Transcriptomic analysis identifies a role of PI3K-Akt signalling in the responses of skeletal muscle to acute hypoxia in vivo

Abstract

Key points: Changes in gene expression that occur within hours of exposure to hypoxia in in vivo skeletal muscles remain unexplored. Two hours of hypoxia caused significant down-regulation of extracellular matrix genes followed by a shift at 6 h to altered expression of genes associated with the nuclear lumen while respiratory and blood gases were stabilized. Enrichment analysis of mRNAs classified by stability rates suggests an attenuation of post-transcriptional regulation within hours of hypoxic exposure, where PI3K-Akt signalling was suggested to have a nodal role by pathway analysis. Experimental measurements and bioinformatic analyses suggested that the dephosphorylation of Akt after 2 h of hypoxic exposure might deactivate RNA-binding protein BRF1, hence resulting in the selective degradation of mRNAs.

Abstract: The effects of acute hypoxia have been widely studied, but there are few studies of transcriptional responses to hours of hypoxia in vivo, especially in hypoxia-tolerant tissues like skeletal muscles. We used RNA-seq to analyse gene expression in plantaris muscles while monitoring respiration, arterial blood gases, and blood glucose in mice exposed to 8% O₂ for 2 or 6 h. Rapid decreases in blood gases and a slower reduction in blood glucose suggest stress, which was accompanied by widespread changes in gene expression. Early down-regulation of genes associated with the extracellular matrix was followed by a shift to genes associated with the nuclear lumen. Most of the early down-regulated genes had mRNA half-lives longer than 2 h, suggesting a role for post-transcriptional regulation. These transcriptional changes were enriched in signalling pathways in which the PI3K-Akt signalling pathway was identified as a hub. Our analyses indicated that gene targets of PI3K-Akt but not HIF were enriched in early transcriptional responses to hypoxia. Among the PI3K-Akt targets, 75% could be explained by a deactivation of adenylate-uridylate-rich element (ARE)-binding protein BRF1, a target of PI3K-Akt. Consistent decreases in the phosphorylation of Akt and BRF1 were experimentally confirmed following 2 h of hypoxia. These results suggest that the PI3K-Akt signalling pathway might play a role in responses induced by acute hypoxia in skeletal muscles, partially through the dephosphorylation of ARE-binding protein BRF1.

Keywords: gene expression; hypoxia; skeletal Muscle.

Figure 1. Timeline of fasting, hypoxia exposure and tissue collection in hypoxic and control animals

The four experimental groups (6 mice per group) were control (normoxia), and 15 min, 2 h and 6 h of hypoxia. Mice were kept in individual tubes with free access to water. The food supply was removed 15 min before 9:00 on the experimental day. The first mouse in the 6 h hypoxia group was exposed to 8% O₂ starting at 9:00, followed by the other 5 mice at 5 min intervals. The first mouse in the 6 h hypoxia group was removed from the hypoxic tube covered with a mask filled with 8% O₂ at 15:00 and rapidly killed by cervical dislocation. Muscles were dissected rapidly within 5 min of removal from the tube, so that killing and tissue collection from successive mice in each group could be performed at 5 min intervals. Mice in the 2 h hypoxia group were exposed to 8% O₂ starting at 11:00, with killing and tissue collection beginning at 13:00. Mice in the 15 min hypoxia group were exposed to 8% O₂ starting at 14:00, with killing and tissue collection started at 14:15. Control mice were exposed to normoxia, until 13:30 when tissue collection was initiated at 5 min intervals for each mouse. Thus, every mouse had a well‐controlled duration of hypoxic or normoxic and a comparable fasting duration. Normoxic exposure, white bars; hypoxic exposure, grey bars. Each vertical line indicates the time when a mouse was moved into a hypoxia tube. Each dotted line shows the time when a mouse was removed from hypoxia.

Figure 2. Effects of hypoxia (8% oxygen) on ventilation in mice (n = 6)

A, oxygen consumption rate (${\dot{V}}_{{\mathrm{O}}_2}$); B, carbon dioxide production rate (${\dot{V}}_{\mathrm{C}{\mathrm{O}}_2}$); C, ventilation rate (${\dot{V}}_{\mathrm{I}}$); D, ${\dot{V}}_{\mathrm{I}}$/ ${\dot{V}}_{{\mathrm{O}}_2}$ ratio; E, ${\dot{V}}_{\mathrm{I}}$/ ${\dot{V}}_{\mathrm{C}{\mathrm{O}}_2}$ ratio. ^*Significantly different from 0 min (P < 0.05).

Figure 3. Effects of sustained hypoxia on blood glucose (A) and blood lactate (B) in mice (n = 6)

^*Significantly different from the control point (P < 0.05).

Figure 4. Lactate levels in plantaris during hypoxic exposure

Plantaris samples were collected from normoxia (n  = 6), 15 min (n  = 3) and 2 h (n  = 6). All values are means ± SD.

Figure 5. Fold‐changes of selected genes measured by RT‐PCR (n = 6) and RNA‐seq (n = 3)

A, fold‐changes in gene expression after 2 h of hypoxia. B, fold‐changes in gene expression after 6 h of hypoxia.

Figure 6. Half‐life distributions of differentially expressed genes that decreased their expression by more than 50% of control after 2 and 6 h of hypoxia

mRNA half‐life ranges of 0–2 h, 2–4 h, 4–6 h, 6–12 h and longer than 12 h are shown. Charts are annotated with the total number of differentially expressed genes in each group.

Figure 7. Hierarchical clustering of differentially expressed genes and corresponding functional annotation clustering the identified sub‐clusters

A, hierarchical clustering highlights four sub‐clusters of differentially expressed genes: fast down‐regulated, slow down‐regulated, fast up‐regulated, and slow up‐regulated. The colour indicates the relative level of gene expression compared with the mean value of all samples. B, GO cellular component annotations of identified sub‐clusters. The colours indicate the four different sub‐clusters, with an enrichment score in the range of 0–10. C, functional annotation of sub‐clusters with enrichment scores greater than 2. The name of the cluster is defined by the name of the child‐cluster with the smallest P value in that cluster. The colour indicates the database resource.

Figure 8. Gene changes in PI3Ks and ECM in 2 and 6 h of hypoxia

The left side of gene boxes indicates the fold‐change of genes in 2 h of hypoxia and the right side of gene boxes indicates the fold‐change in 6 h of hypoxia. The colours stand for the fold‐changes in gene expression.

Figure 9. Western blot analysis of selected signalling molecules and their phospho‐proteins

Plantaris muscles were collected from C57BL/6 mice exposed to normoxia (n  = 3), 8% O₂ for 15 min (n  = 2), 2 h (n  = 3) or 6 h (n  = 3). The abundance and relative phosphorylation levels of AMPK, Akt, ERK, p38 and JNK were measured using antibodies with tubulin as the reference protein. A, Western blot results for pAkt, pP38 and pAMPK. B–D, statistical results for pAkt, pP38 and pAMPK. All values are means ± SD. ^*Significantly different from normoxic control (P < 0.05).

Figure 10. Immunoprecipitation measurements of BRF1

Plantaris samples were collected from normoxia (n  = 10), 15 min (n  = 10), 2 h (n  = 9) and 6 h (n  = 3). All values are means ± SD. ^*Significantly different from normoxic control (P < 0.05).