Profiling of circulating serum exosomal microRNAs in elderly patients with infectious stress hyperglycaemia

1 INTRODUCTION

Stress hyperglycaemia (SHG) has an especially high incidence and mortality in hospitalization for elderly patient particularly.^{1, 2} It is a series of immune-neuro-endocrine responses that appear after stress, including acute myocardial infarction, stroke, perioperative period, severe infection and so on, which mostly manifested as elevated blood glucose level and lead to related changes in body metabolism.³ Especially under infection background, such as patients with coronavirus disease 2019 (COVID-19) are reported to have a greater prevalence of hyperglycaemia.^4-7 Cytokine release as a consequence of severe infection may precipitate the onset of metabolic alterations by affecting glucose homeostasis, which leads to abnormalities in glycometabolic control, insulin resistance (IR) and β cell dysfunction in patients with infection even without any pre-existing history or diagnosis of diabetes. Stress hyperglycaemia seriously affects the stability of the body's internal environment, which cause a profound impact on metabolism and immune function, thereby affecting the patients’ prognosis.^{2, 8} Stress hyperglycaemia is associated with an increased risk of in-hospital mortality regardless of whether the patients have diabetes background or not.^{1, 9} It is suggested that COVID-19-associated new-onset hyperglycaemia may predispose patients to long-term hyperglycaemia, worse clinical outcomes and clinical scores, prolonged hospital stays higher demand for oxygen support or positive-pressure ventilation.¹⁰ Therefore, early identification of stress hyperglycaemia, exploration of its mechanism and precise management are the focus areas that urgently need attention and have become an important public issue of health concern.

Inflammation is one of the main causes of stress hyperglycaemia, especially infectious stress hyperglycaemia (ISH), which are able to predictable deterioration to septic shock, and the fatality rate is as high as 70%–80%.^{1, 11} Moreover, 95% of sepsis patients have abnormal glucose metabolism and severe IR, and the exact molecular mechanism is not yet clear.¹ Intensive insulin therapy (IIT) is currently the main method for the treatment of stress hyperglycaemia.^12-14 However, due to the body's IR and stress state, IIT is not effective in counteract hyperglycaemia, which often leads to undesirable glucose fluctuations, and the risk of hypoglycaemia is greatly increased, which showed no benefit in the medium and long-term survival rate.¹⁵ Therefore, it is important to explore the underlying mechanism of ISH and reveal its molecular pathways and make effort to explore safe and effective therapeutic targets.

Exosomes are small extracellular vesicles (sEVs) secreted by cells and are present in several kinds of body fluids, such as blood, urine, breast milk and ascites.¹⁶ Exosomes contain a variety of molecules that exhibit biological activity, such as mRNAs, miRNAs, lncRNAs, proteins and lipids, of which to some extent reflect the states and types of their cells of origin.¹⁷ As the protection of the membrane, the enclosed miRNAs are more stable and will not be degraded by RNase that presented in biofluids. Thus, exosomes are emerging as non-invasive biomarkers for diagnosis, prognosis and response to treatment.^18-21

MicroRNAs (miRNAs) are endogenous 21–23 nt small non-coding RNAs capable of controlling gene expression through post-transcriptional regulation.²² MiRNAs could degrade complementary target mRNAs in an RNA-induced silencing complex dependent manner, which is essential in embryo development,²³ oncogenesis,¹⁹ immune regulation^{24, 25} and other biological processes. MiRNAs had been identified as exosome-encapsulated form in many biofluids, suggesting the potential of developing circulating miRNAs as biomarkers of cancer and metabolic diseases.²² The previous studies proposed that miRNAs were usually be packaged into exosome that secreted by adipose tissue macrophages and transferred to insulin-target cell and exert robust effects on cellular insulin action and overall glucose homeostasis.²⁶

Due to these advantages, exosomal miRNAs are emerging as promising biomarkers for different kinds of diseases.^27-35 However, the biogenesis of circulating exosomal miRNA is of high heterogeneity, which could derive from different cells.²⁷ There is growing evidence that suggests the potential role for sEVs in early detection of many diseases.^{22, 28-35} However, systematic screening of serum sEVs derived elderly ISH patients’ biomarkers has not yet been reported.

In this study, we detected and screen the total miRNA profile of circulating serum sEVs-enriched fractions in early ISH patients paired with infectious individuals without stress hyperglycaemia (non-ISH) by a next-generation sequencing method. We characterized several sEVs-derived miRNAs, and their combinations could largely distinguish ISH patients from control to some extent. Furthermore, we validated those biomarkers in an independent clinical validation cohort, so as to evaluate their performance to predict negative clinical outcome, including septic shock and death. The significance of our study showed that exosomal miRNAs were able to identify patients with ISH and predict the poor clinical outcome.

2 MATERIALS AND METHODS

3 RESULTS

3.1 Clinical characteristics

In this study, we recruited a discovery cohort of 15 patients and a validation cohort of 52 out of 83 patients, of which 31 patients’ samples were excluded due to haemolysis or incomplete data. First cohort included 10 ISH patients and 5 disease-match non-ISH, whereas the validation cohort included 31 ISH patients and 21 non-ISH (NC). The workflow of the study is shown in Figure 1. The participants’ demographic and clinical information of age, gender, glucose, biochemical marker and hospitalization day are summarized in Table 1. There was no statistically significant difference between the ISH and control groups with respect to age, white blood cell count (WBC), absolute neutrophils and lymphocyte count (LY), C-reactive protein (CRP), length of hospital stay and APACHII score (p > .05).

TABLE 1. The participants’ demographic and clinical information.

Group	non-ISH	ISH	Standardize diff.	p-Value
Female	14 (58.33%)	12 (27.91%)
Male	10 (41.67%)	31 (72.09%)
Age	65.26 ± 16.09	66.02 ± 15.67	0.05 (−0.46, 0.56)	.854
RBG	7.61 ± 1.20	11.93 ± 3.40	1.70 (1.11, 2.28)	<.001
WBC	11.67 ± 4.85	16.44 ± 10.82	0.57 (0.05, 1.09)	.065
NEU	10.12 ± 4.77	14.25 ± 9.83	0.53 (0.01, 1.06)	.064
LY	0.88 ± 0.46	1.10 ± 0.94	0.30 (−0.22, 0.81)	.298
CRP	94.47 ± 85.92	148.67 ± 111.71	0.54 (0.02, 1.06)	.095
Hospitalized days	12.32 ± 13.56	14.18 ± 13.09	0.14 (−0.43, 0.71)	.627
APCHEII	23.35 ± 7.77	26.64 ± 7.93	0.42 (−0.10, 0.94)	.117

• Abbreviations: APACHII score, Acute Physiology and Chronic Health Evaluation; CRP, C-reactive protein; LY, lymphocyte count; NEU, absolute neutrophils; RBG, random blood glucose; WBC, white blood cell count.

3.2 Patients’ serum exosome-derived microRNAs (sE-miRNAs)

The serum exosomes vesicles (sEVs) were extracted from the ISH patients and controls by ultracentrifugation, and the morphology and sizes distribution of sEVs were evaluated by transmission electron microscopy (TEM) and NTA. TEM and NTA analysis showed that sEVs were oval or bowl-shaped with a size range between 80 and 200 nm (Figure 2A,B). Moreover, sEV markers CD9, TSG101 and HSP70 were all detected in the sEVs isolated from the serum (Figure 1C). On the contrary, calnexin, a negative marker of sEVs, was not detectable in our isolated sEVs-enriched fraction samples (Figure 2C). Besides, the protein levels of the sEVs-enriched fraction samples were also evaluated, and an sEV-associated protein concentration of 238.471 μg/mL serum was reported. Total RNA isolated from sEVs-enriched fractions was analysed and quantitated. No significant difference was identified in the RNA concentration of sEVs-enriched fractions (ng/mL plasma) between ISH patients and controls (Figure 2D).

3.3 Comparison of serum exosome-derived microRNAs (sE-miRNAs) signatures between ISH patients and non-ISH controls in the discovery cohort

To get a global profile of the sEVs-enriched fraction-derived miRNA from the ISH patients, we first tested 15 samples (10 ISH patients and 5 controls) in the testing set by miRNA sequencing. For each sample, no less than 6.17 M clean reads were generated. A total of 1220 out of 1500 known miRNAs were detected by sequencing. A total of 276 miRNAs with median TPM >10 were included for the subsequent analysis, so as to avoid bias caused by miRNAs with relatively low expression levels. In differentially expressing miRNA (DEMs) analysis (|log2(FC)| >0.585, p < .05), 11 miRNAs were found up-regulated (with 1 unknown miRNA), and 38 down-regulated in ISH patients than controls with 1 unknown miRNA. Perform hierarchical clustering analysis on the screened differentially expressed miRNAs, and the clustering results are shown in Figure 3A (serum). The difference in the expression levels of miRNAs in two group samples and the statistical significance can be viewed through the volcano plot (Figure 3B, serum). We calculated the specificity and sensitivity of each DEM in the testing set (Figure 3C, serum). Most candidate miRNAs displayed a specificity of 0.6–1.0 and a sensitivity of 0.8–1.0 (Table 2).

TABLE 2. Top 22 miRNAs ranked by sensitivity × specificity (miRNA-seq data) based on serum exosome.

#ID	p-Value	Log2FC	Sensitivity	Specificity	Sen × Spec
hsa-miR-20b-5p	1.534E − 02	−1.0617	1	0.8	0.8
hsa-miR-140-3p	7.858E − 05	−1.6622	1	0.8	0.8
hsa-miR-185-5p	4.922E- − 03	−1.1396	1	0.8	0.8
hsa-miR-486-5p	9.674E − 03	−1.0909	1	0.8	0.8
hsa-miR-2110	7.124E − 03	−1.0911	1	0.8	0.8
hsa-miR-500a-3p	5.023E − 03	−1.6045	1	0.8	0.8
hsa-miR-142-3p	9.995E − 03	0.9660	0.9	0.8	0.72
hsa-miR-484	2.860E − 02	−0.9871	0.9	0.8	0.72
hsa-miR-21-5p	2.622E − 02	0.9336	0.8	0.8	0.64
hsa-miR-335-5p	1.297E − 02	−0.9996	0.8	0.8	0.64
hsa-miR-3613-5p	4.345E − 02	1.0002	0.8	0.8	0.64
hsa-miR-103a-3p	1.970E − 02	−0.8379	1	0.6	0.6
hsa-miR-107	1.989E − 02	−0.8378	1	0.6	0.6
hsa-miR-203a-3p	4.750E − 03	1.6204	1	0.6	0.6
hsa-miR-361-5p	8.833E − 06	−2.0405	1	0.6	0.6
hsa-miR-92a-3p	4.905E − 02	−0.6964	1	0.6	0.6
hsa-miR-106b-3p	2.075E − 03	−1.2227	1	0.6	0.6
hsa-miR-501-3p	4.424E − 03	−1.6974	1	0.6	0.6
hsa-miR-486-3p	1.764E − 02	−0.9894	1	0.6	0.6
hsa-miR-483-5p	3.166E − 02	1.4324	0.9	0.6	0.54
hsa-miR-191-5p	3.406E − 03	−1.1420	1	0.4	0.4
hsa-miR-28-3p	4.173E − 02	−0.8461	1	0.4	0.4

We further collected adipose tissue from severe infection stress hyperglycaemia (ISH) patients who underwent limb amputation surgery. Based on microarray miRNAs gene chip analysis, we analysed the differentially expressed microRNAs profiles in adipose tissue from severe infection amputation-induced hyperglycaemia patients (n = 3) and adipose tissue from the control group, who received hip replacement surgery with normoglycaemic (n = 3). Small RNA sequencing of 6 samples was completed, and 1293 miRNAs were detected, including 1250 known miRNAs and 43 newly predicted miRNAs. Through quality control, the average Clean Data per sample is 15.09 M. A total number of 16 DEMs were detected, of which 6 miRNAs were found up-regulated, and 10 down-regulated (with 2 unknown miRNAs) in ISH patients when compared with controls. The clustering results of differentially expressed miRNAs are shown in Figure 3A (adipose). The difference in the expression levels of two groups miRNAs can be found in the volcano plot (Figure 3B, adipose). We also calculated the specificity and sensitivity of each DEM in the testing set (Figure 3C, adipose).Compared with differentially expressed miRNAs from fat tissue data, there were 2 DEMs (2/49) shared between those two analyses, both are down-regulated (Figure 3D), indicating hsa-mir-486-5p and hsa-mir-486-3p, indicating that they may derive from fat tissue (Figure S1A).

Additionally, in an attempt to reveal the potential function of the serum exosomal miRNAs that were showed up-/down-regulated in ISH patients, we analysed the potential impact of those miRNAs on protein-coding mRNAs. Bioinformatics miRNA target database analysis revealed that 17683 mRNAs were targeted by those DEMs, which were selectively enriched in insulin signalling pathway, insulin secretion, protein procession in endoplasmic reticulum, primary immunodeficiency, fat digestion and absortion pathways, endocytosis, lysosome, cell adhesion molecules (7) (Figure 4A–E). Adipose tissue's bioinformatics miRNA target database analysis results are shown in Figure S2A–D.

3.5 sE-miRNAs as biomarkers of ISH patient in a validation set

To validate the specificity of these seven sE-miRNAs for ISH patients found in the discovery set, expressions of these seven miRNAs were measured by RT-qPCR in the validation set of 52 subjects. Primers’ design was summarized in Table S1. As shown in Figure 5, hsa-miR-21-5p, hsa-miR-335-5p and hsa-miR-28-3p were showed significant up-regulated in serum exosome from ISH patients group by using U6 as an internal calibration when compared with control groups (Figure 5). To verify the potential of miRNA expression, ROC analysis was performed, and AUC (Area Under Curve) was calculated both in discovery and validation set.

Hsa-miR-28-3p, hsa-miR-335-5p and hsa-miR-21-5p exhibited an AUC of 0.76, 0.88 and 0.88 close to the AUC (0.90) calculated from the discovery set (Figure 6A–C). On the other hand, in the validation set, these three miRNAs achieved an AUC of 0.731, 0.772 and 0.777, respectively, which were lower than the AUC calculated from the discovery set (Figure 6A–C). However, integrating three candidate miRNAs could achieve a higher AUC of 0.96 in discovery set, whereas these are limited improvements in validation set with AUC of 0.758 (Figure 6D). Additionally, by using a logistic model, we combined all those candidate miRNAs together with clinical examination index and obtained much improved performance compared with single individual miRNA. Notably, three miRNAs plus WBC and CPR exhibited an AUC of 0.956 (Figure 6E), and hsa-miR-21-5p alone plus WBC or CRP exhibited an AUC of 0.86 (Figure 6F). In summary, three miRNAs, especially hsa-miR-21-5p, provided promising AUC values for discriminating between ISH patients from controls.

To assess the clinical significance of miRNA expression, an ISH survival assay was performed on the validation cohort of 52 patients. Kaplan–Meier analysis showed that patients with high two miRNAs (hsa-miR-28-3p and hsa-miR-335-5p) expression in their serum exosomes had significantly shorter survival than patients with low miRNA expression did (Figure 7A,B). But not for hsa-miR-21-5p (Figure 7C), which indicating that hsa-miR-28-3p and hsa-miR-335-5p may take considerable responsibility for patients’ poor clinical outcome. Combined three miRNAs together still showed significantly shorter survival when compared to patients with low miRNAs expression (Figure 7D).

Bioinformatics miRNA target database (multiMIR) analysis revealed that 4618 mRNAs were targeted by hsa-miR-28-3p, hsa-miR-335-5p and hsa-miR-21-5p, which were selectively enriched in signals of MAPK signalling pathway, endocrine resistance, lipid and atherosclerosis, AGE-RAGE signalling pathway in diabetic complication, metabolic process as well as protein binding or molecular function (Figures 8 and 9).

In conclusion, our study suggested that ISH patients retained specific serum exosome-derived miRNAs profile compared with non-ISH patients. Our study depicted the circulating exosomal miRNAs change in ISH patient, which could be used as a promising biomarker to detect ISH at an early stage and predict patients’ clinical outcome. Of which hsa-miR-21-5p proposed a new promising biomarker category of ISH patients, whereas hsa-miR-28-3p and hsa-miR-335-5p play an important role on patients’ poor clinical outcome.

4 DISCUSSION

Claude Bernard first proposed the concept of SHG in 1877,³⁸ and stress hyperglycaemia usually refers to transient hyperglycaemia during illness and is generally restricted to patients without previous evidence of diabetes.³ However, both physical and psychological stimulation factors are able to cause stress state, and the identification of such patients is complex. No guidelines so far specifically define stress hyperglycaemia. According to American Diabetes Association (ADA)³⁹ and Lancet paper on stress hyperglycaemia review,³ we propose two diagnostic categories of stress hyperglycaemia according to the ADA consensus definition (fasting glucose >7 mmol/L or random glucose >11·1 mmol/L without evidence of previous diabetes) and pre-existing diabetes with the deterioration of preillness glycaemic control. So stress hyperglycaemia was supposed to be diagnosed when peripheral blood sugar is higher than the diagnostic criteria irrespective of whether a patient has pre-existing diabetes or not.³

Stress hyperglycaemia remains a major health emergency issue in our clinical practice especially in-hospital ICU, affecting hundreds of millions of people worldwide.⁴⁰ It is reported that infection such as COVID-19 significantly increased rate of new-onset hyperglycaemia. Surprisingly, 46% of COVID-19 patients showed hyperglycaemia during hospitalization.¹⁰ Among patients who exhibited new-onset hyperglycaemia at hospital admission for COVID-19, persistent hyperglycaemia continued to be observed in the following 6 months in nearly 35% of patients, and 2% of patients were finally diagnosed. Moreover, studies using primary human islets demonstrated that pancreatic β cells are highly permissive to SARS-CoV-2 infection through ACE2. Interestingly, COVID-19 also induces a cytokine storm, an exaggerated immune response with a broad spectrum of cytokine production that establishes a systemic proinflammatory milieu, which may eventually lead to IR and β cell hyperstimulation and death.¹⁰ COVID-19 infection may also enhance the pre-existing proinflammatory status observed in type 2 diabetes, thus worsening patient survival and complications. In conjunction with its much-dreaded complications (e.g. sepsis and septic shock), it substantially reduces the quality of life, increases mortality as well as medical burden among patients.⁴¹ Over the years, oxidative stress and inflammation have been highlighted as key players in the development and progression of stress hyperglycaemia and its associated complications.^{42, 43} However, the underlying mechanism for stress hyperglycaemia caused by the most common reason, infection in particular, remains unknown.

Infection leads to profound alterations in whole-body metabolism, which is characterized by the marked acceleration of glucose, fat and protein. The hyperglycaemia is generally caused by peripheral IR and alterations in hepatic glucose metabolism. MiRNAs of serum exosomes have emerged as biomarkers for many diseases, due to their high stability.²² MiRNAs have been linked to factors associated with inflammation disorder⁴⁴ and glucose metabolism disfunction^{45, 46} especially under stress condition. In this study, we aimed to identify serum exosomal miRNAs as prognostic markers associated with inflammation-related stress hyperglycaemia (ISH) in aged hospitalized patients. Exosomal miRNAs might serve as endocrine and paracrine messengers that facilitate communication between donor cells and tissues with receptor cells or target tissues, thereby potentially having important roles in metabolic organ-cross-talk. In our study, we focus primarily on alterations in whole-body metabolic response to inflammatory stress, and the interactions between the endocrine response to infection and serum exosomal miRNAs change.

Here in our study, RNA sequencing identified a total number of 49 serum exosomal-derived miRNAs with differential expression between ISH and control group in discovery set. Furthermore, compared with differentially expressed miRNAs from adipose tissue data and serum exosome, there were two DEMs (2/49) shared the same pattern, both are down-regulated (Figure 3D), indicating that hsa-mir-486-5p and hsa-mir-486-3p may derive from adipose tissue (AT). Adipose tissue-derived exosomal miRNAs play an important role in glucose metabolism. Carlos Castaño⁴⁷ found that obesity changes the miRNA profile of plasma exosomes in mice, including increases in miR-122, miR-192, miR-27a-3p and miR-27b-3p. Importantly, treatments of lean mice with exosomes isolated from obese mice cause glucose intolerance and IR. Riccarda Granata⁴⁸ observed differential expression of miRNAs in extracellular vesicles (Evs) from healthy and inflamed adipocytes (AT), as well as alteration in signalling pathways and expression of β cell genes, adipokines and cytokines in recipient β cells. Their in vitro results suggest that adipocyte-derived EVs may influence β cell fate and function. As for our study, both miR-486-3p and miR-486-5p, unexpectedly, had no statistical difference in the clinical validation (Figure 5F,G), indicating that they may not play a key role in the pathogenesis of ISH. On the other hand, tissue in situ exosome^{49, 50} instead of total organ should also be taken into consideration for exosome origin verification, but due to clinical sampling and ethical restrictions, more systematic and in-depth tracer experiments are needed in the future.

Moreover, 49 DEMs, including hsa-mir-486-5p, hsa-hsa-mir-486-3p, hsa-hsa-miR-140-3p, hsa-hsa-miR-21-5p, hsa-hsa-miR-335-5p, hsa-hsa-miR-191-5p and hsa-hsa-miR-28-3p, in particularly, were finally selected for clinical large sample validation based on literature review, which were related to glucose metabolism and inflammatory dysfunction. Three miRNAs, including hsa-hsa-miR-21-5p, hsa-hsa-miR-335-5p and hsa-hsa-miR-28-3p, were showed significantly up-regulated in ISH group when compared with control in validation set. As there are gender differences in the baseline data between the two groups, which may act as a confounding factor, for the sake of rigor, we re-examined our data to confirm that whether sex alone or itself can cause differences in miRNA expression or not. It showed that there was no statistical difference between males and females patients when it came to these three miRNAs (Figure S3), indicating that these three miRNAs indeed play a certain role in stress hyperglycaemia regardless of sex.

Literature reviews of these three miRNAs function are summarized as follows. Qin⁵¹ provided evidence that miR-335-5p enhanced IR through VASH1-mediated TGF-β signalling pathway in gestational diabetes mellitus (GDM) mice, which provides more clinical insight on the GDM treatment. Another findings provided evidence that miR-335-5p aggravates type 2 diabetes by inhibiting SLC2A4 expression, and miR-335-5p and SLC2A4 as potentially effective therapeutic targets for patients with T2DM.⁵² Moreover, hsa-miR-335-5p appeared to be involved in T2DM by potentially regulating the expression of various candidate genes, including procollagen C-endopeptidase enhancer 2, connective tissue growth factor and family, which might also be used in the molecular diagnosis and targeted therapy of T2DM.⁵³ Eliasson⁵⁴ reported that the expression of miR-335 negatively correlated with secretion index in human islets of individuals with prediabetes. Overexpression of miR-335 in human EndoC-βH1 and in rat INS-1 832/13 cells (OE335) resulted in decreased glucose-stimulated insulin secretion. As for miR-21-5p and miR-28-3p research, Abulsoud⁵⁵ found that serum levels of miR-148a-5p and miR-21-5p were higher in metabolic syndrome (MetS) patients than in healthy controls; consequently, these serum miRs can serve as novel biomarkers for diagnosis and prognosis of MetS. Xia⁵⁶ found that miR-21 is down-regulated in the serum and placenta of GDM patients compared to normal pregnant women. In the case of IR, miR-21-5p knock-down promoted glucose uptake, but no significant effect was found under physiological condition. Syhan’ study⁵⁷ had measured the circulating levels of cytokines and miRNAs in lean and obese humans with prediabetes. The result reveals that a number of circulating cytokines and miRNAs deregulated in subjects with obesity, prediabetes and T2DM. Specifically, cytokines IL-6, IL-8, IL-10, IL-12 and SFRP4, as well as miRNAs miR-21, miR-28-3p, miR-150, miR-155 and miR-223, significantly changed across the diabetes spectrum and were associated with measures of pancreatic islet β cell function and glycaemic control, and their data suggested that changes in circulating miRNAs and cytokines may have clinical utility as biomarkers of prediabetes. CORDIOPREV study⁵⁸ showed that patients with low levels of miR-29a, miR-28-3p and miR-126 and high plasma levels of miR-150 at baseline showed a higher risk of developing T2DM after the consumption of the Med diet.

Combined with previous literature reports and our research, hsa-miR-28-3p, hsa-miR-335-5p and hsa-miR-21-5p play an important role in both glucose metabolism and inflammation-related diseases. It is worth mentioning that miRNAs which are able to master (NF-κB)-driven inflammatory pathways were called inflammamiRs.⁵⁹ Relevant evidence suggesting that circulating inflammamiRs, along with IL-6, can measure the degree of inflammaging.⁶⁰

Previous study had showed that interleukin-6 (IL-6) is necessary for stress hyperglycaemia,¹⁰ which mediates metabolic reprogramming behaviour of hyperglycaemia under acute stress.⁶¹ Stress-inducible IL-6 is produced from brown adipocytes in a beta-3-adrenergic-receptor-dependent fashion. During stress, endocrine IL-6 is the required instructive signal for mediating hyperglycaemia through hepatic gluconeogenesis, so as to anticipate and fuel ‘fight or flight’ responses.⁶¹ This adaptation comes at the cost of increasing mortality to a subsequent inflammatory challenge. These findings provide a mechanistic understanding of the adaptive purpose of IL-6 as a stress hormone coordinating systemic immunometabolic reprogramming.

Moreover, our supplementary data also show that there was a significant correlation between IL-6 level and blood glucose on ISH patients (Figure S4A). Moreover, when we tracked and monitored continuously in one of the patients, we found that glycaemia levels fluctuate in sync with IL-6 levels (Figure S4B). The results further suggested that stress induced IL-6 is important for hyperglycaemia. However, the underlying mechanism for stress induced IL-6-related hyperglycaemia remains unclear.

Previous study implicated that forkhead box protein O (FOXO) transcription factors in the liver were crucial in stress-induced hyperglycaemia,which acted as promoters of altered glucose homeostasis.⁶² Specifically, FOXO activity was found up-regulated in stress-induced hyperglycaemia, and loss function of FOXO, including subtype of FOXO1, 3 and 4, in the liver attenuated hyperglycaemia and alleviated hyperinsulinemia, which ameliorated insulin sensitivity.^{63, 64}

Mechanistically, the loss of FOXO transcription factors mitigated the stress-induced hyperglycaemia response by suppressing lipolysis in adipose tissue indirectly and altering gene expression and glycogenolysis in the liver directly. It is worth noting that reductions were associated with decreased IL-6, TNF-α and increased FGF21, suggesting that cytokines (such as IL-6) and FOXO-regulated hepatokines contribute to the stress-induced hyperglycaemia response.⁶²

Here, we further used sequence alignment analysis to predict the target genes of these three miRNAs, and it was very exciting to found that both hsa-miR-335-5p and hsa-miR-21-5p had several conserved binding sites with 3′UTR of IL-6 or the FOXO1 (forkhead box O1) (Figure S4C), which implicated that hsa-miR-335-5p and hsa-miR-21-5p may excert effect on transcriptional level on IL-6 or FOXO, both of which were acted as a predominant driver of stress-induced hyperglycaemia through means that include cross-talk between the liver and adipose, highlighting a novel mechanism underlying acute hyperglycaemia and IR in stress.

In summary, our multi-level analyses uncovered and identified three miRNAs that were up-regulated in ISH which may serve as biomarker for stress hyperglycaemia and as key orchestrators promoting aggressiveness in ISH. In particularly, serum exosomal hsa-miR-21-5p might relate to poor clinical outcome, and patients with high expression may deteriorate to septic shock. Further research is needed on the biological functions and mechanisms of these miRNAs and their function on mediating the IR during infection stress-induced hyperglycaemia. Moreover, few therapeutic options currently exist to treat ISH safely so far,⁶⁵ and these three miRNAs may serve as potential therapeutic targets and offer substantial promise in expanding treatment options for patients with ISH. Finally, more research is needed to translate these findings into a human clinical practice.

AUTHOR CONTRIBUTIONS

K.Z. and J.T. as well as G.X.designed the whole project, K.Z. and Z.L.with H.H.aided in conceiving the hypothesis, designed and performed the experiments, analyzed the data, and prepared the manuscript. G.W., Y.S.K., contributed to the analysis and discussion as well as language polishment. This work was also partially supported by M.L., W.W., H.S., H.L., B.C. for sample collection, clinical data analysis as well as data presentation.

CONFLICT OF INTEREST STATEMENT

No potential conflict of interest was reported by the authors.

ETHICS APPROVAL

This study was approved by the ethics committee of Sun Yat-sen Memorial Hospital of Sun Yat-sen University (Ethical approval SYSEC-KY-KS-2021-189).