Aging is associated with structural and functional changes in the organism that result in the declining of its functioning. Postponed parenthood has renewed the interest in age-related decline of testicular function and male fertility. Still, little is known about the molecular mechanisms associated with testicular senescence and related decline of fertility. Here we sought to elucidate the molecular basis of metabolic changes associated with testicular aging and reproductive potential using an NMR-based metabolomics approach. Testicular metabolic profiles of rats from 3 to 24 months-of-age were analysed. An age-associated decrease in most antioxidant metabolites, like betaine, creatine and glutathione was observed. Amino acid content changed as early as 6 months-of-age, with an increase in branched chain and aromatic amino acids, accompanied by decrease of nucleotide synthesis (IMP, CMP, ATP). Testicular content of phospholipid precursors (choline, ethanolamine, myo-inositol, glycerol) increased with advanced age and was accompanied by a decrease in the levels of their phosphorylated products, suggesting compromised spermatogenesis. This is the first metabolomics study of testicular tissue of aged rats and we were able to identify metabolites associated with reproductive maturity from the onset to senescence. Our results provide evidence for an influence of aging on global testicular metabolome, as early as 6 months-of-age, with a profound alteration of several key metabolic pathways associated with the male reproductive potential.
69
Highlights
• Testicular metabolome of Wistar rats is profoundly altered with aging.
• Aging leads to a decrease in testicular content of most antioxidant metabolites.
• Branched chain and aromatic amino acid content increased with aging.
• Increase in testicular phospholipid precursors suggest compromised spermatogenesis.
• Aging changes key metabolic pathways associated with male reproductive potential.
Keywords: aging; male fertility; testicular metabolic profile; metabolomics; nuclear magnetic resonance.
Introduction
There is a growing number of couples that postpone parenthood due to various socioeconomic reasons imposed by modern living [1]. Although changes in male fertility are not as dramatic and definitive as the ones caused by menopause in women, aging promotes genetic and epigenetic changes in spermatozoa, which can lead to decreased sperm quality and impaired fertility. Indeed, increased parental age is reflected in their offspring and growing clinical evidence suggests they are more susceptible to congenital diseases, childhood cancers or neuropsychiatric disorders [2]. The decreased efficiency in testicular function associated with aging has been the subject of a great number of studies [3-6], but the molecular mechanisms by which it occurs remain a matter of debate.
Several age-associated changes have been reported in testicular physiology and hormonal profile, including alterations in testes (seminiferous tubular narrowing, vacuolization of Sertoli cells, decreased Leydig cell number and testosterone production), mitochondrial dysfunction, atherosclerotic alterations in testicular arteries, decreased testicular volume and germ cells depletion [5, 6]. As a result, sperm concentration and total number decrease, and their motility and morphology deteriorate in the aging male. The difficulty to obtain human testicular tissue for the study of age-related changes in the reproductive system, highlights the importance of the use of animal models in the attainment of data on the effects of aging on the testes.
Rodents (and particularly rats) became well and widely accepted models for probing human reproductive aging. Typically, rats live 35–40 months which makes it convenient to examine the mechanisms by which age-related changes occur in the absence of other confounding factors [7]. The interrelationship between age of laboratory rats and human is not yet well defined. In general, it is considered that, in adulthood, every day of the animal is approximately
70 equivalent to 34.8 human days (i.e., one rat month is comparable to three human years). The overall findings indicate that rats grow rapidly during their childhood and become sexually mature at about the sixth weeks of age, but attain social maturity 5-6 months later [7]. Rats reach puberty at an average age of 50 days after birth (50 days) and become sexually mature at 6 weeks (42 days)[7]. Nonetheless, the age of sexual maturity in male rats varies considerably between individuals, ranging from as young as 40 days to as old as 76 days after birth [8]. It is also important to mention that sexual maturity itself does not mark the beginning of adulthood, but rather shows the beginning of adolescence [7]. Thus, we used Wistar rats of different ages, ranging from 3 months-of-age, in order to ensure that the rats have completed at least one spermatogenic cycle and were sexually mature, to 24 months-of-age, to ensure that the study covered the reproductive senescence of rats.
The comprehensive study of metabolome has found growing implementation in aging-related studies. It was successfully applied in models of aging, but so far metabolic studies related to aging were mostly focused on the analysis of tissues like muscle, liver, brain, blood or in whole organisms. Although these studies greatly contributed to our understanding of age-related changes, as well as anti-aging manipulations, the insight into the aging induced metabolic changes in testes was until now not fully addressed in any animal model [11, 12]. Thus, we propose to unveil relevant molecular bases of senescence in the testes by performing a comprehensive study of global testicular metabolome during natural aging of male Wistar rats (Rattus norvegicus). This work represents the first study that applies a metabolomics approach to unveil age-related changes in the total metabolome of testes in a mammalian model.
Methods
Chemicals
Mammalian Protein Extraction Reagent and BCA Protein Assay Kit were purchased from Thermo Scientific (Whalthan, MA, USA). Dried milk was purchased from Regilait (Saint-Martin- Belle-Roche, France). ECF™ substrate was purchased from GE Healthcare (Weßling, Germany). NZYColour Protein Marker II was purchased from NZYTech (Lisbon, Portugal).
Other chemicals were all purchased from Sigma–Aldrich (St. Louis, MO, USA), unless stated otherwise.
Animals
Forty male Wistar rats (Charles River Laboratories, Barcelona, Spain) of different ages were used: 3 months (N = 8), 6 months (N = 8), 9 months (N = 8), 12 months (N = 8), and 24 months (N = 8). As previously mentioned, the starting age of the animal, 3 months old, was chosen in
71 order to ensure that these animals had already completed at least one complete spermatogenic cycle and were sexually mature. Animals were housed in our accredited animal facilities and maintained in type III-H cages (Tecniplast, Italy) with ad libitum food (standard chow diet, 4RF21 certificate, Mucedola, Italy) and water access, and at a constant room temperature (20±2ºC) on a 12-hours cycle of artificial lighting and noise level (<55 dB). All experiments were performed according to the “Guide for the Care and Use of Laboratory Animals” published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) and the European directives for the care and handling of laboratory animals (Directive 2010/63/EU). In accordance with the Portuguese law (Ordinance no. 1005/92 of 23rd October), the research team requested a permission to perform this animal experimentation study to the Portuguese Veterinarian and Food Department (ordinance nº 1005/92 of 23rd October). At the adequate age, animals were anesthetized and sacrificed by decapitation. After sacrifice, both testes of each animal were immediately removed and stored at −80°C until further analyses.
At the time of analysis, 50 mg of the tissue were processed for analysis by 1H-NMR spectrometry and 20 mg for protein extraction.
Protein extraction and quantification
Total protein was extracted from testicular tissue using Mammalian Protein Extraction Reagent (plus 1% protease inhibitor cocktail and 100 mM sodium orthovanadate). Protein concentration was determined by Pierce™ BCA Protein Assay Kit.
Western blot
Protein samples were prepared as described previously [13]. Briefly, proteins (50 μg) were mixed with sample buffer, denatured at 55°C and sonicated at 4°C. After incubation with the primary antibodies immune-reactive proteins were detected separately using the antibodies listed in Supplementary Table S1 using Bio-Rad GelDoc XR (Bio-Rad, Hemel Hempstead, UK). For the analysis of individual protein levels of OXPHOS complexes, 75 μg of proteins were mixed with sample buffer, and stirred for 15 min at 37°C. The protocol was followed as above, except that the proteins were fractionated in 15% polyacrylamide gels and membranes were blocked for 3 h in a 5% non-fat milk solution at room temperature. The membranes were then incubated overnight at 4°C with MitoProfile® Total OXPHOS WB Antibody Cocktail.
Mouse anti-tubulin (1:5000, T9026, Sigma-Aldrich, Rödermark, Germany) was used as protein loading control. In all cases, Quantity One Software (Bio-Rad, Hemel Hempstead, UK) was used to obtain band densities following standard procedures. The band density was divided by the respective tubulin band density and then normalized with the control group value.
72
Evaluation of oxidative damage levels
The levels of oxidative damage were established by evaluating the levels of protein carbonylation, protein nitration, and lipid peroxidation. The content of protein carbonyl group 2,4-dinitrophenylhydrazine adducts (DNPHZ), nitro-tyrosine (NitroT) and 4-hydrononenal (4- HNE) protein adducts in the samples from the different experimental groups was evaluated by Slot Blot using a Hybri-slot manifold system (Biometra, Göttingen, Germany) as described previously [14].
Metabolite extraction for metabolomics analysis
Testicular tissue extracts were prepared using a combined extraction of polar and nonpolar metabolites as previously described [15]. In brief, tissue was homogenized in a mixture of methanol and chloroform (2:1). After that, a combination of chloroform and water (1:1) was added and samples were centrifuged at 10,000g for 15 min at 4°C. The resulting supernatant was lyophilized and dissolved in D2O phosphate buffer (0.2 M, pH=7) for 1H-NMR analysis.
Nuclear magnetic resonance (NMR) spectroscopy and spectral analysis
1H-NMR spectra of the polar cellular extracts were obtained using a Varian Inova 600 MHz (14.1 T) spectrometer equipped with a 3 mm QXI probe with a z-gradient. 1H 1D noesy experiments with water presaturation were acquired at 298 K, using a 7.2 kHz spectral width, 0.1 s mixing time, 4 dummy scans, 4 s relaxation delay with 3 s of water pre-saturation (total relaxation time 7 s), 90o pulse angle, 3 s acquisition time and 128 scans. Pulse durations and water saturation frequencies were optimized for each sample. Spectra were processed by multiplying FIDs with exponential window function (line broadening of 0.3 Hz) and were zero filled to 64k points prior to Fourier transformation using TopSpin (Bruker Biospin, Karlsruhe, Germany). Spectra were manually phased and line corrected using a forth degree polynomial function, with chemical shifts internally referenced to the alanine signal at 1.48 ppm. To help spectral assignment two-dimensional spectra (TOCSY) were also acquired. TOCSY was acquired using sweep width of 5.4 kHz in both dimensions, 48 transients and 400 and 1024 points in t1 and t2 dimensions, respectively. It was processed by applying qsine window function and zero filled to 2048 point in both dimensions. Peaks were assigned by comparing recorded 1D and 2D spectra with reference spectra and public databases such as HMDB [16].
Metabolites are identified according to Metabolomics standards initiative (MSI) guidelines for metabolite identification [17] and the levels of identification are referenced in Supplementary Table S2.
73
Multivariate analysis of NMR data
After processing, 1H noesy spectra were bucketed using one-point bucket (all intensity values) in the region 0.6-9.0 ppm, with signal-free, water and fumarate regions excluded from data matrix used for multivariate analysis. Data matrix was built in Amix Viewer (version 3.9.15, BrukerBiospin, Rheinstetten). Bucketed spectra were aligned using icoshift algorithm [18] and probabilistic quotient normalization (PQN) [19] was applied. Multivariate statistical analysis was applied (SIMCA-P14.1, Umetrics, Sweden) on the aligned and normalized NMR matrix scaled to unit variance [20]. Principal Component Analysis (PCA) was used to provide qualitative information on the observed data set [21]. Partial Least Square-Discriminant Analysis (PLS-DA) was used to assess class separation and to identify the main metabolites that contribute to the class discrimination. PLS-DA models were validated by A 7-fold internal cross-validation and permutation test (n=100) [22]. The corresponding PLS-DA loadings were obtained by multiplying the loading weight factors (w) by the standard deviation of the respective variable, and were colour-coded according to variable importance to the projection (VIP). Qualitative information extracted from the PLS-DA loadings was complemented by the quantitative assessment of metabolite variations between groups through spectral integration.
Spectral peaks contributing to class separation were integrated in Amix-viewer, and normalized by the quotient computed from PQN. Statistical difference between normalized multiplet areas was assessed by Student’s t-test and all p-values were corrected for multiple comparisons using Benjamini-Hochberg False Discovery Rate (significance cut-off value of q=0.01 was used). The results are presented as bar plots as mean ± SEM. Additionally, for each metabolite, the magnitude of variation in each age group relative to 3 months-of-age group was assessed by calculating biologically relevant effect size (ES), adjusted for small sample numbers [23]. The metabolic changes with absolute ES larger than 0.5 (and with confidence interval of 95%) were represented in a heatmap plot (R-statistical software).
Identities of all differentially expressed metabolites were used to identify the most relevant metabolic pathways effected by aging using MetabAnalyst [24]. Pathway enrichment and topology analysis were used for metabolic pathway identification.
Statistical analysis
The statistical significance between the experimental groups was assessed by one-way ANOVA, followed by a Fisher’s LSD test. The experimental results are presented as mean ± SEM. Statistical analysis was performed using GraphPad Prism 6 (GraphPad software, San Diego, CA, USA). Possible outliers were removed using Grubbs’ method, α=0.2. p < 0.05 was considered as significant.
74