Fig 1.
In this study, we investigated the early effects of IH on physiological parameters (CBT, HR, BP, and activity) and tissue-specific transcriptome in male C57BL/6J mice. Physiological parameters were assessed in mice (n = 4 for each condition) after completion of 7 days of normoxia or IH exposure using the HD-X10 DSI telemetry system. We also collected several organs (liver, lung, kidney, muscle, heart, and cerebellum) from mice (n = 3 at each time point for each condition) on the second day of constant darkness for 24 hours with 3-hour intervals following exposure to a normoxic or IH condition. RNA was extracted from these samples and sequenced with NovaSeq S4 flow cell for a minimum of 40 million paired-end reads per sample. The data were then evaluated to assess for differences in mean 24-hour expression and circadian rhythmicity between normoxia and IH. Furthermore, we enriched the observed alterations among associated regulatory targets, hallmark pathways, and mammalian phenotypes. BP, blood pressure; CBT, core body temperature; IH, intermittent hypoxia; MESOR, Midline Estimating Statistics of Rhythm; HR, heart rate.
Fig 2.
Differential mean 24-hour expression.
To determine whether IH impacted mean 24-hour expression, we ran a DESeq2 analysis using the Kallisto R package. (A) The transcriptome changes within liver, lung, kidney, muscle, heart, and cerebellum were examined. Based on q ≤ 0.05 and a 1.5-fold difference, significant differences with up- (orange) or down (blue)-regulated genes have been screened and presented. (B) GSEA of consistently and significantly (q ≤ 0.05) associated regulatory targets with differential gene expression in at least 4 distinct tissues. The results indicated that in the lung, muscle, and heart, MIRs were primarily associated with up-regulated genes, while the same MIRs were associated with down-regulated genes in the kidney and cerebellum. (C) GSEA of hallmark pathways for differentially expressed genes upon IH exposure. Dot size indicates absolute NES. Statistical significance was determined by (q ≤ 0.05, GSEA). Hallmark pathways in the lung and heart are highly impacted compared to those in other organs. Some of these changes in hallmark pathways overlapped with processes in other organs. Significant (p ≤ 0.05, two-way ANOVA) changes in gene expression of PHDs and/or HIFs in (D) the lung and (E) the heart after exposure to normoxia and IH. Data output for Fig 2A can be found in S1 Data, for Fig 2B in S2 Data, for Fig 2C in S3 Data. Raw and processed data files for all figures in Fig 2 accessible through GEO series accession number GSE214530. GSEA, gene set enrichment analysis; HIF, hypoxia-inducible factor; IH, intermittent hypoxia; MIR, microRNA; NES, normalized enrichment score; PHD, prolyl hydroxylase.
Fig 3.
Differential circadian transcriptome.
(A) Differentially rhythmic genes were identified in the liver, lung, kidney, muscle, heart, and cerebellum after exposure to IH. Genes with loss or gain of rhythms were evaluated by combining MetaCycle (q ≤ 0.05) and CircaCompare (p ≤ 0.05). The circadian properties of the differentially rhythmic genes in both conditions were evaluated, including (B) amplitude, (C) acrophase, and (D) MESOR, using a cutoff threshold of q ≤ 0.05 of CircaCompare. Correlation plots were used to illustrate the differences between the 2 conditions. Significant changes in amplitude were observed only in the lung following IH exposure, while significant changes in MESOR were observed in all organs. (E) The lung heatmap illustrates those genes with significant (q ≤ 0.05 of MetaCycle and p ≤ 0.05 of CircaCompare) loss (top panel) or gain (middle panel) of circadian rhythms following IH exposure, as well as rhythmic genes under normoxic and IH conditions (bottom panel). (F and G) Among the canonical clock genes, Arntl, Clock, Npas2, Per3, Nr1d2, Rorc, and Tef in the lung exhibited significant changes (q ≤ 0.05, CircaCompare) in amplitude and MESOR following IH exposure. Additionally, Per1, Per2, Cry2, Hlf, and Bhlhe41 showed significant differences in MESOR. (H) Bioluminescence of PER2 protein in lung tissue from PER2Luc animals exposed to normoxia vs. IH showed significant increases in both amplitude and MESOR (p ≤ 0.01, unpaired t test) following IH exposure. Data output for Fig 3A-3G can be found in S4 Data. Raw and processed data files for Fig 3A-3G accessible through GEO series accession number GSE214530. Data for Fig 3H can be found in S5 Data. IH, intermittent hypoxia; MESOR, Midline Estimating Statistics of Rhythm; TPM, transcripts per million.
Fig 4.
Genes with differences in rhythms and mean 24-hour expression.
IH exposure resulted in significant changes in rhythmicity (q ≤ 0.05 of MetaCycle and p ≤ 0.05 of CircaCompare for presence of rhythmicity; q ≤ 0.05 of CircaCompare for differential circadian properties) and mean 24-hour expression (q ≤ 0.05 with 1.5-fold difference of DESeq2) of only a few genes in the liver, lung, and heart but not in the kidney, muscle, or cerebellum. In comparison with other organs, the lung demonstrated a higher number of altered genes and a significant (p ≤ 0.05, Fisher’s exact test) correlation between circadian rhythmicity and mean 24-hour expression. (A) Up-regulated genes with differential rhythms. (B) Down-regulated genes with differential rhythms. (C) Enrichment of genes to the mammalian phenotype with differential rhythms and changes in mean 24-hour expression across the tissues (p ≤ 0.01). Data output for Fig 4A and 4B associated with S1 Data and S4 Data. Data output associated with Fig 4C can be found in S6 Data. Raw and processed data files associated with Fig 4 accessible through GEO series accession number GSE214530. IgA, immunoglobulin A; IH, intermittent hypoxia; NK, natural killer.
Fig 5.
Physiological properties of mice (n = 4 for each condition) were assessed on the seventh day of their active phase (ZT12 to 22 hours) using an HD-X10 DSI telemetry system. We measured (A) CBT, (B) HR, (C) systolic and diastolic BP, and (D) activity using an average 1-hour bin analysis after exposing the mice to IH. Although systolic and diastolic BP and activity did not show any significant differences in any time bins between normoxia and IH, CBT and HR did show significant differences (p ≤ 0.05, two-way ANOVA) across several time bins. In addition, we also observed an overall significant increase (p ≤ 0.05, two-way ANOVA) in CBT following IH exposure. Data for Fig 5A–5D can be found in S7 Data. BP, blood pressure; CBT, core body temperature; HR, heart rate; IH, intermittent hypoxia; ZT, Zeitgeber time.