If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Gamete Immunobiology, National Institute for Research in Reproductive Health (Indian Council of Medical Research), Jehangir Merwanji Street, Parel, Mumbai 400012, India
Department of Reproductive Endocrinology and Infertility, National Institute for Research in Reproductive Health, Jehangir Merwanji Street, Parel, Mumbai 400012, India
Department of Gamete Immunobiology, National Institute for Research in Reproductive Health (Indian Council of Medical Research), Jehangir Merwanji Street, Parel, Mumbai 400012, India
Antibodies to multiple ovarian antigens have been proposed as markers of ovarian autoimmunity. The role of ovarian autoantibodies has been widely discussed in the pathophysiology of premature ovarian failure and unexplained infertility, but the autoantigens are yet to be identified. Three immunodominant ovarian autoantigens, α-actinin 4 (αACTN4), heat shock 70 protein 5 (HSPA5) and β-actin (ACTB), have been identified using anti-ovarian antibody-positive sera from women with idiopathic premature ovarian failure (n = 50) and women undergoing IVF (n = 695), using mass spectrometry. These autoantigens were subsequently validated using Western blot, immunohistochemistry and enzyme-linked immunosorbent assay. These autoantigens are localized to different components of the ovary such as the ooplasm of the oocyte, theca, granulosa, corpus luteum and zona pellucida. All the above antigens were found to be expressed in the ooplasm throughout follicular development. All the autoantigens are expressed specifically in the oocyte except αACTN4. The three autoantigens could contribute to the array of biomarkers to be used for developing specific and sensitive tests for diagnosis of women at risk of premature ovarian failure and IVF failure due to ovarian autoimmunity and could give an insight into the molecular mechanisms involved in the pathophysiology of these conditions.
Anti-ovarian antibodies (AOA) have been reported in women with premature ovarian failure and unexplained infertility. Among women with infertility, those with evidence of ovarian autoimmunity appear to have poorer IVF–embryo transfer outcomes. Diagnosis of an autoimmune mechanism in these pathologies has relied for a long time on the detection of AOA. Little is known about the precise ovarian antigenic targets in terms of its molecular and cellular identities that are recognized by the antibodies and immune cells in autoimmune diseases of the ovary. In the present study, we observed that 31% of the total women recruited under an IVF–embryo transfer programme (group I) and 46% of women with premature ovarian failure (group II) tested positive for AOA. Three immunodominant ovarian autoantigens, namely non-muscle α-actinin 4, heat shock 70 protein 5 and cytoplasmic β-actin, were identified using mass spectrometry and validated and characterized using AOA-positive sera from women from both groups. Further investigation of the identified targets could give us an insight into the molecular mechanism involved in the pathophysiology of human ovarian autoimmunity.
The influence of the immune system on various aspects of reproduction, especially on the reproductive outcome, has been extensively studied. Autoimmune factors like antiphospholipid antibodies (
Patients with immune-mediated reproductive failure are presented as a clinically heterogeneous group including unexplained infertility, premature ovarian failure (POF), endometriosis and polycystic ovarian syndrome, which have been associated with ovarian autoimmunity (
). Presence of anti-ovarian antibodies has been demonstrated using different assays, which include indirect immunofluorescence (IIF) on monkey ovary sections (
Ovarian antibodies as detected by indirect immunofluorescence are unreliable in the diagnosis of autoimmune premature ovarian failure: a controlled evaluation.
Ovarian failure and autoimmunity. Detection of autoantibodies directed against both the unoccupied luteinizing hormone/human chorionic gonadotropin receptor and the hormone–receptor complex of bovine corpus luteum.
) have also been reported. The results from several studies are conflicting, which could be attributed to methodological differences, multiplicity of potential targets and high incidence of false positives. Therefore, until specific assays with well-characterized ovarian antigens are developed, the ovarian antibody tests will be of little value in determining the aetiology of premature ovarian failure/primary ovarian insufficiency and IVF failure. The problem of the non-specificity of the AOA test has been circumvented by employing a novel blocking protocol (
). Of the several antigens, one 90-kDa antigenic target was identified to be human heat shock protein 90 β (HSPA5-β) and was characterized extensively (
). The present study reports the prevalence, molecular identity and cellular distribution of the other three immunodominant autoantigens reported, α-actinin 4 (αACTN4), heat shock 70 protein 5 (HSPA5) and β-actin (ACTB). The study also highlights their localization during ontogeny and postulates a probable role in fertilization and early embryonic development. Recombinant proteins of these autoantigens were used to confirm the reactivity of both infertile and POF sera with these autoantigens using a standard enzyme-linked immunosorbent assay (ELISA) method. These targets are now being pursued to develop a more specific peptide-based ELISA using the immunodominant regions based on bioinformatics analysis.
Materials and methods
Protease inhibitor cocktail was obtained from Roche Diagnostics GmbH (Mannheim, Germany) and non-fat dry milk (NFDM) from Amul (Mumbai, India). Molecular-weight markers, nitrocellulose membrane, enhanced chemiluminescence detection kit and protein silver-staining kit were obtained from GE Healthcare (Piscataway, NJ, USA). Secondary antibody was obtained from Jackson ImmunoResearch Laboratories (West Grove, PA, USA) and Dako Cytomation (Glostrup, Denmark). ELISA plates were purchased from Nunc (Roskilde, Denmark). Reagents for trypsin digestion and mass spectrometry by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) and matrix-assisted laser desorption/ionization tandem time-of-flight (MALDI-TOF/TOF) were obtained either from Sigma (MO, USA) or Applied Biosystems (Foster City, CA, USA). All other reagents were procured from Qualigens, SRL India and HiMedia (Mumbai, India). Recombinant proteins αACTN4, HSPA5 and ACTB were obtained from Abnova (Taipei, Taiwan). Antibodies to αACTN4 and HSPA5 were obtained from Abcam (UK) and to ACTB from Sigma.
Study population
The study was approved by the Clinical Ethics Committee of the National Institute for Research in Reproductive Health. Three groups participated in the study and gave their informed consent before blood samples were collected. Group I (n = 695) comprised of women of reproductive age (32.5 ± 4.0 years and an average duration of 3.06 years of infertility) recruited for the IVF programme at Jaslok Hospital and INKUS Clinic, Mumbai, India. Group II (n = 50) comprised of idiopathic POF women with an average age of 27.7 ± 6.0 years (<40 years), and presenting with secondary amenorrhoea with mean serum FSH concentrations of 84.24 ± 32.6 mIU/ml (>40 mIU/ml) at the endocrinology clinic at Kasturba Hospital and the infertility clinic at the National Institute for Research in Reproductive Health. Group III (control, n = 100) comprised of age-matched, healthy, regularly menstruating, proven-fertile women.
Women with genetic involvement, abnormal karyotypes, viral oophoritis, galactosaemia-induced POF, endometriosis, polycystic ovarian syndrome or uterine fibroids or having undergone radiation/chemotherapy or any ovarian or pelvic surgeries were excluded from the study. The study population had well-developed secondary sexual characteristics and no dysmorphic clinical symptoms.
Animals
Animals were housed with food and water available ad libitum. All the experimental protocols were approved by the Institutional Ethics Committee for Care and Use of Laboratory Animals for Biomedical Research. Inbred female Holtzman rats (15–20 weeks) were used for tissue specificity, Western blot analysis and immunohistochemistry (IHC) or IIF. Female rats at age 0, 10, 20 and 30 days (n = 3) were used for this developmental study.
Serum and tissue collection
Blood samples were collected from women from all three groups in sterile vacutainers, centrifuged at 336g and 4°C for 10 min, and serum was separated and stored at −20°C until use. Rats were sacrificed and dissected to collect tissues, namely ovary, lung, thyroid, muscle, adrenal, epididymis, testis, gastric mucosa, brain, kidney, liver, pancreas, thymus, heart and spleen. All the tissues were used for IHC/IIF, while the ovary was also processed for protein preparation for sodium dodecyl sulphate (SDS)–polyacrylamide gel electrophoresis (PAGE) Western blotting. Ovaries collected for IHC and Western blot experiments were randomly collected and not designated to a particular phase of the cycle. Normal human ovarian tissue was obtained from a local hospital with prior ethical permission.
Protein extraction
Total rat ovarian protein extracted in 1% SDS solution containing protease inhibitor cocktail was prepared as described earlier (
) and used for Western blot analysis using patient and control sera as well as commercial antibodies.
Protein bands corresponding to the immunoreactive band were electroeluted from ovarian protein gels using the electroelution apparatus from BioRad (Hercules, CA, USA) and used either for Western blot (10 μg/lane) or silver stained and sequenced. Glutathione-S-transferase-tagged recombinant proteins, αACTN4, HSPA5 and ACTB, were procured from Abnova and used for Western blotting and ELISA.
SDS–PAGE Western blotting
Total ovarian protein (40 μg/lane), electroeluted protein (10 μg/lane) or recombinant proteins (αACTN4 (600 ng/lane), HSPA5 (435 ng/lane) and ACTB (2 μg/lane)) were separated electrophoretically on 10% SDS–PAGE gels. Proteins were transferred onto nitrocellulose membrane, blocked with 5% NFDM in 0.01 mol/l phosphate-buffered saline (PBS) pH 7.4 and probed with either neat sera (group I, n = 695; group II, n = 50), target-specific sera (αACTN4, n = 45; HSPA5, n = 63; ACTB, n = 31) for each target or diluted commercial antibodies, namely anti-αACTN4, anti-HSPA5 raised in rabbit, diluted 1:250 and 1:100, respectively, and monoclonal anti-ACTB at 1:3000 dilution overnight at 4°C. The following day, blots were washed with PBS/Tween 20, containing 0.05% (w/v) Tween 20, five times for 10 min each. Secondary antibody varied depending upon the primary antibody used. Horse-radish peroxidase (HRP)-labelled rabbit anti-human, diluted 1:100000, swine anti-rabbit, diluted 1:3000, or rabbit anti-mouse, diluted 1:3000, was added onto the membranes. The peroxidase activity was visualized with a chemiluminescent substrate (ECL Plus Western Blotting Detection Reagents; GE Healthcare). Two sets of negative controls were used; one set was buffer control or secondary alone control where PBS was added in place of primary antibody. For the second set of negative controls, sera of group III subjects or normal rabbit IgG/mouse myeloma supernatant were used. Each experiment was performed thrice at different time intervals. To ensure that the same amount of ovarian protein was subjected to the electrophoresis and Western blot procedures, the same membranes were stripped with the 0.2 mol/l stripping buffer and reprobed using the same protocol using a mouse monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase antibody.
Immunohistological analysis
To determine the cellular targets involved in ovarian autoimmunity, adult rat ovarian tissues were fixed in Bouin’s fixative, paraffin embedded and 5-μm ovarian sections were processed as per the protocol described earlier (
). Briefly, sections were deparaffinized in xylene, quenched by endogenous peroxidase activity and rehydrated. Sections were blocked with 5% NFDM–PBS containing 20% rabbit polyclonal anti-albumin antibodies for 2 h at room temperature. Neat sera of group I (n = 215) and group II (n = 23) positive by Western blot were used as the primary antibody and incubated at 4°C overnight in a humid chamber. PBS and group III sera served as negative control. Next day, sections were washed with 0.01 mol/l PBS thrice and incubated with 1:800 diluted rabbit anti-human HRP-labelled secondary antibody made in 1% NFDM in PBS for 1 h at room temperature. After washing, the immunoperoxidase colour reaction was developed using 3,3′-diaminobenzidine substrate chromogen solution (Sigma), then counterstained for 30 s using Delafield’s haematoxylin, dehydrated through a series of alcohol grades, cleared in xylene and mounted in DPX mounting medium. Slides were examined with an Axioscope microscope (Carl Zeiss, San Marcos, CA, USA) at a X70 magnification.
Sequencing of the target proteins
Sample preparation
Ovarian protein bands showing reactivity at 45, 80, 97 and 120 kDa using sera from group I and II were excised from Coomassie blue-stained gel and subjected to electroelution. The eluted proteins designated as PM45, PM80, PM97 and PM120 were concentrated and separated on 10% gel and stained using a silver-staining kit. Their immunoreactivity was reconfirmed by Western blotting with specific sera (n = 10) for each target.
In-gel trypsin digestion
The bands were manually excised from the silver-stained gels in duplicate. Bovine-serum-albumin gel plugs were used as a protein standard for each set of in-gel tryptic digestion experiment. The gel bands were destained using a 1:1 mixture of 30 mmol/l potassium ferricyanide and 100 mmol/l sodium thiosulphate and digested using sequencing-grade 0.01 mg/ml trypsin (Applied Biosystems) for 16 h at 37°C according to the protocol described earlier (
). The tryptic peptides were eluted using different grades of trifluoroacetic acid. The extracts were pooled, lyophilized, reconstituted in sample diluent and an equal volume of matrix (10 mg/ml α-cyano-4-hydroxycinnamic acid in 70% v/v acetonitrile, 1% v/v trifluoroacetic acid) and spotted in duplicate on a MALDI plate. For each set of MS analysis, trypsin-digested β-galactosidase Escherichia coli and α-cyano-4-hydroxycinnamic acid mixture were also spotted.
MALDI-TOF/TOF and protein identification
The peptides of the eluted proteins were analysed by MALDI-TOF/TOF using a 4700 Proteomics Analyser mass spectrometer (Applied Biosystems) at the proteomics core facility of the National Institute for Research in Reproductive Health, as previously described (
). Most of the observed peaks of trypsin autolysis and keratin were validated and subsequently excluded from monoisotopic precursor ion list generated by tandem mass spectrometry (MS/MS) analysis. A maximum of the 25 strongest precursor ions per sample were chosen for MS/MS analysis. Combined MS and MS/MS spectra were used to search against the taxonomy of Homo sapiens in the Swiss-Prot Protein knowledgebase (Swiss-Prot release 20080226; 356194 sequences; 127836513 residues) and/or protein sequences in the Mass Spectrometry Data Base (MSDB release 20063108; 3239079 sequences; 1079594700 residues) using GPS software (version 3.5; Applied Biosystems) running the Mascot search algorithm (version 2.0; Matrix Science, Boston, MA) for peptide and protein identification. Searches were performed to allow for carbamidomethylation, oxidation, trypsin as an enzyme and a maximum of one missed trypsin cleavage. Confident identification had a statistically significant (P ⩽0.05) protein score (based on combined MS and MS/MS spectra) and best ion score (based on MS/MS spectra). The sequencing experiment was performed at three different times using different batches of eluted proteins.
Validation of the identified proteins
Validation of proteins identified by the proteomic analyses was performed by Western blotting, IHC/IIF and ELISA using recombinant proteins.
Western blotting
In one set of Western blot experiments, eluted proteins were probed with commercial antibodies as described above and the experiment was repeated twice. In the second set of Western blot experiments, recombinant proteins were probed with target-specific patients’ sera (n = 5); i.e. recombinant αACTN4 protein was probed with five representative patients’ sera targeting the 97-kDa αACTN4 protein, recombinant HSPA5 protein was probed with five representative patients’ sera targeting the 80-kDa HSPA5 protein and recombinant ACTB protein was probed with five representative patients’ sera targeting 45-kDa ACTB protein.
Immunohistochemistry/indirect immunofluorescence
Rat ovarian sections were deparaffinized and probed with commercial antibodies to αACTN4 and ACTB for IIF. Anti-HSPA5 was used for IHC after antigen retrieval in citrate buffer pH 6.0 for 15 min at 90–95°C. Sections were blocked with 5% normal goat serum for αACTN4 and ACTB and with 5% normal swine serum for HSPA5 for 1 h at room temperature in a humidified chamber. This was followed by incubation of appropriate primary antibody dilutions (1:100), overnight at 4°C. Next day, slides were washed thrice for 5 min each in 0.01 mol/l modified PBS. Fluorescein isothiocyanate-conjugated goat anti-mouse IgG (for αACTN4 and ACTB, Sigma) and HRP-conjugated swine anti-rabbit IgG (for HSPA5, Dako Cytomation) were used as secondary antibodies. The slides were washed as above and counterstained with 4,6-diamidino-2-phenylindole for αACTN4 and ACTB and haematoxylin for HSPA5 and mounted with Vectashield (Vector Labs, Burlingame, CA, USA) and DPX mounting medium, respectively. Images were acquired with confocal microscopy (LSM-510-META; Carl Zeiss, Jena, Germany) at ×520 magnification and wavelength 488 nm for green (Argon Laser) and 405 nm for blue (Blue diode laser) for αACTN4 and ACTB. Images for HSPA5 were acquired on DMLA Laser Capture Micro-dissection microscope (Leica, Wetzler, Germany) at ×240 magnification.
Enzyme-linked immunosorbent assay
Recombinant protein in carbonate bicarbonate buffer (25 ng; pH 9.2) were coated onto microtitre plates (Nunc) overnight at 4°C. The following day excess antigen was removed and the undiluted patients’ sera of group I and II were used as primary antibody while group III sera served as the control. The wells were then washed five times with PBS/Tween 20. Goat anti-human HRP-conjugated secondary antibody diluted 1:2000 in 1% (w/v) PBS was added and incubated for 1 h at 37°C. Finally, wells were washed with PBS/Tween 20, prior to incubation with 100 μl 3,3′,5,5′-tetramethylbenzidine plus H2O2 solution diluted 1:20 (Bangalore Genei, Bangalore, India). The reaction was terminated by addition of 2 mol/l H2SO4 and the colour was measured at 450 nm on a Universal Micro plate Reader (Bio-Tek Instruments, Winooski, VT, USA). Each sample was assayed in duplicate. Patient sera were considered positive if absorbance was more than two standard deviations above the mean control absorbance value (P <0.05).
Immunohistochemical localization of target proteins in human and rat ovary
Bouin’s fixed human and rat ovarian tissue sections (5 μm) were probed with neat target-specific patients’ sera, i.e. sera targeting 45, 80 or 97 kDa to determine protein localization by IHC as described earlier (
). Group III sera served as negative controls. Images were acquired on DMLA Laser capture Micro-dissection microscope at X70 magnification.
Tissue specificity and developmental expression of target proteins
Sections of Bouin’s fixed rat tissues (5 μm), namely ovary, lung, thyroid, muscle, adrenal, epididymis, testis, gastric mucosa, brain, kidney, liver, pancreas, thymus, heart and spleen, were probed to determine the specificity using neat target-specific sera (n = 3) by IHC as described earlier (
). PBS and group III sera served as negative controls. To analyse the post-natal expression of the autoantigens, rat ovarian sections obtained from animals of age 0, 10, 20 and 30 days (n = 3) were probed with target-specific patient sera. Images were acquired on DMLA Laser Capture Micro-dissection microscope at X240 and X70 magnifications respectively.
Statistical analysis
Statistical analysis was performed using Graphpad Prism 4 software (California, USA). One-way ANOVA was used to identify the significant differences. Statistical significance was set as P <0.05.
Results
Sera of AOA positive women identify cognate molecules involved
To study the frequency distribution of cognate targets responsible for ovarian autoimmunity, sera of women recruited for IVF (group I, n = 695) and idiopathic POF (group II, n = 50) were screened using Western blot against rat ovarian proteins. The sera from both the groups identified ovarian proteins spanning molecular weights 30–220 kDa. Compilation of the data from Western analysis from both groups is presented in Figure 1A. A Coomassie blue-stained profile of total rat ovarian extracts is shown in Figure 1B, corresponding to the representative image of the blot showing positive reactivity in Figure 1C (lanes 1–5). Serum from control group III did not show any specific reactivity; however, a 66-kDa band was identified as albumin by all sera (Figure 1C, lane NC1). No primary antibody (PBS only) (Figure 1B, lane NC2) did not show any reactivity. Of the eight targets identified, 45, 80, 90, 97 and 120 kDa were prevalent in both the groups. Positivity was seen in 31% of women from group I sera, with 90 kDa as the most immunodominant target, while group II showed positivity in 46% of patients with 80-kDa protein as the immunodominant one. Of the positive bands 45, 80, 97 and 120 kDa designated as PM45, PM80, PM97 and PM120 were used for further studies.
Figure 1(A) Bar graph showing percentage of ovarian autoantigens targeted by the sera of women of reproductive age undergoing IVF (group I) and women with idiopathic premature ovarian failure (group II), using data obtained from Western blot analysis. (B) Coomassie blue-stained sodium dodecyl sulphate polyacrylamide gel electrophoresis profile of total rat ovarian protein lysate, showing molecular size of autoantigens (1–5: 45, 80, 90, 97 and 120 kDa, respectively). (C) Representative Western blot of immunodominant autoantigens in rat whole ovarian extract using sera from anti-ovarian antibody-positive patients, targeting 45, 80, 90, 97 and 120 kDa antigens (lanes 1–5). Proven fertile control (lane NC1) and no primary antibody control (lane NC2) did not show any reactivity to any antigenic target. Albumin is seen as a 66-kDa band. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as the loading control.
Cellular distribution of the targets involved in ovarian autoimmunity
Sera of group I and II patients positive by Western blot were analysed for immunohistochemical studies in order to determine the cells targeted by AOA. Rat ovarian sections were probed with target-specific sera to PM45, PM80, PM97 and PM120. A representative image showing the localization pattern is shown in Figure 2A. It was noted that staining was positive in the oocyte, which was the most targeted cell type (panels 1, 2 and 3, showing reactivity as grade 1+, 2+ and 3+, respectively), followed by the theca (panel 4) corpus luteum (panel 5), granulosa (panel 6) and zona pellucida (panel 1). Buffer control and sera from control women (group III) did not show any positive staining (panels 7 and 8). The expression pattern showing positivity of group I and II is presented in Figure 2B. Sera from group I patients (79/215, 36.7%) and in group II (9/22, 41%) were positive for oocyte staining. Sera to PM97 and PM80 exhibited localization in the oocyte, theca and corpus luteum, while PM45 showed expression predominantly in the oocyte. Sera to PM120 showed positive reactivity in the oocyte, theca, corpus luteum and zona pellucida.
Figure 2(A) Representative immunohistochemistry showing immunoreactivity of anti-ovarian antibody-positive patient sera to oocyte ooplasm of secondary follicle and to zona pellucida (1), antral follicle (2), Graafian follicle (3), theca (4), corpus luteum (5) and granulosa (6), indicated with black arrowheads. Length of the bar is 1cm. Magnification is same for all panels and a representative bar is inserted in the first panel. Sera from healthy control women (7) and no primary antibody control (8) showed no immunoreactivity to any cell type. Scale bar represents = 140 μm for all panels. (B) Bar graph showing the frequency of different cell types in rat ovarian tissue targeted by antibodies from the sera of patients from groups I and II.
Protein sequencing of immunoreactive proteins involved in ovarian autoimmunity
The bands identified as immunoreactive were cut from multiple Coomassie blue-stained gels of ovarian extracts and pooled and proteins were electroeluted from these bands for each target protein. A silver-stained gel profile of the eluted target is presented in Figure 3A (lanes 1–4) showing each of the targets PM45, PM80, PM97 and PM120 as a single band, which indicates homogeneity of the eluted targets. The reactivity of these bands was reconfirmed using target-specific sera (n = 5) for each target which is presented in Figure 3B. All the target-specific sera showed reactivity with the respective eluted proteins.
Figure 3(A) Representative silver-stained gel of the eluted proteins PM45 (lane 1), PM80 (lane 2), PM97 (lane 3) and PM120 (lane 4). (B) Western immunoblots showing reactivity of patients’ sera with eluted proteins, showing five different 45-kDa-positive patient sera reactive to eluted PM45 protein (panel a, lanes 1–5), 80-kDa-positive patient sera reactive to eluted PM80 protein (panel b, lanes 1–5), 97-kDa-positive patient sera reactive to eluted PM97 protein (panel c, lanes 1–5) and 120-kDa-positive patient sera reactive to eluted PM120 protein (panel d, lanes 1–5). Sera from control women (panels a, b, c and d, lane 6) showed no reactivity to any of the eluted proteins.
The gel bands were destained and digested with trypsin. The digested proteins were subjected to proteomic analysis. Following MALDI-TOF and MALDI-TOF/TOF, combined MS and MS/MS spectra were used to search against H. sapiens in the MSDB using GPS software running the Mascot search algorithm for peptide and protein identification. Of the four bands, PM45, PM80, PM97 and PM120, three bands gave a single match and the remaining one band did not give any significant protein. PM 45 was identified as ACTB (UniProt ID Q96HG5), PM80 was identified as HSPA5 (UniProt ID Q2KHP4) and PM97 was identified as αACTN4 (UniProt ID AAC17470) (Table 1; protein identification details are provided in Supplements 1 and 2, available online only). Since the proteins were eluted from rat ovary, homology searches were performed to determine the difference between rat and human for the target proteins. NCBI Blast results showed above 90% homology between rat and human forms of αACTN4, HSPA5 and ACTB.
Table 1Autoantigens identified from rat ovary using matrix-assisted laser desorption/ionization tandem time-of-flight mass spectrometry.
Variable
PM45
PM80
PM97
Protein name
β-Actin
Heat shock 70 protein 5
α-Actinin 4
MW (kDa)
40.978
72.377
105.159
pI
5.56
5.56
5.47
Coverage (%)
59
37
27
Mowse score
654
505
199
SwissProt AC
Q96HG5
Q2KHP4
AAC17470
Peptides matched
10
7
5
Putative functions
Structural protein
Molecular chaperone
Structural protein
AC = accession number; Coverage = peptide sequences matched; pI = isoelectric point; MW = molecular weight.
Validation of the target proteins involved in ovarian autoimmunity
To validate that the cognate proteins responsible for ovarian autoimmunity and identified by proteomic analysis were indeed the ones that were immunoreactive, Western blotting, IHC/IIF and recombinant ELISA experiments were performed. Reactivity of rat ovarian extract and eluted proteins using commercial antibodies is presented in Figure 4A(a–c). The monoclonal/polyclonal antibodies reacted with expected molecular-weight proteins in the ovarian extracts (Figure 4A, lane 1). Reactivity of the antibodies with their respective eluted proteins (Figure 4A, lane 2) as well as reactivity of patient sera with the recombinant proteins (Figure 4B, lane 3) prove that the identification obtained from proteomic analysis was definitely authentic. Sera from group III women and buffer control did not show any reactivity to the proteins (Figure 4A lane 3 and Figure 4B lanes 1 and 2).
Figure 4(A) Western blot analysis of rat total ovarian lysate (lane 1) and eluted proteins (lane 2) using: anti-β-actin monoclonal antibody (a), anti-heat shock 70 protein 5 polyclonal antibody (b) and anti-α-actinin 4 polyclonal antibody (c). Myeloma culture supernatant (lane 3) served as negative control and showed no reactivity to any of the eluted proteins. (B) Representative Western blots showing reactivity of patients’ sera with: recombinant glutathione-S-transferase (GST)-tagged β-actin (lane a3); recombinant GST-tagged heat shock 70 protein 5 (lane b3); and recombinant GST-tagged α-actinin 4 (lane c3). A no primary antibody control (lane 1) and proven fertile control sera (lane 2) showed no immunoreactivity to any of the recombinant proteins. (C) Representative immunofluorescent and immunohistochemical localization of autoimmune targets in the ovarian section using: anti-β-actin antibody with ooplasm, granulosa and theca, as indicated with white arrowheads (panels a2 and a3); anti-heat shock 70 protein 5 antibody with ooplasm, theca, granulosa and corpus luteum, as indicated with black arrowheads (panels b2 and b3); and anti-α-actinin 4 antibody with ooplasm, theca and corpus luteum, as indicated with white arrowheads (panels c2 and c3). Myeloma culture supernatant serving as negative control did not show reactivity to any of the rat ovarian proteins (panels a1, b1 and c1). Scale bar represents = 40 μm for all panels. Length of the bar is 1 cm. Green = positive stain; blue = counter stain. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Immunohistological studies using the commercial antibodies show reactivity in rat ovarian sections (Figure 4C). Reactivity of anti-ACTB (panels a2 and a3) and anti-HSPA5 (panels b2 and b3) was seen in the ooplasm, granulosa, theca and corpus luteum, while anti-αACTN4 showed positive staining in the oocyte, theca and corpus luteum (panels c2 and c3). Myeloma culture supernatant (panels a1, b1 and c1), normal rabbit IgG and PBS alone (data not shown) did not show any reactivity.
As it was not possible to validate all the target-specific positive sera by Western blotting using recombinant proteins, an ELISA was performed wherein recombinant proteins were probed with the sera of patients in groups I and II. Sera of group III women served as the control. The number of sera screened were as follows: group I, αACTN4 n = 39, HSPA5 n = 56 and ACTB n = 29; group II αACTN4 n = 6, HSPA5 n = 7 and ACTB n = 2. The optical density values for both the patient groups were significantly higher than the control samples (n = 27, P <0.0001). Using a cut-off value (mean ± 2SD) based on control samples (1.3 for αACTN4, 1.2 for HSPA5 and 1.3 for ACTB), 84.4% (38/45) of total patients’ sera had antibodies against αACTN4, 79.4% (50/63) had anti-HSPA5 antibodies and 64.5% (20/31) had anti-ACTB antibodies.
The results were further analysed for differences in the percentage positivity between the two groups. The percentage of positive sera for the three recombinant antigens in group I and group II were 87.2% (34/39) and 66.7% (4/6) for αACTN4, 80.4% (45/56) and 71.4% (5/7) for HSPA5 and 65.5% (19/29) and 50.0% (1/2) for ACTB. Of the control sera, 14% and 18% had antibodies to αACTN4 and HSPA5, respectively, whereas none had anti-ACTB antibodies (Figure 5). The specificity of sera for a particular antigen was cross-checked using sera targeting other antigens (data not shown).
Figure 5Bar diagram showing the reactivity of group I (black box), group II (red box) patients’ sera and control sera (blue box) against recombinant α-actinin 4 (ACTN4), heat shock 70 protein 5 (HSPA5) and β-actin (ACTB). The optical densities (OD) of group I and group II sera were significantly higher (all P <0.0001) than the control sera.
Immunohistochemical localization of target proteins in human and rat ovarian sections using representative serum positive for each target protein and group III sera (control) is presented in Figure 6. The sera showed positive reactivity to the oocyte in both human and rat ovaries (panels A2, B2, A3, B3, A4 and B4). However, group III sera (control, panels A1 and B1) as well as buffer control (data not shown) did not show any staining in these tissues.
Figure 6Immunohistochemical localization of autoimmune targets in ovary from human (A) and rat (B) using patients’ sera targeting 45 kDa (β-actin, panels A2, B2), 80 kDa (heat shock 70 protein 5, panels A3, B3) and 97 kDa (α-actinin 4, panels A4, B4). Control sera (panels A1, B1) did not show any reactivity to the human and rat ovaries. Scale bar represents = 140 μm for all panels. Length of the bar is 1 cm.
Tissue specificity and developmental expression of the AOA targets
To determine if the expression of the proteins was specific to the ovary, a panel of rat tissues, namely lung, thyroid, muscle, adrenal, epididymis, testis, gastric mucosa, brain, kidney, liver, pancreas, thymus, heart and spleen were probed with patients’ sera specific to each target. A representative image of the reactivity of sera targeting PM97 (αACTN4) is depicted in Figure 7. The sera showed positive reactivity with kidney (panel 15), with no staining in the remaining tissues (panels 1–14). However, patients’ sera targeting PM45 (ACTB), PM80 (HSPA5) and group III sera (control) as well as buffer control did not show any staining in any of these tissues (data not shown).
Figure 7Expression pattern of α-actinin 4 in rat somatic tissues: lung (1), thyroid (2), muscle (3), adrenal (4), epididymis (5), testis (6), gastric mucosa (7), brain (8), kidney (9), liver (10), pancreas (11), thymus (12), heart (13), spleen (14) and kidney (15), as indicated with black arrowheads. α-Actinin 4-positive sera only showed reactivity to kidney (15). No reactivity to any of the somatic tissues tested was observed using β-actin-positive and heat shock 70 protein 5-positive sera and control sera (data not shown). Scale bar represents = 40 μm for all panels. Length of the bar is 1 cm.
To study whether the proteins targeted by AOA were developmentally regulated, ovarian sections from 0, 10, 20 and 30-day-old rats were probed with target-specific sera PM45, PM80 and PM97, as shown in Figure 8. Positive reactivity was seen in the ooplasm of the oocyte from day 0 onwards. No staining was observed in ovarian sections probed with group III sera (Figure 8, panels d1–4) or PBS only (data not shown) which served as negative controls. The experiment was repeated with (n = 3) sera for each target.
Figure 8Immunohistochemical localization of autoimmune targets in rat ovary during ontogeny using patients’ sera targeting β-actin (a), heat shock 70 protein 5 (b) and α-actinin 4 (c) proteins in day 0 (1), day 10 (2), day 20 (3) and day 30 ovaries (4), as indicated with black arrowheads. Control sera showed no reactivity to any cell types at any stage of development (d). Scale bar represents = 140 μm for all panels. Length of the bar is 1cm. All the panels are of the same magnification and a representative bar has been inserted in the panel 1.
Although AOA have been reported to be one of the probable causative factors in cases of unexplained infertility, POF and IVF–embryo transfer failures, none of the autoantigens have been established as ideal serological markers for diagnosis of these infertile cases. Earlier data (
) clearly established the involvement of multiple antigenic targets in women with POF and women recruited under IVF–embryo transfer programmes. The present study was therefore carried out to identify cognate immunodominant antigens that could be used to screen sera for diagnosis and/or prognosis.
Autoantibody detection methods range from bioassays and enzyme immunoassays to proteomics. The majority of studies reported by other groups identified antigens in the range of 40–80 kDa (
), as well as the present data, clearly indicates that the immunoreactivity is not restricted but spans over a wide range of molecular weights, 30–220 kDa, suggesting the involvement of multiple autoantigenic targets in the pathogenesis of the disease. Of these multiple targets, four were immunodominant. Identification and characterization of one of these immunodominant antigens, 90-kDa protein, was previously reported (
). Using immunohistochemistry, the proteins were found to be present across various compartments of the ovary. The present study has also shown reactivity of a representative rat-ovary positive serum for each target identified in the human ovary indicating that rat ovarian tissue can be substituted for human ovarian tissue, which is more difficult to obtain. The present study has also reported the identity and characterization of the other three immunodominant antigens. Sera from patients recruited for IVF–embryo transfer as well as those with confirmed POF was used, unlike the recent report (
) where only sera of unexplained infertile cases were used. It was interesting to note that majority of the targets were common across the groups with minor differences in dominance of one target over another. Moreover, the present study did not concentrate on proteins in a restricted range because the sera from all the groups reacted with proteins in the range of 30–220 kDa.
Immunohistochemical results by several investigators have been reported as just positive or negative, without pictorial depiction of specific sites (
) as well as the present findings clearly indicate that, besides the zona pellucida, oocytes, corpus luteum, theca and granulosa cells are also under attack. It is well known that the zona pellucida has a critical role in species-specific spermatozoa–egg recognition and induction of the acrosome reaction and also protects the early embryo during its passage through the oviduct before implantation. However, close interaction between somatic and germ cells during ovarian folliculogenesis is also extremely vital for regulation of preantral growth of the oocyte and the nuclear events associated with ovulation (
). It is possible that antibodies to oocytes and granulosa cells may damage the bidirectional communication that is critical for proper folliculogenesis. Another point to be noted is, as the oocyte appears to be the major target of the autoimmune attack (
Oolemmal proteomics-identification of highly abundant heat shock proteins and molecular chaperones in the mature mouse egg and their localization on the plasma membrane.
) and therefore would not be suitable as a part of infertility evaluation workup to identify women at a risk of autoimmunity. A test lacking the oocyte targets would not enable the identification of women with probable follicular destruction or be suitable for oocyte retrieval before major damage had set in.
In the present study, sera from both the groups reacted significantly with recombinant αACTN4, HSPA5 and ACTB, confirming the potential use of these proteins as biomarkers in detection of AOA. The presence of all three proteins from day 0 post partum suggests that these proteins seem to be essential for normal ovarian function. With regard to tissue specificity, it was noted that αACTN4 shows reactivity with the kidney besides the ovary, which has been previously reported in human kidney (
ACTB, HSPA5 and αACTN4 have been reported to play an important role in fertilization and development in different species. Actins are highly conserved proteins that are involved in various types of cell motility and in maintenance of the cytoskeleton. A time-course depletion of multiple actin isoforms using RNAi in Caenorhabditis elegans have been shown to cause severe phenotypic responses in the early embryo, including defects in cortical ruffling, pseudo-cleavage, establishment of polarity, polar body extrusion, cytokinesis, chromosome segregation, reduced elongation of the central spindle, aberrant coalescence of telophase chromosomes, failure to displace the spindle towards the posterior and, eventually, egg production. These reports clearly indicate a vital role for actin in oogenesis, fertilization and a wide range of important events during early embryogenesis (
Another autoantigen is the non-muscle α-actinin isoform αACTN4. α-Actinins belong to the spectrin gene superfamily, which represents a diverse group of cytoskeletal proteins. The non-muscle isoforms αACTN1 and αACTN4 are involved in organizing actin filaments into bundles or networks (
) and they also connect the actin cytoskeleton to the membrane via a large number of cytoskeletal proteins and cell-surface receptors and play an important role in promoting cell adhesion and in regulating cell shape and motility (
). High expression of the αACTN4 transcript has been shown in normal human ovary and colon as compared with other tissues using Northern blot analysis (
). αACTN4 expression was shown on the fetal baboon oocyte surface, which is oestrogen regulated and proposed to be required in oocyte microvilli development (
The third target was HSPA5, also referred to as GRP78 (glucose regulatory protein 78 kDa) or immunoglobulin heavy-chain-binding protein (BiP). HSPA5 is among the constitutively expressed HSP70 genes, which accumulates in the lumen of endoplasmic reticulum, and is required for efficient protein processing and export through this organelle and has anti-apoptotic function. It is reported to be one of the nine highly abundant molecular chaperones present in the oolemma of mature mouse egg, as shown by a two-dimensional proteomics approach (
Oolemmal proteomics-identification of highly abundant heat shock proteins and molecular chaperones in the mature mouse egg and their localization on the plasma membrane.
). Presence of the GRP78 transcript in both the trophectoderm and the inner cell mass of E3.5 embryos indicates that GRP78 is essential for embryonic cell growth and pluripotent cell survival (
). It was also demonstrated that embryos devoid of GRP78 exhibit a substantial increase in apoptosis of the inner cell mass and in gene knockout studies: GRP78−/− embryos led to peri-implantation lethality and GRP78 null mice are able to hatch and implant but they quickly degenerate giving rise to empty deciduas.
Another interesting aspect to be noted is that the isoforms of the above-identified proteins have been identified as part of the HSP90 interactome in HEK293 cells by a co-immunoprecipitation approach (
), in a mechanism that might result in the generation of an autoimmune ovarian response in infertile women. It is proposed that autoimmune disease can develop in response to a single inciting antigen and then spread to involve other antigenic molecules of the same organ (
). Further studies are therefore necessary to elucidate how these molecules interact with each other and to ascertain the role of interacting proteins in the generation of an autoimmune response.
Although some of these autoantigens have been shown to be autoimmune targets in the pathogenesis of other autoimmune diseases like systemic lupus erythematosus, multiple sclerosis, rheumatoid arthritis and Sjögren’s syndrome (
), none of them have been reported in sera of POF and unexplained infertile women. Biological fluids of healthy individuals contain αACTN, HSPA5 and ACTB, which are required for various cellular and biochemical processes. Localization or presence of these three antigens on oocytes and at early embryonic stages indicates a significant role in fertilization and early embryonic development. In POF and unexplained infertile cases, it is likely that degradation or minute alterations in otherwise normal proteins leads to their recognition as non-self and thus induction of immune response against these proteins.
In conclusion, this study provides strong evidence to suggest that three proteins, non-muscle αACTN4, HSPA5 and cytoplasmic ACTB, should be targeted in idiopathic POF cases and women recruited under IVF–embryo transfer programmes. Identification of these proteins opens up an important area for future research to study the possible roles of these proteins in ovarian autoimmunity and thereby opening up a possibility for these biomarkers to be used in the diagnosis of infertility with autoimmune aetiology. Results from ELISA using the three recombinant proteins establish the validity of the test for diagnosis of human ovarian autoimmunity. Recombinant antigens represent complex structures containing diagnostically relevant epitopes along with cross-reactive ones, which sometimes compromise the specificity of the system. Therefore efforts are underway to find alternatives to a whole protein by using synthetic peptides of the immunodominant regions of the identified autoantigens. A specific non-invasive diagnostic test is particularly essential for a reliable diagnosis of an autoimmune aetiology and is essential to detect concomitant or future associated disorders, as well as to select the patients in whom immunomodulating therapy may restore, at least temporarily, ovarian function and fertility.
Acknowledgements
This work was supported by grants in part by Department of Biotechnology, Delhi, Government of India (Grant No. BT/IN/US/CRHR/14/VK/2007) and intramural funding from National Institute for Research in Reproductive Health, Delhi, Government of India. Financial support to Ms Purvi Mande and Ms Mamta Gurav is gratefully acknowledged from Department of Biotechnology, Government of India. The technical support of Mr Manish Ghosalkar, staff from central confocal facility and proteomics facility and Mr Hemant Karekar of National Institute for Research in Reproductive Health is gratefully acknowledged. The authors sincerely thank Mrs Shagufta Khan for critically reviewing the manuscript.
Oolemmal proteomics-identification of highly abundant heat shock proteins and molecular chaperones in the mature mouse egg and their localization on the plasma membrane.
Ovarian failure and autoimmunity. Detection of autoantibodies directed against both the unoccupied luteinizing hormone/human chorionic gonadotropin receptor and the hormone–receptor complex of bovine corpus luteum.
Ovarian antibodies as detected by indirect immunofluorescence are unreliable in the diagnosis of autoimmune premature ovarian failure: a controlled evaluation.
Declaration: The authors report no financial or commercial conflicts of interest.
Footnotes
Purvi Mande is a PhD student working in the field of reproductive immunology at the National Institute for Research in Reproductive Health (Indian Council of Medical Research) in Mumbai, India. She obtained her Masters degree in Biochemistry at Devi Ahilya University, Indore, India. Her research interests include female infertility and autoimmunity. Her current research focuses on the autoimmune aspect of premature ovarian failure and IVF–embryo transfer failures.
Purvi Mande is a PhD student working in the field of reproductive immunology at the National Institute for Research in Reproductive Health (Indian Council of Medical Research) in Mumbai, India. She obtained her Masters degree in Biochemistry at Devi Ahilya University, Indore, India. Her research interests include female infertility and autoimmunity. Her current research focuses on the autoimmune aspect of premature ovarian failure and IVF–embryo transfer failures.
00354-3&id=gr1.jpg){kind=link}
00354-3&id=gr2.jpg){kind=link}
00354-3&id=gr3.jpg){kind=link}
00354-3&id=gr4.jpg){kind=link}
00354-3&id=gr5.jpg){kind=link}
00354-3&id=gr6.jpg){kind=link}
00354-3&id=gr7.jpg){kind=link}
00354-3&id=gr8.jpg){kind=link}