{"paper_id":"f5610748-2c08-438f-943f-c2581c2f5840","body_text":"Premature ovarian insufficiency (POI) is characterized\nby hypergonadotropic amenorrhea due to\ncessation of ovarian function before the age of 40\nyears. The diagnosis is based on amenorrhea before\nthe age of 40 associated with follicle stimulating\nhormone (FSH) levels >40 IU/l, detected on\ntwo occasions at least one month apart ( 1 ). POI\ncauses female infertility, while is a significant psychosocial\nburden and a risk to women’s health. It\noccurs in 1% of women, of whom 10-28% have\nprimary and 4-18% secondary amenorrhea ( 2 ,  3 ).\nAlthough there are multiple etiologies of POI\n(genetic, chromosomal, infectious, and iatrogenic\ncauses), the etiology cannot be identified in most\npatients and this is referred to as idiopathic POI;\nup to 30% of idiopathic cases may have an autoimmune\ncause ( 4 ). The most convincing evidence\ncoming from the commonly observed association\nof POI with other autoimmune disorders ( 5 ,  6 ) are\ndemonstration of anti-ovarian antibodies (AOA,\n 7 ,  8 ) and histological findings of ovarian tissue\nfrom affected women. Roughly, one third of POI\npatients have AOA and/or antithyroid antibodies\nin their serum ( 1 ,  9 ). Various organ-specific and\nsystemic autoimmune diseases cause autoimmune\novarian insufficiency in up to 30% of women with\nPOI ( 4 ). According to the literature, 2-10% of POI\ncases are known to be associated with adrenal autoimmunity\n( 10 ). One of the first signs that autoimmunity\nmay be responsible for ovarian function\nfailure came from the observation that ovarian\nfailure may precede the onset of Addison’s disease\nby 8-14 years ( 11 ). Autoimmune Addison’s disease\nseldom develops in isolation, whereas several\nother endocrine glands and organs are generally\naffected, leading to an autoimmune polyglandular\nsyndrome (APGS,  12 ). Two main forms of APGS\ncan be clinically discerned, APGS types 1 and 2.\nAPGS type 1 is characterized by an association of\nmucocutaneous candidiasis, hypoparathyroidism\nand Addison’s disease. In about 60% of cases,\nthere is also an association with ovarian insufficiency.\nBlizzard et al. ( 13 ) and Irvine et al. ( 14 )\nfound that POI commonly presents with adrenal\ncytoplasmic antibodies, called steroid cell antibodies\n(SCA); they react with cytoplasmic antigens of\nother steroid-producing cells present in the ovary,\ntestis and placenta. Alteration of lymphocytes and\ntheir specific subsets, as well as T-cell mediated\ninjury are likely to play an important role in the\npathogenesis of autoimmune oophoritis. Surface\nmarkers of peripheral blood mononuclear cells\n(PBMC) have been shown to be deranged in early\nautoimmune phases and to be persisted through\nthe disease, even after targeted disruption ( 15 ).\nThe presence of pathogenic factors might accelerate\nthe process of apoptosis and atresia of ovarian\nfollicles during the fetal and post-natal period\n( 16 ). This interpretation is based on the dogma that\nthe number of ovarian follicles at birth is final and\nthat there is no possibility of regeneration or renewal\nof reserve follicles in adulthood ( 17 ). Experimental\nwork on animals suggests a possibility\nof renewal of the follicle reserve from proliferative\ngerminal ovarian cells even after birth; verification\nof which is being sought in studies on human ovaries\n( 18 ,  19 ). It has been shown that undifferentiated\novarian stem cells differentiate into structures\nsimilar to egg cells under certain laboratory conditions\n( 20 ). We faced two problems: whether to accept\nthe standard understanding of a final number\nof ovarian follicles or to focus on the hypothetic\npossibility of renewal of the follicular reserve. The\naim of this study was to evaluate the involvement\nof immune abnormality in patients with idiopathic\nPOI.\n\nOur study was a prospective randomized controlled\ntrial. The study group consisted of 20 women\nwith POI (mean age 31.8 years, range 20-39\nyears) and no use of medications or oral contraceptives\nfor at least 4 months prior to the study. The diagnosis\nwas based on the presence of amenorrhea\nbefore the age of 40, associated with two serum\nFSH levels above 40 IU/l at least one month apart.\nAll women with POI underwent karyotyping and\ngenetic testing of the fragile X mental retardation\n1 (FMR1). Women with infectious or iatrogenic\ncauses were excluded from the study. The control\ngroup consisted of 17 healthy women volunteers.\nInclusion criteria were a regular menstrual cycle,\nage between 18 and 39 years (mean 30.8 years),\nand no use of medications or oral contraceptives\nfor at least 4 months prior to the study. They also\nhad to be exempt from autoimmune disease or infertility\nproblems. All women provided a complete\npersonal and family history, with a stress on possible\nimmune-mediated, particularly autoimmune\nprocesses (allergy, asthma, diabetes, thyroiditis,\nrheumatoid arthritis, andatopiceczema), and all\nunderwent physical and vaginal ultrasound examinations.\nSerum levels of luteinizing hormone (LH), FSH,\nestradiol (E 2 ), prolactin (PRL), inhibin B, thyroid-\nstimulating hormone (TSH), anti-Müllerian\nhormone (AMH), antithyroglobulin (aTG) and\nantithyroid peroxidase antibodies (aTPO) were measured, and immunological investigations at\ncellular and humoral levels were performed. The\nadrenocorticotropic hormone (ACTH) stimulation\ntest was performed in the study group, only. The\nstandard Synacthen stimulation test is clinically\nwidely used as a sensitive screening method for\nsymptomatic adrenal insufficiency. Each ampoule\nof Synacthen contains 250 μg of the active ingredient,\ntetracosactrin (Novartis Pharmaceuticals,\nNorth Ryde NSW, Australia). Thirty minutes after\n250 μg Synacthen I.M. (Alliance Pharmaceutical\nWiltshire, UK), blood cortisol was measured by\nelectro-chemiluminescence immunoassay. A normal\ncortisol response to Synacthen was defined\nas a post-stimulation peak cortisol value of >500\nnmol/l at 30 minutes.\nSerum AMH in peripheral blood was determined\nby a Personal Lab analyser using the\nenzyme linked immunosorbent assay (ELISA)\nmethod with Beckman Coulter reagent. The\nnormal range of AMH levels is 0.7-3.5 mg/l.\nValues below 0.3 mg/l indicate a reduced ovarian\nreserve. Hormones were determined on a LIAISON\nanalyser by quantitative direct competitive\nchemiluminescence immunoassays (CLIA).\nEach test is a modified two-step process, in\nwhich the specific antibodies of a certain hormone\nbind to magnetic cells. Normal ranges for\nthe follicular phase of the cycle are as follows:\nFSH: 3.5-9.2 IU/l; LH: 1.1-11.6 IU/l; LH around\novulation: 17-77 IU/l; PRL: 6.2-23.5 μg/l; E 2 :\non day 3 up to 310 pmol/L (0.31 nmol/L); E 2 \npostmenopause: 0-110 pmol/L (0-0.11 nmol/l);\nand TSH: 0.3-3.6 mE/l. Serum Inhibin-B cut-off\nlevel on day 3 was 45 pg/mL. Serum aTG and\naTPO concentrations measured by immunoassay\nusing direct chemiluminometric technology\non an ADVIA CENTAUR analyser (Siemens\nMedical Solutions Diagnostics, Tarrytown,\nUSA). The normal range for serum aTG and\naTPO is <60 KE/l. Detection of AOA was by\nindirect immunofluorescence (IIF) on cryosections\nof normal human ovarian tissue.\nNon-human primate ovaries for the detection\nof AOA are not commercially available, while\nnormal human ovarian tissue is available at\nthe Institute of Pathology, Ljubljana, Slovenia,\nwhen ovaries are removed due to various pathological\nprocesses, particularly tumors, and are\nsent for pathohistological examination. Ovarian\ntissue was incubated with patient serum diluted\n1:10. The second incubation was with fluorescein\nisothiocyanate-labelled anti-human IgG antibody\n(Dakopatts, Copenhagen, Denmark). Positive reactions\nwere semi-quantitatively evaluated (on a\nscale of 1– 4+). Negative control omitting the patient’s\nserum was regularly included ( 21 ). At the\ncellular level fresh PBMC were studied by flow\ncytometry (Becton Dickinson FACS, NJ, USA),\nand percentages of the following blood lymphocyte\npopulations were determined: T cells (CD3+),\nhelper T cells (CD4+), cytotoxic T cells (CD8+),\nnatural killer cells (CD56+CD16+), regulatory T\ncells (CD25 +high ) and B cells (CD19+) ( 22 ). Immunofluorescence\nlabelling was performed by incubating\nPBMCs with monoclonal antibodies to\nCD3, CD4, CD8, CD25, CD56/16 and CD45. A\ndifferential blood count with a standard laboratory\nprocedure was taken to obtain concentrations of\nindividual lymphocyte subtypes.\nNormal data distribution was tested with the\nKolmogorov-Smirnov test. Where variables were\nnormally distributed, we used the Pearson’s chisquare\ntest. If variables were not normally distributed,\nwe used a nonparametric Mann-Whitney\ntest. Statistical analysis was done using Statistical\nPackage for the Social Sciences, version 18 (SPSS\nInc., Chicago, IL, USA). The results were considered\nstatistically significant at p<0.05.\nThe study protocol was approved by the National\nEthics Committee, and all patients gave written\ninformed consent.\n\nThe subjects were comparable with controls by\nage. One patient presenting a mosaic 45X0/46XX\nwas excluded from the final analysis. Other patients\nhad a 46XX karyotype. Anamnestic data of\npatients showed that four had had mumps during\ntheir childhood; they were thus excluded from the\nfinal analysis. Table 1 shows the clinical and endocrine\ncharacteristics of patients and of healthy\nwomen. All endocrine parameters in controls were\nwithin the normal range ( Table 1 ).\nHormone and ovarian peptide levels in patients and in healthy controls\nPOI; Premature ovarian insufficiency, FSH; Follicle stimulating hormone, LH; Luteinizing hormone, E 2 ; Estradiol, PRL; Prolactine and AMH; Anti-Müllerian hormone.\nWe collected targeted history information on\npersonal and familial autoimmune disorders ( Table\n2 ). Sixteen patients (80%) had an associated\nautoimmune disease in their personal and/or familial\nhistory. Four women (20%) with POI had\nfirst grade relatives with ovarian insufficiency\nbefore the age of 40. Thyroid disorders were the\nmost common (15%) of the autoimmune diseases\nassociated with POI in personal histories. Before\nentering the study, three patients had been treated\nfor autoimmune thyroid dysfunction (Hashimoto\nthyroiditis). In the study group, 55% of women\nhad autoimmune disease in the family history; the\nmost frequent autoimmune disorder was diabetes\ntype I. The overall personal and familial incidence\nof autoimmune diseases was lower in the control\ngroup. In the study group, 50% of women had\naTG. One healthy woman had evidence of autoimmune\nthyroid dysfunction, manifested by an elevated\naTG serum level, and was excluded from\nthe final analysis (Tables  2 , 3 ).\nPersonal and family history on autoimmune diseases in patients and healthy controls\nPOI ; Premature ovarian insufficiency.\nAssociated autoimmune abnormalities (some participants showed more than one condition)\nPOI; premature ovarian insufficiency.\nAnalysis of AOA in serum was performed for all\npatients ( Fig 1 ). AOA were detected in 4 (20%)\npatients, and in 3 patients, we found an associated\nautoimmune disease; Hashimoto thyroiditis. There\nwas a clear positive immune fluorescent reaction of\nAOA in the serum of one woman who had mumps\nduring her childhood, and she was excluded from\nthe analysis. AOA were not detected in any of the\ncontrols ( Table 3 ).\nPrevalence of serum anti-ovarian antibodies (AOA)\nin patients.\nCellular autoimmune reaction occurring due to\na changed T cell function was analysed as a potential\ncause of POI. Cell abnormalities were more\nfrequent in women with POI than in healthy women\n( Table 4 ).\nPeripheral T cell count is expressed as a percentage\nof various cell surface markers. In patients\nwith POI, peripheral regulatory T lymphocytes\n(CD25 +high , p=0.015) were low and peripheral\nblood B cells (CD19+) were high (p=0.014); T\nlymphocyte parameters were normal in the control\ngroup.\nPrevalence (in %) of analysed peripheral blood lymphocyte samples for various cell surface markers in patients and in controls\nMarkers for peripheral blood lymphocytes: T lymphocytes (CD3+), helper T lymphocytes (CD4+), cytotoxic T lymphocytes (CD8+), natural killer cells (CD56+CD16+), regulatory T lymphocytes (CD25 +high ) and B cells (CD19+).\n\nTo present POI as a possible autoimmune abnormality,\nwe focused on three potentially interconnected\nfactors: personal and familial history\nof autoimmune disorders, peripheral blood T-lymphocytes\nlevels and presence of AOA.\nWe performed a targeted history of personal and\nfamilial autoimmune disorders and found associated\nautoimmune disorders in our patients. Thyroid\ndisorders were common in personal histories and\ndiabetes type 1 in familial histories, confirming the\nfindings of previous studies ( 5 ,  6 ,  15 ). One of the\nreasons for suspecting an autoimmune etiology\nof POI is its frequent association with nearly all\norgan-specific autoimmune diseases ( 1 ,  3 ). Autoimmune\ndiseases are significantly more frequent in\nyoung women than in men. This phenomenon may\nbe explained by the effect of sex steroids on the\ncomponents of the cellular immune system, which\nmight contribute to the development and progression\nof autoimmune POI ( 23 ).\nHoek et al. ( 24 ) found that 60% of patients with\nsecondary amenorrhea and Addison’s disease have\na detectable SCA serum titre. These antibodies are\nshown by 60-80% of patients with APGS type I. In\n25% of our patients, the Synacthen test revealed\nan abnormal cortisol response. We interpreted this\nas the possibility that patients could be positive for\nSCA, but had a normal cortisol response. Patients\nwith this disorder have an insidious onset; to confirm\nsubclinical autoimmune adrenal insufficiency,\nmeasurement of adrenal antibodies might be a\nmore effective screening method ( 25 ,  26 ).\nThyroid disorders and the presence of antithyroid\nantibodies are often mentioned in association\nwith POI. Thyroid disorders are the most common\nof the autoimmune diseases associated with\nPOI, found in 12-39% of women with POI ( 27 -\n 29 ). Thyroid disorders are often associated with\nendometriosis and polycystic ovary syndrome,\ntwo conditions often resulting in infertility ( 2 ). We\nfound a greater involvement of aTG and aTPO in\nthe study than in the control group, and the explanation\nbeing that ovarian failure may have been\npresent in the latent period of thyroid disease. We\nfound a strong correlation between autoimmune\nthyroid disease and autoimmune POI; 50% of patients\ntested positive for aTG. We concluded that\nnot all patients with positive aTG necessarily have\na clinically expressed thyroid disorder and the\ndisease can have an insidious onset. Furthermore,\ngreater involvement of other immune-mediated\ndiseases, particularly auto immune disorders, such\nas allergy, psoriasis, atopic dermatitis and vitiligo\nin the study group suggests an autoimmune cause\nof POI, consistent also with the findings in the literature\n( 9 ,  15 ,  30 ,  31 ).\nGenetic predisposition is known to be one of the\nprimary causes of autoimmunity and it is generally\nobserved that patients with autoimmune diseases\nhave several types of antibodies, as also confirmed\nin our study. We found AOA in 20% of patients;\nthese results are consistent with other studies ( 9 ,\n 21 ,  32 - 35 ). The prevalence of AOA in women with\nPOI, studied with IF on ovary tissue, varies greatly\nfrom 2 to 70%, which supports their role as a\nmarker of an immune dysfunction process against\novaries. There could be several reasons for the differences\namong study results, including the diverse\norigin of tissue sections, which include human and\nnon-human primate ovaries, as well as different\nstudy inclusion criteria. Many studies have shown\nthat the presence of serum AOA does not correlate\nwith the clinical manifestation of POI. Despite\nthese antibodies being present, their pathogenetic\nrole is highly questionable ( 36 ,  37 ). AOA may occur\nseveral years before the occurrence of clinical\nsymptoms, as detected in 33-61% of women\nwith unexplained infertility; a situation that may\nindicate early stages of autoimmune ovarian insufficiency\n( 38 ,  39 ). For most autoimmune diseases,\nscreening for specific antibodies is probably the\nbest way of evaluating immunological involvement.\nMany ovarian structures are potential targets\nfor autoimmune events with POI. In addition\nto AOA, SCA, aTG and aTPO, antibodies to several\npotential ovarian antigens have been proposed\nas markers of ovarian autoimmunity, which could\npotentially mediate autoimmune damage in POI.\nBetterle et al. ( 40 ) found antibodies against steroid\ngenetic enzymes in some cases with anti-adrenal\nautoimmunity. Antibodies against gonadotropin\nreceptors and gonadotropins have also been found\nin patients with POI, but they are still the subject\nof research. The general opinion is that more research\nis needed ( 41 ,  42 ). Although some studies\nhave managed to identify autoantibodies against\nzona pellucida, corpus luteum and ovarian cells,\nthese findings have no real correlation with the\nclinical picture ( 43 - 45 ). In our preliminary study, AOA were determined semi-quantitatively by the\nIIF method. IIF is a basic method and widely used\nfor determining auto antibodies, while more specific\ntests enabling detection of AOA for specific\nantigen targets are not available in our laboratory.\nInfertile women with AOA also have a decreased\nresponse to gonadotropin stimulation and reduced\npregnancy rate after treatment ( 46 ,  47 ). In 2002,\nit was found that low-responders to gonadotropins\nwith AOA are younger than low-responders without\nAOA ( 48 ). Detection of autoimmune processes\nthat affect the ovarian response should thus be included\nin the diagnostic workup before any infertility\ntreatment, particularly in women with low or\nno response to gonadotropins ( 34 ). Determination\nof AOA, as a specific test, is important in the diagnosis\nof diseases with an autoimmune etiology, but\nit should not be the only reliable diagnostic tool\nfor optimal selection of patients who may benefit\nfrom immune modulatory therapy that could, at\nleast temporarily, re-establish their ovarian function\nand fertility.\nAbnormalities of cellular immunity of T lymphocytes,\nmacrophages and dendritic cells play\nan important role in autoimmune events. Some of\nthese abnormalities have been seen in women with\nPOI, confirming the potential existence of an autoimmune\nmechanism of the disease. Mignot et al.\n( 49 ) found that the absolute number and percentage\nof peripheral blood T lymphocytes, especially\nCD4+ T cells, are increased in patients with POI.\nMoreover, Miyake et al. ( 50 ) found that patients\nwith POI have low levels of CD8+/CD57+ Tcells\n(cytotoxic T lymphocytes) and an increased ratio\nof CD4+ to CD8+ cells, which may reflect cell cytotoxic\ncell migration from blood to inflamed tissue.\nMultiple animal studies have suggested that the basis\nof POI is a cell-mediated autoimmune reaction\ncaused by an alteration in T cell regulation ( 18 ).\nCD4+ T cells with constantly expressed α chain\n(CD25) receptor for IL-2 of were the first detected\nmediators of inhibition of autoimmune diseases in\nmice alteration of suppressor T cell subsets and T\ncell abnormalities are likely to play an important\nrole in the pathogenesis of autoimmune diseases\n( 51 ). Regulatory CD4+25+ T cells show a potent\nimmunosuppressive function in vitro and in vivo,\nand contribute to immunologic self-tolerance by\nsuppressing potentially auto-reactive CD4+ T cells.\nThere is known to be a number of immunological\nmechanisms associated with the failure of immune\ntolerance and the development of autoimmunity.\nAlthough studying regulatory T cells in human\nautoimmune diseases is difficult, and at times,\nfindings have been contradictory, the data suggest\nthat defects in CD4+ CD25+ regulatory T cells\nmediated suppression ( 52 ) are a major subset of\nimmune cells responsible for peripheral immune\nself-tolerance.\nIn agreement with studies in patients with systemic\nlupus erythematosus, multiple sclerosis,\nrheumatoid arthritis and autoimmune vasculitis,\nwe confirmed a reduced number of CD4+CD25 +high \nT cells in the peripheral blood of our patients.\nHigh expressions of CD25 and CD4 surface markers\nhave classically been used for identification of\nregulatory T cells. This may be problematic since\nCD25 is also expressed on antigen-responding\nactivated non-regulatory T cells. The additional\nmeasurement of cellular expression of Foxp3 protein\nallowed a more specific analysis of Treg cells\n(CD4+CD25+Foxp3+ cells). However, Foxp3 is also\ntransiently expressed in activated human effector\nT cells, thus complicating a correct Treg analysis.\nThe large majority of Foxp3-expressing regulatory\nT cells express high levels of the interleukin-2 receptor\nalpha chain (CD25). Since there are no cell\nsurface markers that are uniquely and specifically\nexpressed on all Foxp3-expressing regulatory T\ncells, the measurement of CD4+CD25 +high  T cells\nis still in use in clinical studies of peripheral blood\nlymphocytes.\nWe interpreted the reduced number of\nCD4+CD25 +high  T cells as a possible mechanism\ncontributing to the formation of an autoimmune\nresponse in association with the presence of AOA\nand aTG. We also found an increased number of\nB cells in peripheral blood. A similar picture has\nbeen observed with other autoimmune endocrinopathies;\ntherefore, we interpreted the elevated\nB cell count as activation of the humoral immune\nsystem, crucial for autoantibody production. Some\nauthors, though, have tried estrogen substitution in\nwomen with POI without any effect on peripheral\nB cell count ( 53 ,  54 ).\nThe hormones inhibin B, FSH and AMH have\nbeen proposed as potential markers for determining\nthe functional ovarian reserve ( 55 ,  56 ). In\nyoung ovulatory women, measurement of AMH at\n3-year intervals has shown that the AMH serum level decreases significantly over time, whereas\nother markers associated with ovarian aging, such\nas FSH, inhibin B and antral follicle count (AFC),\ndo not change during this time period. Since it decreases\nat a time when concentrations of FSH and\ninhibin B are still normal, AMH has been proposed\nas the best indicator of ovarian reserve and as a\nmarker of ovarian aging ( 55 - 59 ). In contrast, AMH\nis almost undetectable in women with POI ( 60 ),\nwhich our study also confirmed. In our group of\npatients, there was a small AFC or these structures\nwere no longer seen on ultrasound, which agrees\nwith the data in the literature ( 61 ).\n\nClinical and biological characteristics of women\nwithout known causes of disease suggest a possibility\nof autoimmune pathogenesis. In some patients,\na combination of various autoimmune processes\nhas been found. The presence of AOA and\nanti-thyroid antibodies, together with abnormalities\nof cellular immunity, potentially represent an\nautoimmune mechanism of POI. There is thus an\nincreasing need to find a reliable and simple diagnostic\nprocedure to determine the true prevalence\nof autoimmune ovarian disease. In women with\nPOI, more attention should be paid to evaluation\nof associated autoimmune disorders.","source_license":"CC-BY-4.0","license_restricted":false}