Obesity is associated with increased seminal insulin and leptin alongside reduced fertility parameters in a controlled male cohort

Study design

At the pre-clinical stage, generally done via telephonic or electronic communication, interested participants were only excluded from the study if they had a history of vasectomy, any known reproductive tract pathology (e.g. genital tract infections, prostatitis, epididymitis, etc.), were on any hormonal therapy (e.g. testosterone, insulin, thyroid replacement) in the last six months, if they were hostpitalised or had any surgery in the last six months, had any pre-diagnosed chronic disease (specifically obesity related pathology such as Cushing’s syndrome, hypothyroidism and T2DM) or were on medications associated with increased obesity risk (e.g. antidepressant medications, cortisone, metformin, insulin, etc.) in the last six months. Those on medications for chronic disease risk parameters, such as hypertension, dyslipidaemia and coagulation (such as aspirin) were permitted into the consultation phase. Potential participants with a history of smoking or recreational drug use in the last six months were excluded from the study.

At the clinical stage, following a detailed description of the study and signing of the informed consent, a full standardised medical history and physical examination was conducted by a trained professional. This was followed by serum and semen sample collection and biochemical assessments. If any clinical or biochemical detection of acute or chronic disease was identified, patients were excluded from the study, and were provided with all results and appropriate advice and/or referral. This included clinically apparent reproductive disorders (e.g. varicocele; epididymitis; pelvic pain syndrome). Specific exclusion criteria based on biochemical data included serum glucose of > 7 mmol/L (indicating possible T2DM), azoospermia for any reason and leukocytospermia (defined as > 0.5 × 10⁶/ml) as an objective sign of reproductive tract infection/inflammation.

A trained clinician consulting with the participants recorded their age, body mass index (BMI), waist circumference (WC) and hip circumference (HC). Body weight (Kg) was taken on a digital scale to the first decimal point with patients in underclothes. WC was measured in centimetres around the abdomen at the midpoint between the highest point of the iliac crest and the lowest point of the costal margin. HC was measured in centimetres around the level of the greater trochanter. Both WC and HC were recorded as the mean of three measurements. The waist-to-hip ratio (WHR) was recorded as the WC divided by the HC.

Participants included in data analysis were divided into two groups based on the WHO definitions of obesity [1, 2]. Those with a BMI ≥ 30 were placed in the obese group, whereas participants with a BMI of < 30 were placed in the non-obese group. The non-obese group therefore consisted of normal weight (BMI = 18 – 24.9) and overweight (BMI = 25 – 29.9) participants.

Serum and semen collection

All samples were collected between 07:00 and 10:00. Venous blood samples were collected via venopuncture of superficial vessels in the antecubital fossa or hands by a trained clinician using sodium fluoride and serum separating tubule (SST) vacutainers®. Participants were expected to be fasting for a minimum of 8 hours prior to collection. Appropriate vacutainers were immediately transported to PathCare Laboratories (Pathcare Park, Goodwood, South Africa), a private commercial pathology laboratory servicing clinical practice and research sectors in South Africa, for insulin and glucose analysis using standard methods for clinical practice. The Quantitative Insulin Sensitivity Check Index (QUICKI) was calculated from fasting blood glucose and insulin concentrations [22]. An SST was centrifuged at 5000 × g for 10 minutes, and serum was transferred to cryovials and frozen at -20°C for a maximum of 6 months prior to leptin assays described below.

Semen samples were collection followed serum collection within 30 minutes via masturbation into sterile wide mouthed containers following a minimum of 3 days and maximum of 5 days abstinence from sexual activity or masturbation. Participants were instructed to collect all semen, and report any semen loss. Following semen analysis as described below, remaining sample was centrifuged at 5000 × g for 10 minutes, and seminal fluid was transferred to cryovials and frozen at -20°C for a maximum of 6 months prior to glucose, insulin and leptin assays described below.

Standard semen analysis

Seminal fluid was left for 60 minutes at room temperature to liquefy. After liquefaction, semen was transferred to a test tube in which ejaculate volume was recorded to the nearest decimal point. Sperm count and motility (progressive and total) was assessed using the Motility/Concentration module of the Sperm Class Analyzer® (SCA) CASA system version 4.1.0.1 (Microptic S.L., Barcelona, Spain). For analysis, a Nikon Eclipse 50i microscope (IMP, Cape Town, South Africa) equipped with phase contrast optics and a heated stage (37°C) was used. Sperm vitality was assessed using the eosin-nigrosin staining technique [6]. Morphology was assessed by the preparation of a smear and the application of the Papanicoloaou staining method as outlined by WHO [6], and determined by one person (RM) according to strict criteria as described by Menkveld and colleagues [23]. Leukocytes concentration was determined using the peroxidase staining technique as described by Politch and colleagues [24].

Functional sperm parameters

Spermatozoal mitochondrial membrane potential (MMP) was assessed as described previously [25] using a Zeiss fluorescence microscope (Oberkochen, Germany) for analysis after staining sperm with DePsipher staining kit (R&D Systems Inc., Minneapolis, MN, USA) used as a mitochondrial marker. In brief, semen was diluted 1:5 ratio with human tubal fluid medium (HTFM) prepared according to the method outlined by Quinn and colleagues [26], supplemented with 10 mg/ml Human Serum Albumin (HSA) (Sigma-Aldrich, St. Louis, MO, USA), and centrifuged for 10 minutes at 500 × g. The supernant was discarded, the pellet re-suspended in DePsipher staining solution and incubated for 20 minutes at 37°C in the dark. The DePsipher sperm suspension was then centrifuged at 500 × g, the supernatant was discarded, and the pellet re-suspended in 100 μl pre-warmed 1X reaction buffer. The cells were observed immediately with fluorescence microscopy at 1000-times magnification. Sperm exhibiting a green fluorescence within their mid pieces were regarded as having disturbed MMP, while those sperm showing red fluorescence were regarded as having intact MMP. The percentage of sperm with disturbed MMP was calculated.

Spermatozoa DNA fragmentation (DF) was assessed by the DeadEnd^™ Colorimetric TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labelling) System (Promega Corp., Madison, WI, USA) assay according to Henkel and colleagues [27], which end labels the fragmented DNA of apoptotic cells. A sample of liquefied semen was diluted in a 1:5 ratio with HTFM/BSA and centrifuged for 10 minutes at 500 × g. The pellet was re-suspended in PBS (Oxoid, Basingstoke, Hampshire, UK). A smear on a Superfrost® slide (Mentzel, Braunschweig, Germany) was made and allowed to air dry and accumulated for future analysis. All slides were analysed within 6 weeks of preparation. Prepared slides were fixed in 4% methanol-free formaldehyde (Sigma-Aldrich, St. Louis, MO, USA) in PBS for 25 minutes at 4°C. Slides were washed in fresh PBS for 5 minutes at room temperature, then sperm cells permeabilized in 0.2% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) in PBS for 5 minutes. After adequate rinsing of slides in fresh PBS, cells were allowed to equilibrate using the equilibration buffer (100 μl added to each slide) for 10 minutes. Slides were blotted around the equlibrilated areas and 20 μl TdT incubation buffer was added to an area of 5 cm² and covered with plastic slips. This was incubated in the dark at 37°C for 60 minutes and terminated using SSC diluted appropriately with deionised water for 15 minutes. The slides were washed in fresh PBS at room temperature 5 times for 5 minutes each, before draining excess water. Immediately following washing, DNA fragmentation was assessed by manual counting done using a Zeiss fluorescence microscope (Oberkochen, Germany). A minimum of 100 (those with poor sperm count) and maximum of 200 spermatozoa were counted on each slide and the results expressed as a percentage of cells showing green fluorescence indicating fragmented DNA (TUNEL-positive cells).

Seminal insulin

Seminal insulin was assayed using the Human Insulin ELISA Kit (RayBiotech, Inc., Norcross, Georgia, USA). This is an in-vitro ELISA based assay for the quantitative measurement of insulin. All reagents and frozen seminal fluid samples were thawed and brought to room temperature for analysis. The lower quantitative limit of the ELISA kit is 4 μIU/ml, with an intra- and inter-assay coefficients of variation (CV) of <10% and <12% respectively. All samples were assayed in duplicate, with the mean value recorded for data analysis. Samples were assayed on an ELISA reader obtained from BioTek (Winooski, VT, USA).

Seminal glucose

Seminal glucose was assayed using the Glucose HK Assay Kit (Sigma-Aldrich, St. Louis, MO, USA). This is an in-vitro ELISA based assay for the quantitative measurement of glucose. All reagents and seminal fluid samples were thawed and brought to room temperature for analysis. The CV of the kit based on correspondence with the supplier is 2.0%. If the duplicate samples are within 2% of each other, these were considered accurate and the mean value recorded for data analysis. All samples were assayed in duplicate, with the mean value recorded for data analysis. Samples were assayed on an ELISA reader obtained from BioTek (Winooski, VT, USA).

Serum and seminal leptin

Serum and seminal leptin was assayed using the Human Leptin ELISA Kit (RayBiotech, Inc., Norcross, Georgia, USA). This is an in-vitro ELISA-based assay for the quantitative measurement of leptin. All reagents and serum and seminal samples were thawed and brought to room temperature for analysis. The lower quantitative limit is 2 ng/ml, with an intra- and inter-assay CV of <10% and <12% respectively. All samples were assayed in duplicate, with the mean value recorded for data analysis. Samples were assayed on an ELISA reader obtained from BioTek (Winooski, VT, USA).

Statistical analysis

Statistical analysis was performed using the MedCalc software (Version 12.0; Mariakerke, Belgium). After testing for normal distribution using the Kolmogorov-Smirnov test, appropriate statistical tests, either parametric or non-parametric were performed. Parametric results are presented as mean ± SD, and non-parametric results are presented as median (range). All correlations were done using the Spearman correlation coefficient, with significant correlations reported as r². Fisher’s Exact Test was used in order to determine significant differences between groups based on cohort distributions and potential confounders between the groups. P-value of <0.05 was considered as significant with all statistical analyses.

Obesity and semen parameters

The results generally agree with literature demonstrating that obesity has a negative impact on standard semen parameters, although other studies indicate no relationship [5;12;14;15]. This is particularly evident by decreased sperm concentration and vitality, with BMI correlating negatively with sperm concentration total sperm count (TSC) and vitality. Furthermore, WC and HC correlated negatively with sperm concentration and vitality, but not TSC. In addition, there were lower mean values for TSC, total and progressive motility and morphology in the Ob group, although these did not reach statistical significance. WC and WHR correlated negatively with total motility, however, there was no correlation between morphology and BMI, nor any other sperm parameter, in this study. Details of previous studies are available in the referenced reviews and meta-analyses [5, 12, 14, 15].

Based on the WHO (2010) criteria for semen analysis [6], both groups had mean results above the recommended cut-off values for ejaculation volume (<1.5 ml), sperm concentration (<15 × 10⁶ million/ml) and TSC (<39 × 10⁶ million/ml). However, only the Ob group had mean percentages below the recommended cut-off values for total and progressive motility (<40% and < 32%, respectively) and vitality (<58%). Although the Ob group have higher percentages of the cohort presenting with oligozoospermia, asthenozoospermia and/or necrozoospermia, this difference was not significant (which may be due to the small sample size). Teratozoospermia (<4% normal morphology) was diagnosed in 78.6% of the cohort, and 68.4% of the nOB cohort and 86.9% of the Ob cohort. This high percentage in both groups may explain the lack of correlation between morphology and all other parameters in this study. Interestingly, 21.4% of the cohort had a history of recent couple infertility as defined by WHO (2010) [6]. However, it is important to note that no males or female partners had received any medical investigation for this complaint at the time of the clinical consultation. Although not statistically significant, a higher proportion of the Ob cohort (30.4%) reported this history, in contrast to the nOb group (10.6%). It must be clear, however, that this is defined as couple infertility and no further data on male or female factor infertility is available, and this should not be interpreted as male factor infertility percentages which were not established in the study. Furthermore, no males had consulted a medical or health care professional for infertility related complaints at the time of consultation.

Fewer studies have assessed the impact of obesity on DF and MMP. Both of these parameters were significantly increased in the Ob group compared to the nOb group. Furthermore, MMP and DF both correlated with BMI, WC, HC and WHR. The negative impact of BMI on DF and MMP confirms similar findings in previous studies [5, 14, 29]. A negative impact on MMP and DF in obesity appears to be a consistent variable in the literature when included for analysis. Damage to the sperm mitochondria function is suggested to negatively affect oxidative phosphorylation, reducing ATP synthesis and thus energy availability for motility [9]. MMP is negatively correlated with vitality and total and progressive motility in this study, supporting this well defined relationship. Although it can be hypothesised that obesity related phenomena, such as inflammation, may mediate damage to spermatozoa mitochondria and DNA integrity, the mechanisms of these relationships require further investigation. Serum and seminal insulin and serum glucose (but not seminal glucose) correlated with MMP and not DF, with a negative correlation between QUICKI and MMP.

Serum and seminal insulin and leptin

The concept of insulin resistance, closely associated with abdominal obesity and increased WC, is used to describe the process whereby target tissues develop impaired sensitivity to the action of the hormone, particularly in adipose tissue, liver and skeletal muscle [30–32]. The QUICKI is a useful assessment of insulin sensitivity, correlating closely with Euglycaemic Hyperinsulinaemic Clamp (a gold standard assessment of insulin resistance) across a wide range of glucose and insulin concentrations [22]. The predominantly adipocyte-derived polypeptide hormone Leptin regulates body weight, appetite and energy expenditure via hypothalamic modulation, in addition to modulation of the immune, endocrine, metabolic and reproductive systems [32, 33]. Leptin has been strongly associated with a role in the pathophysiology of obesity and metabolic syndrome, although this role has not been well understood or described [34]. Mean serum insulin and leptin was significantly increased in the Ob group as compared to the nOb group, with a significantly decreased QUICKI. As expected, serum insulin correlated positively with BMI, WC, HC and WHR. QUICKI correlated negatively with these same parameters. However, serum leptin did not correlate with these parameters as expected, possibly due to the relatively small sample size.

Various studies have assessed leptin concentrations in male serum, although an adequate reference range appears elusive. Raised serum leptin is also associated with sperm function changes in males [35, 36]. The serum ranges of leptin in this study generally agree with concentrations found in healthy, obese and infertile male cohorts reported in the literature [35–38].

Insulin and leptin are important regulators of male reproduction via the HPT axis both centrally and peripherally [17], in addition to being present in human semen with important regulatory roles for sperm function and fertilisation [17–20]. Furthermore, both hormones are synthesised and secreted by ejaculated spermatozoa in an autocrine manner [18–20]. Exogenous addition of insulin and leptin to ejaculated semen has been shown to increase total motility, progressive motility, acrosome reaction and nitric oxide production in human spermatozoa [21]. Sertoli cells too have been shown to synthesise and secrete insulin [39].

The results show that obese males have significantly increased seminal insulin and leptin as compared to the nOb group. Strong correlations between serum and seminal insulin (r² = 0.823), as well as serum and seminal leptin (r² = 0.517), are reported. The negative correlations between both seminal insulin and leptin with BMI, WC, HC and WHR may indicate that these parameters are potential predictors of seminal changes of these hormones. Seminal insulin and leptin further correlated negatively with QUICKI (r² = -0.782 and r² = -0.311 respectively). These correlations indicate an important relationship between pathophysiological mediators of obesity and a decrease in male reproductive potential that requires further investigation.

The source of seminal insulin and leptin is not clear. The correlations between serum and semen concentrations may suggest that insulin and leptin in the semen gains access to the reproductive tract via the blood testes barrier (BTB), seminal vesicles or prostate. Since insulin and leptin are strongly associated with increases in obesity, it is plausible that there is a cross over from peripheral circulation to the reproductive tract. Insulin and insulin-like peptides in human semen have previously been suggested to be secreted by the seminal vesicles [40, 41], and insulin appears to freely cross the BTB into the reproductive tract [42]. No apparent data on the source of leptin is available in the literature. In contrast, as insulin and leptin are synthesised and secreted in an autocrine fashion post-ejaculation, and Sertoli cells too secrete insulin within the testes [39], at least a local reproductive tract source of these hormones is also plausible. The source of these hormones in the reproductive tract requires further investigation.

Seminal and serum insulin was negatively correlated, and QUICKI positively correlated, with sperm concentration. The mechanisms for this are unclear. Type-1 diabetes mellitus, in which insulin is absent, is associated with a collapse of spermatogenesis and increased germ cell apoptosis [43]. Insulin resistance may mimic to some degree the loss of insulin in T1DM, and insulin resistance in Sertoli cells may hypothetically be associated with a decrease in spermatogenesis. As increased seminal insulin is associated with insulin resistance and abdominal obesity, increased insulin exposure during spermatogenesis may potentially develop insulin resistance in the Sertoli cells.

Although an acute in vivo increase in insulin and leptin exposure may increase motility and acrosome reaction in the spermatozoa [20, 21], this study did not show significant correlations between seminal insulin and leptin with sperm motility and vitality. Increased insulin exposure during spermatogenesis may plausibly develop insulin resistance within the spermatozoa themselves. Evidence to support this hypothesis may be found in the intracellular molecular cascades associated with insulin receptor stimulation in these cells. Insulin, as well as leptin, exert its effect on spermatozoa via the PI3K/Akt intracelleular signalling pathway, leading to protein kinase B (PKB) phosphorylation [20], which may mediate beneficial effects on ejaculated spermatozoa [17]. This pathway ultimately increases cellular nitric oxide production [17;20]. In human tissues, this intracellular pathway is negatively influenced in insulin resistance [31]. Hypothetically, over the spermatogenic cycle, it is conceivable that spermatozoa may develop insulin resistance in a manner similar to other tissue cell via a breakdown of the PI3K/Akt intracelleular signalling pathway. This hypothesis would provide an explanation as to the potential negative association between increased seminal insulin and reduced motility of ejaculated sperm. Although Lampiao & du Plessis [21] found an increase in motility of ejaculated spermatozoa exposed to leptin, this was not found by Li and colleagues [44]. However, if this intracellular pathway does breakdown in spermatozoa, we would also expect to see a reduced or even negative correlation between seminal leptin and ejaculated sperm function. A model in which insulin and leptin resistance is induced in spermatozoa and Sertoli cells is required to further investigate this potential relationship.

Based on the seminal-to-serum insulin ratio, insulin was found to be highly concentrated in human semen in both groups. The concentration of insulin in human semen is somewhat supported by a limited number of studies across numerous groups, including fertile and infertile normoglycaemic subjects, carbohydrate intolerant subjects and excretory and secretary azoospermic subjects [42, 45, 46]. No plausible explanation for a physiological concentration of insulin in semen is apparent on a search of the literature. As a result of the insulin concentration, the seminal glucose-to-insulin ratio was lower than the serum glucose-to-insulin ratios. Both the serum and seminal glucose-to-insulin ratios where significantly lower in the Ob group due to a significant increase in serum and seminal insulin compared to a more subtle increase in serum glucose and a non-significant decrease in seminal glucose. This is likely due to the exclusion of participants with a high fasting glucose.

Serum and seminal glucose

Glucose has been identified in human semen, with more than half the sugar consumed by ejaculated spermatozoa being in the form of glucose glycolysis [47, 48]. Reports on normal concentrations vary widely, from 1.02 mmol/L – 5.7 mmol/L [47]. Seminal glucose concentration ranges in this cohort (0.69 – 3.69 mmol/L) were within these reported ranges. Although there was a lower mean in the Ob group compared to the nOb group, this did not reach statistical significance. A small sample size may be the reason for this not reaching statistical significance. Sampling indicates possible significance with n = 44 in the nOb group and n = 54 in the Ob group.

Serum glucose correlated positively serum insulin and negatively with QUICKI as would be expected, and further correlated negatively with spermatozoa motility and positively with MMP, indicating an association negative relationship between serum glucose and energy production in spermatozoa. Seminal glucose correlated negatively to BMI only. There was no correlation between serum and seminal glucose either.

Seminal glucose concentrations were relatively lower compared to serum levels, as indicated in the seminal-to-serum glucose ratio. This may be due to a tight control mechanism for glucose to pass from the peripheral circulation into the reproductive tract through the BTB, reducing glucose concentrations in order to optimally support and maintain spermatogensis. Testicular cells have glucose sensing machinary which enable them to react and adapt to hormonal fluctuations and counteract hyper- or hypoglycaemic events, as spermatogenesis maintainance in vivo is dependent on adequate glucose metabolism [49]. Glucose transport across the BTB is mediated by various glucose transport molecules (GLUT’s), such as GLUT1, GLUT3 and GLUT8, and are sensitive to various hormones (including insulin), inflammatory cytokines and growth factors [49]. With changes in glucose or insulin, glucose transport machinary adapts in order to maintain lactate production [49]. Insulin deprived Sertoli cells in culture show decreased glucose uptake via the BTB barrier [49]. Therefore, it may be biologically pausible that insulin resistance in the setting of obesity may be associated with a decrease in glucose uptake across the BTB. This is however hypothetical, and further research in the physiology and pathophysiology of the BTB in relation to glucose is required. However, there was a non-significant trend for seminal glucose to be decreased in the Ob group, whereas serum glucose was significantly increased, which may warrent further insight.