Preliminary miRNA profiling of plasma from patients infected with distinct SARS-CoV-2 variants

Abstract

Background

Epigenetic regulation can play a dual role in viral infection, acting as part of the host defense, while being hijacked to control viral latency, replication, and persistence. During the COVID-19 pandemic, substantial evidence indicated that different SARS-CoV-2 variants may lead to distinct outcomes by affecting different molecular pathways, including epigenetic mechanisms.

Methods and Results

In this study, we assessed miRNA expression using high-throughput microarray technology to simultaneously measure the expression of thousands of miRNAs. We focused on: comparing Delta and Omicron infections with healthy controls; identifying specific miRNAs linked to symptom severity; analyzing miRNA expression across different stages of Delta variant infection; and predicting miRNA target genes and evaluating miRNA–mRNA interactions. We identified 65 miRNAs that were significantly differentially expressed simultaneously in both SARS-CoV-2 variants compared to healthy samples. These miRNAs were categorized into distinct functional groups based on their roles in viral infection. We observed an association between the expression of specific miRNAs and virus-induced symptom severity and patients’ outcomes in both variants. Furthermore, we examined miRNA expression across the various phases of the Delta variant and identified 38 miRNAs that were significantly differentially expressed. Finally, mRNA-miRNA interaction analysis revealed PTEN, IGF1R, MYC, and STAT3 as the most interacting genes.

Conclusion

Profiling thousands of miRNAs in response to different SARS-CoV-2 variants offers new opportunities for diagnosis, prognosis, and therapy. Although comorbidities and pathway cross-talk may influence the obtained results, our findings provide novel insights into the host response to SARS-CoV-2 infection at the epigenetic level.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11033-026-12632-x.

Introduction

Even several years after the initial wave of the COVID-19 pandemic was brought under control, the disease continues to pose a global threat to human health. Although vaccination, natural immunity, and improved treatment and clinical intervention have significantly reduced mortality rates [1–3], SARS-CoV-2 continues to circulate seasonally within the population in evolving subvariants. The virus remains unpredictable, especially concerning its long-term effects, and the emergence of new, more virulent variants is a very real concern. Ongoing, detailed research at the molecular level is essential to deepen our understanding of the interactions between viral variants and the human immune system, and to better prepare for potential future outbreaks.

Among the numerous variants that have shaped the course of the pandemic, the Delta (B.1.617.2) and Omicron (B.1.1.529 and its sublineages) variants stand out for their distinct virological and clinical characteristics. Delta, which dominated in mid-2021, was associated with increased disease severity, resulting in higher hospitalization rates and the need for intensive care, as well as a greater risk of long-term complications compared to earlier variants, including the original strain [4]. Multiple studies reported that the Delta variant led to a two- to threefold increase in hospitalization and mortality rates relative to Alpha and earlier variants, with estimated hospitalization rates reaching approximately 10% and case fatality rates between 0.7 and 2.4% depending on population [5, 6]. The observed increase in hospitalization and mortality rates has been attributed to higher viral load, faster viral replication, and greater lung involvement [7]. Omicron, which quickly replaced Delta in late 2021, exhibited enhanced transmissibility and immune escape. However, due to its preferential replication in the upper airways, it has generally caused less severe disease, particularly in vaccinated individuals [8]. A meta-analysis of 33 studies found that, in the general population, Omicron was associated with approximately 2.5 times lower risks of hospitalization and more than five times lower risk of death compared to Delta [6]. Although these variants are no longer dominant, they remain relevant models for studying how different viral traits, such as inflammatory potential or immune escape, affect disease severity and recovery.

One promising approach to better understanding the differences in clinical outcomes —such as the course of hospitalization, severity of illness, and risk of death — at the molecular level might be microRNA (miRNA) profiling of samples from patients infected with COVID-19. miRNAs are small, non-coding RNAs that post-transcriptionally regulate gene expression, playing pivotal roles in immune responses and viral pathogenesis. During viral infection, the expression of specific miRNAs can be significantly altered, resulting in distinct miRNA profiles that reflect the host’s immune response, dysregulated inflammation processes, and disease severity, and may vary depending on the viral variant or individual host factors [9]. These different profiles have great potential to serve as prognostic and predictive biomarkers.

In this study, we utilized high-throughput microarray technology to perform an initial, comprehensive profiling of miRNA expression patterns across different stages of infection and varying disease severity in individuals infected with the Delta and Omicron variants of SARS-CoV-2, as these variants are known to trigger distinct immune dynamics.

Materials and methods

Subjects

The study was approved by the Ethics Committee of the Jessenius Faculty of Medicine in Martin, Comenius University Bratislava, with identification number EK 65/2021. All participants in the study signed the informed consent form. Biospecimens were collected from 160 patients hospitalized due to severe respiratory symptoms of COVID-19 in the University Hospital Martin from November 2021 to April 2022. The presence of SARS-CoV-2 was confirmed at admission using a LAMP assay and subsequently verified by RT-PCR analysis within this project. Specific viral variants were routinely sequenced in cooperation with the Public Health Authority of the Slovak Republic, identifying first the Delta variant (B.1.617.2) and, from February 2022 to April 2022, the Omicron variant (B.1.1.529). Patients in both viral variant groups were further classified by clinicians into the following categories based on the clinical manifestations of SARS-CoV-2 infection, according to the COVID-19 Treatment Guidelines: asymptomatic, mild, moderate, severe, critical, and fatal.

The exclusion criteria for the microarray study included age under 18 years, pregnancy, inability or unwillingness to provide written informed consent, incomplete trio of the samples taken at three time points in case of the Delta cohort, and insufficient biospecimen quality standards at any stage of processing (e.g., hemolysis, low RNA yield, or other indicators of inadequate sample integrity). A detailed characterization of the patient groups is presented in Table 1. Blood samples from 17 healthy volunteers (median age, 57.0 years; IQR, 54.0–60.0; 7 females and 10 males) with no prior history of COVID-19 infection were used as controls.

Table 1.

Characteristics of COVID-19 patients included in the study

Median age (IQR)	Delta (n = 23)	Omicron (n = 16)
	68.0 (54.0–74.0)	72.6 (66.0-86.3)
Sex (F/M)	11/12	8/8
Smokers/Ex-smokers	3/5	1/4
Pneumonia	18	9
Asthma	1	2
IHD/CVD	9	11
COPD	2	2
Hypertension	15	12
Obesity	8	7
Diabetes mellitus	8	4
Cancer	1	6
Thyroid disease	6	4
Kidney disease	6	5
Chronic/Acute GID	4/9	2/6
Vaccination	2	5
Course of the COVID-19 disease Mild Moderate Severe Critical Fatal	5 5 1 4 8	8 0 0 0 8

Abbreviations: IQR, Interquartile range; F, Female; M, Male; IHD, Ischemic heart disease; CVD, Cardiovascular disease; COPD, Chronic obstructive pulmonary disease; GID, Gastrointestinal disorders

Samples collection and plasma processing

Peripheral blood samples were collected into EDTA-coated tubes under aseptic conditions and subsequently processed within two hours after collection. Samples in Delta cohort were collected from the patients at three time points: (1) within 24 h after hospital admission (Phase A, assuming fully developed disease); (2) between the 5th and 8th day of hospitalization, depending on the course of the disease and the time of discharge (Phase B, the beginning of the recovery phase); and (3) on average on the 42nd day after the first blood collection (Phase C, assuming full recovery). Samples from Phase B and Phase C were used exclusively for differential analysis across three phases of the Delta variant. All samples for the Omicron cohort were collected within 24 h of hospital admission (Phase A). The whole blood was centrifuged at 4 °C and 380 × g for 20 min. The obtained plasma was stored at − 80 °C until further analysis. Hemolysis levels in plasma samples were measured on a NanoDrop OneC (Thermo Fisher Scientific, Waltham, MA, USA) spectrophotometer at a wavelength of 414 nm, employing a “custom method for hemoglobin measurements” provided by Thermo Fisher Scientific. Prior to isolation, non-hemolysed plasma samples underwent an additional centrifugation step at 16,000 x g and 4 °C for 10 min. The second centrifugation ensured the complete removal of blood fragments and cryoprecipitates for subsequent analyses.

Total RNA isolation

RNA extraction from plasma samples was performed using the ExoRNeasy Midi Kit (Qiagen, Hilden, Germany) following the manufacturer’s protocol. The quality and quantity of the isolated RNA were determined using the Agilent 2100 system (Agilent Technologies, Santa Clara, CA, USA) and the Small RNA kit (Agilent Technologies) according to standard protocol. This kit allows for the analysis of small RNA and miRNA samples ranging from 15 to 200 nucleotides, providing information on both miRNA and other small RNA percentages and quantities. The extracted RNA was stored at − 80 °C for further analysis.

Microarray analysis

Two microliters of total RNA, including miRNAs, were used as input, with all steps performed according to the manufacturer’s protocol (miRNA Microarray System with miRNA Complete Labeling and Hyb Kit, Version 4.1, October 2021). Briefly, spike-ins from the microRNA Spike-in kit (Agilent Technologies) were added to the samples, which were then labeled with cyanine 3 cytidine bisphosphate (pCp-Cy3) using the Agilent miRNA Labeling and Hybridization Kit. To enhance hybridization sensitivity and efficiency, 1 µL of HPLC-purified synthetic DNA poly-A oligonucleotide (TIB Molbiol, Genoa, Italy), 3’ labeled with pCp-Cy3, was incorporated into the hybridization mix at a concentration of 50 amol/µL, following the protocol described by Coco and Vanni [10]. The subsequent hybridization lasted 20 h at 55 °C. During this process, the samples were hybridized to a microarray slide (SurePrint G3 Human miRNA r21, 8 × 60 K, Array kit, Agilent Technologies), where the probes complementary to individual miRNAs were spotted. This system enables the determination of 2,549 miRNA expression levels in a single run. After hybridization, the arrays were washed with two different washing buffers, drained, and scanned using the Agilent SureScan Dx Microarray Scanner. During scanning, emission in the green spectrum was recorded, producing a comprehensive TIFF image. The visual data from these TIFF images were then converted into text files using the Agilent Feature Extraction software (v12.1.1.1), which also generated quality reports used for subsequent sample selection. Samples meeting all required parameters were included in further statistical analyses using the GeneSpring Multi-Omic Analysis program (v14.9.1).

Statistical analysis

Statistical analyses were conducted using the Agilent GeneSpring GX software (Agilent Technologies). Data were normalized to the 75th percentile, which provides a more robust intensity value for normalization. Using a higher percentile intensity effectively captures the median of probes with reliable signals, while probes with low intensity are filtered out. The normality of miRNA expression data was assessed with the Shapiro-Wilk test. Differentially expressed miRNAs between the two groups were identified using the Mann-Whitney test with Benjamini-Hochberg correction. When comparing miRNA expression levels across multiple groups, the Kruskal-Wallis test was applied. Differences in miRNA expression with corrected p-values less than 0.05 and fold change (FC) greater than 2 or less than − 2 were considered statistically significant.

To evaluate global expression patterns and group discrimination, principal component analysis (PCA) on the expression profiles of significantly deregulated miRNAs was performed using the BioChemCalc application (https://biochemcalc.com/e_pca). The input matrix was standardized using Z-score scaling across features to ensure equal weight during variance optimization and dimensionality reduction.

Unsupervised hierarchical clustering of statistically significant deregulated miRNAs was performed using the online tool Morpheus (https://software.broadinstitute.org/morpheus). Clustering of both rows (miRNAs) and columns (samples) was calculated based on the 1 − Pearson correlation metric using the Average linkage method. The resulting expression matrix was visualized as a heatmap, with color-coded expression levels ranging from blue (low expression/row minimum) to red (high expression/row maximum).

Results

Differentially expressed miRNAs in SARS-CoV-2 variants compared to healthy controls

From a total of 2,549 miRNAs spotted on the microarray slide, 798 miRNAs were successfully detected above the background threshold. The top 20 miRNAs with the highest fold change in both variants are presented in Fig. 1. To evaluate the discriminative power and global variance of statistically significant deregulated miRNAs across the groups, we performed a 2D PCA score plot analysis (Fig. 2). Unsupervised hierarchical clustering in Morpheus confirmed the clear separation observed in the PCA (Fig. 3). Additionally, differential expression across specific group comparisons was visualized using volcano plots (Fig. 4).

Fig. 1.

Top 20 differentially expressed miRNAs with the highest absolute fold change values. The bar plots represent the top 20 most deregulated miRNAs in patients infected with the Delta variant (left) and the Omicron variant (right) compared to healthy controls. Green bars indicate upregulated miRNAs, and red bars indicate downregulated miRNAs

Fig. 2.

2D Principal Component Analysis (PCA) score plots of miRNA expression profiles. The plots show the spatial segregation between healthy controls (green dots) and infected patients (red dots) for the Delta variant (left) and the Omicron variant (right)

Fig. 3.

Unsupervised hierarchical clustering and heatmap analysis of deregulated miRNAs. The heatmaps illustrate individual miRNA expression profiles for the Delta variant (left) and the Omicron variant (right). Columns represent individual samples (infected patients Del1-Del23, Om1-Om16, and healthy controls Ctrl1-Ctrl17) and rows represent individual miRNAs. The color scale indicates row-standardized relative expression levels from minimum (blue) to maximum (red)

Fig. 4.

Volcano plots demonstrating differential expression of all detected miRNAs. Compared groups: (A) Delta variant versus healthy controls, (B) Omicron variant versus healthy controls, (C) Delta variant versus Omicron variant, and (D) Delta variant severe/critical/death cases versus mild/moderate cases. Blue dots in the upper left part of each chart represent significantly downregulated miRNAs, while red dots in the upper right part represent significantly upregulated miRNAs (fold change ≥ 2, p < 0.05). The x-axis represents the log2 fold change, and the y-axis represents the -log10 (p-value)

In the group of patients with the Delta variant, we identified 92 differentially expressed miRNAs, of which 87 were upregulated, and five were downregulated (Fig. 4A; Table S1). The most deregulated were miR-4455 (FC = 33.67, p = 5.36 × 10⁻⁷), miR-6785-5p (FC = 25.82, p = 9.20 × 10⁻⁵), miR-32-3p (FC = 23.33, p = 4.47 × 10⁻⁷), miR-3149 (FC = 20.51, p = 5.67 × 10⁻⁶) and miR-574-5p (FC = 19.75, p = 4.05 × 10⁻⁶).

When comparing miRNA expression in the Omicron variant group with healthy controls, we identified 106 upregulated and five downregulated miRNAs (Fig. 4B; Table S2). The most altered expression was in miR-6785-5p (FC = 24.33, p = 7.86 × 10⁻⁴), miR-3652 (FC = 12.50, p = 0.027), miR-30d-5p (FC = 11.82, p = 0.014), miR-4745-5p (FC = 10.30, p = 0.008), and miR-4485-5p (FC = 10.06, p = 0.013).

The comparison between the Delta variant and Omicron variant revealed one upregulated miR-4644 (FC = 3.24, p = 0.039) and four downregulated miRNAs: miR-4745-5p (FC=-6.18, p = 0.026), miR-211-3p (FC=-2.88, p = 0.043), miR-99b-5p (FC=-2.46, p = 0.032), and miR-17-3p (FC=-2.01, p = 0.023) (Fig. 4C).

To understand how these distinct variant profiles behave on a global scale, we examined their individual variance relative to healthy controls in our multivariate model. As visualized in the 2D PCA score plot (Fig. 2), for the Delta variant, the separation trend along the PC1 is clearly visible (PC1 = 50.32%, PC2 = 14.68%); however, the cohort exhibited substantial inter-individual dispersion and a moderate spatial overlap with controls, indicating a heterogeneous expression profile. For the Omicron variant, the PCA plot showed a tighter and clearer separation along the PC1 axis (PC1 = 51.09%, PC2 = 13.00%), with healthy controls tightly clustering in the negative domain and patients localizing to the positive domain.

This distinct separation pattern and the global variance observed in the PCA were further evaluated at the individual level using unsupervised hierarchical clustering (Fig. 3). For both variants, the sample dendrograms successfully achieved a non-overlapping segregation, partitioning all infected patients from healthy controls into two primary branches. The heatmaps clearly mirrored the quantitative Fold Change data, displaying a tightly clustered set of highly upregulated miRNAs (red domain) dominated by top candidates such as miR-6785-5p and miR-4455. Reciprocally, the top downregulated targets, including miR-150-5p and miR-1183, formed a distinct suppressed block (blue domain) in patient samples. These visual findings demonstrate that despite individual heterogeneities captured by PCA, the deregulated miRNA profiles allow clear differentiation between patients and healthy controls.

miRNAs associated with the severity of the symptoms

For the following comparison, we divided patients with the Delta variant into two distinct groups based on the severity of their symptoms. The first group included patients with only mild or moderate symptoms (n = 10), while the second one comprised patients with severe and critical symptoms, including fatal outcomes (n = 13). Comparing the group with more severe symptoms to the first group, we identified 10 significantly altered miRNAs: seven upregulated and three downregulated (Fig. 4D). The upregulated miRNAs were: miR-6808-5p (FC = 9.54, p = 0.004), miR-6723-5p (FC = 7.60, p = 0.049), miR-6826-5p (FC = 3.94, p = 0.009), miR-3663-3p (FC = 3.58, p = 0.021), miR-4270 (FC = 3.65, p = 0.032), miR-4516 (FC = 3.13, p = 0.005), and miR-4281 (FC = 2.93, p = 0.006). The downregulated miRNAs: miR-150-5p (FC=-4.91, p = 0.039), miR-371b-5p (FC=-4.84, p = 0.020), and miR-4672 (FC=-4.22, p = 0.018).

The patients with the Omicron variant were stratified into a group that exhibited mild symptoms (n = 8) and a group with a fatal outcome (n = 8). The comparison did not reveal any significantly deregulated miRNAs.

Differential miRNA expression in the various phases of the Delta variant

In the group of patients who recovered from the COVID-19 Delta variant, we were able to compare miRNA expression levels in three different phases of the disease versus healthy controls: assumed fully developed disease upon hospital admission, the beginning of the recovery phase during hospitalization, and assumed full recovery approximately one month after hospital discharge. By comparing the three time points, we identified 38 significantly altered miRNAs, 37 upregulated and one downregulated (Table S3). The most altered expression was in miR-4455 (Phase A: FC = 63.82, Phase B: FC = 36.09, Phase C: FC = 2.81; p = 1.89 × 10⁻⁴), miR-595 (A: FC = 46.04, B: FC = 19.60, C. FC = 1.75; p = 2.21 × 10⁻⁴) or miR-144-3p (A: FC= -17.48, B: FC= -40.15, C: FC= -1.15; p = 0.001). The top 10 miRNAs with the highest difference between the three phases are presented in Fig. 5.

Fig. 5.

Expression profiles of the top 10 differentially expressed miRNAs across three clinical phases of the Delta variant. Data are presented as fold change relative to the control group of healthy volunteers. Phase A: blood collected within 24 h of hospital admission; Phase B: blood collected after approximately one week of hospitalization; Phase C: control sampling one month after hospital discharge. Statistical significance across the three phases is indicated above the bars (*p < 0.05, **p < 0.01, ***p < 0.001)

miRNAs associated with the clinical parameters

Alongside the changes in miRNA expression observed between the Delta and Omicron variants of SARS-CoV-2 and healthy controls, we further investigated alterations in miRNA expression levels in COVID-19 patients in relation to their coexisting diseases and clinical parameters. We detected significant changes in expression levels across a spectrum of 9 clinical conditions in a total of 232 miRNAs: (1) Diabetes mellitus (n = 12), 90 deregulated miRNAs, (2) cardiovascular diseases (n = 20), 66 miRNAs, (3) kidney diseases (n = 11), 41 miRNAs, (4) obesity (n = 15), nine miRNAs, (5) chronic gastrointestinal diseases (n = 15), 11 miRNAs, (6) cancer (n = 7), nine miRNAs, (7) thyroid disorders, n = 10, 3 miRNAs, (8) gastrointestinal symptoms upon hospital admission, (n = 15), two miRNAs, and (9) hypertension (n = 27), one miRNA. More detailed information regarding variations in miRNA expression in relation to clinical parameters and conditions of patients with SARS-CoV-2 can be found in Supplementary Tables S4-S7.

microRNA-mRNA interactions

Within all analyzed groups, we identified differences in the expression levels of a total of 256 miRNAs. However, some miRNAs exhibited varying expression levels across multiple groups simultaneously. Upon eliminating these instances, the number of uniquely differentially expressed miRNAs was reduced to 142. These miRNAs underwent further analysis using the online tool miRTargetLink 2.0 [11], a multifunctional tool primarily designed for predicting miRNA-mRNA interactions. It utilizes data from 8 resources: miRbase, miRTarBase, mirDIP, miRDB, miRPathDB, miRATBase, miEAA, and GeneTrail. By accepting strongly validated interactions, the tool assigned a total of 571 target mRNAs to 48 of these miRNAs. Weak and predicted interactions were filtered out. The list of all strongly validated target genes and their associated miRNAs from our set of deregulated mRNAs is provided in Supplementary Table S8.

Subsequently, we focused on targets that interacted with the highest number of miRNAs in at least two previously conducted comparisons (50 genes). The PTEN gene emerged as the most frequently interacting gene, engaging eight specific miRNAs in five different comparisons: miR-21-5p, miR-92a-3p, miR-103a-3p, miR-144-3p, miR-221-3p, miR-222-3p, miR-377-3p, and miR-494-3p. Following was the IGF1R gene interacting with six miRNAs: let-7b-5p, let-7c-5p, let-7e-5p, miR-21-5p, miR-99b-5p, and miR-494-3p; and MYC and STAT3 genes with five interacting miRNAs, respectively: let-7c-5p, miR-24-3p, miR-320b, miR-451a, and miR-494-3p (MYC) and let-7c-5p, miR-21-5p, miR-92a-3p, miR-125a-5p, and miR-4516 (STAT3). The interaction network of genes targeted by a minimum of 4 different miRNAs, including both upregulated and downregulated miRNAs, is shown in Fig. 6.

Fig. 6.

Interaction network generated by miRTargetLink between the deregulated miRNAs and the genes with a minimum of four shared targets. Upregulated miRNAs are represented by blue circles, downregulated miRNAs are indicated by red circles, and their corresponding target genes are shown as green circles

Protein-protein interactions

Functional enrichment analysis of targeted genes was conducted using the STRING interactive web tool [12]. STRING is a database and web resource that predicts protein-protein interactions, enabling the creation and analysis of interaction networks for a genome of interest. We input all the previously assigned target genes into the Multiple protein search tool. The search tool returned 12,530 predicted associations with an average clustering coefficient of 0.487 and 6,210 significantly enriched terms across 14 categories covering Gene Ontologies, pathways, and domains. Among those outputs, we searched for relevant terms such as “sars”, “covid”, “coronavirus”, “cytokine”, “interleukin”, and others. The most relevant search results are listed in Supplementary Table S9.

Due to the overwhelming number of outputs, we reduced the number of genes undergoing STRING analysis. We prioritized genes interacting with at least four from our set of deregulated miRNAs and repeated the protein search with 50 genes. The number of predicted associations was reduced to 493 (Fig. 7), with a higher average clustering coefficient of 0.745. The functional enrichments in our network listed 2,382 enriched terms in 12 categories. Again, we searched for the same keywords as previously. Table 2 summarizes the key findings of the search, with specific indications for genes targeted by both upregulated and downregulated miRNAs.

Fig. 7.

Network of predicted protein-protein interactions for 50 selected targets generated by STRING. Nodes represent proteins; network edges indicate functional and physical associations derived from multiple data sources. Single disconnected nodes display proteins with no predicted interactions within this specific set

Table 2.

The search results from functional enrichments of 50 selected proteins associated with at least four of the deregulated miRNAs

Category	Description	Observed / background genes count	Observed genes
Reactome	SARS-CoV-1 targets host intracellular signalling and regulatory pathways	3/15	EP300*, SP1*, SMAD3
WikiPathways	T-cell activation SARS-CoV-2	7/88	MAPK1*, CCND1, TP53*, FOXO3, MTOR*, PTEN*, AKT0
WikiPathways	Extrafollicular and follicular B cell activation by SARS-CoV-2	3/72	ETS1*, CXCR4, ZEB1*
WikiPathways	SARS-CoV-2 innate immunity evasion and cell-specific immune response	2/66	TGFB1*, EP299
GO Process	Response to virus	5/356	TGFB1*, CDK6, BCL2*, CXCR4, HMGA2
WikiPathways	Host-pathogen interaction of human coronaviruses - apoptosis	3/21	BCL2L11, BCL2*, AKT1*
WikiPathways	Host-pathogen interaction of human coronaviruses - MAPK signaling	2/35	MAPK1*, BCL2*
UniProt Keywords	Host-virus interaction	10/540	MAPK1*, EP300*, STAT3, ICAM1, TP53*, SP1, CXCR4, CREB1*, H2AX, CCNA2
GO Process	Response to cytokine	14/804	MAPK1*, CORO1A, SFRP1, STAT3, ICAM1, TIMP3, TP53*, FOXO3, SMAD4*, BCL2*, CXCR4, CREB1*, AKT1*, NOTCH1*
GO Process	Cellular response to cytokine stimulus	12/711	MAPK1*, CORO1A, SFRP1, STAT3, ICAM1, TP53*, FOXO3, SMAD4*, CXCR4, CREB1*, AKT1*, NOTCH1*
GO Process	Cytokine-mediated signaling pathway	7/369	STAT3, TP53*, FOXO3, SMAD4*, CXCR4, AKT1*, NOTCH1*
Reactome	Cytokine Signaling in Immune system	15/706	MAPK1*, TGFB1*, CCND1, CDKN1B, STAT3, ICAM1, TP53*, BIRC5*, FOXO3, ZEB1*, BCL2L11, BCL2*, CREB1*, AKT1*, MYC*
GO Process	Cellular response to interleukin-17	2/13	STAT3, NOTCH1*
GO Process	Interleukin-6-mediated signaling pathway	2/14	STAT3, SMAD4*
Reactome	Interleukin-4 and Interleukin-13 signaling	11/107	TGFB1*, CCND1, STAT3, ICAM1, TP53*, BIRC5*, FOXO3, ZEB1*, BCL2*, AKT1*, MYC*
Reactome	Signaling by Interleukins	13/453	MAPK1*, TGFB1*, CCND1, STAT3, ICAM1, TP53*, BIRC5*, FOXO3, ZEB1*, BCL2*, CREB1*, AKT1*, MYC*
WikiPathways	Interleukin-11 signaling pathway	8/44	MAPK1*, TGFB1*, STAT3, ICAM1, BIRC5*, BCL2*, CREB1*, AKT1
GO Process	Response to vitamin D	2/28	SFRP1, TGFB1*

Note: No genes were targeted exclusively by downregulated miRNAs; genes with no mark interact with upregulated miRNAs only; genes marked with asterisk (*) interact with both upregulated and downregulated miRNAs

Discussion

The biological consequences of SARS-CoV-2 infection arise from physiological, structural, biochemical, morphological, and genetic alterations in infected cells [13]. Since the emergence of SARS-CoV-2, several mechanisms underlying COVID-19 pathogenesis and how the virus interacts with host cells have been identified, along with multiple strategies the virus employs to evade the immune system [14]. During the pandemic, considerable evidence suggested that different SARS-CoV-2 variants can cause varying outcomes because they affect distinct molecular pathways, including epigenetic machinery [15]. Recently, several studies have demonstrated a strong link between gene-specific epigenetic modifications in host cells and the disease’s pathogenesis and its severity [16]. Identifying and determining epigenetic changes is crucial for understanding the disease, as they can serve as early indicators of risk, diagnostic or prognostic markers, and therapeutic targets. In our study, we examined the changes in miRNA expression during infection with two different variants of SARS-CoV-2, specifically the Delta and Omicron variants, using high-throughput microarray technology to simultaneously measure the expression of thousands of miRNAs.

As epigenetic regulators of gene expression, miRNAs play a crucial role during viral infection. There are several mechanisms by which miRNAs interact with viruses; overall, these miRNAs can either facilitate or impede viral replication, depending on various factors [17]. Additionally, numerous miRNAs can regulate various pathways associated with viral infection. Through these mechanisms, miRNAs can affect the pro-inflammatory cascade, immune regulation, and the activity of immune cells, ultimately influencing the course and outcome of the infection itself. In our study, we identified numerous miRNAs with statistically significant differences in expression associated with a specific SARS-CoV-2 variant. Interestingly, the same 65 miRNAs showing significant changes were observed in both SARS-CoV-2 variants, suggesting a shared pattern of inflammatory processes regardless of whether the infection was caused by the Omicron or Delta variant. Regarding their role in viral infection, all significantly altered miRNAs were categorized based on their regulatory activity into three groups: (1) inflammation-associated miRNAs, (2) miRNAs involved in host immune responses, viral replication, and immune evasion mechanisms, and (3) miRNAs that regulate immune cells.

Inflammation is characterized as a biological response of the immune system that can be triggered by viral infection, and miRNAs play a significant regulatory role in these processes [18]. Our analysis revealed significant changes in miR-32-3p, miR-574-5p, miR-320a, miR-320b, or miR-320e in patients infected with either the Delta or the Omicron variant. These miRNAs play a crucial role in regulating inflammatory pathways activated during infection, as well as in chronic inflammation [19–21]. In contrast, only patients infected with the Delta variant exhibited significant changes in miRNAs involved in regulating inflammation, such as miR-21-5p and miR-221-3p. These miRNAs can act as regulators of inflammatory processes in the context of various diseases associated with chronic inflammation (i.e., Crohn´s disease and other autoimmune diseases) or pathologies closely linked to the production of inflammatory cytokines and interferons [22–25]. In the analysis of the Omicron group of patients, we identified a different set of inflammation-associated miRNAs, including miR-125a-5p, miR-222-3p, miR-331-3p, miR-377-3p, and miR-484. Deregulation of the mentioned miRNAs is supported by findings from other studies that confirm them as regulators of the inflammatory pathway [26–30].

Beyond their role in inflammation, certain miRNAs can modulate host responses to infection, influencing viral replication and immune evasion. We detected significant changes in let-7b-5p, let-7d-3p, miR-24-3p, miR-32-3p, miR-144-3p, miR-150-5p or miR-197-3p in both variants. Analysis of samples obtained from Omicron patients revealed variant-specific alterations in miR-125a-5p and miR-574-3p, while analysis of Delta patients identified variant-specific differences in miR-4299. The mentioned miRNAs have also been documented in experimental studies, where they play a crucial role in virus infection [17, 31–38].

Another key aspect of miRNA function during viral infection is their ability to modulate immune cells [39]. These changes in miRNA expression impact various aspects of the immune system, including the activation of immune cells, the regulation of inflammation, and the prevention of autoimmune reactions. Notably, in both the Delta and Omicron variants of SARS-CoV-2, we found significantly upregulated let-7b-5p, let-7d-3p, or miR-24-3p. It has been documented that these miRNAs influence the function of immune cells [40–42].

Our analysis yielded extensive data, identifying a large set of differentially expressed miRNAs in both variants compared to healthy controls. These results demonstrate the overall view of the issue of miRNA expression associated with SARS-CoV-2 infection. Most of the discussed miRNAs share similar or the same pathways affecting different processes linked to viral replication, immune modulation, and inflammation.

To explore the clinical implications of these findings, we also focused on identifying specific miRNAs associated with symptom severity. Several studies focusing on individual miRNAs have identified a correlation between a specific miRNA signature and the severity of COVID-19 and patients’ outcomes [43–45]. In our study, we compared miRNA expression changes between groups with mild or moderate symptoms and those with severe and critical clinical manifestations in patients with Delta variants. Our analysis revealed the upregulation of miRNAs associated with immune regulation and inflammatory responses, as well as viral interaction, and a reduced immune response. Specifically, miRNAs such as miR-4270 and miR-4281 are strongly linked to the regulation of immune response, miR-150-5p is implicated in viral interaction, and miRNA-3663-3p is involved in inflammatory pathways [34, 46–48]. On the contrary, analysis of miRNA expression differences in Omicron patients between the group with mild symptoms and those with fatal outcomes showed no significantly altered miRNAs. There are several reasons for these variations, including differences in viral pathogenesis and immune evasion strategies. The Delta variant has been shown to cause more severe illness compared to earlier variants and triggers a stronger immune response, which may be related to the upregulation of specific miRNAs involved in immune regulation and inflammation [15, 49]. Another reason could lie in differences in viral load and replication, as well as variability in disease progression [50, 51].

In our further analysis, we examined the differences in miRNA expression across the various phases of the Delta variant. The results acquired from these experiments showed differences in the expression of 38 miRNAs across three distinct phases of disease: fully developed disease, the beginning of recovery, and assumed full recovery. The most altered expression was documented in miR-4455, miR-595, and miR-144-3p. As previously discussed, these miRNAs are strongly linked to viral replication, the immune response, immune evasion by viruses, or inflammation [52, 53].

Based on the differentially expressed miRNAs from our analysis, we predicted their target genes and analyzed miRNA–mRNA interactions, followed by functional annotation to gain insight into the biological processes and pathways potentially regulated by these miRNAs. Within all analyzed groups, we identified differences in the expression levels of a total of 142 miRNAs. Our data showed that the most interacting genes were PTEN, IGF1R, MYC, and STAT3. All these genes are involved in immune regulation, viral infection, and inflammation [54–57]. Thus, given the key roles of these genes in regulating immune functions and inflammatory pathways, their involvement in SARS-CoV-2 infection is expected.

In conclusion, the simultaneous analysis of thousands of miRNAs involved in the host response to SARS-CoV-2 infection, mediated by different variants, opens new opportunities in diagnosis, prognosis, and therapy. Recent evidence suggests differences between the Delta and Omicron variants; however, their effects on the epigenetic machinery in host cells have not been well investigated. Our study revealed differences in miRNA expression between two SARS-CoV-2 variants, which show distinct clinical manifestations and patient outcomes. Using a miRNA microarray enables the identification of specific miRNA signatures among COVID-19 variants and healthy controls, as well as their associations with symptom severity, the monitoring of different disease stages, and the detection of miRNA-mRNA or protein-protein interactions. Nonetheless, a comprehensive analysis is sensitive to the influence of comorbidities or cross-talk within regulatory pathways, owing to overlapping effects on miRNA expression. Despite these limitations, our research offers innovative and new insights into SARS-CoV-2 infection.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

COVID-19

Coronavirus Disease 19

CVD

Cardiovascular disease

EDTA

Ethylenediaminetetraacetic Acid

FC

Fold Change

GO

Gene Ontology

GID

Gastrointestinal disorders

HPLC

High-Performance Liquid Chromatography

IHD

Ischemic heart disease

IQR

Interquartile Range

LAMP

Loop-Mediated Isothermal Amplification

miRNA

microRNA

mRNA

Messenger Ribonucleic Acid

PC1

First Principal Component

PC2

Second Principal Component

PCA

Principal Component Analysis

PPI

Protein-Protein Interaction

RT-PCR

Reverse Transcription Polymerase Chain Reaction

SARS-CoV-2

Severe Acute Respiratory Syndrome Coronavirus 2

Author contributions

I.B, D.D, D.B.; Investigation, Analysis microarray; statistical analyses, Software, Writing – original draft, Writing – review & editing. A.K.; Conceptualization, Analysis – microarray; Writing – review & editing. M.S, E.B.; Conceptualization, Writing – review & editing. V.H, Z.K, M.S, D.L.; Analysis – qPCR, Writing – review & editing, Conceptualization. E.N.; Conceptualization, Resources, Validation. E.H.; Supervision, Validation, Writing – review & editing, Resources, Funding. P.L, P.B, A.B, R.R.; Data curation, Supervision, Writing – review & editing. Z.D.; Investigation, Resources, Writing – review & editing, Supervision.

Funding

This study and publication was produced with the support of the Integrated Infrastructure Operational Program for the project: New possibilities for laboratory diagnostics and massive screening of SARS-CoV-2 and identification of mechanisms of virus behaviour in human body, ITMS: 313011AUA4, co-financed by the European Regional Development Fund and with the support of the Integrated Infrastructure Operational Program for the project: Research and development of a telemedicine system to support the monitoring of a possible spread of COVID-19 in order to develop analytical tools used to reduce the risk of infection, ITMS: 313011ASX4, co-financed by the European Regional Development Fund.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

Authors declare that they have no conflict of interest.

Ethics approval and consent to participate

The study was approved by the Regional Ethics Committee of the Jessenius Faculty of Medicine (code 65/2021), and the research was performed in compliance with the Declaration of Helsinki. Informed consent was obtained from all participants.