Novel expression of zona pellucida 3 protein in normal testis; potential functional implications

The expression of the zona pellucida glycoprotein 3 (ZP3), originally thought to be specific for oocytes, was recently extended to ovarian, prostate, colorectal and lung cancers. Earlier successful ZP3 immunization of a transgenic mouse model carrying a ZP3 positive ovarian tumor emphasized the suitability of ZP3 for cancer immunotherapy. This study was carried out to determine whether any other normal tissues besides the ovary in healthy human and mouse tissues may express ZP3, considered important to exclude off-target effects of ZP3 cancer immunotherapy. Strong ZP3 expression was found in normal human and mouse testis. ZP3 protein and mRNA transcripts were localized in spermatogonia, spermatocytes and round and elongated spermatids of both human and mouse testis, as well as in a mouse spermatogonial cell line, but absent in testicular Sertoli, Leydig, spermatogonial stem and progenitor cells. All other normal human and mouse tissues (n = 14) were ZP3 negative. This surprising testicular ZP3 expression has implications for the development of ZP3 cancer immunotherapies, and it also alludes to the potential of using ZP3 as a target for the development of a male immunocontraceptive.


Introduction
The extracellular layer of the zona pellucida (ZP) surrounds mammalian oocytes (Wassarman 1999). It plays a crucial role in oogenesis, fertilization, prevention of polyspermy and embryo protection at the early stages of development (Wassarman 1999). ZP is composed of three or four glycoproteins that form a characteristic fibro-granular structure around the oocytes by non-covalent interactions (Wassarman and Mortillo 1991). In humans, ZP is composed of all the four glycoproteins named ZP1, ZP2, ZP3, and ZP4, whereas in mice only ZP1, ZP2, ZP3 are expressed (Harris et al., 1994;Lefievre et al., 2004).
ZP3 is essential for successful fertilization and plays a key role in the species-specific spermatozoa-oocyte binding and activation of the acrosome reaction (Bleil and Wassarman 1983;Wassarman 1999). ZP3 knockout female mice are infertile (Liu et al., 1996). Recent studies demonstrated that missense mutations in the ZP3 sequence in ZP-layer caused empty follicle syndrome in humans (Chen et al., 2017;Zhou et al., 2019). It has been suggested that ZP3 plays an important role during germinal vesicle breakdown in the early stage of meiosis and in further oocyte maturation (Gao et al., 2017). Higher ZP3 levels have been found in immature oocytes compared to mature ones (Canosa et al., 2017). However, there are very limited data on the ZP3 function other than its classical role as a component of the ZP. Due to its critical role in the fertilization process, ZP3 became a natural candidate for female immunocontraception (Shabanowitz 1990), and has already been successfully used as a target antigen for active immunization of certain wild-life populations (Millar et al., 1989;Gupta et al., 1997). However, the loss of ovarian function caused by autoimmune oophoritis mediated by ZP3-specific antibodies and autoreactive T cells in various animal models (Mahi-Brown et al., 1988;Paterson et al., 1998) has precluded further contraceptive development in humans. Recently, a careful selection of epitope combinations has been found to induce infertility in animal models without any ovarian pathologies (Paterson et al., 2002;Mohd-Lila et al., 2019). Despite the great interest in ZP3, there is a distinct shortage of studies investigating its potential expression outside the ovary. Recently, ectopic ZP3 expression was demonstrated in ovarian granulosa and epithelial cell cancers (Rahman et al., 2012), in transgenic mice presenting with granulosa cell tumors and in cell lines derived from these tumors as well as in human cancer cell lines KGN). ZP3 expression has also been shown in prostate, colorectal and lung cancers (Buschmann et al., 2004;Bennink 2016, Costa et al., 2018Coelingh Bennink and RahmanInventor, 2019). Any future clinical cancer development with ZP3 immunization would require the characterization of the ZP3 expression profile in normal tissues. For this reason, we found it important to investigate the potential extraovarian ZP3 expression profile in healthy human and wild-type mouse tissues.

Human tissue samples
We examined 94 normal human tissue samples (subjects age 30-45 years) from archival paraffin blocks or frozen tissues from the Medical University of Bialystok Department of Pathomorphology and the University Biobank, in Poland (n = 10/ovary; n = 8/testis; n = 7/myometrium smooth muscle; n = 7/skeletal muscle; n = 6/prostate and epididymis and n = 5/small intestine, large intestine, liver, pancreas, stomach, lymph node, brain, lung, kidney and breast for each tissue type). Samples were taken during the surgery and checked by the pathologist to ensure that they were healthy tissues and devoid of any cancer cells. The human testes samples removed during the surgery (n = 6), were received from the Medical University of Bialystok Biobank. None of the male patients had any case of infertility or spermatogenic maturation arrest. Testicular structure and spermatogonia, spermatocytes and spermatids were identified and distinguished based on the following characteristics: 1) typical size of the nuclear structure; 2) position in the epithelium; 3) typical pachytene morphology; 4) chromatin condensation patterns. Human semen samples (n = 6) were obtained from male patients treated for couple infertility problems (diagnosed female factors) recruited from the IVF program at the Department of Reproduction and Gynecological Endocrinology of the Medical University of Bialystok, Poland. Semen samples were obtained by masturbation after 3 days of sexual abstinence. After liquefaction at room temperature for 30-60 min, semen parameters were assessed according to the WHO guidelines (WHO 2010) and samples were used for the IVF procedure. The remaining samples were used to collect spermatozoa using the swim-up method. Samples with less than 15 × 10 6 spermatozoa/ml were excluded from this study. Written informed consent was obtained from all patients before inclusion. The local Human Investigations Ethics Committee at the Medical University of Bialystok, Bialystok Poland approved the study.

Mouse tissues samples
We tested a total of n = 40 normal 3 month-old wild-type C57BL mouse fresh or fixed tissue samples from paraffin blocks (n = 10 for ovary, testis, spleen and skeletal muscle each).

RNA isolation
The TRIzol-based extraction method was used for total RNA isolation (Invitrogen, Carlsbad, CA). The quantity and quality of extracted RNA were determined by the absorbance measurement using NanoDrop (Thermo Scientific, Waltham, MA). Gel electrophoresis was performed to determine the integrity of isolated RNA.

RT-PCR analysis
Before the reverse transcription (RT) reaction, 1 μg of total RNA was incubated for 30 min with deoxyribonuclease (DNase I) (Invitrogen) at room temperature and then inhibited by 25 mM EDTA solution for 10 min at 65 • C. The RT reaction was performed with SensiFAST cDNA SynthesisKit (Bioline Reagents Ltd, London, UK) according to the manufacturer's protocol. First-strand cDNA was used as template in PCR (initial denaturation of 96 • C for 3 min, then 35 cycles of 94 • C for 1 min, 57 • C for 45 s, and 72 • C for 45 s, with a final extension period of 5 min at 72 • C). Primer sequences were as follows: mouse Zp3 gene F: GAGCTTTTCGGCATTTCAAG, R: AGCTTATCGGGGATCTGGTT and mouse Ppia gene F: CATCCTAAAGCATACAGGTCCTG, R: TCCATGGCTTCCACAATGTT; human ZP3 gene F: ATGCAGGTAACT-GACGATGC, R: CCATCAGACGCAGAGAGAAA, human FSHR gene F: TGGGCTCAGGATGTCATCATCGGA, R: TGGATGACTCGAAGCTTGGT-GAGG, human LHR gene F: CTGAGTGGCTGGGACTATGA, R: CCAAATCAGGACCCTAAGGA; and human PPIA gene F: GCCAA-GACTGAGTGGTTGGATG, R: GAGTTGTCCACAGTCAGCAATGG.

Real-time quantitative PCR (qPCR)
For Real-time qPCR, SYBR Green PCR master mix (Applied Biosystems, Foster City, CA) and the thermo-cycler 7500 Real-Time PCR System (Applied Biosystems) were used. Reaction conditions were as follows: 2 min at 50 • C, 10 min at 95 • C, 15 s at 95 • C, and 1 min at 60 • C up to 40 amplification cycles. The PCR products were analyzed by melting curve analysis and agarose gel electrophoresis to ensure the amplification of a single product. Every reaction product (both for RT-PCR and qPCR) was separated and verified by sequencing analysis. Expression levels of the investigated genes were normalized to the housekeeping gene peptidylprolyl isomerase A (PPIA). The primer sequences were as follow: mouse Zp3 F: CCAACGACCAGACTGTGGAA, R: AGGACTATAGCTGCCAGGGT; mouse Ppia F: CATA-CAGGTCCTGGCATCTTGTC, R: AGACCACATGCTTGCCATCCAG; and human ZP3 F: TGGCAACAGCATGCAGGTA, R: CTGAGTGGCTGGGAC-TATGA, human FSHR F: GCCAAGAGAGCAAGGTGACA, R: CTCGAAGCTTGGTGAGGACA, human LHR F: CCGGTCTCACTCGAC-TATCACT, R: AAGCTTGAGATGGGATCACTTTG, and human PPIA F: GTTCTTCGACATTGCCGTCG, R: TGTCTGCAAACAGCTCAAAGG.

RNAscope in situ hybridization
Formalin-fixed paraffin-embedded tissue samples were handled according to the manufacturer's protocol using RNAscope 2.0 HD Assay (catalog number 310033, Advanced Cell Diagnostics [ACD], Hayward, CA). In brief, slides were deparaffinized in xylene (2 × 5 min), 100% EtOH (2 × 1 min) and air-dried for 5 min at room temperature. Each section was treated with hydrogen peroxide for 10 min at RT, then washed twice in distilled water. Slides were boiled in antigen retrieval buffer for 15 min and submerged in distilled water immediately thereafter. Next, slides were washed in 100% EtOH and air-dried. For each section, barriers were drawn with the hydrophobic pen and protease was applied for 30 min in 40 • C in HybEZ™ Oven (ACD). Slides were washed twice in distilled water. Probes for the targeted transcripts (mouse ZP3: ACD-447551, human ZP3: ACD-442631) were applied as well as probes for positive (cyclophilin B-PPIB, a housekeeping gene, human ACD-313901; mouse ACD-313911) and negative controls (DapB -negative control probe targeting bacteria gene, ACD-310043). Then, slides were incubated at 40 • C for 2 h in the oven and washed 2 × 2 min in the wash buffer. Thereafter, hybridization amplifiers (AMPs) were applied for 30 min (AMP 1, 3, 5) or 15 min (AMP 2, 4, 6) at 40 • C (AMP 1-4) or at room temperature (AMP 5 and 6) with double washing in between every step. After the last washing, equal volumes of BROWN-A and BROWN-B reagents were combined and applied onto the sections for 10 min at RT. After double washing with distilled water, slides were counterstained in 50% Gill's hematoxylin (Vector Laboratories, Burlingame, CA, USA) for 2 min, then washed in 0.02% ammonia water for 10 s and twice in distilled water. Dehydrated slides (2 × 2 min in 70% EtOH, 2 × 2 min in 100% EtOH and 5 min in xylene) were mounted with Pertex (Histolab Products, Göteborg, Sweden).
Formalin-or Bouin -fixed paraffin-embedded samples were deparaffinized and hydrated. Slides were boiled for 15 min in antigen retrieval buffer (10 mM citric acid buffer with 0.05% Tween20, pH = 6). Then, the sections were incubated in a humidified chamber for 1 h at RT with 3% BSA for reducing nonspecific background staining. Next, slides were incubated overnight in a humidified chamber at 4 • C with the primary antibody anti-ZP3 (0.125 μg/mL) or anti-vimentin (ab28028; Abcam, Cambridge, UK; dilution 1:300). Endogenous peroxidase activity was blocked by incubating slides in 0,5% H 2 O 2 in PBS at room temperature for 20 min. DAKO polymer (DAKO EnVision + System -HRP labeled polymer; Agilent, Santa Clara, CA) was applied onto each section and incubated in a humidified chamber for 30 min at room temperature. DAB + Chromogen (DAKO) was applied for 5 min. Slides were washed in dH 2 O, counterstained in Mayer's hematoxylin (Sigma-Aldrich, Saint Louis, MO), dehydrated and mounted with Pertex (Histolab Products). As a control for the antibodies, tissues were incubated with 3% BSA and DAKO polymer to differentiate unspecific from specific staining.

Immunocytochemical staining
GC-2(spd)ts cells cultured on collagen-coated microscope slide coverslips were washed twice with PBS and fixed in 3.7% paraformaldehyde (PFA) for 20 min. For spermatozoa staining, sperm smear was made and allowed to air-dry at RT for a minimum of 1 h. Next, slides with spermatozoa were fixed in 4% PFA in PBS for 1 h. Slides were rinsed in PBS and treated with 0.5% Triton X-100 in PBS for 15 min. To block unspecific binding sites, cells were incubated in blocking solution (2% BSA in PBS with 0.05% Tween 20). Thereafter, cells were incubated for 1 h with primary mouse monoclonal anti-ZP3 antibody or for dual staining anti-ZP3 antibody with anti-α-tubulin antibody (ab52866, Abcam; dilution 1:300) diluted in blocking solution. After washing, cells were incubated with Alexa Fluor 488 goat anti-mouse IgG (A11029; Thermo Fisher, Waltham, MA; dilution 1:500) or Alexa Fluor 488 goat antirabbit IgG (A11034; Thermo Fisher; dilution 1:500) and Alexa Fluor 546 goat anti-mouse IgG (A11030; Thermo Fisher; dilution 1:500) for 45 min. DAPI dye was used as a counterstain to detect cell nuclei. As a control for the antibodies, the cells were incubated with either 2% BSA or Alexa Fluor-488 goat anti-mouse IgG as a primary antibody to differentiate unspecific from specific staining.

Immunofluorescence staining of FFPE sections
PFA-fixed paraffin-embedded samples were deparaffinized and hydrated. Slides were boiled for 15 min in antigen retrieval buffer (10 mM citric acid buffer with 0.05% Tween20, pH = 6) and cooled down. Slides were washed three times in PBST for 5 min. The autofluorescence was blocked with 100 mM NH 4 Cl for 3 min at RT in a humidified chamber, then the slides were washed three times in PBST for 5 min. Sections were incubated in a humidified chamber for 1 h at RT with a solution of combined 5% NGS and 1% BSA to reduce nonspecific background staining. Next, sections were incubated in a humidified chamber for 1,5 h at RT with the mix of anti-vimentin (1:1000; ab92547, Abcam, Cambridge, UK) and anti-ZP3 (1 μg/ml) primary antibodies suspended in PBST with 1% NGS and 0,5% BSA and washed three times in PBST for 5 min. After washing, cells were incubated with Alexa Fluor 488 goat antimouse IgG (A11029; Thermo Fisher, Waltham, MA; dilution 1:500) and Alexa Fluor 594 goat anti-rabbit IgG (A11012; Thermo Fisher; dilution 1:500) for 1,5 h at RT. DAPI dye was used as a counterstain to detect cell nuclei. As a control for the antibodies, the cells were incubated with either 2% BSA or Alexa Fluor-488 goat anti-mouse IgG as a primary antibody to differentiate unspecific from specific staining.

Apart from the ovary, ZP3 is also expressed in normal human and wild-type (WT) mouse testis in spermatogenic cells
Quantitative PCR analysis showed Zp3 expression in WT mouse ovary and testis (Fig. 1A -1B, Supplementary Table S1). Zp3 expression was not detectable in the WT mouse spleen and muscle used as negative controls (Fig. 1A-B). Immunohistochemical staining and RNAscope in situ hybridization localized abundant Zp3 protein and mRNA transcripts in oocytes from primary, secondary and antral follicles of the mouse ovary ( Fig. 2A and C, respectively). Similarly, abundant Zp3 protein and mRNA were localized in spermatogonia, spermatocytes, round and elongated spermatids in the mouse testis ( Fig. 2B and D, respectively). The mouse primordial follicle oocytes and testicular Sertoli and Leydig cells, as well as the spermatozoa, did not express Zp3 ( Fig. 2A-D). In humans, abundant ZP3 protein and mRNA transcripts were also localized in the oocytes from ovarian follicles ( Fig. 3A and C, respectively) and in testicular spermatogonia, spermatocytes, round and elongated spermatids ( Fig. 3B and D, respectively). ZP3 expression was not detected in the human testicular Sertoli and Leydig cells, either at mRNA or protein levels ( Fig. 3B and D). Cyclophilin B was used as a positive control, which was expressed abundantly in mouse and human ovary and testis (Supplemental Fig. 1, left column). DapB was used as a negative control, and it was not detected in any of the gonadal samples (Supplemental Fig. 1, right column). As a control for immunohistochemical analysis, the samples were incubated with a blocking solution with an omission of the primary antibody (Supplemental Fig. 2A and B) as well as with non-immune isotype control (IgG2A) for the primary anti-ZP3 antibody (Supplemental Fig. 3A and B), and no staining in mouse or human testis was detected. To identify Sertoli cells we immunostained testicular tissue with vimentin, a Sertoli cell marker. We observed positive immunoreactivity for vimentin in the cytoplasm of Sertoli cells with a thin layer of the cytoplasm passing through the whole epithelium (Supplemental Figs. 4A-B). We performed double immunostaining with vimentin and ZP3 to confirm that Sertoli cells did not express ZP3 (Supplemental Fig. 5). ZP3 mRNA transcripts were not found in human primordial follicle oocytes (Supplemental Fig. 6A), whereas abundant expression was localized in oocytes from primary (Supplemental Fig. 6B), secondary (Supplemental Fig. 6C) and antral (Supplemental Fig. 6D) follicles. Zp3 expression was also found in the mouse GC-2spd(ts) immortalized spermatogenic cell line and WT mouse ovary as a positive control (Fig. 4A). No Zp3 expression was detected in Leydig tumor cell line BLTK-1 and Sertoli MSC-1 cells (Fig. 4A). WT mouse skeletal muscle was used as Zp3 negative control (Fig. 4A). Immunocytochemical studies localized Zp3 expression in the cytoplasms of mouse GC-2spd(ts) cells (Fig. 4B). No staining was detected in the control analysis without a primary antibody (Fig. 4B). ZP3 mRNA expression was not detectable in either human primary Sertoli (HSerC) and Leydig (HLC) cells (Fig. 5A -5B). FSHR and LHR expression were analyzed as a control for HSerC and HLC cells, respectively. HSerC expressed FSHR and no LHR, whereas HLC expressed LHR and no FSHR (Fig. 5A-B). ZP3, LHR, FSHR expression were found in human ovary used as a positive control, whereas was not detectable in human muscle used as a negative control (Fig. 5A-B). GFRa1-positive (GDNF family receptor alpha-1) spermatogonia are considered stem and progenitor cells (SSPCs) (differentiating spermatogonia, cells that precede A1 spermatogonia that are differentiated and committed to sperm development) of the male germline and thus responsible for the life-long sperm production (Makela and Hobbs 2019). In order to check Zp3 expression in SSPCs, we performed a whole-mount seminiferous tubule staining for adult mouse testis. We did not observe immunoreactivity for mouse monoclonal anti-ZP3 antibody on the basement membrane of adult mouse seminiferous epithelium suggesting that GFRa1/SALL4-positive (Spalt-like 4, a pan-spermatogonial marker (Chan et al., 2017)) SSPCs and differentiating progenitor spermatogonia (GFRa1-negative/SALL4-positive) do not express ZP3 (Fig. 6). Neither was ZP3 mRNA expression detectable in human spermatozoa (Supplemental Fig. 7). ZP3 expression was found in the human ovary used as a positive control, whereas it was not detectable in human muscle used as a negative control (Supplemental Fig. 7A). No ZP3 staining was detected using immunocytochemical analysis in human spermatozoa (Supplemental Fig. 7B).

Discussion
ZP3 has been shown to be expressed in several cancers (ovarian, prostate, colon and lung) (Buschmann et al., 2004;Rahman et al., 2012;Costa et al., 2018;Coelingh Bennink and RahmanInventor, 2019) as well as in ovarian cancer metastases and recurrences (Rahman et al., 2012), giving rise to the concept of its use as a target for cancer immunotherapy. We have provided the first evidence that ZP3 immunization is an effective treatment of ovarian cancer, using a transgenic mouse model of ovarian granulosa cell tumors (Rahman et al., 2012). The immunization strategy used has the potential of being effective in the treatment of other cancer types expressing ZP3. We designed the present study to characterize ZP3 expression in healthy tissues, with the potential of off-target effects in cancer immunization.
Our results showed the novel expression of ZP3 in spermatogonia, spermatocytes and spermatids in the human and mouse testis. However, no expression was found in mature spermatozoa, spermatogonial stem and progenitor cells and in Sertoli or Leydig cells. In the ovary, ZP3 is required for successful oocyte fertilization, but there is no data on any ZP3 function outside the ovary. High-throughput screening data from   Human Protein Atlas showed ZP3 RNA expression in different normal tissues (in females in the ovary, breast and fallopian tubes and in males in the epididymis, seminal vesicles, prostate, gastrointestinal tract, muscle, brain, adrenal, pituitary, bone and lymphoid tissue) and ZP3 protein expression in the duodenum and small intestine, but no expression in the testis (https://www.proteinatlas.org/ENSG000001 88372-ZP3). For the human protein atlas, they used HPA054061 polyclonal antibody (Sigma-Aldrich), where liver and plasma were negative controls for the ZP3 presence and the U2-OS human osteosarcoma cell line was a positive control, not the ovary, classically expressing the highest amount of ZP3. An osteosarcoma cell line was used as the positive control for ZP3 protein expression, instead of the reference ovarian tissue, which could have led to the inappropriate assignment of ZP3 expression to other tissues. In contrast to the Human Protein Atlas, we did not find protein ZP3 expression in any other healthy normal tissues, except for the ovary and the testis, confirming that results from highthroughput projects should be taken with caution and the need of their verification by independent studies and validation by additional experimental methods (Dammeyer and Arner 2011).
Human protein atlas single-cell RNA sequencing (scRNAseq) data showed male spermatogenic cells express substantially higher levels of ZP3 mRNA than any other cell types for which data were reported, but oocytes were not included in the list (https://www.proteinatlas.org/ ENSG00000188372-ZP3). It has been shown that the data obtained from single-cell RNA-sequencing may often give rise to an abundance of negative data, increased variability, and complex expression distributions. Incorrect normalization may cause over-normalization of the weakly/moderately expressed genes (Bacher and Kendziorski 2016). Due to these limitations of the single-cell RNA-sequencing, proper validation of achieved results by other methods (like qPCR, RT-PCR, Western blots, ICC/IHC, in situ hybridization etc), as well as by statistical/computational methods are needed. In this paper, we used several techniques to check/confirm the ZP3 expression. We analyzed the ZP3 protein localization with a specific monoclonal antibody and confirmed its presence by RNA transcripts in the testis using a sensitive RNAscope in situ hybridization method (Wang et al., 2012).
In female animals, ZP3 immunization induces autoimmune oophoritis and atrophy of the ovaries (Lou et al., 2000). When primordial ovarian follicles are selected for growth and become primary follicles, they start to express ZP antigens in the glycoprotein layer around the oocyte and become susceptible to ZP3-specific antibodies and autoreactive immune cells. It has been demonstrated that antibody binding to ZP3 protein in growing or mature follicles is followed by induction of T cell-mediated inflammation resulting in degeneration of developing follicles (Lou et al., 2000). Recently, it has been shown that in actively immunized mice after careful selection of B-cell epitopes, ZP3 vaccines may induce reversible infertility without adverse effects on the ovaries (Paterson et al., 2002). With a chimeric peptide vaccine, the defined  B-cell epitope had a single critical amino acid substitution to prevent cross-reaction with a native ZP3 T-cell epitope, but not B-cell response to ZP3 (Lou et al., 1995). This ZP3 vaccine was able to successfully prevent pregnancy without causing ovarian pathology.
Only one study reported ZP3 expression in human oocytes and granulosa cells as early as the primordial follicle stage (Gook et al., 2008), suggesting ZP protein expression since oogenesis. However, this study used a polyclonal antibody (Harris et al., 1994) not validated by others. In contrast to this study, and in line with the rest of the literature (Bleil and Wassarman 1980;Litscher 2013, Canosa et al., 2017), we did not find any ZP3 expression (either protein or mRNA transcripts) in the primordial follicles or granulosa cells in the present study, using monoclonal ZP3 antibodies. This result confirmed our earlier study with the Zonagen/Repros polyclonal antibody Fig. 6. Immunohistochemical localization of ZP3 in mouse spermatogonial stem and progenitor cells, SSPCs. A confocal microscopy image of stage VI-VIII adult mouse whole-mount seminiferous tubule staining using antibodies against SSPC marker GFRa1 (red), ZP3 (green) and a pan-spermatogonial marker SALL4 (blue). White arrowheads indicate GFRa1/SALL4-positive A-single (A s ) undifferentiated spermatogonia. ZP3 is not expressed in A s SSPCs. ZP3-positive cells in the middle of the image are preleptotenes spermatocytes (white arrows).
Women have a variety of contraceptive methods to choose from, but male contraceptive choices are still limited (Amory 2016). Condoms are the most popular option, but are unreliable (Duerr et al., 2011). A second option is a vasectomy, but this method is invasive and largely irreversible (Kanakis and Goulis 2015). Hormone-based (androgen alone or in combination with a progestin) male contraceptive methods have been tested extensively, but have failed to suppress spermatogenesis in all men completely (Oduwole and Huhtaniemi 2014). These hormonal methods often cause side effects due to the administration of high doses of testosterone and progestin preparations, such as acne, mood changes, weight gain, night sweats and altered libido (Meriggiola et al., 2003). Currently, several projects (such as combined nestorone-testosterone gel or male contraceptive pill using a progestogenic androgendimethandrolone 17 beta-undecatone) to develop a hormone-based male contraceptive are ongoing, but adequate efficacy and safety have not been proven yet (Serfaty 2020). There is no doubt that breakthroughs are needed to revive the stagnant field of male contraceptive development.
ZP3 antigen expression in the testis introduces the surprising possibility to use ZP3 immunization for male contraception. The absence of ZP3 expression in any other normal male tissues, as well as in somatic testicular cells, reduces the risk of concomitant damage to other organs. However, the ZP3 epitopes should be carefully selected to avoid the destruction of testicular tissue and allow the ability to cross the bloodtestis barrier (Cheng and Mruk 2012). Another important feature of such immunization would be its reversibility. Also, since ZP3 is not expressed by the testosterone synthesizing Leydig cells in the testis, unfavorable endocrine side effects may not occur, provided no general auto-immune response occurs due to adequate selection of the epitopes. Furthermore, since ZP3 expression was not detected in the SSPCs, as well as in Sertoli cells, it is unlikely that anti-ZP3 male contraceptive therapy would destroy spermatogenic stem cells to induce irreversible infertility; hence this male contraceptive strategy could be reversible upon discontinuation.
In conclusion, we report here the novel extraovarian ZP3 expression in human and mouse spermatogenic cells, and the absence of ZP3 expression in a series of normal human and mouse tissues. These findings on absent ZP3 expression in normal nongonadal tissues provide crucial background information for any ZP3 cancer immunotherapy. Any future ZP3 cancer immunotherapy clinical trials in males (such as prostate, lung or colorectal cancer) should proceed with caution due to the ZP3 expression in testis. Besides cancer immunotherapy, another therapeutic strategy was proposed by the finding on ZP3 expression in the testis. The exclusive localization of testicular ZP3 expression in spermatogonia, spermatocytes and spermatids make ZP3 immunization a promising new target for male immunocontraception. Further studies should also address the functional role of ZP3 in the testis, and in particular whether ZP3 immunization has a long-term or reversible effect on male fertility.

Declaration of competing interest
The other authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study. HJTCB is the owner of the Pandora Endocrine Innovation.

Simple summary
Ovarian oocyte-specific zona pellucida glycoprotein 3 (ZP3) expression was found recently in several cancer tissues, including those of the prostate and lung. The efficacy of ZP3 immunization has been demonstrated in the treatment of ZP3-positive ovarian cancer in transgenic mice, suggesting the efficacy of similar therapy for other ZP3positive cancers. Hence, we examined whether any other normal tissue besides the ovary may express ZP3, to exclude potential off-target effects of cancer ZP3 immunotherapy. Unexpectedly, ZP3 expression was found in spermatogonia, spermatocytes and round and elongated spermatids of human and mouse testis, but absent in mature spermatozoa, Sertoli, Leydig, spermatogonial stem and progenitor cells, and in an array of 14 different tissues examined. The novel ZP3 expression in testis needs to be taken into consideration during putative ZP3 cancer immunotherapy. ZP3 expression in testicular germ cells, but absence in spermatogenic stem and somatic cells, makes ZP3 a potential target for male immunocontraception.

Declaration of competing interest
HJTCB is the owner of the Pandora Endocrine Innovation. The other authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study.