Association Between Abnormal DNA Methylation and Altered Transcriptome in Muscle Five Years After Critical Illness

Critically ill patients requiring intensive care unit (ICU) admission suffer from muscle weakness that persists for years. Recently, altered RNA expression was documented in muscle of former ICU patients 5 years after critical illness that suggested disrupted mitochondrial function, disturbed lipid metabolism and fibrosis, of which many associated with the former patients' long-term loss of muscle strength. We hypothesized that abnormal DNA methylation detectable years after critical illness associates with these abnormal RNA expression patterns, as a potential biological basis for the persistent loss of muscle strength.

Genome-wide DNA methylation was assessed (Infiniumv2-HumanMethylationEPIC-BeadChips) in skeletal muscle biopsies from 118 former ICU patients harvested 5 years after critical illness (79.6% male, median 58 years, median BMI 27.3 kg/m²) and 30 controls who never required ICU admission (76.7% male, median 61 years, median BMI 26.4 kg/m²). Differentially methylated positions (DMPs) in former patients versus controls were identified, adjusting for age, sex, and BMI (minfi-package in R, Benjamini-Hochberg false-discovery-rate < 0.05), followed by pathway over-representation of affected genes. Spearman correlations between DMP methylation and RNA expression were compared among groups of RNA with Z-test and Kolmogorov-Smirnov test. Risk factors for abnormal DNA methylation were identified with multivariable linear regression.

As compared with controls, former ICU patients showed 7379 DMPs (average difference 2.6% ranging up to 24.9%). They were associated with 1334 unique genes, enriched for muscle contraction, vascular development, cell differentiation and signal transduction. DMPs correlated more strongly with differentially expressed RNAs (DERNAs) than with non-differentially expressed RNAs (18.1% vs. 1.7% correlations with |rho| > 0.3, p < 2.2 × 10^-16). Such correlations were more abundant among DERNAs associated with reduced muscle strength vs. those not associated (24.4% vs. 12.5%), also within the previously identified disrupted pathways (mitochondrial function 23.3% vs. 10.9%, lipid metabolism 15.9% vs. 7.2%, fibrosis 44.3% vs. 5.8%, all p < 2.2 × 10^-16). Older age, female sex, in-ICU treatment with glucocorticoids, benzodiazepines, early parenteral nutrition and opioids and insulin and antipsychotic medication at follow-up were most notably associated with more abnormal DNA methylation.

Abnormal DNA methylation in muscle biopsied 5 years after critical illness associated with long-term altered RNA expression that has been linked to lower muscle strength. These data suggest a possible epigenetic basis for this long-term sequel after critical illness. Abnormal DNA methylation was also found to associate with (possibly) avoidable risk factors during and after ICU stay. These findings may open perspectives for prevention and possibly treatment of long-term muscle weakness after critical illness.