Prenatal androgen exposure and transgenerational susceptibility to polycystic ovary syndrome

How obesity and elevated androgen levels in women with polycystic ovary syndrome (PCOS) affect their offspring is unclear. In a Swedish nationwide register-based cohort and a clinical case–control study from Chile, we found that daughters of mothers with PCOS were more likely to be diagnosed with PCOS. Furthermore, female mice (F0) with PCOS-like traits induced by late-gestation injection of dihydrotestosterone, with and without obesity, produced female F1–F3 offspring with PCOS-like reproductive and metabolic phenotypes. Sequencing of single metaphase II oocytes from F1–F3 offspring revealed common and unique altered gene expression across all generations. Notably, four genes were also differentially expressed in serum samples from daughters in the case–control study and unrelated women with PCOS. Our findings provide evidence of transgenerational effects in female offspring of mothers with PCOS and identify possible candidate genes for the prediction of a PCOS phenotype in future generations. Prenatal androgen exposure causes transgenerational increases in offspring susceptibility to polycystic ovary syndrome in adulthood.

Prenatal androgen exposure in rodents leads to reproductive and metabolic dysfunction in first-generation (F 1 ) female offspring 10,11 , but whether elevated maternal androgen levels potentiate transgenerational susceptibility to PCOS and related diseases in adulthood is not known. To be considered transgenerational transmission, the inherited traits should be displayed in the third generation (F 3 ), because F 1 fetuses and the germ cells of the second generation (F 2 ) are directly exposed to the maternal intrauterine milieu. To date, only endocrine disruptors have been shown to result in the transgenerational transmission of reproductive and metabolic dysfunction in adult life [12][13][14][15] .
In this study, we showed that daughters of mothers with PCOS had a fivefold-increased risk of being diagnosed with PCOS. Furthermore, we showed that PCOS-like traits were passed on to F 3 female mice, and we identified several genes with altered expression in mouse oocytes and in the serum of daughters with PCOS that might serve as biomarkers for the transgenerational effect of PCOS 16,17 .

NaturE MEDICINE
with PCOS were more prone to developing neuropsychiatric disorders 18 and to having longer anogenital distance 7 , which strongly indicate prenatal androgen exposure, we conducted a Swedish nationwide register-based cohort study to assess whether daughters of women with PCOS are more often diagnosed with PCOS ( Fig. 1a and Supplementary Table 1). We also followed 21 daughters of women with PCOS from birth to adulthood and compared them to 14 daughters of women without PCOS of similar age from a casecontrol study in Chile (Supplementary Table 2). Using the Swedish Multi-Generational Register and the National Patient Register, a total of 29,736 daughters were identified, of whom 2,275 (7.7%) were born to mothers diagnosed with PCOS. Overall, 3.4% (n = 78) had a diagnosis of PCOS compared to only 0.6% (n = 159) of daughters born to mothers without a PCOS diagnosis. Daughters born to women with PCOS had a fivefold-increased risk of a PCOS diagnosis compared to daughters born to women without PCOS (crude hazard ratio (HR) = 5.39, 95% confidence interval (CI), 3.38-8.60; adjusted model HR = 5.12, 95% CI, 3.17-8.26; Fig. 1a). In the longitudinal study of the 21 daughters of women with PCOS, 15 had hyperandrogenism (Ferriman-Gallwey score > 6) (Fig. 1b), 12 had irregular cycles (Fig. 1c) and 11 had polycystic ovarian morphology (Fig. 1d), compared to 2, 0 and 2 of the daughters in the control group, respectively. Notably, 15 daughters (71%) of women with PCOS met the Rotterdam criteria for PCOS (Fig. 1e), with 6 classified as phenotype A, 5 classified as phenotype B and 4 classified as phenotype C (Supplementary Table 2). One daughter of a mother without PCOS was classified as having phenotype C. Daughters of women with PCOS also had higher body mass index (BMI), larger waist circumference and higher diastolic blood pressure, demonstrating metabolic dysfunction (Supplementary Table 2). Daughters of women with PCOS had higher Ferriman-Gallwey scores and a higher free androgen index, which are markers of clinical and biochemical hyperandrogenism, and higher anti-Müllerian hormone (AMH) levels, which indicate increased ovarian follicular mass.
Prenatal androgen exposure but not obesity causes transgenerational reproductive and metabolic dysfunction. To test whether female offspring (F 1 ) of prenatally androgen-exposed mothers (F 0 ) are susceptible to developing PCOS-like reproductive and metabolic traits in adulthood and whether these phenotypes are transmitted to F 2 (intergenerational) and F 3 (transgenerational) offspring, we fed F 0 female mice with a control diet or a high-fat, high-sucrose diet for 6 weeks and measured body weight development, body composition and glucose homeostasis before mating. They were then mated with male mice fed with the control diet. Pregnant F 0 dams were injected subcutaneously with dihydrotestosterone dissolved in a mixture of sesame oil and benzyl benzoate or vehicle alone from embryonic day (E) 16.5 to E18. 5 (refs. 16,19,20 ). Thus, four experimental lineages were studied: control diet + vehicle (control); control diet + dihydrotestosterone (androgenized); high-fat, high-sucrose diet + vehicle (obese); and high-fat, high-sucrose diet + dihydrotestosterone (obese and androgenized) (Fig. 2a).
F 0 dams fed the high-fat, high-sucrose diet were obese. They gained more weight but had similar food and calorie intake as dams fed the control diet (Extended Data Fig. 1a-c). They had more fat and less lean mass (Extended Data Fig. 1d,e) but showed no difference in weight gain during pregnancy. Obese F 0 dams also displayed impaired glucose homeostasis as demonstrated by the larger area under the curve during an oral glucose tolerance test (OGTT) compared to F 0 dams that were fed the control diet before mating (Extended Data Fig. 1f,g), but there was no difference in insulin levels (Extended Data Fig. 1h). F 1 dams in the androgenized lineage and F 1 and F 2 dams in the obese lineage had smaller litters, whereas F 1 dams in the obese and androgenized lineage delivered only one female F 2 pup (Extended Data Fig. 2a). This F 2 female delivered two male offspring that died at weaning. Phenotypic data of this F 2 female are presented together with other groups when possible. To determine why there were so few F 2 offspring in the obese and androgenized lineage, the experiment was repeated, and F 2 litter size and fetal growth were evaluated by measuring the crown-rump length at E12.5 and E18.5. Although differences were not significant, there were fewer and smaller live F 2 embryos at E12.5 in the obese and androgenized lineage than in the other groups ( Fig. 2b and Extended Data Fig. 2b). Moreover, the E12.5 embryos in the androgenized lineage were significantly smaller than the E12.5 embryos in the control lineage (Fig. 2c). At E18.5, F 1 females in the obese and androgenized lineage had miscarriages and no live F 2 fetuses were obtained (Fig. 2d), and live fetuses in the androgenized lineage remained smaller than those in the control lineage (Fig. 2e). For the crownrump length, there was a main effect in the androgenized lineage in F 2 embryos at E12.5 (two-way ANOVA: F 1,64 = 9.81, P = 0.0026) and E18.5 (one-way ANOVA: F 2,58 = 11.84, P < 0.0001). Collectively, these results demonstrate that dihydrotestosterone exposure in F 0 mothers, independently of diet, results in compromised fetal development and that obesity and androgenization in F 0 mothers leads to the death of F 2 offspring at E12.5-E18.5. To determine why this was, we first analyzed the gene expression of the germ cell markers Mvh and Dppa3 in the gonads of F 2 fetuses at E12.5 and E18.5 but found no significant differences (Extended Data Fig. 2c,d). We then analyzed the expression of genes known to affect placental function at E12. 5 (refs. 21-23 ). Expression of transcription factor AP-2, gamma (Tfap2c) was decreased in the obese and androgenized lineage, and expression of thymoma viral proto-oncogene 1 (Akt1) and achaete-scute family bHLH transcription factor 2 (Ascl2) was decreased in the androgenized lineage (Fig. 2f). There was an interaction between the androgenized lineage and the obesity lineage on Tfap2c mRNA expression (F 1,20 = 11.74, P = 0.002) and a main effect in the androgenized lineage on Akt1 (F 1,19 = 15.59, P = 0.0009) and Ascl2 (F 1,20 = 5.048, P = 0.036) mRNA expression.
At weaning, F 1 , F 2 and F 3 female offspring in the androgenized lineage exhibited longer anogenital distance than control offspring (Fig. 2g), indicating a transgenerational effect due to prenatal androgen exposure. There was an interaction between the androgenized lineage and the obesity lineage on anogenital distance in F 1 offspring (two-way ANOVA: F 1,121 = 6.38, P = 0.013) and a main effect in the androgenized lineage in F 2 offspring (oneway ANOVA: F 2,49 = 14.19, P < 0.0001) and F 3 offspring (one-way ANOVA: F 2,25 = 4.55, P = 0.021). Independently of diet, F 1 and F 2 female offspring in the androgenized lineage displayed disrupted estrus cycles compared to the control lineage (Extended Data Fig. 3a,b). Although F3 offspring in the androgenized lineage tended to have impaired estrus cycles, the difference was not significant compared to controls (Extended Data Fig. 3a,b).
Independently of diet, circulating testosterone and androstenedione levels were lower in F 1 female offspring in the dihydrotestosterone-exposed lineages, with no difference in dihydrotestosterone levels or any circulating androgen levels in F 2 and F 3 offspring (Extended Data Fig. 4a-c). This suggests that transgenerational effects were caused by transient prenatal androgen exposure instead of abnormal circulating steroid levels per se in F 1 -F 3 female offspring. Although daughters of women with PCOS had elevated serum AMH (Supplementary Table 2), no changes were seen in F 1 -F 3 female offspring in mice (Extended Data Fig. 4d). F 1 female offspring in the androgenized lineage had similar body weight (Extended Data Fig. 5a,b) but more fat mass (Fig. 3a) than controls. There was no such difference in F 2 offspring, whereas F 3 female offspring in the androgenized lineage weighed more (Extended Data Fig. 5a-c) and had more fat mass (Fig. 3a). For fat mass, there was an interaction between the androgenized lineage and the diet-induced obesity lineage in F 1 offspring (two-way ANOVA: F 1,78 = 10.68, P = 0.002) and a main effect in the Articles NaturE MEDICINE obesity lineage in F 1 offspring (F 1,78 = 4.29, P = 0.041). In F 3 offspring, there was a main effect in the androgenized lineage (one-way ANOVA: F 2,20 = 5.92; P = 0.01). In the obese and androgenized lineage, F 1 female offspring weighed more than offspring in the obese lineage (Extended Data Fig. 5c). These observations were further supported by the increased adipocyte size in F 1 -F 3 female offspring (Fig. 3b,c) and altered expression of genes affecting adipogenesis in F 1 and F 3 offspring (Extended Data Fig. 5d). Furthermore, in the androgenized lineage and obese lineage, we found accumulation of neutral lipids and triglycerides in the liver of F 1 and F 3 female offspring (Fig. 3d,e). Hepatic expression of some genes involved in lipid biosynthesis was also affected in F 1 and F 2 offspring (Extended Data Fig. 5e). For liver triglycerides, there was an interaction between the androgenized and the obesity lineage in F 1 offspring (two-way ANOVA: F 1,36 = 12.81, P = 0.001). In F 3 offspring, there was a main effect in the androgenized lineage (one-way ANOVA: F 2,47 = 8.023; P = 0.001). Despite these differences, there was no change in lean mass or glucose metabolism (Extended Data Figs. 5c and 6a-c) in F 1 -F 3 offspring. To further understand the metabolic effects, we used indirect calorimetry for metabolic phenotyping.   . c,e, Embryos (left) and embryo crown-rump length (right) at E12.5 (c) and E18.5 (e). CD, control diet; HFHS, high-fat, high-sucrose diet; DHT, dihydrotestosterone. f, mRNA levels of Tfap2c (n = 6 placentas per group), Ascl2 (n = 5 placentas in control diet + vehicle; n = 6 placentas for the other groups) and Akt1 (n = 5 placentas in HFHS + DHT; n = 6 placentas for the other groups). g, Prenatal androgen exposure causes a transgenerational increase in anogenital distance in F 1 -F 3 female offspring (n is the number of offspring). All data are presented as mean ± s.e.m. Numbers of mice are stated in the bars of each group. F 1 : two-way ANOVA, Tukey's post hoc analysis; F 2 and F 3 : one-way ANOVA, Dunnett's post hoc analysis. Mat, maternal; GMat, grand-maternal; GGMat, great-grand-maternal. Articles NaturE MEDICINE F 1 and F 3 female offspring in the androgenized and obese lineages had lower respiration exchange ratios (RERs; VCO 2 /VO 2 ) and reduced energy expenditure (EE) during the dark phase than offspring in the control lineage (Fig. 4a,c). Food intake and total activity were similar among the groups (Fig. 4a,c). These results demonstrated a shift from carbohydrate to fatty acid use and coincided with the increased adiposity in F 1 and F 3 female offspring in the androgenized and obese lineages (Fig. 4a,c). In F 1 , there was an interaction between the androgenized and the obesity lineages in RER during the night (two-way ANOVA: F 1,764 = 14.82, P = 0.0001) and a main effect in the obesity (F 1,764 = 50.75, P < 0.0001) and androgenized (F 1,764 = 22.01, P < 0.0001) lineages. There was an interaction between the androgenized lineage and the obesity lineage in EE during the day (two-way ANOVA: F 1,764 = 9.168, P = 0.0025) and night (F 1,764 = 24.48, P < 0.0001) and a main effect in the obesity (F 1,764 = 41.7, P < 0.0001) and androgenized (F 1,764 = 21.6, P < 0.0001) lineages. During the night, there was a main effect in the obesity (F 1,764 = 75.63, P < 0.0001) and androgenized (F 1,780 = 40.7, P < 0.0001) lineages. In F 3 , in the obesity lineage, RER was higher during the day (one-way ANOVA: F 2,573 = 25.96, P < 0.0001], and during the night it was lower in the obesity and androgenized lineages (F 2, 585 = 7.454, P = 0.0006). EE was lower during the day (F 2, 573 = 17.87, P < 0.0001) and night (F 2,573 = 12.58, P < 0.0001) in the obesity and androgenized lineages. F 2 female offspring in the androgenized lineage had increased RER and EE compared to the control lineage but no change in food intake or total activity in the dark phase (Fig. 4b). During the day, RER was higher in the obesity lineage (one-way ANOVA: F 2,573 = 11.69, P < 0.0001), and during the night it was higher in the androgenized lineage (F 2,573 = 8.125, P < 0.0003). EE was higher during the day (F 2,573 = 9.352, P = 0.0001) and night (F 2,573 = 11.81, P < 0.0001) in the androgenized lineage. These findings suggest that PCOS-like reproductive and metabolic dysfunctions were transmitted across generations in female offspring from androgenized F 0 mothers, whereas female offspring from obese F 0 mothers showed substantial metabolic dysfunction only in F 1 offspring, which was less pronounced in F 2 and F 3 offspring.
Prenatal androgen exposure and obesity cause transgenerational alteration of oocyte mitochondria. In F 1 -F 3 offspring, transmission electron microscopy analyses of metaphase II (MII) oocytes in the control lineage showed small rounded mitochondria with cristae distributed evenly in the cytoplasm (Extended Data Fig. 7a,e,i). In contrast, F 1 -F 3 oocytes in both the androgenized and obese lineages, as well as F 1 offspring in the obese and androgenized lineage, showed swollen mitochondria with fewer developed cristae, resembling vacuole-like structures (Extended Data Fig. 7a,e,i). These findings suggest that mitochondrial morphology was affected by dihydrotestosterone exposure and diet. There were also more mitochondria in the F 1 offspring of the obese lineage and the F 3 offspring of the androgenized lineage (Extended Data Fig. 7b,j), increased mitochondrial DNA (mtDNA) in the F 1 offspring of the androgenized lineage (Extended Data Fig. 7c) and more lipid droplets in the F 1 offspring of the obese and androgenized lineage (Extended Data Fig. 7d).
Prenatal androgen exposure and obesity resulted in altered transcriptomic profiles of MII oocytes across generations. To address 'fetal reprogramming' via the maternal germ line, we performed offspring. RER and EE were measured by indirect calorimetry by using the TSE system in F 1 -F 3 adult female offspring. Gray areas indicate dark (night) periods (18:00 to 6:00). Bar graphs represent RER, EE, food intake and total activity average values during the day and night (n = 4 animals per group). F 1 : two-way ANOVA, Tukey's post hoc analysis; F 2 and F 3 : one-way ANOVA, Dunnett's post hoc analysis. All data are presented as mean ± s.e.m.

NaturE MEDICINE
single-cell RNA sequencing (scRNA-seq) of 202 MII oocytes from all F 1 , F 2 and F 3 female offspring, except for the obese and androgenized lineage (Fig. 5a). After quality control of the scRNA-seq data, 156 MII oocytes with an average of 7,488 expressed genes (reads per kilobase of transcript per million mapped reads (RPKM) >1) were used for downstream analysis (Supplementary Table 3). Oocytes were collected from each generation of females after they had been phenotyped and thus showed a batch effect-that is, clustering by generation (Extended Data Fig. 8a). To determine the biological effects of dihydrotestosterone exposure and diet on all F 1 -F 3 MII oocytes collected and sequenced at different times, we first performed batch correction using Seurat v2.0 (ref. 24 ) to remove technical variances across generations (Extended Data Fig. 8b). We confirmed that all MII oocytes expressed similar levels of oocyte-specific genes, for example, developmental pluripotency associated 3 (Dppa3), mos -proto-oncogene serine/threonine-protein kinase (Mos), zona pellucida glycoprotein 3 (Zp3), KIT proto-oncogene receptor tyrosine kinase (Kit), NLR family, pyrin domain containing 3 (Nlrp5), growth differentiation factor 9 (Gdf9), bone morphogenetic protein 15 (Bmp15) and DEAD (Asp-Glu-Ala-Asp) box polypeptide 4 (Ddx4), independently of their lineages across the three generations (Fig. 5b).
To understand how dihydrotestosterone exposure and diet affect oocytes across generations, we performed differential gene expression analysis using DESeq2 and identified 410 and 231 differentially expressed genes (DEGs) in the androgenized and obese lineages compared to the control lineage, respectively, with 44 overlapping  Table 4C). Oocytes from each generation (F 1 -F 3 ) were better clustered according to their lineage as shown in the principal-component analysis plot (Extended Data Fig. 8c), which enabled us to further explore their potential role in the transgenerational transmission of PCOS-like phenotypes.
Next, we created heat maps showing the expression patterns of the top 20 DEGs across all generations (F 1 -F 3 ) from each comparison ( Fig. 5f,g). Many genes are involved in embryonic development, DNA repair and metabolic processes. To further understand gene function, we annotated the DEGs by Gene Ontology enrichment (Fig. 5h,i and Supplementary Table 5a,b). The top enrichment terms for the androgenized lineage involved DNA repair, germ cell and reproductive processes, glucose homeostasis and steroid hormone signaling pathways, whereas many metabolic processes, including ATP metabolic processes, energy metabolites and mitochondrial transport signaling pathways, were enriched in the obese lineage.

Common gene signatures in tissues of women with PCOS and serum of daughters from mothers with PCOS.
To further explore the potential implication of the DEGs in MII oocytes of the androgenized and obese lineages, we compared them with previously identified DEGs from subcutaneous adipose tissue of women with and without PCOS 28 using 1:1 orthologs. There was an overlap of 10,894 expressed genes (77.93%, 1:1 orthologs) in human 29 and mouse MII oocytes, indicating a conserved function (Extended Data Fig. 9). We also found 33 and 23 DEGs from MII oocytes in the androgenized lineage and 28 the obese lineage that overlapped with DEGs in human adipose tissue, respectively, and 2 genes shared among all conditions (Fig. 6a) (Fig. 6b).
Notably, we had a unique collection of serum samples from daughters of women with and without PCOS from Chile, as well as serum samples of 20 women with PCOS and 20 women without PCOS from the genome-wide adipose tissue expression profile study 28 . Because there was a strong indication of transgenerational effects in the androgenized lineage, we examined the expression of several DEGs in the human serum samples (Supplementary Table 6). Twenty-five of 44 selected genes were identified in human serum. Of these, the expression of TIAL1 was significantly upregulated in the serum of women with PCOS (unrelated and daughters), in the adipose tissue of women with PCOS 28 and in the MII oocytes of the androgenized lineage. Fatty acid binding protein 5 (FABP5) expression was also upregulated, whereas ring finger protein 141 (RNF141) and INTS3 and NABP interacting protein (INIP) expression was downregulated in human serum, similarly to MII oocytes from the androgenized lineage (Fig. 6c,d).

Discussion
Despite decades of research in PCOS pathology, it remains unknown whether offspring from women with PCOS might develop PCOS owing to prenatal androgen exposure and further pass it on to subsequent generations. We showed, in a large Swedish nationwide   . Two-way ANOVA was followed by Tukey's post hoc analysis, and one-way ANOVA was followed by Dunnett's post hoc analysis. For mRNA expression in serum, Student's t-test was used. All data are presented as mean ± s.e.m.

NaturE MEDICINE
register-based cohort study and a case-control study from Chile, that daughters of women with PCOS are more likely to be diagnosed with PCOS. The daughters in the case-control study showed typical PCOS reproductive phenotypes, including hyperandrogenism, irregular menstrual cycles and polycystic ovarian morphology, along with specific metabolic perturbations, including increased waist circumference, higher BMI and higher diastolic blood pressure. Although physiological discrepancies exist between mice and humans, we provided several lines of evidence that prenatal androgen exposure alone, representing lean women with PCOS, predisposes female offspring to a PCOS-like phenotype, which passes on to subsequent generations. Our comprehensive assessments identified strong transgenerational PCOS-like reproductive and metabolic phenotypes due solely to prenatal androgen exposure in F 0 mothers and milder transgenerational effects due to maternal obesity, suggesting that maternal hormonal disruption has more severe effects on the health of future offspring. Notably, the combination of prenatal androgen exposure and obesity, representing obese woman with PCOS, greatly affected the development of F 2 fetuses, which agrees with clinical findings that obese women with PCOS are at higher risk of pre-eclampsia, gestational diabetes, miscarriages, preterm birth and perinatal mortality 37, 38 .
Previous studies showed that daughters of women with PCOS are at higher risk of reproductive, metabolic and neuropsychiatric disorders 18,39-41 but failed to draw causal conclusions 37 . Nevertheless, investigations into fetal development in women with PCOS are lacking because this is difficult to study in humans. Family-based study designs with sibling comparisons can be used to account for unmeasurable factors shared in families, such as genetic influences and environmental factors, when siblings are discordantly exposed to the disruptors of interest. However, siblings born to women with PCOS have similar in utero androgen exposures. To circumvent this, cousins were instead used as a comparison group in an 'offspring of sibling' study 18 , which showed that prenatal androgen exposure due to maternal PCOS was associated with a higher risk of neuropsychiatric disorders among female offspring, even after adjusting for genetic and familial factors shared by cousins 18 . Our register-based and clinical case-control studies demonstrated that first-generation female offspring were more likely to develop PCOS in adulthood. However, in these samples, we cannot disentangle whether this association reflects a causal association or whether it is because of confounding genetic factors. Additionally, it is challenging to perform a longitudinal human cohort study to investigate whether granddaughters and great-granddaughters of women with PCOS are also more likely to develop PCOS. The daughters of women with PCOS in the current registers were too young to assess their second and third generations. Therefore, we addressed whether a PCOS phenotype could be transgenerationally transmitted across generations in female mouse offspring derived from F 0 mothers in which PCOS was modeled by prenatal androgen exposure with or without obesity. Specifically, we demonstrated that female F 1 -F 3 offspring from the androgenized lineage had longer anogenital distance and irregular estrous cycles, thereby supporting and expanding previous clinical observations of daughters of women with PCOS 7 . Transgenerational transmission of metabolic dysfunction in the androgenized lineage was demonstrated by increased fat mass, larger adipocytes and disturbed adipogenesis. These findings are in line with clinical studies showing that enlarged adipocytes, together with increased waist circumference, are strong markers for insulin resistance in women with PCOS 42 and even in normal-weight women with PCOS 43 . Therefore, the enlarged waist circumference in daughters of women with PCOS in this study strongly supports the notion that the adipose tissue dysfunction might be due to prenatal androgen exposure. We also found higher liver triglyceride concentrations together with larger lipid droplets and dysfunctional expression of liver enzymes related to triglyceride metabolism in the androgenized lineage. These findings are consistent with clinical studies that reported liver dysfunction in women with PCOS, including steatosis and fibrosis, caused by hyperandrogenism 44 . However, milder effects for several phenotypic traits were observed in F 2 female offspring in the androgenized lineage. These results imply alterations of some phenotypes between generations, as previously observed 45 . The transgenerational effects occurred despite lower circulating testosterone levels in F 1 female offspring and no difference in the subsequent generations. Thus, the observed transgenerational effects are due to the initial (F 0 ) prenatal androgen exposure. Human studies from our group demonstrated that daughters of women with PCOS have increased ovarian follicular mass during infancy and childhood and higher insulin levels before puberty, despite normal androgen levels when the gonadal axis is still quiescent 46,47 . Some of these features were no longer present when maternal androgen levels were normalized, supporting the participation of a fetal programming effect 48 .
Developmental delay and growth restriction have previously been reported in fetuses from obese dams 17 . Although F 1 -F 3 females in the obese lineage did not have a reproductive phenotype, they displayed a metabolic phenotype, albeit much milder than that in the androgenized lineage. F 1 and some F 2 females had more fat mass, enlarged subcutaneous adipocytes and increased liver triglyceride and lipid content, but none of these variables was affected in F 3 female offspring from the obese lineage. These results reiterate that prenatal androgen exposure, compared to maternal obesity, has long-term fetal programming effects.
We also found that the combination of dihydrotestosterone exposure and diet in F 0 females severely compromised embryonic development of F 2 offspring. Only one F 2 female survived but failed to give rise to any F 3 females in the obese and androgenized lineage, which indicates a strong combinatorial detrimental effect.
Our repeated experiments confirmed that this was due to severely compromised embryonic development of F 2 offspring, resulting in smaller and fewer live embryos at E12.5, and no live embryos could be obtained at E18.5 from a total of 19 F 1 females (6 were pregnant, and 13 were plug positive). This late embryonic death is likely due to placental dysfunction, because expression of Tfap2c, a key controller of placental growth that is linked to fetal death 22,23 , was significantly decreased together with that of the downstream markers Ascl2 and Akt1. In addition, the androgenized lineage, representing lean women with PCOS, also produced smaller embryo sizes at E12.5 and E18.5. Several clinical studies support these findings that women with PCOS are at increased risk of poor pregnancy outcomes, including miscarriage, which are further aggravated by obesity 38,49,50 .
Genetic and epigenetic modifications to the germline are a proposed mechanism for transgenerational inheritance of disease 51,52 . Oocytes contain the stored cellular material required for early embryo development, which can have long-lasting effects on the offspring and potentially also on subsequent generations. Our scRNA-seq analysis provided evidence that the gene expression of oocytes is indeed altered transgenerationally. Despite differences in ovarian morphology between mice and humans, the global gene expression of MII oocytes is highly conserved. Notably, oocytes respond differently to prenatal androgen exposure and maternal diet, and oocytes in the androgenized lineage were more affected than those in the obese lineage. All these findings suggest that prenatal androgen exposure has a more adverse and long-term effect on oocyte function than maternal obesity, which corroborated the phenotypic results.
One of the commonly upregulated genes in F 1 -F 3 oocytes, Tial1, is also upregulated in the adipose tissue of women with PCOS 28 , as well as in the serum of the daughters diagnosed with PCOS in adulthood and in unrelated women with PCOS from our previous study 28 . TIAL1 encodes a member of a family of RNA-binding Articles NaturE MEDICINE proteins that play roles in various biological activities by regulating RNA stability, localization and post-transcriptional regulation 53,54 . Interestingly, alternative splicing of androgen receptor and the PCOS candidate gene DENND1A 55 has been shown to be a pathogenic mechanism for hyperandrogenism and abnormal folliculogenesis in women with PCOS [56][57][58] . Therefore, it is intriguing to speculate that upregulation of Tial1 expression in MII oocytes of the androgenized lineage might serve as a stress response mechanism for oocytes to cope with aberrant RNAs. Meanwhile, excessive TIAL1 might also cause aberrant splicing to generate abnormal gene variants. The differential expression of FABP5, RNF141 and INIP is also of interest because these were regulated similarly in human serum and mouse MII oocytes. FABP5 is expressed in adipocytes and macrophages and is involved in the development of insulin resistance, diabetes mellitus and cardiovascular diseases 59 ; RNF141 encodes a male germ cell-specific transcription factor required for spermatogenesis 60 ; and INIP is involved in maintaining genomic stability.
Taken together, our analyses of transgenerational phenotypes in PCOS-like mouse models suggest a strong adverse effect of androgen excess during gestation. Such transgenerational effects are manifested in both reproductive and metabolic dysfunctions mediated by in utero and/or oocyte-derived factors. Notably, common transcriptional signatures were found in different tissues, pointing toward a possibility of evaluating transgenerational effects in different tissues.

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41591-019-0666-1.

Methods
Ethical approvals. All procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the 1975 Declaration of Helsinki, as revised in 2008. The register-based study was approved by the regional ethical review board in Stockholm, Sweden (diary number 2013/862-31/5; 2016/1214-32). The requirement for informed consent was waived because the study was register based, and the included individuals were not identifiable at any time. The case-control protocol was approved by the institutional review board of the University of Chile (Approval of Research Project N°032-2015). The unrelated case-control study used for serum gene analyses was approved by the regional ethical review board in Gothenburg, Sweden (diary number 307-05) 28 . All mothers and daughters and women with and without PCOS in the unrelated case-control study gave written informed consent before any examination. Register-based study population. Using the unique personal identification number assigned to each person in Sweden at birth or at immigration, several nationwide longitudinal registers containing health and sociodemographic data were linked, including the Swedish Medical Birth Register, the National Patient Register, the Total Population Register, the Migration Register and the Cause of Death Register. The Multigenerational Register was used to identify first-generation female offspring born to mothers with PCOS and matched unaffected mothers.
The first diagnostic criteria for PCOS were established in 1990. Women with PCOS were identified by having at least one PCOS International Classification of Diseases (ICD) code (ICD-9: 256E; ICD-10: E28.2) recorded in the Medical Birth Register or the National Patient Register after the age of 13 and between 1990 and December 2013. Concurrent diagnosed conditions that could give PCOS-like symptoms were excluded to ensure specificity (see Cesta et al. for ICD codes 18 ). Each mother with PCOS was then matched on her birth year and county of residence within the year of diagnosis to ten comparison mothers without a PCOS diagnosis who were randomly selected from the general population. Using the Multigenerational Register, female offspring born from 1973 to 2000 to mothers with PCOS (PCOS-exposed daughters) and comparison mothers without PCOS (PCOS-unexposed daughters) were identified. All offspring from each mother were included. Offspring were excluded if they were born outside of Sweden, adopted, stillborn or died on the day of birth or had congenital malformations. This yielded a total of 2,275 PCOS-exposed daughters and 27,461 PCOS-unexposed daughters from the general population. Daughters were followed until 31 December 2013, and had an age range of 13-41 years (mean ± s.d. = 19.7 ± 5.2 years) at the end of follow-up.
Human case-control study population. From March 2000 to December 2018, women were screened for a PCOS diagnosis at the Unit of Endocrinology and Reproductive Medicine, University of Chile, Santiago, Chile, as previously reported 47 and were recruited during and after pregnancy for several follow-up studies regarding their offspring. Ninety-five of these women gave birth to female offspring and participated with their children in different studies at our unit. Only women fulfilling all three PCOS diagnostic criteria-chronic oligomenorrhea or amenorrhea, hyperandrogenism and polycystic ovarian morphology-were selected for these studies. Control mothers were of similar socioeconomic status, with regular 28-to 32-d menstrual cycles and no signs of hirsutism or other clinical manifestations of hyperandrogenism, infertility or pregnancy complications. For this study, we selected 21 daughters born to women with PCOS (PCOS-D) and 14 daughters born to control mothers (Control-D) between 17 and 25 years of age. All daughters were born from singleton pregnancies. Six PCOS-D women were using oral contraceptives as a treatment for PCOS symptoms, and their medication was not stopped. The rest of the daughters included in the study were not taking oral contraceptives or any other medication during the 6 months before examination. Daughters with any chronic diseases were excluded. Most of the daughters had been part of our previous studies 46,47 , but each daughter participated in the present study only once.
The recruitment and clinical examination of the women in the unrelated case-control study were previously described in detail 28 . In brief, subcutaneous abdominal adipose tissue biopsies were obtained under local anesthesia and immediately snap frozen for later analyses of global genome-wide gene expression and DNA methylation patterns 28 . Fasting blood samples from 20 women with PCOS and 20 controls (BMI <26 kg m −2 ) in that study 28 were used for the gene expression analyses in the present study.
Clinical examination. Physical examination of the daughters included anthropometric measures of weight, height, and waist and hip circumference and was performed at the Policlinic of the Endocrinology and Metabolism Laboratory, West Division, School of Medicine, University of Chile. The age at menarche was registered, and the mean cycle length of the last 12 months was reported. Gynecologic age was defined as the number of years after menarche at the examination. Hirsutism was evaluated by determining the presence of terminal hair using the modified Ferriman-Gallwey score, with hyperandrogenism defined as a value of ≥6. Daughters with a history of early puberty and precocious pubarche were excluded. Serum glucose, insulin testosterone, sex-hormonebinding globulin (SHBG), luteinizing hormone (LH), follicle-stimulating hormone (FSH) and AMH concentrations and serum lipids-total cholesterol, triglycerides, low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C)-were determined in a fasting sample. Serum SHBG and testosterone were used to calculate the free androgen index as serum testosterone/ SHBG × 100.
On the following day, a transabdominal ultrasound was performed and analyzed by an observer blinded to the condition of the subject. The examination was performed with a 5-MHz abdominal probe using Medison Sonoace X 8 equipment (Medison). Ovarian volume was calculated using the simplified formula for a prolate ellipsoid 61 . The larger ovary was used to evaluate the ovarian size. All examinations were performed during the early follicular phase of the menstrual cycle (days 3-7) or whenever feasible in women with amenorrhea.
Assays. Serum AMH was assayed by ELISA (Immunotech Beckman Coulter). The analytical sensitivity was 2.1 pM, and the intra-and interassay coefficients of variation were 5.3% and 8.7%, respectively. Serum glucose and lipid profiles were determined by the glucose oxidase method (Biosystems). The intra-and interassay coefficients of variation of this method were <1.2% and <2.7%, respectively. LDL-C was calculated according to the Friedewald formula. Serum insulin was assayed by IRMA (DIAsource Immunoassay) with a sensitivity of 1.0 µIU ml −1 and intra-and interassay coefficients of variation of 3% and 7%, respectively. Serum LH, FSH and SHBG were determined by IRMA Izotop (Institute of Isotopes (Izotop)). Assay sensitivities were 0.05 mIU ml −1 , 0.02 mIU ml −1 and 0.22 nM, respectively. The intra-and interassay coefficients of variation were 1.0% and 3.1% for LH; 2.5% and 2.7% for FSH; and 4.9% and 3.8% for SHBG, respectively. Serum testosterone, androstenedione and 17-hydroxyprogesterone were assayed by RIA (DIAsource Immunoassay), and the limits of detection were 0.1 ng ml −1 , 0.1 ng ml −1 and 0.03 ng ml −1 , respectively. The intra-and interassay coefficients of variation were 3.5% and 5.5%, 3.8% and 7.5%, and 4.6% and 7.7%, respectively. Serum estradiol was assayed by RIA (Pantex) with a sensitivity of 10 pg ml −1 and intra-and interassay coefficients of variation of 4.7% and 5.4%, respectively.
Animals. Three-week-old female C57BL/6J mice were obtained from Janvier Labs. All mice were maintained under a 12-h light/dark cycle and in a temperaturecontrolled room with ad libitum access to water and a diet. After 1 week of acclimatization, female mice were randomly divided into two groups and fed either (1) a control diet (Research Diets, D12328) comprising 11% fat, 73% carbohydrates (0% sucrose) and 16% proteins or (2) a high-fat, high-sucrose diet (Research Diets, D12331) comprising 58% fat, 26% carbohydrates (17% sucrose) and 16% protein for 6 weeks 17 . Body weight was measured once every 2 weeks. Food intake was calculated as mean food consumed by five mice per cage per day. Before mating, mice (F 0 dams) were subjected to an OGTT and magnetic resonance imaging (EchoMRI-100 system, EchoMRI) to measure total fat and lean mass. Eight-to 12-week-old male mice were used for mating and fed an in-house chow diet (R34, Lantmännen). A female in proestrus or estrus phase, as determined by vaginal cytology, was mated overnight with a male. Females were checked daily for post-copulatory plugs, and a plug on the morning after mating was considered E0.5. Before starting the experiments, the number of animals required for the experiments was estimated from our previous work, based on the same model where the success of the breeding was about 60% of the F 0 dams 16,62,63 .
Mouse breeding scheme and feeding paradigm to generate F 1 -F 3 offspring. At postnatal day (P) 21, all female offspring (F 1 , F 2 and F 3 ) in the control, androgenized, obese, and obese and androgenized lineages were weaned onto the control diet. A subset of F 1 female offspring were mated with unrelated males fed chow diet to generate F 2 offspring, and a subset of F 2 female offspring were mated with unrelated males fed chow diet to generate F 3 offspring. The remaining F 1 , F 2 and F 3 female offspring were subjected to phenotypic testing as described below. The exact number of mice used for each procedure and their sex and age are given in the figure legends and/or text. Details of the number of mice used for (1) phenotypic testing and (2) breeding to generate F 1 , F 2 and F 3 offspring in each Articles NaturE MEDICINE group are specified in Supplementary Table 7. To ensure variability within each group, offspring in each generation were randomly allocated for phenotypic testing or breeding.
Assessment of reproductive phenotype. At weaning, anogenital distance was measured in F 1 , F2 and F 3 female offspring. Body weight development was recorded weekly from 3 weeks of age to 15 weeks of age, and estrous cyclicity was assessed by daily vaginal smears for ten consecutive days (two cycles). Estrous cycle changes were determined by vaginal cytology as previously described 64 .
Assessment of metabolic phenotype. Body composition (lean and fat mass) was assessed in conscious mice by EchoMRI at 16-18 weeks of age. Food intake, gas exchange and spontaneous locomotor activity were recorded in metabolic cages (TSE PhenoMaster, TSE Systems). Animals were kept in metabolic cages individually for three consecutive days, with the first day being considered the adaptation period (not analyzed), and 24-h readings were used for analysis after the adaptation period. Parameters for each mouse were recorded at 3-min intervals, and the mean of every 15 min was reported. EE was determined by indirect gas calorimetry and adjusted for total body mass. RER was calculated as the ratio between the volumes of CO 2 produced and O 2 consumed. Spontaneous locomotor activity was measured by recording interruptions of infrared light beams emitted along the x axis and y axis of each cage (expressed in counts). Glucose metabolism was measured by an OGTT after a 6-h fast at 17-19 weeks of age. D-glucose was administrated orally by gavage (2 g kg −1 ), and blood glucose was measured at time 0 (before glucose administration) and at 15, 30, 60 and 90 min (Free Style Precision). Blood was collected at 0 and 15 min for insulin measurement by tail bleeding.
Biochemical assessment of insulin, sex steroids and AMH. Serum insulin from the OGTT was analyzed by an ELISA kit (Crystal Chem). Serum testosterone, dihydrotestosterone and androstenedione were measured in serum of F 1 -F 3 female offspring using a gas chromatography and tandem mass spectrometry assay as previously described 65 . Serum AMH level was measured using a mouse AMH ELISA kit (Ansh Labs).
Adipocyte size measurement. Subcutaneous adipose tissue was dissected and prepared as previously described 64 . Tissue was sectioned at 10-µm thickness; one section was taken every 30 µm, and a total six sections were taken for each sample. Three animals were used in each group (the F 2 obese and androgenized lineage had only one animal). One or two representative images were taken per section with a light microscope at ×20 magnification (Zeiss Axioplan). Adipocyte size was quantified using CellProfiler (https://cellprofiler.org). Identified cells with an area >4,000 AU or >0.97 eccentricities were removed. The ×40 representative images were taken with an inverted microscope (Olympus IX73; Olympus).
Liver triglyceride quantification. Liver triglycerides were extracted as follows. Briefly, 100 mg of liver was homogenized in the Tissue Lyzer (Qiagen) for 3 min in 5% NP-40 alternative (492016, Merck) in distilled water. Next, the lysate was heated at 95 °C for 5 min in a heating block until the samples became cloudy and was then cooled to room temperature. The previous step was repeated, and the samples were centrifuged at 13,000 g for 2 min. The supernatant was removed, diluted tenfold and analyzed using the TG kit Randox (TR210, Crumlin) according to the manufacturer's instructions.
Liver Oil Red O staining. For Oil Red O staining, liver tissue was collected from F 1 -F 3 female offspring and immediately flash frozen in liquid nitrogen. A piece of liver was embedded in OCT (Sakura Finetek) on the mold and dropped into 2-methylbutane (277258, Merck) on dry ice. OCT-embedded tissues were sectioned into 10-µm sections and stained as described 66 .

RNA isolation and mRNA expression.
Mouse subcutaneous adipose tissue, liver and placenta were homogenized in 1 ml TRI reagent (T9424, Sigma-Aldrich) with a Tissue Lyzer, and total RNA was isolated according to the manufacturer's instructions. RNA samples were treated with DNase (18068-015, Invitrogen), and cDNA was synthesized from 500 ng of total RNA using the High Capacity RNAto-cDNA kit (4387406, Applied Biosystems). Quantitative real-time PCR for gene analysis was performed in a ViiA 7 Real-Time PCR system thermal cycler with SYBR Green PCR Master Mix (both Applied Biosystems). The primer sequences can be found in Supplementary Table 8. The relative gene expression was calculated using the comparative cycle threshold method. Gapdh was used as the endogenous control, and mRNA expression is presented as 2 −ΔΔCt .
From a subset of the daughters and women with and without PCOS in the case-control study 28 , RNA was extracted from serum using an miRNeasy Serum/ Plasma Kit (21784, Qiagen) combining phenol/guanidine-based lysis and silica membrane purification of total RNA. The RNA samples were treated with DNase, and cDNA was synthesized as described above. The 7900HT Fast Real-Time PCR System (Applied Biosystems) was used for analyzing the custom-designed TaqMan Array cards (Applied Biosystems) that included 44 target genes and 4 endogenous controls (Supplementary Table 6). The selection criteria for genes to be analyzed in serum were transgenerational DEGs with a >1.25or <0.75-fold change and RPKM > 8 in mouse MII oocytes from the control versus androgenized lineage and DEGs in human adipose tissue 28 overlapping with these genes. ACTB was used as the best endogenous control out of four endogenous controls identified by NormFinder (https://moma.dk/normfinder-software). The expression of mRNA was calculated as 2 −ΔΔCt .

Collection of MII oocytes, organs and embryos.
To collect MII oocytes, 20-weekold F 1 , F 2 and F 3 female offspring were superovulated by injecting them with 5 IU of pregnant mare's serum gonadotropin (PMSG; Folligon, MSD Animal Health Care) followed by 5 IU of human chorionic gonadotropin (hCG; Pregnyl 5000IE, Merck Sharp & Dohme) 48 h after PMSG priming. Cumulus-oocyte complexes were isolated at 16 h after hCG injection from oviduct ampulla. Denuded single MII oocytes were then obtained by removing the cumulus mass in M2 medium (M7167, Merck) containing 0.3 mg ml −1 hyaluronidase (H3884, Merck) at room temperature. At finalization, mice were fasted for 2 h before blood, oocyte and tissue collection. Briefly, mice were anesthetized with isoflurane, and blood was collected through the axillar vein. The subcutaneous adipose tissue and liver were quickly dissected on ice, snap frozen and stored at −80 °C.
When investigating embryonic development in F 1 female offspring, pregnant mice were dissected at E12.5, E15.5 and E18.5. The crown-rump length of the embryos was measured with ImageJ software to assess development.
Transmission electron microscopy for mitochondrial morphology. MII oocytes from F 1 , F 2 and F 3 female offspring were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) at room temperature for 30 min. Oocytes were rinsed in 0.1 M phosphate buffer before post-fixation using 2% osmium tetroxide in 0.1 M phosphate buffer (pH 7.4) at 4 °C for 2 h. The oocytes were subsequently dehydrated in ethanol followed by acetone and finally embedded in LX-112. Ultrathin sections were prepared using a Leica EM UC7 (Leica Microsystems) and contrasted with uranyl acetate followed by Reynolds lead citrate. The sections were examined on a Hitachi HT 7700 electron microscope (Hitachi High-Technologies) at 80 kV, and images were acquired using a 2,000 × 2,000 Veleta CCD camera (Olympus Soft Imaging Solutions).
Mitochondrial DNA in MII oocytes. Single MII oocytes were loaded into PCR tubes with 15 μl lysis buffer (50 mM Tris-HCl, 0.1 mM EDTA, 0.5% Tween-20, 200 µg ml −1 Proteinase K, pH 8.5) and incubated at 55 °C for 1 h. Proteinase K was heat inactivated at 95 °C for 10 min, and then the samples were used directly for PCR analysis. Quantitative real-time PCR was performed using the 7900HT Fast Real-Time PCR System (Applied Biosystems) for mtDNA by using mouse-mtDNA-specific primers: B6 forward, AACCTGGCACTGAGTCACCA, and B6 reverse, GGGTCTGAGTGTATATATCATGAAGAGAAT. For nuclear DNA, Gapdh was used (forward, TGG AGC CAA AAG GGT CAT CA, and reverse, TCG TGG TTC ACA CCC ATC AC).
Sequencing library preparation. Next, 1-2 ng cDNA was used for tagmentation using a Tn5 transposase and amplified for ten cycles using the Nextera XT DNA Sample Preparation kit (Illumina). The prepared samples were sequenced using an Illumina HiSeq 2500 sequencer.
RNA-seq data processing. Raw reads were mapped to the mouse reference genome (GRCm38/mm10) using STAR default mapping arguments 68 . We used 'rpkm for genes' to calculate RPKM on the basis of the uniquely mapped read counts. Cells with low complexity (fewer than 3,000 expressed genes), low mapping reads (fewer than 200,000) and lower mapping ratios (less than 70%) were excluded.

NaturE MEDICINE
Normalization and batch effect correction. After removing the low-quality cells, the Seurat normalization method 'LogNormalize' was applied to quality-controlled single-cell data. The 'ScaleData' function in Seurat was then implemented to remove the batch effect caused by technical variation.

DEG analysis and Gene Ontology analysis.
After batch effect correction, to identify the DEGs, the read counts were analyzed using the R package 'DESeq2' two-sided Wald test (version 1.20.0) 69 . DEGs were qualified by P < 0.05. Gene Ontology analysis of DEGs was performed using the 'Clusterprofiler' (version 3.8.1) R package 70 .
Global gene expression analyses in adipose tissue. Gene expression was analyzed using the Illumina HumanHT-12 v4 Expression BeadChips in women with PCOS and controls in the previously published unrelated case-control study 28 . Each chip had 47,000 probes that represent well-characterized genes and unknown splice variants. A false-discovery rate < 5% (q < 0.05) was used to correct for multiple testing, and 1,720 DEGs were identified 28 . Here, there was an overlap between DEGs in MII oocytes and DEGs in human subcutaneous adipose tissue 28 .

Statistical analyses.
In the register-based cohort study, associations between maternal PCOS and a diagnosis of PCOS in the daughters was estimated as HRs with 95% CIs using stratified Cox regression models with attained age as the underlying time scale. In addition to the maternal matching criteria (maternal birth year and county of residence within the year of PCOS diagnosis), potential confounding was addressed in a fully adjusted model controlling for daughter's year of birth, maternal age at daughter's birth, maternal education and maternal region of birth. In the case-control study and mouse experiments, data were assessed for normality and variance (Kolmogorov-Smirnov test). Group allocation in the experiments was not performed with blinding of the investigators. However, analyses were repeated by two or more investigators and controlled by an independent statistician not involved in the study. Sample size in the mouse experiments was based on differences in anogenital distance between the control and androgenized lineages in our previous studies 62,63 . Nine animals per group were required to detect a mean difference in anogenital distance of 40.6% with an s.d. of 0.1, a significance level of 0.05 and a power of 0.8. All data are presented as mean ± s.e.m. or as median and range. Differences between two groups were determined by Student's t-test or Fisher's exact test. Differences between more than two groups were determined by two-way ANOVA followed by Tukey's post hoc test, one-way ANOVA followed by Dunnett's post hoc test or the Kruskal-Wallis test followed by Dunn's post hoc test. Differences were considered statistically significant at P < 0.05. Statistical analyses in the register-based study were performed using Stata statistical software version 14.0 (Stata Corp.) and in the case-control and animal studies using GraphPad Prism 8 (GraphPad Software).
Reporting Summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Raw data of F 1 -F 3 females and clinical characteristics of daughters are available through Dryad: https://doi.org/10.5061/dryad.jwstqjq4m. All raw and analyzed scRNA-seq data of mouse MII oocytes from F 1 -F 3 females are available at the Gene Expression Omnibus database via accession number GSE133100. For the Swedish register-based cohort, original data are held by the Swedish National Board of Health and Welfare and Statistics Sweden, and because of Swedish data privacy laws we cannot make the data publicly available. Any researcher can access the data by obtaining an ethical approval from a regional ethical review board and thereafter asking the Swedish National Board of Health and Welfare and Statistics Sweden for the original data. However, aggregated data used in the analysis of this study are available from the authors upon reasonable request and with approved data sharing and data processing agreements in line with the General Data Protection Regulation. Further use of these data must be authorized by the local ethics committee regarding the merit of the project involved. A detailed description of the unrelated case-control study, including global gene expression analyses in subcutaneous adipose tissue, has previously been published 28

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted

Software and code
Policy information about availability of computer code

Data collection
The register-based data was provided in SAS files.

Data analysis
Analyses in the register-based part were performed using Stata version 15.1 (Stata Corps, Texas, USA). The code is available upon request. The adipocyte size was quantified using CellProfiler: https://cellprofiler.org NormFinder (https://moma.dk/normfinder-software) was used to identify the best endogenous control(s) to be used in mRNA analyses. The STAR default mapping arguments, see ref 69 and: http://labshare.cshl.edu/shares/gingeraslab/www-data/dobin/STAR/STAR.posix/ doc/STARmanual.pdf To get the differentially expressed genes of single cell sequencing, the read counts were performed by R package "DESeq2". Gene ontology analysis of differentially expressed genes was performed using "Clusterprofiler" R package. The register-based data was provided in SAS files: SAS 9.4 GraphPad Prism 8 (GraphPad Software Inc., CA, USA).
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: raw and analysed single cell RNA-sequencing data of mouse MII oocytes from F1-F3 females is available at GEO database via accession number GSE133100. Both will be released on publication. For the Swedish register-based cohort, original data is held by Swedish National Board of Health and Welfare and Statistics Sweden and because of Swedish data privacy laws we cannot make the data publicly available. Any researcher can access the data by obtaining an ethical approval from a regional ethical review board and thereafter asking the Swedish National Board of Health and Welfare and Statistics Sweden for the original data. However, aggregated data used in the analysis of this study are available from the authors upon a reasonable request and with approved data sharing and data processing agreements in line with the General Data Protection Regulation (GDPR). Further use of this data must be authorized by the local ethics committee regarding the merit of the project involved. A detailed description of the unrelated case-control study including global gene expression analyzes is subcutaneous adipose tissue has previously been published (Kokosar et al. Scientific Report, 6, 22883 (2016).

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Field-collected samples The study did not involve samples collected from the field.

Ethics oversight
All animal experiments were approved by the Stockholm Ethical Committee for Animal Research (10798-2017) in accordance with the legal requirements of the European Community (SJVFS 2017:40) and the directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. Animal care and procedures were performed in accordance with guidelines specified by European Council Directive and controlled by Comparative Medicine Biomedicum (KM-B), Karolinska Institutet, Stockholm, Sweden.
Note that full information on the approval of the study protocol must also be provided in the manuscript.

Human research participants
Policy information about studies involving human research participants

Population characteristics
Case-control study: Only women fulfilling all three PCOS diagnostic criteria: chronic oligomenorrhea or amenorrhea, hyperandrogenism and polycystic ovarian morphology have been selected for this study. Control mothers were of similar socioeconomic status, with regular 28 to 32 day menstrual cycles, with no signs of hirsutism or other clinical manifestations of hyperandrogenism, infertility or pregnancy complications. For the present study we selected 21 daughters born to women with PCOS (PCOS-D) and 14 daughters born to control mothers (Control-D) between 17 and 25 years of age. All daughters were born from singleton pregnancies.
The unrelated case-control study has previously been described in detail (

Recruitment
Case-control study: Only women fulfilling all three PCOS diagnostic criteria: chronic oligomenorrhea or amenorrhea, hyperandrogenism and polycystic ovarian morphology have been selected for these studies. Control mothers were of similar socioeconomic status, with regular 28 to 32 day menstrual cycles, with no signs of hirsutism or other clinical manifestations of hyperandrogenism, infertility or pregnancy complications. For the present study we selected 21 daughters born to women with PCOS (PCOS-D) and 14 daughters born to control mothers (Control-D) between 17 and 25 years of age. All daughters were born from singleton pregnancies.
The unrelated case-control study: In brief, they were recruited by advertising in local newspapers and in frequently visited places in the community. All women were ≥18 and ≤ 38 years of age and with a BMI ≤ 25. The eligibility criteria for women with PCOS were at least two of the following three signs; polycystic ovarian morphology (PCO), clinical signs of hyperandrogenism and oligo/amenorrhea. To avoid potential bias all statistics has been made by persons not involved in the actual patient handling. Other potential bias is that the recruitment in the case-control cohorts were either done via advertisement in the local community or by selecting women fulfilling all three PCOS diagnostic criteria (see page 23) and can thus not be directly be translated to the general population.

Ethics oversight
The register-based study was approved by the regional ethical review board in Stockholm, Sweden (diary number 2013/862-31/5; 2016/1214-32). The case-control protocol was approved by the institutional review Board of the University of Chile and all mothers and daughters gave written informed consent prior any examination (IRB, Faculty of Medicine, University of Chile; Approval of Research Project Nº032-2015). The unrelated case-control study used for serum gene analyses was approved by the regional ethical review board in Gothenburg, Sweden (diary number: 307-05) Note that full information on the approval of the study protocol must also be provided in the manuscript.

Clinical data
Policy information about clinical studies All manuscripts should comply with the ICMJE guidelines for publication of clinical research and a completed CONSORT checklist must be included with all submissions.

Clinical trial registration
No registration for register-based and case-control study.

Study protocol
Not applicable.

Data collection
Register-based population study: Women with PCOS were identified by having at least one PCOS International Classification of Diseases (ICD) code (ICD-9: 256E; ICD-10: E28.2) recorded in the MBR or in the NPR after the age of 13 and between 1990 and until December 2013. Concurrent diagnosed condition that could give PCOS-like symptoms were excluded to ensure specificity, see Cesta et al study (ref number '14' the manuscript) for ICD codes15. Each mother with PCOS was then matched on her birth year and county of residence within the year of diagnosis to 10 comparison mothers without a PCOS diagnosis randomly selected from the general population. Using the Multigeneration Register (MGR), female offspring born from 1973 to 2000 to mothers with PCOS (PCOS exposed daughters) and comparison mothers without PCOS (PCOS unexposed daughters) were identified. All offspring from each mother were included. Offspring were excluded if they were born outside of Sweden, adopted, stillborn or died on the day of birth, or had congenital malformations. Outcomes Primary outcome measure: PCOS diagnosis.
Register-based cohort study: Associations between maternal PCOS and a diagnosis of PCOS in the daughters was estimated as hazard ratios (HR) with 95% confidence intervals (CI) using stratified Cox regression models with attained age as the underlying