively), while TEF were lower compared to control ( p  = 0.05) as well as compared to CX ( p  = 0.026), and SEF were also lower, but not significantly (Fig.  3 F–I). These results suggest that inhibition of follicular activation by AMHR2BP is the primary mechanism underlying its ovarian protection. CL/mm 3 were significantly decreased in AMHR2BP compared to CX ( p  = 0.005 on day 3 and p  = 0.039 on day 14, respectively) and control group ( p  = 0.008 on day 3 and p  = 0.018 on day 14, respectively) at both time points, confirming AMHR2BP’s inhibitory effect on ovulation (J, K). Fig. 3 The effect of AMHR2BP on ovarian cortex treated with cyclophosphamide. A H&E-stained sections of ovarian cortex from C57BL/6 J mice treated with cyclophosphamide only (CX) or with prior administration of AMHR2BP (AMHR2BP + CX) and sacrificed on day 3 (D3) or day 14 (D14) after treatment. Note the presence of more primordial follicles (PDF) on D3, and the more solid appearance of the cortex, with fewer tertiary follicles (TEF), on D14 in the AMHR2BP + CX sections. Follicle counts from placebo (black), CX (red), and AMHR2BP + CX (blue) groups for B – E day 3 (D3) and F – I day 14 (D14). Measurements of corpora lutea in the three groups on J D3 and K , L D14, TUNEL assay showing increased apoptosis in the CX group and decreased apoptosis in AMHR2BP groups, compared to controls. Additionally, the more solid appearance of the ovarian cortex in the AMHR2BP group compared to the CX and control groups can be appreciated. The asterisks indicate significance of the comparisons between the different treatments using Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005) The effect of AMHR2BP on ovarian cortex treated with cyclophosphamide. A H&E-stained sections of ovarian cortex from C57BL/6 J mice treated with cyclophosphamide only (CX) or with prior administration of AMHR2BP (AMHR2BP + CX) and sacrificed on day 3 (D3) or day 14 (D14) after treatment. Note the presence of more primordial follicles (PDF) on D3, and the more solid appearance of the cortex, with fewer tertiary follicles (TEF), on D14 in the AMHR2BP + CX sections. Follicle counts from placebo (black), CX (red), and AMHR2BP + CX (blue) groups for B – E day 3 (D3) and F – I day 14 (D14). Measurements of corpora lutea in the three groups on J D3 and K , L D14, TUNEL assay showing increased apoptosis in the CX group and decreased apoptosis in AMHR2BP groups, compared to controls. Additionally, the more solid appearance of the ovarian cortex in the AMHR2BP group compared to the CX and control groups can be appreciated. The asterisks indicate significance of the comparisons between the different treatments using Kruskal–Wallis test for comparison of medians (* p  < 0.05; ** p  < 0.005) TUNEL assay showed increased apoptosis in the CX group and decreased apoptosis in AMHR2BP groups, compared to controls (L). In summary, AMHR2BP protected the PDF’s pool from CX effects in the immediate (day 3) and extended (day 14) windows.

With both in vitro and in vivo experiments, we explored how this peptide specifically binds to AMHR2 with a strong affinity and functional efficacy . Table 1 provides a summary of the experiments performed and their objectives. Table 1 Summary of the in silico , in vitro , and in vivo experiments and their objectives Objectives Materials Experiments in silico Identification of AMHR2BP AMHR2 epitope mapping AMHR2BP’s specificity for AMHR2 UniProt-reviewed sequences and AlphaFold3-predicted structures Experiments in vitro AMHR2BP’s binding specificity and affinity for AMHR2 Ovarian cancer (OC) lines AMHR2BP’s mechanism of action and SMAD cascade activation Primary granulosa cell line AMHR2BP’s effects on cell proliferation and cell death Primary granulosa cell and ovarian cancer (OC) lines Experiments in vivo AMHR2BP for prevention of physiological ovarian follicle depletion C57BL/6 J female mice AMHR2BP for prevention of cyclophosphamide (CX)-induced ovarian follicle depletion C57BL/6 J female mice Summary of the in silico , in vitro , and in vivo experiments and their objectives Epitope mapping of AMHR2 was performed using the immunosignatures technique, as previously described [ 34 ]. AMHR2 was analyzed with a microarray of 330,000 random sequence peptides. The pattern of binding to these peptides is the immunosignature of the epitope. A NCBI protein BLAST search was performed for each peptide sequence identified by epitope mapping, and peptides resulting in an 80–100% match were selected to form a list of candidate targets. Candidate protein targets were then cross-referenced with the proteins identified by mass spectrometry leading to the identification of twenty AMHR2-specific binding peptides (US Patent # 11,535,676, 12/2022). To confirm specific binding in silico, we modeled the human and mouse AMH-R2 extracellular domains (ECD, residues ~ 18–149) using UniProt-reviewed sequences and AlphaFold3-predicted structures, confirming a conserved three-finger cystine-knot topology and more than 75% sequence identity [ 35 ]. The ECD mediates the recognition of binding to AMH and is distinct from TGF-β family type 2 receptors, both in structure and ligand specificity [ 36 ]. In fact, AMHR2-ECD and AMHR2BP were screened against ECDs of Bone Morphogenetic Protein Receptor Type 2, Activin A Receptor Type 2 A, Activin A Receptor Type 2B, and Transforming Growth Factor Beta Receptor Type 2, and these potential interfaces were excluded. Based on the peptide ranking of the in silico analysis and the AlphaFold3-predicted structures, all experiments were conducted with the peptide with the highest predicted affinity for AMHR2, designated as AMHR2BP. AMHR2BP and its analogs were synthesized by WatsonBio (Houston, TX, USA) and sent as a desiccated powder of > 90% purity after standard high-quality control. For all experiments, AMHR2BP was dissolved in phosphate-buffered saline (PBS; HyClone) at the concentration required for each experiment as noted below. Concentrations of AMHR2BP were extrapolated based on prior experiments utilizing rAMH [ 15 – 17 , 19 , 21 ]. Multiple cell lines were used for in vitro experiments: Primary granulosa cell (GC) lines were isolated from human follicular fluid obtained at the time of oocyte retrieval from three patients who underwent ovarian stimulation for in vitro fertilization, as previously described [ 15 , 19 ]. The A2780 human ovarian cancer (OC) line was obtained from Sigma-Aldrich (St. Louis, MO) and cultured in RPMI-1640 (ThermoFisher Scientific). Epithelial OC cell lines MDAH-2774 (CRL-10303) and SKOV-3 (HTB-77), as well as human macrophage cells (EL-1, CRL-9854), were obtained from American Type Culture Collection (ATCC, Manassas, VA). All cells were cultured with media supplemented with fetal bovine serum (FBS, Innovative Research, Novi, MI) and penicillin/streptomycin which was replaced every 2 days. Contamination from mycoplasma was assessed by growing cells on coverslips with DAPI stain. OC cells were utilized between passages 3 and 5. Cells (5 × 106) were seeded in 150-mm dishes and were collected for isolation of RNA and protein extraction as described below. We utilized real-time RT-PCR to determine tissue mRNA levels for all markers under investigation in each experiment, adhering to the protocols provided by the manufacturers and as previously described [ 15 – 17 , 19 , 21 ]. Briefly, total RNA was extracted from tissues with the RNeasy Mini Kit (Qiagen). Sample dilution was the same for all protein specimens (50 µg) and was corrected to beta-actin and a fixed amount of RNA. Real-time RT-PCR was performed with the Express SYBR GreenER qPCR SuperMix RT-PCR kit (Life Technologies) and a Cepheid 1.2f Detection System (Cepheid, Sunnyvale, CA). Human oligonucleotide primers that amplify variable portions of the protein coding regions were selected with the aid of the Beacon Designer software (Premier Biosoft, Palo Alto, CA) as previously described [ 19 ] and are reported in Table  2 . The housekeeping gene was β-actin (79 base pairs (bp)), and all PCR analyses were executed in triplicate. Table 2 Real-time RT-PCR oligonucleotide primer sequences and cycling conditions Gene Sense (3′−5′) Antisense (5′−3′) Amplicon length (bp) Annealing temperature (°C) Annealing time (s) B-actin ATGACTTAGTTGCGTTACAC AATAAAGCCATGCCAATCTC 79 58 10 AMH GTGCTGCTGCTGAAGATG CTCCGACAGGCTGATGAG 102 62 10 AMH-R2 CCAGAAGCACGGCTGACAG TGGAAAGGGGTGGCTCTCT 81 62 10 Ki67 TCCTTTGGTGGGCACCTAAGACCTG TGATGGTTGAGGCTGTTCCTTGATG 156 55 30 SMAD 1 CTGCATGTGTATTCGTGAGTTCGC AACCCAGTCAGCACAAAGACA 76 72 63 SMAD 5 CAGGAGTTTGCTCAGCTTCTGG GGTGCTGGTTACATCCTGCCG 141 72 62 SMAD 8/9 AGTGGCCAACCTGTAGATGC TGGGGCTCCTCGTAACAAAC 92 63 72 GDF9 TCGCATTACTACCGTTGAA CACACATTTGACAGCAGAG 91 53 10 BMP 15 TCAAGCCAAACACAAACAGC AGCCACAATCCAGTCATTCC 92 54 10 FSH-R GTCCACAACACCCATCCAAGG GGGCTAAATGACTTAGAGGGACAA 97 66 10 miRNA181a1 GTGAACATTCAACGCTGTCGG CATAGGGTACAATCAACGGTCG 70 56 72 miRNA181a2 GGAACATTCAACGCTGTCGG TAAGGACCCCAAGGTACAGT 74 57 72 miRNA181b1 AACATTCATTGCTGTCGGTGGG GGGGCCACAGTTGCATTCATTG 71 62 72 miRNA181b2 CATTCATTGCTGTCGGTGGG TTTGGTCCGCAGTTTGCATT 70 56 72 Real-time RT-PCR oligonucleotide primer sequences and cycling conditions In sandwich ELISA assays, a specific antibody for the target protein was pre-coated onto a microplate. Standards and samples were added into the wells, and any target protein present is bound by the immobilized antibody. An enzyme-linked polyclonal antibody specific for the target protein was added followed by the addition of a substrate solution, and color was developed in proportion to the amount of target protein bound and measured at 450 nm. We measured colorimetric ELISA for FSH-R (Bioassay technology laboratory), Ki67 (MYBioSource), BMP15 (ABclonal Technology), GDF9 (Elabscience), SMAD 5 and SMAD 8/9 (MYBioSource; SMAD 5, MBS9714237, and SMAD 8/9, MBS9714235), following the manufacturer protocol. The Colorimetric Biotin Assay Kit (#MAK171-1KT Sigma Aldrich) with amino terminus biotin labelled AMHR2BP was used to quantify total biotin concentration and estimate the biotin-protein molar ratio in biotin-protein conjugates [ 37 , 38 ]. A2780 total protein (50 μg) was incubated with increasing concentrations of the biotin-labeled AMHR2BP at 27 °C overnight. We extrapolated Kd as previously described [ 37 ] from a plot of the relationship between increasing concentration of the peptide with a fixed amount of protein. Western blot analyses of AMHR2 and AMHR2BP were performed as previously described [ 36 ]. Briefly, pre-cleared lysates were incubated with 2–5 µg of AMHR2 Polyclonal Antibody (Proteintech, Catalog no. 25888–1-AP) overnight at 4 °C, followed by incubation with 30 µL of Protein A/G PLUS-Agarose beads for 2–4 h, and the eluted proteins were transferred to PVDF membranes. Membranes were hybridized with either AMHR2 antibody or biotin-labeled AMHR2BP. Detection was performed using a secondary antibody for AMHR2 and HRP-conjugated streptavidin for the biotin-labeled peptide. Signals were visualized with chemiluminescent substrate to assess co-binding of the antibody and peptide to AMHR2. These were used to further assess target specificity on both membrane and cytoplasmic fractions isolated from human ovarian cell lines, using an AMHR2-specific monoclonal antibody [data not shown]. A 62-kDa band identified the presence of AMHR2 in the cell membrane, and a similar 62-kDa band was observed adding biotin-tagged AMHR2BP, confirming peptide binding to AMHR2. The TACS® MTT Cell Proliferation Assay (Trevigen, Gaithersburg, MD, USA) was performed on the OC cell lines MDAH-2774, SKOV-3, and A2780 to determine cell proliferation in response to AMHR2BP following the manufacturer’s protocol and as previously described [ 33 ]. Briefly, cells were seeded into 96-well plates in a fixed volume of 100 µl at a density of 100,000 cells/well. Cells were treated with increasing concentrations (5, 10, 20 µg/ml) of AMHR2BP for 24 h. This time was chosen based on these cell lines undergoing a replication cycle between 16 and 28 h [ 39 ]. A negative peptide (a random sequence peptide containing 17 amino acids) was a control. Cell proliferation was determined by measuring absorbance at 570 nm. A blank containing only medium was subtracted from all test samples. To further corroborate specificity of AMHR2BP effects, we repeated the experiment using two analogs with two or three amino acid substitutions within the sequence (analog#1, VG YAN YNQFrVSGVaIS; analog#2 VGSNAYNQFrV V G S aIS; the bolded amino acids indicate the substitutions). MDAH-2774 and SKOV-3 were treated with 10 µg/ml of AMHR2BP, a negative peptide, an analog, or left untreated (control), for 24 h, and proliferation was measured. The reported studies were approved by the Institutional Animal Care and Use Committees (IACUC) at the University of Tennessee Health Science Center in Memphis, TN, PI: Laura Detti [15–089.0 and 2015–2020] and at Baylor College of Medicine in Houston, TX, PI: Feng Li [AN-6589]. Adult C57BL/6 J female mice were used for all in vivo experiments. In C57BL/6 J, puberty is characterized by vaginal opening at approximately 25 days of life and by estrus 3–4 days later. Estrous cycles are every 4–5 days, and maturity is at 90 days of life. Littermates were randomly assigned to different experimental groups. The animals were bred as per Jackson Lab guidelines and maintained in ventilated cages under a 12-h light/dark cycle with food and water ad libitum in the Animal Centers at UTHSC, in Memphis, TN, or BCM, in Houston, TX. Because of the low biological variability of inbred, sex-matched mice maintained under identical environmental conditions, we accepted the statistical power afforded by a sample size of six mice per group. To assess AMHR2BP prevention of physiological ovarian follicle depletion, 24 12-week-old mice were randomized to four groups: baseline (euthanized just prior to the experiment), AMHR2BP (AMHR2BP, 50 µg/day), rAMH (rAMH, 1.8 µg/day), and placebo group (normal saline). Extrapolating from our prior work with AMH delivered by continuous infusion [ 21 ] and accounting for differences between AMH, which is a large dimeric protein, and our peptide, we calculated 50 µg/day × 28 days for continuous infusion or 500 µg for a single weekly dose. AMHR2BP was administered via intraperitoneal micro-osmotic pumps (Alzet Pumps, Durect Corporation, CA) as previously described [ 21 ] which gradually delivered the solution for 28 days. Mice were euthanized 3 weeks after pump placement, and the ovaries were explanted for histological analysis as described above. To assess AMHR2BP prevention of cyclophosphamide (CX)-induced ovarian follicle depletion, 12-week-old mice were randomized to three groups: PBS control, CX, and AMHR2BP. On day 1, the control and CX groups received PBS, and the AMHR2BP group received 500 µg AMHR2BP in a 0.5 ml solution volume via intraperitoneal injection. On day 2, the control group received PBS, while the AMHR2BP and CX groups received 250 mg/kg CX. Because severe ovarian damage was desired, the CX dose was increased from previous mouse studies using 200 mg/kg, which induced moderate damage [ 40 ]. Mice were euthanized 14 days after CX/PBS. To evaluate the immediate effects of CX and AMHR2BP, on day 3 (24 h after CX/PBS), only three mice in each group were euthanized. This step was taken to ensure that CX administration had an effect and to confirm that subsequent results would not be misinterpreted; it was not intended for inclusion in the statistical analysis. In all mice, the ovaries were explanted for histological analysis as described below. Explanted ovarian specimens were fixed in formalin, embedded in paraffin, and stained with hematoxylin and eosin (H&E) and with immune-stains. Five-µm-thick sections were serially cut at the level of the largest ovarian diameter not including the hilum, and every 5th section was analyzed for follicular counts, as previously described [ 21 , 41 ]. Only follicles with a visible oocyte with a nucleus (germinal vesicle) were counted. To avoid double counting of oocytes in large follicles that could be present in multiple sections (oocytes > 25 µm), a minimum of four sections were analyzed simultaneously on H&E-stained slides. Follicle counts from individual sections (spanning a total thickness of 25 µm) were normalized to a 1 mm 2 area and then multiplied by 40 µm to estimate the total number of follicles per mm 3 . Primordial (PDF), primary (PRF), secondary (SEF), and tertiary follicles (TEF) were counted in the ovarian specimens by investigators blinded to treatment group, while the immune-stained slides were used for examination of follicular development. Spectrum (Version 10.2.2.2314; by Aperio, Vista, California, USA) and Fiji (Version 2.9.0, GPL https://imagej.net/software/fiji/downloads ) were used for image analysis and counting. For some mice, the collected ovarian specimens were not adequate to accurately account for all the different ovarian follicle types and were excluded from the statistical analyses. Terminal deoxynucleotidyl transferase (TdT) dUTP Nick-End Labeling (TUNEL) assay was used to detect apoptotic cells as previously described [ 21 ]. Apoptosis was quantified per ovarian section using the Fiji software. Given the difficulty of objectively assessing TUNEL staining across the entire ovary, results were reported qualitatively in the assessed sections as “increased” or “decreased,” relative to controls. All data were expressed as the means ± standard deviations. Analysis of variance with Tukey’s multiple-comparisons post hoc testing and Kruskal–Wallis test were used to compare groups. Raw data for each analysis was imported and processed using R (version 4.4.1). For RT-PCR test results, the log₂ fold change of gene expression levels for the treatments was calculated and compared to placebo. The workflow involved data preprocessing, statistical summarization, calculation of fold changes, and visualization through bar plots with error bars. The data were grouped by test, gene, treatment (tmt), and biological replicate (bio_rep). The mean of technical replicates was calculated for each group to obtain a single value per biological replicate. The mean, standard deviation, sample size, and standard error were then calculated for each combination of analysis, gene, and treatment. From here, we calculated the log₂ fold change (log₂FC) of each treatment compared to the baseline mean values for each test and gene, computing the standard error using error propagation for ratios and logarithms. The final dataset containing log₂ fold changes and associated statistics was then plotted onto a bar plot using the ggplot2 package and exported as a table for further analysis. In Fig.  3 A–F, outliers in the dataset were identified and filtered using a robust method based on the median absolute deviation (MAD). For each unique combination of group and assay, the median and MAD of the value variable were calculated. The MAD was determined as the median of the absolute deviations from the group median. Outliers defined as observations that deviated more than 2 MADs from the group median were identified and subsequently filtered from the analysis.

We developed and patented a 17-amino acid peptide, AMHR2BP, which we show here mimics the biological effects of AMH, both in vitro and in vivo. Our data confirmed that AMHR2BP acts by specifically binding to the AMHR2 receptor with high affinity, confirming the in silico results and AlphaFold3-predicted structures [ 36 ]. Further, we showed that AMHR2BP activates the same SMAD signaling pathway as AMH and mirrors observed effects reported for rAMH [ 15 – 17 , 21 ], regulating key ovarian factors (BMP15, GDF9) and miRNA181 family members involved in granulosa cell function and epigenetic modulation [ 19 ]. Reproducing these downstream effects further supports that AMHR2BP is acting through AMHR2. Finally, AMHR2BP maintained the follicle pool through natural and chemotherapy-induced models of ovarian follicle depletion. AMHR2BP would therefore prevent the massive follicle activation and “burn-out” [ 23 , 24 ] caused by gonadotoxic agents, but not the oocytes’ DNA double-strand breaks and the extensive stromal apoptosis [ 22 ]. Other medical compounds have previously been proposed to protect the ovaries from chemotherapy-related damage. The first were GnRH-agonists, which showed promise in both meta-analyses and randomized controlled trials [ 42 , 43 ]. Based on these studies, the American Society of Clinical Oncology currently recommends using GnRH-agonists in combination with surgical preservation [ 44 ]. We speculate that a mechanism by which GnRH-agonists may protect against ovarian damage is maintenance of AMH levels through sustained production by secondary and small tertiary follicles [ 17 , 19 ], which spans approximately 3–6 months in humans [ 45 ]. However, because of the profound hypoestrogenism, GnRH-agonists cause deterioration of bone and overall health, making their long-term application limited. Further, animal studies showed GnRH-agonists providing only partial protection and for only a limited time [ 40 ]. Other pharmacological interventions targeting the PTEN/AKT/FOXO3a and PI3K/AKT/mTOR signaling pathways have been pursued, such as melatonin [ 46 ], quercetin [ 47 ], rapamycin [ 48 , 49 ], luteinizing hormone [ 50 ], metformin [ 51 ], sphingosine-1-phosphate (S1P) [ 52 ], and ceramide-1-phosphate (C1P) [ 53 ]. Although it remains unclear whether follicle protection was mediated through activation or inhibition of those pathways, interestingly AMH levels were consistently maintained across all these treatments. AMH also influences key downstream regulators, such as PI3K/AKT/FOXO3a, which supports follicle quiescence [ 29 , 54 ]. Although the downstream pathway was not directly assessed, our study corroborates histological findings that have previously been associated with activation of the PI3K/AKT/FOXO3a pathway [ 29 , 54 ]. Only one study found that in transgenic mice that moderately overexpress AMH, the mild increase in AMH levels would not protect the PDF pool [ 55 ]. However, the AMH concentration achieved in that experiment was only moderately increased over the physiologic levels, undermining the significance of the results. Like AMH, AMHR2BP may function through a parallel and intersecting pathway that could be strategically targeted to improve fertility preservation and support future reproductive potential. Because AMH is not FDA-approved for use in humans, AMHR2BP offers a compelling and potentially superior alternative to AMH. Additionally, like AMH, AMHR2BP could reduce follicular loss following ovarian tissue autotransplantation, a setting in which early follicular activation and ischemic damage contribute substantially to graft attrition [ 21 ]. The strength of our study lies in the multimodal experimental design, yielding consistent results across both in vitro and in vivo models, and encompassing analysis of both gene activation and protein translation. A few key limitations were also identified. First, we did not assess other unrelated RNA changes to exclude off-target effects. Secondly, the number of mice in one in vivo experiment was small. We fully recognize that the sample sizes of 2–3 in the day 3 groups of the CX experiment are insufficient, and it was not our original intention to include them in the analysis. However, we believe it was important to present the complete set of experiments and to offer a plausible interpretation of these significant results. Thirdly, the AMHR2BP dosing and treatment duration might have been adequate to obtain significant results in the in vitro and mouse experiments, but the translational implications for human fertility preservation are currently speculative. Although the absolute number of follicles per mm 3 may not have been reflective of the whole ovary follicle density, follicle counts were conducted in all ovaries using a consistent methodology, ensuring that relative changes were reliably assessed. Future studies should evaluate ovarian cortex and granulosa cell function, also through transcriptomics, following discontinuation of AMHR2BP, blockade with AMHR2BP-specific antibodies, or administration of an AMHR2BP antagonist, to assess the timeline for restoration of normal function, ovulation, fertility, and reproduction. The findings of this study provide foundational evidence that AMHR2BP, like AMH, will be an effective and specific treatment to prevent ovarian follicle depletion. These findings establish AMHR2BP as a potent, mechanism-specific AMH-analog with promising therapeutic potential not only for the protection of the follicle pool during chemotherapy treatments but also for preempting the slow but inexorable follicle loss due to AMH depletion during other pathological and natural processes, such as endometriosis and natural aging. Thus, AMHR2BP stands out as a transformative therapy, combining a targeted mechanism with a favorable pharmacological profile to address a critical, unmet medical need.

Reproductive competence can be dramatically impacted by cancer treatments by permanently depleting the ovarian follicle reserve [ 1 ]. Chemotherapy, predominantly alkylating agents such as cyclophosphamide, accelerates ovarian aging, ultimately causing infertility from diminished ovarian reserve, premature ovarian insufficiency, and early menopause in cancer survivors [ 1 ]. Similarly to natural menopause, characterized by a complete loss of follicles, iatrogenic menopause causes decline of cardiovascular, cognitive, and overall health, which is even more impactful as it occurs more abruptly in younger women, overlaid on the health detriment caused by cancer and its treatment [ 2 , 3 ]. Despite the increasing incidence of female cancer survivors [ 4 ], at present, there is no effective medical treatment to prevent chemotherapy’s devastating side effects [ 1 – 3 , 5 ]. Anti-Müllerian hormone (AMH), best known as a biomarker that declines with diminishing ovarian follicles, is currently FDA (Food and Drug Administration)-approved for assessing ovarian reserve [ 6 ]. However, it has also emerged as a key inhibitor of follicle development and ovarian aging [ 7 – 12 ]. AMH is secreted by ovarian granulosa cells (GC) with its primary paracrine action on other GC within the ovarian cortex by specifically binding to AMH receptor 2 (AMHR2), a transmembrane serine/threonine kinase localized on GC and thecal cells, but not on the oocyte [ 13 , 14 ]. After binding to AMHR2, AMH function is mediated by SMAD 1 and SMAD 5, and SMAD 8/9 [ 13 , 14 ]. AMH controls the pacing of follicle development by precluding primordial follicles’ activation and progression through their natural evolutionary process, thus preventing ovarian aging [ 9 ]. Experiments using synthetic recombinant AMH (rAMH) support AMH as a key therapeutic agent for preventing ovarian aging in both human in vitro and animal in vivo studies [ 15 – 19 ]. Specifically, rAMH administration reduced tissue expression of AMH itself and inhibin B, as well as AMHR2, FSH-R (follicle-stimulating hormone receptor), LH-R (luteinizing hormone receptor), and IGF1-R1 (insulin-like growth factor 1-receptor 1) in both human ovarian cortex [ 15 ] and isolated human luteinized GC [ 16 ]. Remarkably, rAMH also downregulated human ovarian cortex hormone production and reduced cell proliferation and apoptosis and regulated stemness potential [ 15 , 16 ]. Furthermore, rAMH regulated the oocyte-derived growth factors that regulate GC’s mitotic proliferation and differentiation, BMP15 and GDF9 [ 17 – 19 ], and regulated cell function and differentiation by controlling epigenetic regulators such as micro-RNAs (miRNA) in GC [ 19 , 20 ]. Through these mechanisms, rAMH reduced primordial follicle activation and apoptosis after xenotransplantation of human ovarian cortex into nude mice, while not impacting the initial hypoxia-driven primordial follicle loss [ 21 ]. Gonadotoxic agents cause accelerated ovarian aging by severely depleting ovarian follicle reserve, and chemotherapy has been shown to induce direct toxicity with DNA double-strand breaks in oocytes [ 22 ], as well as indirect toxicity from massive primordial follicle activation and subsequent “burn-out,” overall resulting in a sharp decline of follicle numbers [ 23 , 24 ]. Multiple preclinical studies have again identified AMH as a key therapeutic candidate for safeguarding the ovarian follicular reserve against chemotherapy-induced damage [ 25 – 28 ]. RAMH was found to protect from chemotherapy-induced ovarian follicle depletion by potently suppressing primordial follicle activation in mouse models, thereby protecting fertility and reproduction during chemotherapy treatment [ 25 , 28 ]. More recent mouse studies have confirmed that rAMH was able to effectively protect both fertility and ovarian reserve from cyclophosphamide, whether isolated or in combination with other follicle activation inhibitors [ 29 – 31 ]. Although our findings [ 15 – 17 , 19 , 21 ], along with those of others [ 25 – 28 ], unequivocally support a role for AMH as a key therapeutic agent for the prevention of ovarian damage from chemotherapy [ 32 , 33 ], neither AMH nor rAMH have FDA approval for therapeutic human use. This may be due, in part, to the challenges associated with producing biologically active rAMH, given its large size and complex dimeric structure. To address this gap, in 2018, our group designed a peptide that mimics natural AMH function by specifically binding to AMHR2. The peptide, called AMHR2BP, has a unique amino acid sequence that contains D-amino acids for added stability in plasma. This peptide was awarded a patent in the USA (US Patent # 11,535,676, 12/2022) and in other countries. We hypothesized that binding to AMHR2 like AMH, AMHR2BP could inhibit GC’s replication and function, ultimately inducing a quiescence that would protect the follicle reservoir from chemotherapy-induced ovarian damage and subsequent premature ovarian aging.