Multi-organ gene expression analysis and network modeling reveal regulatory control cascades during the development of hypertension in female spontaneously hypertensive rat

Fig 1. High throughput multi-organ multi-pathway time series gene expression data analysis and network modeling during the development of hypertension. A schematic of experimental design and computational workflow. 3 female SHR rats were compared to 3 age-matched female WKY rats at 8, 10, 12, 16, and 24 weeks of age during the development of hypertension. High throughput qRT-PCR data were obtained from the adrenal gland, kidney, liver, lung, and spleen for genes involved in sympathetic (Symp.) pathways, renin-angiotensin (RAS), inflammation, fibrosis, and metabolism. These data were contrasted with previously published male SHR data for the genes in common between the datasets [13]. Computational modeling using Harley Modulating Functions (HMF) to inform peak expression time course analysis and the generation of multi-organ gene regulatory networks incorporated previously published data from brainstem regions of the same female SHR animals [28].

Fig 2. Stage-specific organ-specific drivers of transcriptomic variation across hypertension development in female SHR. A: Principal components (PCs) 1–3 for the female SHR and WKY multi-organ time series dataset containing 146 samples and 92 genes, colored for strain (Red: SHR, Blue: WKY). B: PC2 versus PC1 of the data subset for the 8-week time point, colored for strain, with the adrenal gland samples highlighted. C: Heat map of the top 20 genes positively and negatively contributing to PC2 at 8 weeks. D: PC2 versus PC1 of the data subset for the 10-week time point, colored for strain, with the kidney samples highlighted. E: Heat map of the top 20 genes positively and negatively contributing to PC1 at 10 weeks. F: PC2 versus PC1 of the data subset for the 16-week time point, colored for strain, with the kidney samples highlighted. G: Heat map of the top 20 genes positively and negatively contributing to PC2 at 16 weeks.

Fig 3. Model-predicted time course of multi-organ gene expression changes during the development of hypertension. The multi-organ gene regulatory network models for male and female SHR and WKY were simulated to predict the dynamic expression of the 92 genes across the inflammatory, RAS, metabolic, fibrotic and sympathetic pathways. The simulations were performed to span the ages from 8 to 24 weeks. A: Heat map of model-predicted expression levels of 92 genes across the 24-week time course in female SHR and WKY. The genes are ordered from top to bottom according to the time point of peak expression in SHR. B: Heat map of model-predicted expression levels of 17 genes in common between male and female datasets across the 24-week time course. The genes are ordered from top to bottom according to the time point of peak expression in female SHR.

Fig 4. Multi-organ gene regulatory networks in female SHR vs. WKY. A,B: Network of organ-organ interactions in SHR (panel A) and WKY (panel B). The edge thickness and color are mapped to the number of interaction coefficients that are at least two standard deviations from the mean, as a heuristic threshold for illustrative purposes. C: Heat map of regulatory matrix containing the interaction coefficients that are at least two standard deviations from the mean in female SHR (top) and WKY (bottom). The columns of the matrix correspond to the regulatory genes in the network in any given organ. The rows correspond to the target genes within and across organs. Positive interaction coefficients are colored red in the matrix, while negative interaction coefficients are colored blue. The highlighted columns represent selected key regulators in the network (described in the Results section). D: Total number of times each organ, adrenal gland (top) and CVLM (bottom), shows up as a regulator in panel C in SHR (left), WKY (right) or both (middle). E,F: Network of gene-gene interaction coefficients in SHR (panel E) and WKY (panel F) across adrenal gland and kidney. The edges shown correspond to the…

Fig 5. Comparison of multi-organ gene regulatory networks in male and female SHR vs. WKY. A,B: Network of organ-organ interactions in SHR (left) and WKY (right) (A: female, B: male). The edge thickness and color are mapped to the number of interactions within two standard deviations of the mean. C-D: Heat map of interaction coefficients within two standard deviations of the mean in SHR (top) and WKY (bottom) (C: female, D: male). E,F: Network of organ-organ interactions in SHR (top) and WKY (bottom) in female (panel E) and male (panel F). The edges shown correspond to the absolute interaction coefficients greater than 75% of maximum value (female SHR) and 25% of maximum value (female WKY, as well as male SHR and WKY). The edge thickness and color are mapped to the level of the interaction coefficients.

Fig 6. Multi-organ gene expression changes in catecholaminergic processes across the development of hypertension. A: Heat map of catecholaminergic-related gene expression across 146 samples in the female SHR and WKY dataset. B-E: Relative gene expression (- ΔΔCt) of Th (panel B) and Nts (panel C) (*strain p = 0.001) in adrenal gland as well as Adra1a (panel D) (strain p = 0.0508) and Adrb1 (panel E) (strain p = 0.02) in spleen in SHR (red) and WKY (blue). Th is significantly enriched in SHR adrenal gland at 8 (p = 0.002) and 12 weeks (p = 4 x 10−6) as compared to WKY. F: Differential expression of Adra1a in the SHR kidney relative to WKY in female (solid) and male (striped). Adra1a is significantly upregulated in male SHR kidney compared to female at 8 (p = 1 x 10−3), 12 (2 x 10−6), and 16 (0.01) weeks. Error Bars represent +/- standard error of the mean. * p < = 0.01.

Fig 7. Multi-organ gene expression dynamics related to the Renin-Angiotensin System across the development of hypertension. A: Heat map of RAS-related gene expression across 146 female samples. B-E: Relative expression (–ΔΔCt) of Ren (panel B) and Agtr1a (panel C) in adrenal gland as well as Agt in liver (panel D), and Ace in kidney (panel E), annotated for strain (SHR: red, WKY: blue). Adrenal Ren is significantly higher in SHR at 16 weeks compared to WKY (p = 0.03). Agtr1a is significantly higher in SHR adrenal gland compared to WKY (strain p = 2 x 10−4). Agt is significantly higher in SHR liver compared to WKY (strain p = 0.004). Ace is significantly lower in SHR kidney compared to WKY at 8 (p = 1 x 10−4), 12 (p = 0.02), 16 (p = 0.01), and 24 week (p = 0.003). F-I: Ren expression in SHR adrenal gland (panel F) (sex p = 4 x 10−4) and kidney (panel G), as well as Ace expression in SHR kidney (panel H) and Agt expression in SHR liver (panel I), all relative to WKY in female (solid) and male (striped). Ren is significantly upregulated in female SHR compared to male SHR adrenal gland (sex p = 4 x 10−4) and Ren is significantly upregulated in male SHR kidney as compared to female at 16…

Fig 8. Multi-organ gene expression dynamics related to inflammatory processes in female SHR. A: Heat map of inflammation-related gene expression across 146 samples in female SHR and WKY. B: Relative gene expression (–ΔΔCt) of Ephx2 across 5 organs, annotated for strain (SHR: red, WKY: blue). Error Bars represent +/- standard error of the mean. (*p < 0.05, ANOVA Tukey HSD post-hoc p-value).

Fig 9. Multi-organ differential gene expression statistically significant at the pre-hypertensive stage. A: Relative expression (–ΔΔCt) of Esr1 in adrenal gland (left) and lung (right), annotated for strain (SHR: red, WKY: blue). Esr1 is significantly upregulated in SHR adrenal gland compared to WKY (strain p = 0.002) and in SHR lung compared to WKY (strain p = 0.001). B: Relative expression (–ΔΔCt) of Tlr3 (left) and Bcl6 (right) in adrenal gland. Adrenal Tlr3 is significantly higher in SHR at 8 weeks as compared to WKY (p = 0.04). Adrenal Bcl6 is significantly higher in SHR at 8 weeks compared to WKY (p = 0.002). C: Relative expression (–ΔΔCt) of Hmgb1 (left) and Rhoa (right) in liver. Hmgb1 is significantly higher in SHR liver at 8 weeks as compared to WKY (p = 0.002). Rhoa is significantly higher in SHR liver at 8 weeks as compared to WKY (p = 0.04). D: Adrenal Il1b expression relative to WKY in female (solid) and male (striped) SHR. Il1b is significantly downregulated in male SHR adrenal gland compared to female at 8 weeks (p = 0.02). Error Bars represent +/- standard error of the mean. All p-values indicated are from a post-hoc Tukey Honest Significance test.

Fig 10. Multi-organ differential gene expression changes statistically significant at hypertension onset. A-B: Relative expression (–ΔΔCt) of Tnf (panel A, left), Il1b (panel A, middle), and Ccr1 (panel A, right) in kidney as well as of Ccr5 (panel B, left), Ilrn (panel B, middle), and Ltb4r (panel B, right) in kidney. Tnf is significantly upregulated in SHR at 10 weeks (p = 0.02). Il1b is significantly upregulated in SHR at 16 weeks (p = 0.01). Ccr1 is significantly upregulated in SHR at 10 weeks (p = 0.006). Ccr5 is significantly upregulated in SHR at 10 weeks (p = 0.03). Ilrn is significantly upregulated in SHR at 10 weeks (p = 0.002). Ltb4r is significantly upregulated in SHR at 10 weeks (p = 8 x 10−4). C: Il1b (left), Tnf (middle), and Tgfb1 (right) expression relative to WKY in female (solid) and male (striped) SHR kidney. Il1b is significantly downregulated in male SHR kidney compared to female at 16 weeks (p = 6 x 10−5). Tnf is significantly upregulated in male SHR kidney compared to female at 8 (p = 7 x 10−4) and 12 (p = 0.006) weeks, but downregulated at 16 weeks (p = 6 x 10−6). Tgfb1 is significantly decreased in male SHR kidney compared to female at 16 weeks of age (p =…