Introduction

Sepsis is a life-threatening dysregulated host immune response to infection that leads to multisystem organ damage1. Pathophysiological features include a complex network of pro-inflammatory responses, immune suppression, and endothelial dysfunction and cell damage from uncontrolled inflammation, which includes myocardial inflammatory responses2,3. This systemic immune response leads to septic shock with circulatory maldistribution associated with peripheral vasodilation, arterial and capillary shunting, and sepsis-induced myocardial dysfunction leading to organ damage and high risk of death. Septic cardiomyopathy has been defined as reversible acute uni- or bi-ventricular systolic or diastolic dysfunction with reduced contractility not caused by ischemic coronary heart disease4.

MicroRNAs (miRNAs) are a class of non-coding RNA genes that regulate mRNA expression to affect various physiological processes through the adjustment of protein levels5. MiRNAs regulate multiple pathways that form a complex network. Single miRNAs can target hundreds of genes, inhibiting or promoting inflammatory and innate immune responses, a key role in sepsis-induced endothelial and organ dysfunction, including sepsis-related myocardial dysfunction6,7,8. In addition, circulating miRNAs can act as signaling molecules and induce hormone-like responses in target cells and tissues9,10.

To date, human studies on cardiac miRNA tissue levels in sepsis are lacking. We were interested in the expression of miRNA in postmortem myocardial specimens of critically ill septic patients and non-septic controls.

Methods

Subjects

This retrospective observational study was conducted at Oulu University Hospital, Oulu, Finland, an academic tertiary referral hospital. The study protocol was approved by the hospital administration and the ethics committee of Oulu University Hospital and the Northern Ostrobothnia Hospital District (72/2021; 33/2021). All research was performed in accordance with relevant guidelines and regulations in accordance with the Declaration of Helsinki. The use of myocardial tissue obtained from the medicolegal autopsies as case control samples was approved by the Finnish Institute for Health and Welfare (THL/873/5.05.00/2023; THL/697/5.05.00/2017). Consent from next of kin or from the donor was waived by the Ethics Committee since according to the Finnish law, during post-mortem examinations, samples removed can also be used for medical research other than that related to investigation of the cause of death.

All adult non-cardiac surgical non-neurosurgical intensive care unit (ICU) patients with emergency admission who died with sepsis during the years 2015–2020 and underwent clinical autopsy with available myocardial specimens were included in this study. Those sepsis cases that died with malignancies, cardiac complications, or had forensic investigation performed were excluded. The specimens were collected during autopsy for routine histological examination and preserved in the Biobank Borealis of Northern Finland.

Myocardial tissue samples for non-septic controls were collected during medicolegal autopsies at the Finnish Institute for Health and Welfare. Finnish law requires a medicolegal autopsy to be carried out when the death is caused by non-natural conditions, such as accident, suicide, or homicide, or when the death is unexpected. The selected control subjects had either trauma or coronary artery disease (sudden cardiac death) as the underlying cause of death, and they were matched for age, sex, and presence of coronary artery disease.

Sepsis was defined as life-threating organ dysfunction caused by dysregulated host response to infection according to the American College of Chest Physicians/Society of Critical Care Medicine Criteria1. All ICU sepsis patients were treated by a multidisciplinary team of intensivists and infectious disease specialists in our 26-bed, closed adult ICU. Intensive care treatment was performed according to normal ICU protocols and sepsis guidelines. The decision to perform a postmortem examination was made at the discretion of the treating physician or following a request from the next of kin.

Clinical data collection

Clinical parameters were collected from the ICU clinical data management system database (Centricity Critical Care, Clinisoft, GE Healthcare). Illness severity was determined by the Acute Physiology and Chronic Health (APACHE II) score, which was recorded on admission. Organ failure was defined by the Sequential Organ Failure Assessment (SOFA) score, which is commonly used in ICU settings11,12. Data were obtained regarding age, sex, focus of infection, chronic illnesses, vasoactive use, and other ICU treatments, as well as the length of stay in the ICU and hospital. Clinical laboratory and microbiological samples were analyzed by commercially available laboratory methods in the hospital’s accredited central laboratory (NordLab, Oulu University Hospital, Oulu, Finland). The control tissue samples were linked to the subjects’ limited health data, which included age, sex, and major chronic diseases.

Autopsy material collection and tissue sample preparation

The autopsies of the sepsis cases were performed by the Department of Pathology in Oulu University Hospital. Following the cessation of vital functions, post-mortem preparations were conducted within the confines of the intensive care ward prior to the cadaver’s transfer to the morgue. The time at ambient room temperature typically lasted a couple of hours. At the morgue, cadavers were stored at a temperature of 4 °C. The median time from death to autopsy was 5 days (range: 3–6 days). During the autopsy, cardiac specimens were routinely collected from the septum and left ventricular free wall. Standard procedures were used for preparation. Specimens were fixed in 10% formalin for approximately 24 h. After fixation, the specimens were processed in a tissue processor, dehydrated in 100% alcohol, and cleared with xylene before being embedded in paraffin and stored at room temperature. The medicolegal autopsies of the control cases were performed in the Forensic Medicine Unit, Finnish Institute for Health and Welfare, Oulu, Finland, and the tissue preparations at the Department of Forensic Medicine, University of Oulu, Oulu, Finland. The specimen preparation was done according to the standard procedures mentioned above with the main difference of an estimated time in temperatures above 4 °C before forensic investigation of the cadaver. Control cases’ time of death was estimated by forensic medicine. In this study, we present postmortem intervals as the estimated time when bodies have been at ambient room temperature before moving to the cold room.

Tissue samples were preserved as formalin-fixed paraffin-embedded (FFPE) blocks due to their excellent tissue morphology preservation over time. This method prevents tissue autolysis, enabling histological analysis. Isolating nucleic acids from these samples is important for understanding disease mechanisms at the genomic level. MiRNAs are very short (18–23 nucleotides) and less affected by fixation/embedding, allowing their recovery from FFPE samples even after many years, possibly decades. The expression levels of isolated miRNAs are directly comparable in both frozen and FFPE tissue samples13,14,15,16,17. FFPE cardiac tissue specimens and the quality of cardiac tissue specimens were verified by hematoxylin-eosin staining and microscopic analysis before being sectioned for miRNA analysis, using a fresh microtome blade for each block.

RNA isolation and RNA sequencing

RNA was isolated from FFPE cardiac samples using the RecoverAll Total Nucleic Acid Isolation Kit for FFPE (Invitrogen). For miRNA sequencing the RealSeq Single Index kit (RealSeq Biosciences) was used for library preparation, and sequencing was performed on Novaseq SP100 (Illumina) using single end reads. Sequencing data were analyzed using the miND® analysis pipeline for miRNA data and the meND v1.2.8 analysis pipeline for the 3’mRNA data18. Shortly following miRNA sequencing the reads were adapter trimmed using cutadapt v3.3 and mapped to GRCh38.p12 and miRBase v22.1 with bowtie v1.3.0 and miRDeep2 v2.0.1.2.

For 3’mRNA sequencing, the QuantSeq 3’ mRNA-Seq FWD and UMI Add-on (Illumina) was used for library preparation, and sequencing was performed on NovaSeq SP (Illumina). UMI sequences were extracted from 3’mRNA sequencing data and reads were trimmed and quality filtered using bbduk from the bbmap package v38.69 and filtered for a minimum length of 17nt and phred quality of 30. Alignment steps were performed with STAR v2.7 using samtools v1.9 for indexing and reads were mapped against the genomic reference GRCh38.p13. Deduplication of reads was performed after mapping using the UMI-tools package v1.1.2. Assignment of features to the mapped reads was done with htseq-count v0.13. The quality of the NGS data was evaluated using fastQC v0.11.9 and multiQC v1.14.

Differential expression analysis

Differentially expressed miRNAs were identified using edgeR v3.32. The target genes for each miRNA were identified utilizing the miRNA target filter within Ingenuity Pathway Analysis (IPA) software (Qiagen) by incorporating target predictions from TargetScan human, Ingenuity Expert Findings, miRecords, and Tarbase databases. Predicted miRNA targets were retrieved by IPA software, and pathway enrichment and classification were performed using PantherDB (www.pantherdb.org/, Qiagen Ingenuity Pathway Analysis, Version 12210362, https://digitalinsights.qiagen.com/products-overview/discovery-insights-portfolio/analysis-and-visualization/qiagen-ipa/, edgeR, (“Empirical Analysis of Digital gene Expression Data in R”), v3.40 https://bioconductor.org/packages/release/bioc/html/edgeR.html). The target filter prioritized targets that were either experimentally observed or predicted with high confidence, ensuring the exclusion of less reliable predictions from the target gene lists. Enrichment analysis for the target genes of each miRNA was performed using PantherDB’s overrepresentation test (GO biological process complete).

Differential expression analysis of 3’mRNA sequencing was performed with edgeR v3.40 and independent filtering method of DESeq2 was adapted for use with edgeR to remove low abundance genes and thus optimize the false discovery rate (FDR) correction.

.

miRNA-mRNA interaction analysis

Sets of down- and up-regulated miRNAs and mRNAs served as input for the multi-omics analysis, identifying patterns of miRNA regulation. To integrate information of up-and downregulated miRNAs and mRNAs, the miRNA target prediction tool miRNAtap (Pajak M, Simpson TI. miRNAtap: miRNAtap: microRNA Targets - Aggregated Predictions. R package version 1.38.0., 2024) provided predicted mRNA targets for the set of regulated miRNAs. Predicted mRNAs were subsequently mapped to the regulated mRNAs, revealing miRNA-mRNA interaction patterns.

Statistical analysis

For continuous variables, we used medians with 25th -75th percentiles or means with standard deviations; and counts (%) were used for categorical variables. Comparisons of continuous variables were analyzed using the t-test for independent samples. We also report the 95% confidence intervals (CIs) of the differences between the mean miRNA counts.

MiRNAs were considered up- or down-regulated if the FDR was < 0.05 and logarithmic fold change (log2FC) ≥ 1 for up-regulated or log2FC ≤ -1 for down-regulated miRNAs. The mean difference and 95% CI was calculated for normalized read counts. Analyses were performed using SPSS for Windows software (IBM SPSS Statistics for Windows, version 25.0; IBM Corp., Armonk, New York).

This study was designed to be descriptive, and the sample size was feasible for the autopsy material available and judged to be adequate to provide reasonable data. Two-tailed P-values were calculated and P < 0.05 considered significant in all analyses.

Results

We identified a total of 44 patients died with sepsis and who had autopsies performed. Of those 19 were excluded and finally 15 patients deceased due to sepsis had autopsies performed and had myocardial tissue samples available (Fig. 1.) All myocardial specimens for this study’s analysis were obtained from the left ventricular free wall.

Fig. 1
figure 1

Emergency admissions to ICU from year 2015 to 2020.

Full size image

In the sepsis group, the foci of infection were pulmonary (47%), gastrointestinal (33%), skin (13%), or urinary (7%). Ten (67%) of the sepsis cases had a laboratory confirmed positive blood culture. The median stay in the ICU was 1.3 days, whereas hospital stay was 2.6 days. Cause of death for the 15 control cases was accidental trauma (n = 3) or self-inflicted injury (n = 2) leading to death on-site, or sudden cardiac arrest (n = 10) with coronary artery disease being the underlying cause of death. None had signs of acute infection at autopsy. There were no significant differences regarding gender, age, or major chronic illnesses between the septic and control subjects (Table 1). In both groups, 67% of the subjects had coronary artery disease and the median age was 72 years (Table 2).

Table 1 Demographics of the controls and patients with myocardial autopsy specimens.
Full size table
Table 2 Demographics and clinical data for sepsis patients with myocardial autopsy specimens.
Full size table

A total of 1753 miRNAs were detected in the miRNA sequencing, 267 of which were abundantly expressed in the samples and subjected to differential expression analysis. We identified 32 differentially expressed miRNAs in the myocardial specimens of sepsis cases compared to controls (Table 3).

Table 3 Differentially expressed miRNAs in the myocardia of sepsis patients and controls.
Full size table

Eight miRNAs had log2FC > 1 and a significant change in the mean difference based on the 95% CI: miR-12136, miR-146b-5p, miR-155-5p, miR-21-3p, miR-320c, miR-320d, miR-3960, miR-4488, and miR-4787-5p. Of these, miR-12136 had the largest increase in mean counts in sepsis hearts, and miR-146b-5p had the highest fold change. Two miRs, miR-144-5p and miR-451a, had log2FC < -1 and a significant change in the mean difference based on the 95% CI. MiR-451a had the largest difference in mean counts and highest fold decrease. Enrichment analysis was performed for the predicted target genes of the 32 differentially expressed miRNAs to identify regulated biological processes. The pathways targeted by the differentially expressed miRNAs included cell adhesion, cell death and apoptosis, cardiovascular growth and development, and inflammation and immune response (Table 4).

Table 4 miRNAs categorized according to the enriched biological pathways of target genes.
Full size table

Finally, we performed 3’ mRNA sequencing from the same samples to identify mRNAs potentially regulated by the miRNAs. 17 out of the 30 samples met the quality criteria and were included in the analysis. The analysis identified 14 differentially expressed miRNAs, and 11 of those were among the 32 miRNAs identified in the full data set. The interaction analysis identified eight miRNAs with target genes also differentially expressed in septic hearts. Out of the eight miRNAs, only hsa-miR-21-5p was not among the 32 miRNAs differentially regulated between control hearts and septic hearts. The highest number of potential targets were identified for hsa-miR-363-3p, downregulated in septic hearts. Five of the predicted target genes for hsa-miR-363-3p were upregulated and three target genes were downregulated in septic hearts.

Fig. 2
figure 2

miRNA – mRNA interaction analysis.

Full size image

Figure 2. Interaction analysis of differentially expressed miRNAs and their predicted target mRNAs in a subset of samples (N = 6 for control, N = 11 for sepsis case). Up- and downregulation is expressed in Log2 fold change (logFC) and is color coded according to the figure inset.

Discussion

In this study using postmortem myocardial specimens, we identified several miRNAs differentially expressed in septic patients compared to non-septic controls. To the best of our knowledge, the present study is the first to investigate miRNA expression using FFPE cardiac tissue from patients with sepsis. A previous study investigated myocardial specimens obtained from patients who died due to sepsis but focused on messenger RNA (mRNA) expression in tissue samples that were snap-frozen or placed on dry ice19. In that study, the mRNA expression pattern in the hearts of septic patients showed significant reductions in the levels of mRNAs associated with proteins responsible for cardiac energy production and contractility. In our study, the predicted miRNA target genes were associated with cardiovascular functions, cell adhesion, cell viability, inflammation, and immune response, suggesting their potential involvement in septic cardiomyopathy. Our findings support the growing evidence that miRNAs play a pivotal role in the molecular regulation of cardiac dysfunction during sepsis, adding new insights into the genetic and intracellular pathways that mediate this pathological condition.

Key miRNA findings in septic myocardium

The most prominent statistically significant miRNA observed in this study was miR-146b-5p, which exhibited the highest fold increase in septic myocardial samples. miR-146b-5p is known to be upregulated in fibroblasts, endothelial cells, and macrophages under hypoxic conditions during post-MI cardiac injury and repair in mice and in the serum of myocardial ischemia patients20. Its predicted target genes regulate cell adhesion, apoptosis, as well as inflammation and immune response. MiR-146-5p up-regulation has been found to be cardioprotective and local inhibition of the miRNA significantly restores cardiac remodeling and function in both mouse and porcine infarct models20. In contrast to a mouse sepsis model of cecal ligation and puncture, a previous study reported that miR-146b protected against sepsis-induced myocardial injury by suppressing the expression of IL-1β and myocardium apoptosis21.

Another notable miRNA, miR-12136, showed the largest increase in mean counts among septic myocardial specimens. Although its specific function in sepsis or septic cardiomyopathy is not known, miR-12136 is involved in the process of mRNA translation22.

miR-155-5p was significantly up-regulated in septic patients’ myocardial tissue samples compared to controls. The biological function regulated by the predicted target genes of miR-155-5p include cardiovascular growth and development, as well as inflammation and immune response. Previous data indicate that miR-155-5p exerts a significant influence on multiple pathways associated with sepsis and renal damage, and a direct relationship exists between the activation of miR-155-5p and increased expression of pro-inflammatory IL-6 cytokine and chemoattractant IL-8 cytokine23,24. Earlier studies showed that up-regulation of miR-155 has cardioprotective properties25. In rat hearts, miR-155 levels were increased after intraperitoneal lipopolysaccharide injection. Inhibition of miR-155 significantly down-regulated the apoptosis of cardiomyocytes, whereas overexpression of miR-155 significantly up-regulated the apoptosis of cardiomyocytes and in vivo ejection fraction and significantly increased fractional shortening and heart weight26. MiR-155 has been shown to be highly up-regulated in endotoxemic mice, as well as in human endothelial cells, increasing endothelial leakage in a tight junction protein-dependent manner27. Septic endothelial dysfunction and capillary leakage cause tissue edema, and myocardial edema has been suggested to play a role in the pathophysiology of septic cardiomyopathy4. The upregulation of miR-155 in the present study may support these previous findings. In a study with elderly severe septic shock patients, miR-155 and miR-143 were found to be potential useful biomarkers with relation to serum TNF-α, IL-6, CK-MB, and cTnI levels28.

We also identified increased levels of miR-21-3p in septic hearts. In a previous study, miR-21-3p was shown to be up-regulated in mouse hearts with LPS-induced cardiac dysfunction, and pharmacological inhibition of miR-21-3p led to the preservation of cardiac systolic function and improved survival29. The same study also found that miR-21-3p levels were significantly increased in the plasma of sepsis patients with cardiac dysfunction compared to patients without septic cardiac dysfunction, but the mechanistic link between miR-21-3p and septic cardiac dysfunction in humans has not been shown. Overexpression of miR-21-3p has also been shown to exacerbate myocardial inflammation, whereas downregulation suppresses cardiomyocyte apoptosis in LPS-treated rats30. Our findings reinforce the involvement of miR-21-3p in the inflammatory pathways contributing to myocardial injury in septic patients.

In our series, the expression of miR-4787-5p was also up-regulated compared to controls. MiR-4787-5p regulates vascular smooth muscle cell apoptosis, and its overexpression has been shown to have diagnostic value in patients with acute aortic dissection31. Our current data suggests that miR-4787-5p may play a role in the vascular pathology observed during sepsis, potentially contributing to the vascular dysfunction associated with septic cardiomyopathy.

Several additional miRNAs, including miR-320c, miR-320d, miR-3960, and miR-4488, were upregulated in heart tissue specimens in patients with sepsis, and their predicted target genes related to cell adhesion. Although sparse data exist on these miRNAs, miR-4488 has been shown to inhibit inflammatory protein accumulation in endothelial cells32, suggesting a potential role in the endothelial dysfunction observed in sepsis. Up- or down regulated miRNAs with little previous data may offer interesting potential for future studies in septic cardiomyopathy.

The expression of miR-451a showed largest difference in mean counts and the highest fold decrease in septic heart specimens compared to controls. A rat model has shown that miR-451a expression decreases in ischemia-reperfusion injury, and up-regulation of miR-451a in myocardial tissue reduces the area of myocardial infarction, attenuates myocardial injury, and reduces myocardial cell apoptosis33. Human studies have shown that miR-451a expression is elevated in the plasma of patients with acute myocardial infarction compared with unstable coronary disease and healthy control groups34, and that miR-451 has a regulatory role in ischemic heart injury35. The decreased miR-451a count in septic hearts compared with controls in the present study may support the assumption that myocardial dysfunction in sepsis is not caused by ischemia. MiR-144-5p was also down-regulated in sepsis and is associated with the macrophage response to vascular inflammation36, further suggesting the involvement of immune and inflammatory pathways in septic myocardial dysfunction.

miRNA-mRNA interactions and pathway analysis

Analysis for miRNA - mRNA interactions identified eight miRNAs with target genes differentially expressed in septic myocardium with potential functional miRNA–target interaction for most miRNAs. For example, the downregulation of hsa-miR-486-5p was associated with the upregulation of LILRB4 and SERPINE1, both involved in cellular proliferation, migration, angiogenesis, and fibrosis37. LILRB4 also modulates inflammation and immune cell activity by regulating immune responses during infectious diseases38. It has been shown to protect against cardiac hypertrophy and fibrosis, helping to reduce inflammation and pathological remodeling in the heart37,38,39. Plasminogen activator inhibitor type I (PAI-1), encoded by SERPINE1, is a regulator of fibrinolysis and is closely linked to inflammation and coagulopathy in sepsis40. While upregulation of PAI-1 protects from myocardial fibrosis and hypertrophy41,42, it is associated with worse clinical outcomes40,43,44.

Downregulation of hsa-miR-363-3p and hsa-miR-144-3p was associated with upregulation of their target genes ADAM19, BTG2, DUSP5, GRAMD1B, ELL2, and PI15, which are involved in stress responses and inflammatory pathways45,46. ADAM19 is a mediator of cell signaling, cytokine activity, and inflammation, which are key processes in cardiac remodeling and dysfunction. It also plays a role in cell adhesion and migration, processes that are often disrupted in sepsis47. DUSP5 is involved in regulating mitogen-activated protein kinase (MAPK) pathways, which are crucial for controlling inflammation and cellular stress responses48. In sepsis, dysregulation of MAPK signaling can intensify the inflammatory response, contributing to mitochondrial dysfunction and cardiomyocyte stress48.

Upregulation of miR-21-5p and miR-21-3p was associated with regulation of target genes involved in cardiac hypertrophy, inflammation, and regulation of left ventricular function, further supporting their roles in myocardial dysfunction during sepsis49,50.

Target mRNAs involvement in pathways related to cell adhesion and cardiovascular system regulation and development reflects results shown in Table 4. ADAM19, NR4A3, SFRP1, and LILRB4 are linked to positive regulation of cell adhesion pathways. KLF5 and NPPB are part of the circulatory system development pathway. SFRP1, NPPB, and SERPINE1 are involved in the regulation of the angiogenesis pathway. In some predicted interactions both the miRNA and mRNA were regulated in the same direction, a common finding also in other miRNA-mRNA interaction studies51. In cases where the directions of regulation are opposite, it is more likely that the miRNA-mRNA interaction is direct.

These findings underscore the importance of miRNA regulation in the complex network of intracellular signaling pathways during septic cardiomyopathy. These genes are involved in key processes like cellular stress, coagulation, immune suppression, and inflammation and may contribute to the development of sepsis-induced cardiac dysfunction.

Clinical significance

Potential clinical use in diagnostics or prognostication of miRNA blood levels in the ICU setting would require accurate, rapid, and low-cost assays with multimarker panels of numerous miRNAs because of the heterogeneous nature of the septic response. However, as miRNAs may have tissue-specific expression and can be secreted into the blood, they could offer a biomarker tool for detecting myocardial dysfunction or possibly help guide treatment and monitor the treatment response in sepsis. For example, miR-155 and miR-146 have been proposed as potential biomarkers7,28, also with notable differential expression found in the present study. It is logical to think that the use of miRNAs as biomarkers will become routine following technological development [52]. In addition to biomarkers, the first animal studies imply that miRNAs may offer possible therapeutic interventions through restoration of the dysregulated immune system53. Further studies are needed to explore whether soluble miRNAs may serve as biomarkers or prognostic tools for septic myocardial dysfunction.

Limitations

We were not able to differentiate between distributive shock and septic cardiomyopathy. It would have required systematic echocardiography studies or other methods to record cardiac function, such as routine use of a pulmonary artery catheter. However, these two forms of cardiovascular dysfunction often coexist. Furthermore, the collection of myocardial tissue samples was not standardized, as the autopsies were performed for clinical diagnostic purposes.

The selected control subjects were heterogeneous, deceased due to traffic accidents, self-inflicted injury, or sudden cardiac death. They were matched for age, gender, and history of coronary artery disease, and there was no evidence of infection. Due to the small sample size and wide 95% CIs for differences in the mean miRNA counts, we chose a 2-fold increase as significant. Furthermore, a more conservative 2-fold increase was chosen, not because of pure statistical reasons, but to detect a clinically meaningful increase in expression levels that might lead to changes in cell level. However, as far as we know, there is no scientifically established clinically meaningful difference in each miRNA and it is not known how big of a change is needed to alter gene expression that causes alterations in cell functions. miRNA profiling studies have demonstrated that even subtle alterations in miRNA expression, such as a 1.5-fold difference, may exert a notable influence on cellular biology54.

Due to the small sample size, the reported associations cannot reliably be drawn as causal. There are several concerns regarding small sample sizes and lack of power. For one, in small sample sizes, the number of outliers can have a relatively high contribution to the effect increasing the false positive findings. Due to the low power, the true effect might be exaggerated so that mainly large effects are significant. Also, low power can lead to a higher rate of false negative findings and true findings might be missed55. The main reason for the absence of power analysis in the present study was that we included all patients who died due to sepsis, had an autopsy done in the years 2015–2020, had FFPE cardiac tissue samples available, and met our inclusion criteria. However, we calculated a post hoc sample size to detect a 1.5-fold change, with 80% power, a 5% false discovery rate (FDR), and an estimated 88% proportion of non-differentially expressed miRNAs. For this, an estimated sample size of 12 cases per group would be needed. This supports the sample size of the present study.

The study groups were not matched with a postmortem interval which can be a concern. In previous studies with FFPE cardiac tissue samples, the stability of miRNAs after long-term fixation has been noticed15,56. miRNAs are resistant and exhibit stability following the removal of tissue from a patient, after freezing, fixation, or after slicing, etc., However, the ischemic time can potentially alter miRNA expression analysis. Studies have shown that miRNA expression can change during ischemia, potentially leading to degradation or altered expression patterns [16,57]. It has also been noticed that the changes are very small, if any15. However, the total effect of warm ischemia or postmortem interval on miRNAs has not been established. The uncertainty regarding the total ischemic duration can be a concern, particularly in retrospective analyses of archived tissue samples, whereas prospective studies can implement standardized protocols for timekeeping and documentation [57]. In our study, time-dependent factors were especially confounding in deceased controls who were found dead at home as warm tissue ischemia usually has a greater overall impact on tissue integrity compared to cold ischemia. Warm ischemia time was shorter and more controlled in sepsis cases as the deceased were moved to a morgue and a cold room soon after death. In the present study, the proportion of non-differentially expressed miRNAs were roughly 88%. From this, a cautious interpretation can be made that miRNAs are stable as expressions are quite similar in both groups regardless of postmortem intervals.

We investigated cardiac tissue specimens, but as far as we know, none of the miRNAs in the present study are cardiac-specific. We cannot be certain whether microRNAs are expressed in the myocardium or if they originate from other tissues and are delivered to the myocardium by blood circulation. It has been noticed that circulating miRNAs can act as signaling molecules and induce hormone-like responses in target cells and tissues9,10.

The results must be interpreted as observational and hypothesis-generating. Despite its limitations, our study brings new insight into the cardiac response in sepsis at the organ level. Several miRNAs seem to be integrated with the pathophysiology of the septic myocardial response, either promoting or inhibiting cardiac damage. The up-regulated miRNAs were related to endothelial adhesion, inflammation, and apoptosis, and the down-regulated miRNAs to cardiovascular functions, cell viability, and immune response.

Future studies are needed in understanding the role of miRNAs in sepsis-associated cardiomyopathy. To demonstrate the function of miRNA biomarkers for sepsis-associated cardiomyopathy, several types of functional studies used in animal or in-vitro studies are needed also in human studies58,59.

Conclusions

Several regulatory miRNAs were up- or down-regulated in the myocardial tissue of septic patients compared to non-septic subjects. All the discussed miRNAs have documented roles in inflammation or pathological cardiovascular conditions. The predicted target genes of miRNAs and miRNA-mRNA interaction analysis indicated the possible association with biological functions related to cardiovascular functions, cell viability, cell adhesion, regulation of inflammatory and immune response, and endothelial response. These findings might be associated with clinically meaningful circulatory responses and with the development of septic cardiomyopathy. However, further studies are needed to explore whether these findings are associated with the development of septic cardiomyopathy.