This study was conducted in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996). The experimental protocol was approved by the Animal Care and Use Committee of the Institute of Molecular Genetics, CAS and carried out in accordance with the relevant guidelines and regulations. We used a T2D mouse model generated by a combination of high-fat diet and a low dose of streptozotocin (28, 30). Male inbred C57BL/6J mice (purchased from Charles River, Germany), aged 6 weeks, were divided into two groups and were fed either a high-fat diet (HFD: 60% of the metabolizable energy coming from fat, 20% from carbohydrates, and 20% from protein; ssniff^® EF acc. D12492 from Ssniff Spezialdiäten GmbH, Germany) or a control standard chow diet (11% of the metabolizable energy coming from fat, 65% from carbohydrates, and 20% from protein, #1320 diet, Altromin, Germany). After 9 weeks of HFD, intraperitoneal injection of 100 mg/kg body weight of streptozotocin (STZ; Sigma) was applied. Body weight, and blood samples were taken after 6-h fasting at indicated time points during the experiment (Experimental schematics in Figure 1 ). Blood glucose levels were measured in animals by a glucometer (COUNTOUR TS, Bayer, Switzerland); blood glucose levels maintained above 13.9 mmol/L are classified as diabetic. Mice were kept under standard experimental conditions with constant temperature (23–24°C). Glucose tolerance tests (GTT) were performed in 6h-fasted males. Mice were injected intraperitoneally with glucose (2 g/kg). Blood glucose was measured at 0, 15, 30, 60, and 120 min after glucose administration. After 12 weeks in the experiment, T2D males were glucose intolerant with blood glucose levels maintained above 13.9 mmol/L.

At 18 weeks of age, male founders (P generation) were mated with females fed with the standard chow diet. Individual males were placed in a cage with 1 healthy adult female. The animals were kept together for a week with minimalized disruptions and therefore, a vaginal plug was not checked in female mice. During this week all mice were fed with a standard chow diet. Immediately after separation, the males were killed, tissues and blood were collected for analyses. The females were housed individually during the gestation period and the litter size was recorded. Only male pups, the F₁ offspring, were left in the litters to be used for following up transgenerational studies. At 9 weeks of age, randomly selected F₁ males were mated with females to produce the F₂ offspring. The offspring were fed only by the regular control diet. To minimize any additional environmental influence, molecular and cellular analyses were performed in both F₁ and F₂ male offspring at 10 weeks of age.

Blood serum was collected from non-diabetic and T2D males following a 6-h fast and was analyzed using a fully automated chemistry analyzer, Beckman Coulter AU480 (Beckman) according to the manufacturer’s protocol in the Core Facility for Phenogenomics, Biocev. System Beckman Coulter AU480 reagents used for the quantification: glucose (OSR6621), AST (OSR6009), ALT (OSR6007), TG (OSR60118), insulin (33410), PAI-1 (plasminogen kit 20009000), and LDL (ODC0024).

Mouse sperm cells were obtained from the left and right cauda epididymis. Spermatozoa from both caudae were left to release into warm (37 °C) PBS in a CO₂ incubator for 15 min. The resulting mixture was poured through a 30 μm filter (Partec) to obtain only the sperm fraction and PBS was added to 1 mL of final volume (31). Sperm were counted in a Bürker chamber under the light microscope, and sperm concentration, sperm morphology, and sperm head separation were determined as described previously (32). Sperm morphology was evaluated in 200 cells per each animal. Sperm viability was analyzed with SYBR14 from Live/Dead sperm viability kit (Molecular Probes, USA) and by flow cytometry (LSRII, blue laser 488 nm, Becton Dickinson, USA), minimum 15,000 events were evaluated. An annexin V–FITC apoptosis kit (Sigma) and chromomycin A₃ staining (CMA3, Sigma) were used to assess the sperm damage (29). The molecular and functional quality of sperm was further evaluated by the expression of two molecular biomarkers, sperm acrosomal transitional endoplasmic reticulum ATPase (TER ATPase, TERA) and glyceraldehyde-3-phosphate dehydrogenase-S (GAPDS), using monoclonal antibodies Hs-14 (33) and Hs-8 (34), respectively (EXBIO Ltd., Czechia). These antibodies are used for pathological human sperm and reproductive diagnostics, and evaluation of dysfunctional spermatogenesis (33, 35–37). A minimum of 200 spermatozoa per each animal was examined with a Nikon Eclipse E400 fluorescent microscope equipped with a Nikon Plan Apo VC 60/1.40 oil objective (Nikon Corporation Instruments Company) and photographed with a ProgRes MF CCD camera (Jenoptik, Germany) with the aid of NIS-ELEMENTS imaging software (Laboratory Imaging, Ltd.). Protamine 1 and 2 ratio was obtained from 4 x 10⁶ sperm cells per each sample, as described previously (29).

Dissected testes were fixed with 4% paraformaldehyde in PBS (pH 7.4) at 4°C overnight, and embedded in paraffin. Paraffin-embedded tissues were cut into 8 µm sections, de-paraffinized and rehydrated sections were stained with hematoxylin & eosin or used for immunohistochemistry. Citrate buffer (pH 6.0) was used as a heat-induced antigen retriever. Evaluation of meiotic cells and the 12 stages of production of spermatozoa in the mouse seminiferous epithelium was done using immunolabeling of the axial element of the synaptonemal complex (SYCP3, mouse anti-SCP3 #ab97672, 1:1000 dilution, Abcam). Rabbit anti-CX43 (#C6219, 1:2000 dilution, Sigma) was used to evaluate gap junctions. The secondary antibodies Alexa Fluor^® 488 AffiniPure Goat Anti-Mouse IgG (Jackson ImmunoResearch Laboratories 115-545-146), and Alexa Fluor^® 594 AffiniPure Goat Anti-Rabbit IgG (Jackson ImmunoResearch Laboratories 111-585-144) were used in a 1:400 dilution. Samples were incubated with primary antibodies at 4°C for 24 hours. After several PBS washes, secondary antibodies were added and incubated at 4°C for 24 hours. The sections were counterstained with Hoechst 33342 (#14533 Sigma) and imaged with a confocal microscope (ZEISS LSM 880 NLO) or with a fluorescence microscope (Nikon Eclipse E400). A relative quantification of Cx43 staining in Leydig cells was done using ImageJ software. Morphometric evaluations of seminiferous tubule diameter and thickness of the seminiferous epithelium were performed. The diameter of a seminiferous tubule was defined as the shortest distance between two parallel tangent lines of the outer edge of the tubule. Paraffin sections (8 μm) were stained with Periodic acid–Schiff to visualize advanced glycation end products (Periodic acid–Schiff (PAS) kit; 395B-1KT, Sigma).

RT-qPCR was performed as previously described (29). Briefly, cDNA was prepared using 2 μg of total RNA isolated from the testes. We used RevertAid Reverse Transcriptase (Thermo Fisher Scientific), 1 μL Oligo(dT) and random hexamer primers (Thermo Fisher Scientific) in reverse transcription (RT). Quantitative real-time PCR (qPCR) was performed with a final concentration of cDNA 10 ng/μL using a CFX 384 – qPCR cycler (BioRad). The relative expression of a target gene was calculated, based on the quantification cycle (Cq) difference (Δ) of an experimental sample versus control. The control was set at 100% and experimental samples were compared to the control (29). The β-actin (Actb) and Peptidylprolyl isomerase A (Ppia) genes were used as the reference genes. Primer sequences are listed in Supplementary Table S1 .

The comparisons between diabetic and control mice were done using STATISTICA 7.0. (Statsoft, Czech Republic) and GraphPad Prism 7.0 (GraphPad Software, Inc., USA). Differences in organ weights were tested by ANCOVA with body weight as covariate. Differences in sperm parameters, testis morphology, and gene expression between control and T2D groups were assessed by t-test. One-way ANOVA, follow by Tukey’s post hoc tests for multiple comparisons was used to assess differences among offspring generations. Repeated measure-based parameters (such as weight gain over time, GTT, or blood glucose levels) were analyzed using two-way ANOVA for repeated measures. P value < 0.05 was considered to be statistically significant. Sample sizes and individual statistical results for all analyses are provided in the figure legends and tables.

We used a non-genetic type 2 diabetes (T2D) mouse model that manifests the metabolic abnormalities associated with human prediabetes and type 2 diabetes (28, 30). This model involves a combination of a high fat diet (HFD) to induce insulin resistance and hyperinsulinemia, and a low dose STZ, which reduces β-cell mass but does not cause diabetes in control-fed mice, to bring about glucose intolerance and hyperglycemia (28, 38) ( Figure 1 Experimental schematics). Together these two stressors have been designed to mimic the pathology and the multi-genetic/environmental background of T2D in humans (39–41). We continuously monitored the metabolic changes in our T2D model. As expected, after 8 weeks of HFD, all male C57BL/6J male mice in the experiment had significantly higher fasting plasma insulin levels and were glucose intolerant, as shown by glucose tolerance tests, indicating compromised insulin response ( Figure 2A ). Body weight of the T2D mice increased progressively over time compared to control mice maintained on a normal standard chow diet ( Figure 2B ). After STZ treatment, mice became hyperglycemic with blood glucose levels maintained above 13.9 mmol/l, classified as diabetic ( Figure 2C ). Fasting plasma levels of glucose and insulin were significantly higher in the T2D group ( Figure 3A ). Correspondingly to the diabetic phenotype, after STZ treatment, plasma glucose levels were increased. Plasma insulin levels, which were initially increased in response to the fat-enriched diet in T2D mice, decreased after STZ-induced β-cell reduction but remained at higher levels than insulin levels in control mice. After 12 weeks, compared to control males, T2D mice displayed significant changes in serum metabolic biomarkers, including increased serum glucose, triglycerides, low-density lipoprotein (LDL), plasminogen activator inhibitor 1 (PAI-1, a marker of insulin resistance), metabolic liver enzymes, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) ( Figure 3B ). Metabolic differences were also manifested by increased body weight, and liver weight of T2D males ( Figure 3C ). These physiological and metabolic changes confirmed decreased insulin sensitivity and impaired glucose tolerance associated with the development of type 2 diabetes mellitus. To analyze the effects of the T2D environment on the reproductive system and reproductive outcomes, we mated these T2D males or control males to female mice maintained on a control chow diet through the course of the experiment ( Figure 1 ). After 1 week of mating, males were removed, limiting their influence on their progeny to the mating itself.

All diabetic males, with the exception of one mouse, were able to mate in the period of one week. The rate of pregnancy was in the limits of the rate of pregnancy in our C57BL/6J colony (typically 70–80%, unpublished observation), indicating that overall reproductive performance of T2D males was comparable to controls. However, the litter size of T2D males was more variable (5.63 ± 1.35, n = 6 litters) compared to controls (7 ± 0.53, n = 7 litters, Table 1 ). The weights of the reproductive organs, epididymis and seminal vesicles, and the anogenital distance (AGD), as an androgen-responsive outcome, were unaffected in T2D mice, except for the prostate ( Supplementary Table S2 ).

We performed a sperm quality assessment of collected caudal epididymal sperm. No significant effects were found in sperm concentration, viability, apoptosis, or in the packaging quality of the chromatin (as assessed by chromomycin A₃ staining) between T2D and control males ( Supplementary Table S3 ). The abnormalities in sperm head morphology were increased in T2D males compared to controls ( Figure 4A ). To further assess sperm quality, we analyzed protamine 1 and protamine 2 ratios, which affect DNA stability, and play a role in the establishment of epigenetic marks (42, 43). The protamine 1 and protamine 2 ratios were altered in the paternal T2D generation ( Figure 4A ). Additionally, the expression of biomarkers of sperm dysfunction, transitional endoplasmic reticulum ATPase (TERA) and glyceraldehyde-3-phosphate dehydrogenase-S (GAPDS) was evaluated. TERA is expressed in the acrosomal part of the sperm head (33), whereas GAPDS is located mainly in the principal piece of the sperm flagellum and in lesser amount in the acrosome (34, 44). The expression of both TERA and GAPDS was reduced in sperm from T2D males compared to sperm from control males ( Figures 4B, C ).

In mature testes, sperm are produced in the seminiferous epithelium during the cycle of I to XII stages of spermatogenesis, which are defined by the differentiation steps of germ cells, and occur continuously (45). In order to evaluate the morphology of the seminiferous epithelium at different stages of the differentiation of spermatogenic cells, we focused on four broad categories of the differentiation - spermatogonia, spermatocytes, spermatids, and spermatozoa (Schematic drawing in Figure 5A ). We did not find any pronounced morphological abnormalities, such as testicular atrophy, deficiency, or loss of the seminiferous epithelium between the control and T2D testes. Immunolabeling of a meiotic marker, synaptonemal complex protein 3 (SYCP3) (46), revealed that T2D testes contained all stages of meiotic cells similar to controls ( Figures 5B–D ). Although all the distinct developmental stages of spermatogenesis were identified in similar percentages in both control and T2D testes, a significantly thinner germinal epithelium was found in all cycle stages in the testes of diabetic mice and the diameter of seminiferous tubules was reduced in diabetic testes compared to controls ( Tables 2 , 3 and Supplementary Table S4 ). To evaluate early markers of diabetes-induced tissue damage, the expression of gap junction protein connexin 43 (Cx43) and production of advanced glycation end products (AGE) in the testes was compared between control and T2D mice. Cx43 levels were increased specifically in the interstitial tissue containing Leydig cells, which are adjacent to the seminiferous tubules in the testes, suggesting altered intercellular communications between neighboring Leydig cells in the T2D testes ( Figures 5E–G ). Leydig cells are the primary source of testosterone or androgens in males, under the regulation of the hypothalamic-pituitary axis, and are vital for spermatogenesis (47). We also found an increased production of AGE in the testicular tissue from T2D males ( Figures 5H–I” ). As AGE are implicated in diabetes related complications, consequently, the formation of AGE may contribute to testicular dysfunction.

Having recognized the effects of T2D on testis morphology, we next wanted to assess the effects of the diabetic environment on mRNA expression in testicular tissue. Out of a selection of germ cell marker genes, we found two with altered expression: the expression of Sycp1 in spermatocytes was reduced and Prm1 levels were increased correspondingly with changes in the protamine ratio found in sperm from T2D males ( Figure 6A ). We also measured expression of a selection of blood testis barrier markers amongst which we found a significant increase in the expression of Cldn11, and Cdh2 in the testes from T2D males compared to control testes ( Figures 6B, C ).

To assess whether the paternal response to a diabetic environment affects the reproductive system of male offspring, males on either diet were mated to females. Fathers were removed after mating, limiting influence on their offspring only to the mating itself. The females were housed individually during the gestation period and the litter size was recorded. All females and the offspring generations were fed by the standard, control diet. Only male pups, the F₁ offspring, were left in the litters to be used for following up transgenerational studies. The F₂ offspring were produced from the mating of randomly selected F₁ males with control females (n = 10 males/group). To evaluate paternal effects and minimize any additional environmental influence, all molecular and cellular analyses of both F₁ and F₂ offspring were performed at 10 weeks of age, when adult male mice are considered sexually mature (48, 49).

First, to determine the potential effect of paternal T2D on the F₁ and F₂ generations, we examined physiological and metabolic changes in the offspring. At 10 weeks of age, the body and liver weights were significantly increased in the F₁ offspring of T2D males ( Figures 7A, B ). Most analyzed serum biomarkers were unaffected in the F₁ and F₂ offspring with the exception of glucose and ALT ( Figures 7C, D ). Plasma glucose levels were increased in the F₁ males at 10 weeks of age, however, these F₁ males had no impairment of glucose tolerance, as shown by a glucose tolerance test. Interestingly, ALT was increased in F₂ males, indicating a transgenerational effect of T2D exposure.

Next, we evaluated the effects of paternal T2D on the reproductive system and reproductive parameters of the F₁ and F₂ male offspring. The testicular morphology of the F₁ offspring from control and T2D fathers was comparable, including the thickness of the germinal epithelium and diameter of the seminiferous tubules. Accordingly, the mRNA expression of selected genes in the testes was not different between the F₁ offspring groups ( Supplementary Figure S1 ). However, the weight of testes of the F₁ offspring from T2D fathers was significantly smaller compared to the control F₁ offspring ( Supplementary Table S5 ). Additionally, significantly reduced fertility was found in the F₁ males from the T2D parental generation with a pregnancy rate of 63% compared to a 90% pregnancy rate in the F₁ control generation ( Table 4 ). Of eight F₁ males in the experiment, only five males were able to mate in the period of one week and produce progeny with the average litter size of 4.63 ± 1.41 (n = 5 litters) compared to the F₁ control cohort with the litter size of 6.44 ± 1.14 (n = 9 litters). The sex ratio of born pups was not significantly different ( Table 4 ).

Caudal epididymal sperm were collected from the F₁ and F₂ offspring for a sperm quality assessment. Similar to the paternal T2D males, we found no changes in sperm concentration, viability, and cell apoptosis in the F₁ and F₂ offspring generations ( Supplementary Table S6 ). The packaging quality of the sperm chromatin in the F₁ and F₂ offspring was not affected by the T2D parental exposure, as evaluated by chromomycin A₃ staining of sperm, and by protamine 1 and protamine 2 ratios ( Supplementary Table S6 ). However, abnormalities in sperm head morphology were increased in both the F₁ and F₂ generations from T2D fathers compared to the control offspring generations ( Figure 8A ). Additionally, the levels of TERA and GAPDS were reduced in the offspring of diabetic males in both F₁ and F₂ generations ( Figures 8B, C ), indicating not only intergenerational but also transgenerational transmission of negative effects of the diabetic environment.