Biological sex, sex steroids and sex chromosomes contribute to mouse cardiac aging

Animals

C57BL6/J male and female mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA) or the aging colony of the Quebec Aging Research Network (RQRV) and kept for up to 24 months of age. They were housed on a 12 h light/12 h dark cycle with free access to chow and water.

Four-core genotypes (FCG) (B6.Cg-Tg(Sry)2Ei Srydl1Rlb/ArnoJ) breeders were purchased from Jackson Laboratory. These mice had already been on a C56Bl/6J background for at least eighteen generations. The breeding strategy consisted of mating a C57Bl6/J female with an XY Srydl1Rlb Tg(Sry)2Ei male [25, 26]. The progeny of this cross was either XX normal females (+/+), XY phenotypic females, XX phenotypic males or XY phenotypic fertile males (Sry/Tg(Sry2)). The four possible genotypes in FCG mice were characterized as follows: After gonadectomy, Xist expression by quantitative PCR was determined in the ovaries of females to discriminate between XY and XX animals. The size of testicles was used to identify XX males (small; about half of the standard size) and XY males (normal). The genotype was confirmed in all animals by measuring Xist expression in the left ventricle (LV) after euthanasia. Sequences of Xist primers were forward, TAA GGA CTA CTT AAC GGG CT and reverse, TAC TCA GAC ATT CCC TGG CA.

The protocol was approved by the Université Laval’s animal protection committee and followed the recommendations of the Canadian Council on Laboratory Animal Care (#2019-075 and #2020-603). This study was conducted according to ARRIVE guidelines.

Experimental design

The experimental design is schematized in Figure 1.

Figure 1. Schematic representation of the experimental design of the study.

Three-month-old mice: Sixteen 7-week-old mice (8 males and eight females) were used to form the young control groups. These mice were euthanized at the age of three months.

Twelve-month-old C57Bl6/J mice: Thirty 7-week-old mice (14 males and 16 females) were randomly distributed between 4 groups. Half of the animals were gonadectomized (Gx) at seven weeks [27].

Twenty-four-month-old C57Bl6/J mice: Forty-one 12-month-old mice (24 males and 17 females) were randomly distributed between 4 groups: males and females, Gx at 12 months or not.

Twenty-nine FCG 5-week-old mice were Gx and distributed between the four possible genotypes (Males: XY and XX, females: XX and XY). All mice had an echocardiography exam at two months and the day before euthanasia at six months of age.

Forty-six FCG mice were Gx at six months and distributed between the four genotypes. All mice had an echocardiography exam at 6, 12 and 20 months the day before euthanasia.

Animals were monitored daily by experienced technicians for health and changes in behaviour during the protocol. Mice display signs associated with poor prognosis of quality of life or specific signs of severe suffering or distress. Among those signs are significant loss or gain of weight, palpable tumours and grooming habits. All older mice over 12 months of age had available in their cage a running saucer (Innowheel™, Innovive, Billerica, MA, USA) as an environmental enrichment to avoid unwanted behaviours such as barbering.

Euthanasia was performed by total exsanguination of isoflurane-anesthetized animals, followed by cervical dislocation. Hearts were quickly removed, rinsed in phosphate-buffered saline, and weighed. The left atrium was then dissected and weighed. Lungs were also weighed.

Echocardiography

Unless specified above, all mice have an echocardiography (Echo) exam performed the day before euthanasia. The same investigator, blinded for mouse identification, acquired Echo images on a Vevo 3100 imaging system (VisualSonics, FujiFilm, Toronto, Canada). Transthoracic echocardiography was performed using a 40 MHz image transducer (MX550S) as previously described [27, 28]. Briefly, animals were anesthetized using isoflurane. The concentration of isoflurane was maintained around 2.5–3.5%. Images were taken from the LV parasternal long axis (PSLAX; B-mode and M-mode) view. PSLAX views (at least three consecutive heartbeats) were used to delineate the LV trace end-diastole and end-systole to obtain the averaged measurements of heart rate (HR), end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), ejection fraction (EF), and cardiac output (CO) with the VevoLab software (VisualSonics). The wall thickness (septal, IVSW; posterior, PW) and internal end-diastolic and end-systolic diameters (EDD and ESD) of the LV were measured in at least three consecutive cardiac cycles in PSLAX M-mode that were averaged. Fractional shortening (FS), relative wall thickness (RWT) and corrected LV mass were calculated using the Vevo Lab software. Images were also taken in pulse wave Doppler and tissue Doppler modes to measure blood velocity through the mitral valve and velocity of the mitral annulus, respectively. E wave, A wave and E/A ratio were estimated using the VevoLab software.

Histological analysis

Tissue preparation. OCT-embedded frozen cardiac tissue sections (long axis) were cut 10 μm thick and fixed on a microscope slide. LV sections were fixed using Bouin’s solution overnight, and the nuclei were stained using Wiegert’s hematoxylin for 10 minutes. Fibrosis was then stained using Picrosirius Red dye for 1 hour. Images of each section of the LV (interventricular septal wall (IVS), apex, posterior or free wall, and LVPW) were acquired using a wide-field microscope (Zeiss). Each of these sections was then analyzed using GIMP (GNU Image manipulation program; https://www.gimp.org/). Both microscopy and image analyses were performed by investigators who were blinded for mouse identification.

LV fibrosis assessment. Three sections of the LV (septum, apex, and free wall) were independently analyzed, and the mean percentage of interstitial fibrosis of the three sections was determined. Briefly, LV fibrosis of each section was measured by calculating the ratio of stained (red) pixels to the total number of pixels representing cardiomyocytes and fibrosis. The number of pixels corresponding to white sections (holes in the tissue) was measured and deducted from the total number of pixels of the image to get the total number of pixels representing the myocardium. Measures were acquired using the “by colour select” tool in the software.

Cardiomyocyte cross-sectional area (CSA). When acquiring the images, an LV image on which cardiomyocytes were easily visible and in the “en face” position was explicitly taken for cardiomyocyte CSA measurements. At least fifteen cardiomyocytes were measured for the three sections described above. The mean was calculated for each sample. Cardiomyocyte CSA was measured in pixels using the “free select” tool in the software. The pixel area was then converted to μm² by multiplying the number of pixels for one cardiomyocyte by the size in μm² of a single pixel. The microscope software automatically calculates a single pixel’s width and length in μm.

RNA preparation for mRNA sequencing

Total RNA was extracted from cardiac tissue samples using TRI Reagent (Sigma, Mississauga, Ontario, Canada). RNA from LV was removed for 3 to 4-month-old males and females, 12-month-old males and females, and all 24-month-old experimental groups. Samples were then treated with DNase I (RapidOut DNA removal kit, Thermo Fisher Scientific, Waltham, MA, USA), and RNA integrity was analyzed with the Agilent 2100 Bioanalyzer and RNA 6000 Nano kit.

Bulk mRNA-sequencing

LV samples, three to four pools of total RNA from two mice for each experimental group, were used to create libraries using Stranded mRNA Prep and Ligation kit (Illumina, San Diego, CA, USA) following the manufacturer’s recommendations. Libraries were analyzed using the Agilent 2100 Bioanalyzer and DNA 1000 kit. All samples were then quantified using the KAPA Library Quantification Kit (Kapa Biosystems). The average size of libraries was 335 base pairs. Indexed libraries were pooled and sequenced on the Illumina NextSeq 2000 (paired-end 150 bp) using sequencing-by-synthesis chemistry v4 according to the manufacturer’s protocols. We averaged 8.0 × 10⁷ (3.7 to 15.3 × 10⁷) paired-end reads, with an average of >95% of the reads achieving a quality score equal to or greater than Q30. Illumina BaseSpace RNA-Seq Alignment (STAR) RNA-Seq Differential Expression (DESeq2) software for sequence alignments and identifying differentially expressed genes (DEG) using the Mus musculus UCSC 10 mm reference genome. The DEGs were determined by adjusting the p-value for multiple tests using Benjamini–Hochberg correction with false discovery rate (FDR) ≤0.05 and Log2 fold change (Log2FC), |Log2 FC| ≥1.0. Gene set enrichment analysis was performed using the Panther GO Enrichment Analysis (https://geneontology.org/) [29–31]. A differentially expressed genes list (FDR <0.05 and |Log2FC| ≥1.0) was used to perform this analysis. Significant GO terms were determined using Fischer’s Exact test using Benjamini–Hochberg correction with FDR.

Statistical analysis

All data are expressed as mean ± standard error of the mean (SEM). Intergroup comparisons were conducted using the Student’s T-test using GraphPad Prism 10.0 (GraphPad Software Inc., La Jolla, CA, USA). Comparisons of more than two groups were analyzed using one-way or two-way ANOVA and Holm-Sidak post-test. P < 0.05 was considered statistically significant.

The heart gradually gains mass during life in female mice

Groups of male and female C57Bl6/J mice were sacrificed at the age of 3 (young), 12 (adult) and 24 months (old). As illustrated in Figure 2A, young male mice gained weight during the first ten months of follow-up (+46%) and then lost 12% of their body weight during the second year. In females, the 44% increase in body weight during the first year of life was followed by relatively stable values. We used tibial length as a surrogate of body growth. For both sexes, tibial growth was present up to 24 months. The overall growth was less in males (2.3%) than in females (5.7%) (Figure 2B). Body weight and tibial length were similar between male and female mice at 24 months. Heart weight (HW) in males increased by 30% (22% for indexed HW (HW/TL)) until 12 months, then remained stable after that. In females, HW gained 19% (17% for indexed HW) at 12 months and an additional 17% (21% for iHW) in the second year (Figure 2C, 2D).

Figure 2. Effects of aging on the heart. For each panel, graphs of males (blue and green) and females (orange and red) are represented side by side. (A) Body weight, (B) Tibial length, (C) Heart weight, (D) indexed heart weight for tibial length (HW/TL), (E) Diastolic thickness of the interventricular septum (IVSd) by echo, (F) End-diastolic LV diameter (EDD), (G) Left ventricular relative wall thickness (RWT) and (H) End-diastolic LV volume (EDV). Data are represented as mean +SEM. One-way ANOVA followed by Holm-Sidak post-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 and ^****p < 0.0001 between indicated groups.

Aging is associated with left ventricle concentric remodelling

An echo exam was performed the day before the euthanasia at 3, 12 and 24 months. Detailed echo data are described in Tables 1 and 2. In males, as illustrated in Figure 2E, end-diastolic interventricular septum (IVSd) thickness was markedly increased in adults compared to younger animals (+20%) and remained stable until 24 months. In females, the IVSd thickness increase was gradual up to 24 months (+17% at 12 months and +12% at 24 months). End diastolic LV diameter (EDD) was larger in males at 12 months (+7%) and remained unchanged in older animals. EDD had increased by 8% in females in two-year-old mice compared to young females (Figure 2F). In males, relative wall thickness (RWT), an index of LV remodelling, increased until 12 months and then remained unchanged. In females, the RWT increase was gradual over two years. Interestingly, RWT at 24 months was significantly higher in females than males (p = 0.024), suggesting a more concentric LV remodelling in females (Figure 2G).

End diastolic LV volumes (EDV) followed a different trend. In males, EDV was stable during the first year and then increased in 24-month-old mice (+26%). In females, a 20% increase in EDV was observed up to 12 months, remaining unchanged (Figure 2H). Detailed echocardiography data are listed in Tables 1 and 2.

Sex and age differences in the heart response to losing gonadal steroids

We then investigated how gonadectomy (Gx), either in young (7-week-old) or adult mice (12 months), would influence heart growth and LV morphology. We compared 12- and 24-month-old male and female mice after losing gonadal sex steroids for 10 and 12 months, respectively, to same-age control animals.

As illustrated in Figure 3A, 3I, the body weight (BW) of 12-month-old ovariectomized (Gx) males tended to be lower than for control adult males, whereas, in old males, BW was higher in Gx animals at 24 months. In females, Ovariectomy (Gx) of young female mice resulted in a higher BW 10 months later, whereas late Gx did not affect this parameter. Gx increased tibial length for both males and females at 12 months. Interestingly, this effect was also present in old Gx males (Figure 3B, 3I). The evolution of body weight in older mice after Gx is illustrated in Supplementary Figure 1A. It shows apparent sex differences. Cardiac growth was reduced in younger Gx males by almost 20%, whereas Gx had no effects on HW or indexed HW (iHW). In 2-year-old males or females, loss of sex hormones at 12 months resulted in lower HW compared to age-matched controls (Figure 3C, 3D and Supplementary Table 1).

Figure 3. Loss of gonadal steroid hormones in young and adult mice and its effects on the heart later in life. (A–H) Results are illustrated as a ratio of individual values of indicated parameters in gonadectomized (Gx) animals on the mean value for this parameter from a group of age-matched controls. The dotted line indicates a ratio of 1 or the absence of difference between Gx and non-Gx animals. Blue and green (males) and orange and red (females) indicate the ratio in adult (12 months) and old mice (24 months). (A) Body weight, (B) Tibial length, (C) Heart weight, (D) indexed heart weight for tibial length (HW/TL), (E) Diastolic thickness of the interventricular septum (IVSd) by echo, (F) End-diastolic LV diameter (EDD), (G) Left ventricular relative wall thickness (RWT) and (H) End-diastolic LV volume (EDV). (I) Representative pictures of picrosirius-stained LV sections for the indicated groups. (J) Interstitial myocardial fibrosis in aging male and female mice, intact or Gx. Data are represented as mean +SEM. (A–H) Student T-test comparison with age-matched controls. (J) One-way ANOVA followed by Holm-Sidak post-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 and ^****p < 0.0001.

Using cardiac ultrasound, we studied LV morphology changes caused by Gx. In younger males, LV septal wall thickness and diastolic diameter were reduced by Gx, resulting in a relatively stable RWT and reduced EDV. In older males, Gx had no effect during the second year of life (Figure 3E–3H). Gx increased LV septal wall thickness in younger females without affecting LV diameter. This resulted in increased RWT and a smaller LV chamber, as indicated by the reduced EDV. In older females, Gx slowed the continuing LV concentric remodelling during the second year of life, as evidenced by the decreased IVSd, RWT and EDV (p = 0.11) of about 10%. Supplementary Figure 1B illustrates the evolution of LV wall thickness in older male and female mice after Gx. Detailed echocardiography data for Gx animals are shown in Tables 1 and 2.

As illustrated in Figure 3I, 3J, aging was associated with increased interstitial LV fibrosis at 24 months. In males, Gx at 12 months reduced the development of interstitial myocardial fibrosis as observed a year later at 25 months, whereas in females, loss of sex steroids had no effect. At 12 months, interstitial fibrosis was not significantly higher than in young mice (not shown).

Sex chromosomes have limited effects on cardiac morphological changes associated with aging

To study the effects of sex chromosomes, we used the transgenic four-core genotypes (FCG) mouse line where the sex chromosome complement is dissociated from the sex phenotype. In these mice, the Sry gene, which is involved in developing the male phenotype, is no longer located on the Y chromosome but on an autosome (24). This allows male and female XX or XY mice to be obtained.

We studied these mice by echocardiography from 2 months up to 20 months. Mice that were euthanized at six months were gonadectomized (Gx) at the age of 5 weeks to remove the contribution of sex steroids during most of their life. Non-Gx FCG mice of the same age were euthanized as controls. In mice euthanized at 20 months, gonadectomy was performed at the age of 6 months to remove the effects of sex steroids later in life and to simulate, at least for females, menopause. Non-Gx mice were also studied for up to 20 months.

As illustrated in Figure 4A, in 6-month-old non-Gx mice, female animals had smaller bodies and indexed heart weights (HW/TL) than males, as expected. The number of X chromosomes did not affect these parameters. A small contribution of X dosage was observed for body weight and tibial length. Mice with an abnormal sex chromosome complement (XX for males and XY for females) tend to have lower body weight and shorter tibial weight.

Figure 4. Sex chromosomes significantly affect body growth but little on cardiac growth and LV remodelling during aging. Body weight, Tibial length, and indexed Heart weight for tibial length (HW/TL). A and B: Young mice (6 months). (A) non-Gx 6-month-old mice with one (X) or two X chromosomes (XX). Phenotypic males (solid blue dots) and females (solid red dots). (B) Gx 6-month-old mice (Gx at five weeks). Phenotypic males (blue circles) and females (red circles). (C, D) Old mice (20 months). (C) non-Gx 20-month-old mice. Phenotypic males (solid green dots) and females (solid orange dots). (D) Gx 20-month-old mice (Gx at six months). Phenotypic males (green circles) and females (orange circles). Data are represented as mean +SEM. Two-way ANOVA followed by Holm-Sidak post-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 and ^****p < 0.0001 between indicated groups. Variables are phenotypical sex (Sex) and one or two X chromosomes (X dosage). P-values below 0.05 are indicated on the graphs for the two variables. (E–H) LV wall thickness, End-diastolic LV diameter (EDD) and LV relative wall thickness (RWT) during aging by echocardiography. (E) Non-Gx males with one X chromosome (left; solid dots) or two X chromosomes (right; empty dots and gray background). (F) Gx males (MGx) at the age of 6 months. (G) Non-Gx females with two X chromosomes (left; solid dots) or one X chromosome (right; empty dots and gray background). (H) Gx females (FGx) at the age of 6 months. Data are represented as mean +SEM. One-way ANOVA followed by Holm-Sidak post-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 and ^****p < 0.0001 between indicated groups.

In Gx mice, having two X chromosomes resulted in higher body weight and longer tibial length. However, this effect did not translate to indexed heart weight, which remained identical for all four genotypes (Figure 4B).

In older mice at 20 months, X dosage effects were present in both Gx and non-Gx mice. In Figure 4C, non-Gx male mice had higher body weights, but those with two X chromosomes were heavier than their normal XY counterparts. The tibial length was also longer in XX mice. Indexed heart weights remained unchanged in male or female mice, notwithstanding their chromosome complement. In older Gx mice, having two X chromosomes resulted in heavier body weight and longer tibial length, but again, no differences were observed for heart weight (Figure 4D).

We then investigated by echocardiography if sex chromosomes affected LV morphology or if they could influence its remodelling during FCG mice aging. The animals had an echocardiography exam at 2, 6, 12, and 20 months. As illustrated in Figure 4E, in non-Gx males, X dosage did not affect LV wall thickening during aging and the associated concentric LV remodelling as indicated by LV relative wall thickness (RWT). This was also true in Gx males (Figure 4F). In FCG females, LV wall thickening during aging was slowed in XY mice, and this effect of X dosage was surprisingly more evident in non-Gx than in Gx females (Figure 4G, 4H). More complete echo data (at 2 and 20 months) are described in Supplementary Tables 2–5.

A transcriptomic study of left ventricular genes modulated during aging

We performed a bulk RNA sequencing (RNA-Seq) analysis from a piece of the LV mid-posterior wall from mice of both sexes at 3 (young) and 24 months (old; Gx or not). As mentioned above, we defined DEGs as genes having a |Log2 FC| ≥1.0. As illustrated in Figure 5A, 5B, 154 LV genes in males (Non-Gx and Gx) and 232 genes (Non-Gx and Gx) in females met this criterion when comparing 3-month-old mice to 24-month-old ones. Over 80% of these modulated genes were upregulated in older animals. We then identified the shared genes upregulated during aging in either male or female mice at 24 months of age (non-Gx vs. Gx) compared to young animals. Venn diagrams in Figure 5A, 5B list the 20 most-expressed genes for each comparison. Male and female mice shared 38 up-regulated genes by aging and six down-regulated (Figure 5C).

Figure 5. Transcriptomic changes in the left ventricle induced by aging and gonadectomy. Differentially expressed genes (DEGs, |Log2 fold change| ≥1) between 3-month-old (3 m) and 24-month-old males (24 m) (A) or between 3-month-old and 24-month-old female (B) mice, intact or Gx. (A) Venn diagrams showing DEGs specific to aging intact or Gx males as well as shared genes either up (left) or down-regulated (right). Venn diagrams showing DEGs specific to aging intact or Gx females and common genes either up (left) or down-regulated (right). Below the Venn diagrams are lists of up to twenty genes ranked by their mean expression levels for each comparison from highest to least expressed. Legend for the background colouring of the gene list based on mean expression count is illustrated. In boldface are listed genes that are common between males and females. (C) Venn diagrams for the common upregulated genes (left) and downregulated genes (right) by aging between males and females. Only genes having a mean normalized count over 50 from RNA-Seq data were considered.

LV aging highlighted the enrichment of differentiated expressed genes (DEG) for several biological processes and cellular components. Figure 6A, 6B describes a list of biological processes where enrichment of DEG is observed for comparison between young and old mice (Gx or not). Interestingly, in males, only one biological process is shared between intact and Gx old males compared to their young counterparts. Ion transport processes are more present in non-Gx males, whereas, in Gx males, it was linked to inflammation. In females, biological processes were similar between Gx and non-Gx old female mice. The cellular components described in Figure 6C, 6D for male and female old mice were associated with the cell membrane and the extracellular region. Only a few genes were differentially expressed when comparing old mice groups (Gx vs. non-Gx), as listed in Supplementary Tables 6–8.

Figure 6. Gene ontology (GO) lists biological processes (A, B) and cellular components (C, D) from differentially expressed left ventricle genes between young and older mice represented as bubble plots. All genes significantly modulated (|Log2 fold change| ≥1) were included. The list is limited to the top 10 processes (listed below the graph) having the lowest false discovery rate. Number (Nb) of genes included for each category (bubble size, legend on the right of each graph), fold enrichment over expected representation (Y-axis) and false discovery rate (X-axis).

Myosin high chain beta (Myh7), Myotilin (Myot), 3-Hydroxybutyrate Dehydrogenase 1 (Bdh1), periostin (Postn) and the carboxypeptidase x, m14 family member 2 (Cpxm2) were five up-regulated genes shared between male and female mice having the highest levels of expression in the LV. (Figure 7A, 7B). Slc38a1 (Solute Carrier Family 38 Member 1) was the most abundant down-regulated mRNA at 24 months in both sexes (Figures 5A, 5B and 7C, 7D).

Figure 7. Gene expression of common LV genes up-regulated during cardiac aging in male (A) and female (B) mice. Myh7, Myosin heavy chain beta, Bdh1, 3-Hydroxybutyrate Dehydrogenase 1, Myot, Myotilin, Postn, Periostin, Cpxm2, Inactive Carboxypeptidase-Like Protein X2. (C) (males) and (D) (females). Downregulated genes and genes have different modulations with aging between Gx and non-Gx animals or between males and females. Slc38a1, Solute Carrier Family 38 Member 1, CD74, CD molecule, Fmo2, Flavin Containing Dimethylaniline Monooxygenase 2 and Zbtb16, Zinc Finger and BTB Domain Containing 16. Results are expressed as the mean ± SEM (n = 4). One-way ANOVA followed by Holm-Sidak post-test. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001 and ^****p < 0.0001 between indicated groups.

Several highly expressed genes had levels not only modulated by age but also by the loss of sex steroids in elderly mice, such as CD74, which was more modulated in Gx animals, and Fmo2 and Zbtb16, showing the opposite trend (Figure 7C, 7D).