Abstract
The COVID-19 pandemic has created an urgency to study the host gene response that leads to variable clinical presentations of the disease, particularly the critical illness response. miRNAs have been implicated in the mechanism of host immune dysregulation and thus hold potential as biomarkers and/or therapeutic agents with clinical application. Hence, further analyses of their altered expression in COVID-19 is warranted. An important basis for this is identifying appropriate reference genes for high quality expression analysis studies. In the current report, NanoString technology was used to study the expression of 798 miRNAs in the peripheral blood of 24 critically ill patients, 12 had COVID-19 and 12 were COVID-19 negative. A list of potentially stable candidate reference genes was generated that included ten miRNAs. The top six were analyzed using reverse transcription quantitative polymerase chain reaction (RT-qPCR) in a total of 41 patients so as to apply standard computational algorithms for validating reference genes, namely geNorm, NormFinder, BestKeeper and RefFinder. There was general agreement among all four algorithms in the ranking of four stable miRNAs: miR-186-5p, miR-148b-3p, miR-194-5p and miR-448. A detailed analysis of their output rankings led to the conclusion that miR-186-5p and miR-148b-3p are appropriate reference genes for miRNA expression studies using PaxGene tubes in the peripheral blood of patients critically ill with COVID-19 disease.
Introduction
The World Health Organization declared the coronavirus disease 2019 (COVID-19) pandemic in March 2020 [1]. As of August 2022, COVID-19 has accounted for more than 577 million confirmed cases and more than six million confirmed deaths [2]. The course of illness of COVID-19 ranges from mild to severe and can be vastly unpredictable [3, 4], however certain groups are particularly vulnerable to severe illness such as the elderly and comorbid patient population [5]. Mortality rates in the literature range from 17–39% among critically ill patients [5]. While the advent of vaccination has had a remarkable effect in preventing serious illness [6, 7], vaccinated individuals can suffer breakthrough infections with high illness severity [4]. Given the variable clinical course of this disease, clinicians face challenges in assessing (i) a patient’s vulnerability and (ii) deciding when and if to initiate appropriate therapeutic interventions.
Scientists have diverted a unique focus to understanding the dysregulated host immune response in COVID-19 responsible for variable disease phenotypes described above. Both the innate and adaptive immune responses contribute to an overactive inflammatory process, which may lead to end organ damage [8]. The virus enters respiratory epithelial cells through the ACE2 receptor and can evade intracellular immune responses and affect an altered type I interferon response [9, 10]. It is not yet known why some people have asymptomatic infection while others progress toward potentially lethal immune activation, however, genetic variants in immune components have been proposed [11, 12].
A complete picture on how COVID-19 leads to severe illness likely includes various mechanisms such as host immune response, inflammation and damage repair. Several studies have identified potential peripheral blood markers of COVID-19 related to those processes [13–16]. Other reports have turned to blood transcriptomics to gain large scale data on potential mechanisms attributable to COVID-19 outcomes [17–21]. Blood transcriptomics has emerged in the last decade as a rich source of biomarker discovery with applications in various fields including sepsis [22], cancer [23] and cardiovascular disease [24].
The non-coding RNAome is a key aspect of transcriptomic work alongside analyses of messenger RNA. miRNAs are endogenous non-coding RNA molecules of approximately 21 nucleotides in length [25]. They affect gene expression by binding to an mRNA’s 3’-untranslated region (UTR) to degrade or inhibit protein translation [25]. The circulating non-coding transcriptome, including miRNAs, has been shown to provide beneficial information as a biomarker of disease diagnosis and prognostication for adverse outcomes, with potential for therapeutic intervention that has reached clinical trials [26–30].
An important aspect of high-quality transcriptomic work involves technical verification of differentially expressed nucleic acids using reverse transcription quantitative polymerase chain reaction (RT-qPCR) as a gold standard. Accurate and reliable results rely heavily on the selection of appropriate reference genes. Several computational programs have been designed to aid in the selection of stable housekeeping genes across a range of test and control samples for a given experimental context [31–34]. The importance of such programs as a starting point for achieving reliable results in differential expression analyses has been well established [35, 36].
In the current analysis, reference genes have been systematically selected for miRNA studies involving the peripheral blood of COVID-19 patients collected in PAXgene tubes. PAXgene tubes allow for ease of sample collection and processing with prolonged maintenance of nucleic acid stability, and as such their use in transcriptomic work is widespread [20]. The objective of this study was to 1) identify potential housekeeping genes across a range of clinical samples from patients with critical illness and 2) verify a subset of housekeeping genes for direct use in miRNA expression analyses in the particular clinical context presented here–critically unwell patients with and without COVID-19.
Materials and methods
Ethics and blood collection
Approval for the collection of peripheral blood from critically ill patients was obtained from the St. Michael’s Hospital Research Ethics Board (REB number 20–078). Written consent from substitute decision makers was obtained for enrollment in our study as per institution protocol. Blood was collected within 48 hours of ICU admission into PAXgene Blood RNA Tubes (PreAnalytiX, Hombrechtikon, Switzerland, catalog number 762165) as per manufacturer instructions. Briefly, approximately 2.5mL of blood was drawn directly into the PAXgene tube and stored at -80°C until analysis.
Nucleic acid extraction
Before analysis, samples were thawed and kept at room temperature with PAXgene reagent for approximately 1.5hrs to increase yield. Total RNA, including miRNAs, was extracted using the PAXgene blood miRNA kit (PreAnalytiX catalogue number 763134) according to manufacturer instructions. RNA quantification was carried out with the Qubit 4 Fluorometer (Invitrogen, catalogue number Q33226) ThermoFisher Scientific, MA, USA, using the RNA HS assay kit (same company; product number Q32852). A subset of samples were analyzed using the Agilent 2100 Bioanalyzer (Agilent Technologies; catalogue number G2939BA), CA, USA, with the RNA 6000 NanoKit (same company, PN #5067–1511) to ensure a minimum RIN of 8.
NanoString quantification
100ng of total RNA was used as an input for the NanoString nCounter human v3 miRNA expression panel (CSO-MIR3-24) (NanoString, Washington, United States), which profiles 798 human miRNAs. Samples (24 in total) were prepared as recommended under the nCounter miRNA Expression Assay with the following modification: an attenuation probe master mix was created using miR-451a (40nM) and miR-16-5p (4nM) attenuating oligonucleotides in order to remove 99% and 90% of these miRNAs, respectively. The oligo sequences were as follows: hsa-miR-451a – GCCCATAGTTATTAATCTCTGTTTCAGCAATGAAAACCACCGCAAAGAAG; hsa-miR-16-5p – ATGCTAACGCTTTAGAGTATTTTGATGCGCGTTTAAAAGAGATTTTAGAC. Two microlitres of the Attenuation Probe Master Mix was mixed in with each sample after adding the Reporter/Hybridization mix.
Data analysis, selection and validation of reference genes
Normalized counts for each gene per sample were generated with ROSALIND® NanoString miRNA Expression Methods (https://rosalind.onramp.bio/), whereby normalization is carried out using criteria as stated in the NanoString nSolver 4.0 User Manual.
Two methods were used to generate a list of candidate reference genes (Method 1 and Method 2). In Method 1, the standard deviation (STDEV) of normalized counts per gene across all 24 samples (analyzed in NanoString experiments) was computed. Given this study searches for appropriate reference genes and good expression in all samples is needed for a reference gene to be successful, only those genes expressed in all samples, with normalized counts greater than or equal to 8 were included. Then, a ranking of genes on the basis of their STDEV was done and the top three genes with the smallest STDEV were carried forward for validation assays.
For Method 2, candidate in silico miRNA controls were identified using NanoString.RCC files generated from the same 24 samples, which were read into R [37] using the NanoStringQCPro [38] package, and each sample was evaluated for quality control metrics related to imaging, binding density, limit of detection, and positive linearity of controls using the NanoNormIter package [39]. All samples passed the quality control metric based on previously established guidelines [39]. As the housekeeping genes B2M (p<0.05), GAPDH (p<0.05), and ACTB (p<0.1) were found to be differentially expressed between COVID-19 + and COVID-19 –groups using a negative binomial generalized linear model, we sought to find in silico controls that were not differentially expressed. Upper quartile normalization was performed on the endogenous probes, followed by estimating the common and tagwise dispersions before running a negative binomial generalized log-linear model using functions from the edgeR package [40] to test for differential expression between COVID-19 + and COVID-19—patients. In silico controls were selected from 100 least differentially expressed probes, further filtered for a mean expression level of log2 > 6.5 (~90) across samples, and a coefficient of variation (CV) < = 0.50. Again, the top three selected probes across all samples, were carried forward in RT-qPCR assays.
Reverse transcription quantitative PCR (RT-qPCR)
10ng of total RNA per sample were used for the reverse transcription reaction using the TaqMan Advanced miRNA cDNA Synthesis Kit (Applied Biosystems, Massachusettes, United States, A28007) as per manufacturer instructions. Real time PCR was carried out using TaqMan Advanced miRNA Assays (Applied Biosystems, A25576) and the TaqMan Fast Advanced Master Mix (Applied Biosystems, 4444963) on duplicate technical replicates for a total of 41 patients (please note these included the 24 nanostring patients) with and without COVID-19 as per manufacturer instructions, however, the final qPCR reaction was carried out in a total of 10uL of reaction volume (volumes for each reagent were scaled down proportionally). The sample maximization method was used, whereby all the patient samples and technical replicates were run in the same plate for a single gene to avoid inter-plate variability [41]. The assays used were (top 3 from Method 1 and top 3 from Method 2): hsa-miR-148b-3p, hsa-miR-2116-5p; hsa-miR-216b-5p; hsa-miR-448; hsa-miR-194-5p; hsa-miR-186-5p; Real time qPCR (RT-qPCR) was carried out using QuantStudio™ 7 Flex Real-Time PCR System, 384-well, desktop (Applied Biosystems, catalog number 4485701).
The averaged Ct values for technical replicates (Ct duplicate data less than or equal to 0.5 cycles only was included) per miRNA assay for each of 41 patient samples were input into Reffinder (accessed at http://blooge.cn/RefFinder/?type=reference) [42]. The Pearson correlation for technical replicates for the four candidate reference genes used for downstream analysis were as follows: hsa-miR-148b-3p (0.98), hsa-miR-448 (0.98), hsa-miR-194-5p (0.76) and hsa-miR-186-5p (0.95).
The averaged expression levels of the 4 candidate reference miRNAs were tested for associations with the following study population characteristics in the 41 patient cohort: sex, age, the presence of comorbidities (hypertension, type II diabetes, dyslipidemia) and treatments (antibiotics, steroid use, respiratory support). For binary metadata (2 groups), miRNA expression was compared using a Welch’s 2-sided t-test (S1 Table). Categorical metadata with 3 or more groups was modelled against miRNA expression using ANOVA. Continuous parameters were modelled using linear regression (S2 Table). No significant associations (p < 0.05) were found between any reference miRNA and the study population characteristics tested.
Software programs used for expression stability analysis of reference genes
The six candidate reference genes hsa-miR-148b-3p, hsa-miR-2116-5p, hsa-miR-216b-5p, hsa-miR-448, hsa-miR-194-5p and hsa-miR-186-5p were analyzed by RT-qPCR as described above in 41 patients with and without COVID-19 for their expression stability. Average Ct values for four candidate reference genes that had adequate expression in RT-qPCR (hsa-miR-148b-3p, hsa-miR-448, hsa-miR-194-5p and hsa-miR-186-5p) were input into Genorm [33], Normfinder [32], BestKeeper [31] and RefFinder [42] (raw Ct data can be found in S3 Table).
GeNorm and NormFinder algorithms were used as implemented in the Bioconductor packages ReadqPCR v 1.42.0 and NormqPCR v 1.42.0 [43]. The quantification cycle (Ct) values (also known as the threshold-cycle (Ct) values) of technical replicates are imported into R (v 4.2.0) using the read.qPCR function to generate an r-object of class “qPCRBatch” to contain the miRNA gene expression and associated clinical metadata. To identify potential housekeeping genes, the selectHKs function from NormqPCR is subsequently used to find the optimum reference genes, with minimum number of housekeeping genes set to 2 and log = TRUE. With the geNorm method, stability values M is calculated sequentially to narrow down to a final set of two genes. For the NormFinder algorithm, the gene expression stability value rho is calculated on the same qPCRBatch object.
BestKeeper (version 1) was used through the Excel-based tool obtainable from https://www.gene-quantification.de/bestkeeper.html#download (accessed August 10th 2022).
The algorithm is based on the concept that lower variation in Ct values (when cDNA amount is constant) reflects higher gene expression stability. Ct values are used as input data and the algorithm then calculates various values, many of which are markers of variation in data and include: geometric (Geo Mean), arithmetic mean (Ar Mean), minimum and maximum Ct, Ct standard deviation (Std Dev) and coefficient of variation (CV). Ct values with Std Dev higher than 1 are considered inconsistent. The authors suggest using more than one reference gene in order to get more reliable results. For all possible pairs of reference genes, the algorithm performs pair-wise correlation analyses and for every calculated correlation, the Pearson correlation coefficient (r) and the probability p value are calculated. The highly correlated House Keeping Genes (HKGs) are combined into an index, and then a comparison is made between each candidate HKG and the index yielding a Pearson correlation coefficient (r), coefficient of determination (r2) and p-value.
RefFinder was used through http://blooge.cn/RefFinder/?type=reference (accessed on August 10th 2022). It provides a user-friendly web-based platform to analyze large datasets of reference genes and integrates the major computational programs (GeNorm, NormFinder, BestKeeper and the comparative delta Ct method) in its analysis. The output from each software program is then used to assign an appropriate weight to each gene, calculate the geomean of weights and use this to output an overall final ranking.
Results
Candidate reference gene selection
NanoString technology was used to profile 798 human miRNAs in the peripheral blood of 24 patients, 12 of which were COVID-19 positive and 12 which were negative.
The miRNA expression data was then processed in two distinct manners–Method 1 and Method 2 –so as to obtain an array of candidate reference genes to test in RT-qPCR and major computational programs that analyze reference genes. RT-qPCR/computational analysis was done in a total of 41 patients (of note, this included 16 of the NanoString patients already mentioned). Table 1 lists the top five miRNAs deemed as most stable by each method.
Table 1. Five top ranking candidate miRNA reference genes generated by two separate methods using NanoString data.
| Method 1 | Method 2 |
|---|---|
| Human miRNA | Human miRNA |
| miR-148b-3p | miR-2116-5p |
| miR-194-5p | miR-216b-5p |
| miR-186-5p | miR-448 |
| miR-331-3p | miR-944 |
| miR-30e-5p | miR-146a-5p |
Validation of candidate reference gene expression stability in the peripheral blood of patients with COVID-19
The top three stable miRNAs resulting from each of Method 1 and Method 2 were used in an RT-qPCR experiment to generate data that can be subject to appropriate reference gene validation by four gold standard computational algorithms. This was done in a total of 41 patients. Table 2 displays the patient characteristics.
Table 2. Patient characteristics.
| Characteristic | Nanostring cohort (n = 24) | RT-qPCR cohort (n = 41, including 16 patients from nanostring cohort) |
|---|---|---|
| Age, median years (IQR) | 60 (24.8) | 61 (21) |
| Sex, n (%) | ||
| Female | 9 (37.5) | 8 (19.5) |
| Male | 15 (62.5) | 33 (80.4) |
| COVID-19 status, n (%) | ||
| Positive | 12 (50) | 20 (51.7) |
| Negative | 12 (50) | 21 (48.3) |
| Initial SOFA score (within 48 hours of admission), mean (STDEV) | 7.7 (4.1) | 8.3 (4.0) |
The top six miRNAs, as shown in Table 2, were as follows: miR-148b-3p, miR-194-5p, miR-186-5p, miR-2116-5p, miR-216b-5p and miR-448. RT-qPCR analysis yielded poor results for miR-2116-5p and miR-216b-5p –a large number of samples had poor expression with “undetermined” values. Given that reference gene quality is incumbent on high expression in the sample population at hand, these two candidate reference genes were eliminated from further analysis. Samples were included in the analysis only if technical replicate Ct values were equal to or less than 0.5 cycles. For the candidate miRNAs carried forward for analysis by reference gene selection algorithms, the minimum and maximum values of averaged Ct values across all patients were as follows: miR-194-5p (20.02 to 24.76); miR-186-5p (20.96 to 25.01); miR-148b-3p (20.13 to 25.18); miR-448 (26.44 to 34.04).
Table 3 shows the reference gene rankings for GeNorm, NormFinder and BestKeeper, along with associated stability values. The accepted gene stability thresholds for the three algorithms are M<1.5, SV<1.0 and SD<1.0, respectively [44]. The GeNorm and NormFinder analysis suggests that all four reference genes selected would be considered stable enough, whereas the BestKeeper analysis shows that miR-448 would not meet the rejection cutoff (St Dev about 1), however, miR-186-5p, miR-194-5p and miR-148b-3p would be considered adequate. Importantly, the three algorithms mainly agreed on the ranking of gene stability–all three identified miR-186-5p as the top stable reference gene. GeNorm and BestKeeper both resulted in miR-148b-3p as the second-best ranking gene. All three algorithms placed miR-448 as the least stable reference gene.
Table 3. Reference gene ranking by GeNorm, NormFinder and BestKeeper and associated stability values (from most to least stable).
| GeNorm | GeNorm M value | NormFinder | NormFinder Stability value rho | BestKeeper | BestKeeper Std Dev |
|---|---|---|---|---|---|
| miR-186-5p/miR-148b-3p | 0.536 | miR-186-5p | 0.129 | miR-186-5p | 0.86 |
| miR-194-5p | 0.670 | miR-194-5p | 0.151 | miR-148b-3p | 0.9 |
| miR-448 | 1.11 | miR-148b-3p | 0.286 | miR-194-5p | 0.9 |
| miR-448 | 0.447 | miR-448 | 1.15 |
The use of a single reference gene in experiments is discouraged given that is can introduce bias, and as such multiple reference genes are suggested as standard practice [33]. The pairwise variation (V) analysis calculated with geNorm helps to identify the ideal numbers of reference genes to be used (a lower V number means including the additional reference gene in question is beneficial). The ideal cutoff value of V <0.15 determines when to stop adding additional reference genes. However, this number is considered to be a suggested threshold criteria rather than a universal cutoff and the experimental context should be taken into account, referring mainly to the biological system at hand and the level of variation expected (for example, less variable cell line extraction data versus animal tissues or human biological samples) [45]. The samples under analysis in this study are highly variable and thus we considered the V2/V3 at 0.229 to be sufficient, as V3/V4 was higher at 0.377. Several studies have also indicated the use of slightly higher V values [45–48]. Chasing more reference genes to target a perfect V value when these results are near to the cutoff would be highly impractical, since the experiments described here require a large throughput of data in many patient samples, oftentimes including multiple target genes. Having greater than four reference genes would be expensive and labor intensive for likely no or little benefit in accuracy of data processing.
Table 4 shows the output data for BestKeeper, which assigns higher stability to lower Ct variation. The most stable reference genes are selected on the basis of the lowest Ct standard deviation values [31, 49]. As already alluded to, all reference genes except miR-448 met the cutoff criteria of 1 for stability. Furthermore, the software produces pair-wise correlation results to estimate the relationship between reference gene pairs (data not shown). P-values for all reference genes are less than 0.05 are considered to be correlated with the BestKeeper index and hence are good reference genes.
Table 4. Analysis of reference genes by BestKeeper algorithm.
| miR-194-5p | miR-186-5p | miR-148b-3p | miR-448 | |
|---|---|---|---|---|
| n | 41 | 41 | 41 | 41 |
| geo Mean [CP] | 22.29 | 22.77 | 22.54 | 31.01 |
| ar Mean [CP] | 22.32 | 22.79 | 22.56 | 31.05 |
| min [CP] | 20.02 | 20.96 | 20.13 | 26.45 |
| max [CP} | 24.76 | 25.01 | 25.18 | 34.04 |
| std dev [+/- CP] | 0.90 | 0.86 | 0.90 | 1.15 |
| CV [% CP] | 4.05 | 3.77 | 3.97 | 3.70 |
| p-value | 0.001 | 0.001 | 0.001 | 0.001 |
Lastly, raw Ct data were analyzed by RefFinder software. Fig 1 displays the output graphics from RefFinder, illustrating results in agreement with those described above for the previous three algorithms. The comprehensive ranking also yields highly agreeable results, with miR-186-5p and miR-148b-3p as the top two reference genes.
Fig 1. RefFinder output for candidate reference genes across 41 patients.
(A) geNorm ranking by stability value (B) normFinder ranking by stability value (C) delta CT method ranking by average STDEV (D) BestKeeper std dev ranking and (E) comprehensive ranking (geomean of ranking values).
Discussion
Identifying stable reference genes is of the utmost importance in correctly interpreting gene expression in any experimental design. The choice of a wrong reference gene can greatly affect the target gene results, and the use of multiple reference genes is considered to be important for accurate analysis [33, 50]. The current report sought to identify reliable reference genes for use in RT-qPCR miRNA expression studies in the peripheral blood of critically ill patients with and without COVID-19. The results are particularly applicable for blood collected in PaxGene tubes, which is becoming an increasing common way to sample human specimens due to its ease of use and reliability in preserving nucleic acid species [21, 51, 52]. The methodology employed here, with the use of geNorm, NormFinder, BestKeeper and RefFinder, represents a comprehensive review of standard approaches used numerous times in the literature to reliably identify appropriate reference genes [44, 45, 49].
The search for reference genes for a new experimental design usually involves a detailed literature search to obtain a list of candidate genes used in similar studies that are then formally tested in the samples at hand. Given that our cohort is unique in several ways (COVID negative patients as controls and whole blood processed in PaxGene tubes), we did not feel our original literature search would produce an adequate list of candidate reference genes that would have a high chance of success in standard reference gene computational analyses. Thus, the first part of our analysis involved a large-scale study of the expression of 798 miRNAs in a smaller cohort of patients using NanoString technology. This was carried out to obtain a list of six potential reference gene miRNAs for appropriate testing in RT-qPCR/computational program analysis: miR-186-5p, miR-148b-3p, miR-194-5p, miR-2116-5p, miR-216b-5p and miR-448. Two of the miRNAs, miR-2116-5p and miR-216b-5p, had low expression in RT-qPCR analysis and thus the associated expression data was not carried forward.
Four miRNAs have been deemed stable enough as reference genes by most of the computational programs employed here: miR-186-5p, miR-148b-3p, miR-194-5p and miR-448. To the best of our knowledge, none of these candidates have been underscored as biomarkers or significantly deregulated in COVID-19 critical illness. One candidate (miR-186-5p) has been shown in one study to be decreased in the plasma of COVID-19 patients, however, the study population was small and very different to what we present here: COVID-19 patients (n = 12) were mainly mildly ill and results were compared to healthy controls [53]. Our results suggest that the top two miRNAs, miR-186-5p and miR-148b-3p, used in conjunction (as suggested by GeNorm technology) offer the best means to normalize target gene expression data. Importantly, all software programs were in agreement the vast majority of the time on the ranking of stable reference genes, which suggests that our experimental design and data analysis was robust. There was a minor difference in ranking with NormFinder, in that the top two reference genes listed were miR-186-5p and miR-194-5p. Minor differences in ranking across algorithms are not unexpected, since their approaches have inherent differences, as described above [31–33, 42]. Candidate reference miRNAs were not significantly associated (p<0.05) with the study population characteristics of sex, age, comorbidities (type II diabetes, hypertension, dyslipidemia) or treatments (antibiotics, steroid use, respiratory support) (S1, S2 Tables).
This study has certain limitations. The use of paxgene tubes provide a high level of convenience for obtaining peripheral blood at the bedside. The caveat is that the whole sample is immersed in a proprietary reagent which lyses cells and stabilizes nucleic acids and as such the source of miRNA signal is not known. Our choice of a COVID-negative critically ill control group was deliberate in that it offers the most clinically relevant comparison. However, many studies continue to use healthy controls as a comparison group and in that regard the utility of applying these miRNAs candidates in that particular experimental scenario is unknown. We would like to re-iterate that while the reference genes identified here are appropriately validated for our sample population, they should be considered candidate reference genes that need validation for any new experimental context, which may indeed include healthy controls or different sample types such as plasma or serum.
The reference genes identified here can likely be considered appropriate candidates for miRNA studies in non-COVID-19 critical illness scenarios, since the non-COVID-19 group represent heterogenous diseases lending to Intensive Care Unit admission, such as acute respiratory distress syndrome and sepsis. The initial phase of creating an appropriate list of potential miRNAs is expensive and time consuming, and the current report may offer authors a means to bypass that screening step by offering ten reasonable reference gene choices as shown in Table 1. Again, we strongly suggest verifying such candidates in the particular experimental analysis under consideration.
Conclusion
The COVID-19 pandemic has created an urgency to study the host gene response that leads to variable clinical presentations of the disease, particularly the critical illness response. miRNAs have been implicated in the mechanism of host immune dysregulation thus necessitating further analyses of their altered expression in COVID-19, and an important basis for this is identifying appropriate reference genes for high quality expression analysis studies. The current report aimed to identify reference genes for use in miRNA expression analysis in the peripheral blood of critically ill patients with and without COVID-19 disease. A list of ten candidate reference genes was generated using NanoString technology and six were investigated in validation studies by RT-qPCR. We have validated four stably expressed miRNAs and conclude that miR-186-5p and miR-148b-3p used in combination are appropriate reference genes for miRNA expression studies using PaxGene tubes in the peripheral blood of patients critically ill with COVID-19 disease.
Supporting information
(XLSX)
(XLSX)
(XLSX)
(XLSX)
Acknowledgments
We thank our research coordinators Gyan Sandhu and Dr. Valeria Di Giovanni for their contributions to human sample collection and organizing of our biobank, respectively. We also thank patients that donated samples for this project.
Data Availability
All relevant data are within the paper and its Supporting Information files.
Funding Statement
A NanoString grant provided this study with reagents for experiments and technical support. The St. Michael’s Hospital Foundation provided financial support to collect patient samples.
References
- 1.Zhu N, Zhang D, Wang W, Li X, Yang B, Song J, et al. A Novel Coronavirus from Patients with Pneumonia in China, 2019. N Engl J Med. 2020;382(8):727–33. doi: 10.1056/NEJMoa2001017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.WHO. WHO coronavirus (COVID-19) dashboard 2022 [Available from: https://covid19.who.int/.
- 3.Uddin M, Mustafa F, Rizvi TA, Loney T, Suwaidi HA, Al-Marzouqi AHH, et al. SARS-CoV-2/COVID-19: Viral Genomics, Epidemiology, Vaccines, and Therapeutic Interventions. Viruses. 2020;12(5). doi: 10.3390/v12050526 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Johnson S, Mielke N, Mathew T, Maine GN, Chen NW, Bahl A. Predictors of hospitalization and severe disease due to breakthrough SARS-CoV-2 infection in fully vaccinated individuals. J Am Coll Emerg Physicians Open. 2022;3(4):e12793. doi: 10.1002/emp2.12793 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Auld SC, Harrington KRV, Adelman MW, Robichaux CJ, Overton EC, Caridi-Scheible M, et al. Trends in ICU Mortality From Coronavirus Disease 2019: A Tale of Three Surges. Crit Care Med. 2022;50(2):245–55. doi: 10.1097/CCM.0000000000005185 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Polack FP, Thomas SJ, Kitchin N, Absalon J, Gurtman A, Lockhart S, et al. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N Engl J Med. 2020;383(27):2603–15. doi: 10.1056/NEJMoa2034577 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bahl A, Johnson S, Maine G, Garcia MH, Nimmagadda S, Qu L, et al. Vaccination reduces need for emergency care in breakthrough COVID-19 infections: A multicenter cohort study. Lancet Reg Health Am. 2021;4:100065. doi: 10.1016/j.lana.2021.100065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Laing AG, Lorenc A, Del Molino Del Barrio I, Das A, Fish M, Monin L, et al. A dynamic COVID-19 immune signature includes associations with poor prognosis. Nat Med. 2020;26(10):1623–35. doi: 10.1038/s41591-020-1038-6 [DOI] [PubMed] [Google Scholar]
- 9.Sirpilla O, Bauss J, Gupta R, Underwood A, Qutob D, Freeland T, et al. SARS-CoV-2-Encoded Proteome and Human Genetics: From Interaction-Based to Ribosomal Biology Impact on Disease and Risk Processes. J Proteome Res. 2020;19(11):4275–90. doi: 10.1021/acs.jproteome.0c00421 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Acharya D, Liu G, Gack MU. Dysregulation of type I interferon responses in COVID-19. Nat Rev Immunol. 2020;20(7):397–8. doi: 10.1038/s41577-020-0346-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bastard P, Rosen LB, Zhang Q, Michailidis E, Hoffmann HH, Zhang Y, et al. Autoantibodies against type I IFNs in patients with life-threatening COVID-19. Science. 2020;370(6515). doi: 10.1126/science.abd4585 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Initiative C-HG. The COVID-19 Host Genetics Initiative, a global initiative to elucidate the role of host genetic factors in susceptibility and severity of the SARS-CoV-2 virus pandemic. Eur J Hum Genet. 2020;28(6):715–8. doi: 10.1038/s41431-020-0636-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Shen B, Yi X, Sun Y, Bi X, Du J, Zhang C, et al. Proteomic and Metabolomic Characterization of COVID-19 Patient Sera. Cell. 2020;182(1):59–72 e15. doi: 10.1016/j.cell.2020.05.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gardinassi LG, Souza COS, Sales-Campos H, Fonseca SG. Immune and Metabolic Signatures of COVID-19 Revealed by Transcriptomics Data Reuse. Front Immunol. 2020;11:1636. doi: 10.3389/fimmu.2020.01636 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Xiong Y, Liu Y, Cao L, Wang D, Guo M, Jiang A, et al. Transcriptomic characteristics of bronchoalveolar lavage fluid and peripheral blood mononuclear cells in COVID-19 patients. Emerg Microbes Infect. 2020;9(1):761–70. doi: 10.1080/22221751.2020.1747363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Alsamman AM, Zayed H. The transcriptomic profiling of SARS-CoV-2 compared to SARS, MERS, EBOV, and H1N1. PLoS One. 2020;15(12):e0243270. doi: 10.1371/journal.pone.0243270 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Blanco-Melo D, Nilsson-Payant BE, Liu WC, Uhl S, Hoagland D, Moller R, et al. Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19. Cell. 2020;181(5):1036–45 e9. doi: 10.1016/j.cell.2020.04.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mukhopadhyay S, Sinha S, Mohapatra SK. Analysis of transcriptomic data sets supports the role of IL-6 in NETosis and immunothrombosis in severe COVID-19. BMC Genom Data. 2021;22(1):49. doi: 10.1186/s12863-021-01001-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Li CX, Chen J, Lv SK, Li JH, Li LL, Hu X. Whole-Transcriptome RNA Sequencing Reveals Significant Differentially Expressed mRNAs, miRNAs, and lncRNAs and Related Regulating Biological Pathways in the Peripheral Blood of COVID-19 Patients. Mediators Inflamm. 2021;2021:6635925. doi: 10.1155/2021/6635925 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Prokop JW, Hartog NL, Chesla D, Faber W, Love CP, Karam R, et al. High-Density Blood Transcriptomics Reveals Precision Immune Signatures of SARS-CoV-2 Infection in Hospitalized Individuals. Front Immunol. 2021;12:694243. doi: 10.3389/fimmu.2021.694243 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jackson H, Rivero Calle I, Broderick C, Habgood-Coote D, D’Souza G, Nichols S, et al. Characterisation of the blood RNA host response underpinning severity in COVID-19 patients. Sci Rep. 2022;12(1):12216. doi: 10.1038/s41598-022-15547-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Baghela A, Pena OM, Lee AH, Baquir B, Falsafi R, An A, et al. Predicting sepsis severity at first clinical presentation: The role of endotypes and mechanistic signatures. EBioMedicine. 2022;75:103776. doi: 10.1016/j.ebiom.2021.103776 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Best MG, Sol N, Kooi I, Tannous J, Westerman BA, Rustenburg F, et al. RNA-Seq of Tumor-Educated Platelets Enables Blood-Based Pan-Cancer, Multiclass, and Molecular Pathway Cancer Diagnostics. Cancer Cell. 2015;28(5):666–76. doi: 10.1016/j.ccell.2015.09.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Vargas J, Lima JA, Kraus WE, Douglas PS, Rosenberg S. Use of the Corus(R) CAD Gene Expression Test for Assessment of Obstructive Coronary Artery Disease Likelihood in Symptomatic Non-Diabetic Patients. PLoS Curr. 2013;5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;116(2):281–97. doi: 10.1016/s0092-8674(04)00045-5 [DOI] [PubMed] [Google Scholar]
- 26.de Gonzalo-Calvo D, Vea A, Bar C, Fiedler J, Couch LS, Brotons C, et al. Circulating non-coding RNAs in biomarker-guided cardiovascular therapy: a novel tool for personalized medicine? Eur Heart J. 2019;40(20):1643–50. doi: 10.1093/eurheartj/ehy234 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Guay C, Regazzi R. Circulating microRNAs as novel biomarkers for diabetes mellitus. Nat Rev Endocrinol. 2013;9(9):513–21. doi: 10.1038/nrendo.2013.86 [DOI] [PubMed] [Google Scholar]
- 28.Valihrach L, Androvic P, Kubista M. Circulating miRNA analysis for cancer diagnostics and therapy. Mol Aspects Med. 2020;72:100825. doi: 10.1016/j.mam.2019.10.002 [DOI] [PubMed] [Google Scholar]
- 29.Rupaimoole R, Slack FJ. MicroRNA therapeutics: towards a new era for the management of cancer and other diseases. Nat Rev Drug Discov. 2017;16(3):203–22. doi: 10.1038/nrd.2016.246 [DOI] [PubMed] [Google Scholar]
- 30.Pinilla L, Benitez ID, Gonzalez J, Torres G, Barbe F, de Gonzalo-Calvo D. Peripheral blood microRNAs and the COVID-19 patient: methodological considerations, technical challenges and practice points. RNA Biol. 2021;18(5):688–95. doi: 10.1080/15476286.2021.1885188 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Pfaffl MW, Tichopad A, Prgomet C, Neuvians TP. Determination of stable housekeeping genes, differentially regulated target genes and sample integrity: BestKeeper—Excel-based tool using pair-wise correlations. Biotechnol Lett. 2004;26(6):509–15. doi: 10.1023/b:bile.0000019559.84305.47 [DOI] [PubMed] [Google Scholar]
- 32.Andersen CL, Jensen JL, Orntoft TF. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res. 2004;64(15):5245–50. doi: 10.1158/0008-5472.CAN-04-0496 [DOI] [PubMed] [Google Scholar]
- 33.Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002;3(7):RESEARCH0034. doi: 10.1186/gb-2002-3-7-research0034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Silver N, Best S, Jiang J, Thein SL. Selection of housekeeping genes for gene expression studies in human reticulocytes using real-time PCR. BMC Mol Biol. 2006;7:33. doi: 10.1186/1471-2199-7-33 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Faheem M, Khaliq S. Validation of Housekeeping Genes for Gene Expression Profiling in Fish: A Necessity. Crit Rev Eukaryot Gene Expr. 2019;29(6):565–79. doi: 10.1615/CritRevEukaryotGeneExpr.2019028964 [DOI] [PubMed] [Google Scholar]
- 36.Gorji-Bahri G, Moradtabrizi N, Hashemi A. Uncovering the stability status of the reputed reference genes in breast and hepatic cancer cell lines. PLoS One. 2021;16(11):e0259669. doi: 10.1371/journal.pone.0259669 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Team RC. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria: 2021. [Available from: https://www.R-project.org/. [Google Scholar]
- 38.NanoStringQCPro: Quality metrics and data processing methods for NanoString mRNA gene expression data. In: Nickles D ST, Ziman R, Bourgon R, editor. 2021.
- 39.Bhattacharya A, Hamilton AM, Furberg H, Pietzak E, Purdue MP, Troester MA, et al. An approach for normalization and quality control for NanoString RNA expression data. Brief Bioinform. 2021;22(3). doi: 10.1093/bib/bbaa163 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(1):139–40. doi: 10.1093/bioinformatics/btp616 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Derveaux S, Vandesompele J, Hellemans J. How to do successful gene expression analysis using real-time PCR. Methods. 2010;50(4):227–30. doi: 10.1016/j.ymeth.2009.11.001 [DOI] [PubMed] [Google Scholar]
- 42.Xie F, Xiao P, Chen D, Xu L, Zhang B. miRDeepFinder: a miRNA analysis tool for deep sequencing of plant small RNAs. Plant Mol Biol. 2012. doi: 10.1007/s11103-012-9885-2 [DOI] [PubMed] [Google Scholar]
- 43.Perkins JR, Dawes JM, McMahon SB, Bennett DL, Orengo C, Kohl M. ReadqPCR and NormqPCR: R packages for the reading, quality checking and normalisation of RT-qPCR quantification cycle (Cq) data. BMC Genomics. 2012;13:296. doi: 10.1186/1471-2164-13-296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Linardic M, Braybrook SA. Identification and selection of optimal reference genes for qPCR-based gene expression analysis in Fucus distichus under various abiotic stresses. PLoS One. 2021;16(4):e0233249. doi: 10.1371/journal.pone.0233249 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Kaur R, Sodhi M, Sharma A, Sharma VL, Verma P, Swami SK, et al. Selection of suitable reference genes for normalization of quantitative RT-PCR (RT-qPCR) expression data across twelve tissues of riverine buffaloes (Bubalus bubalis). PLoS One. 2018;13(3):e0191558. doi: 10.1371/journal.pone.0191558 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Chen J, Huang Z, Huang H, Wei S, Liu Y, Jiang C, et al. Selection of relatively exact reference genes for gene expression studies in goosegrass (Eleusine indica) under herbicide stress. Sci Rep. 2017;7:46494. doi: 10.1038/srep46494 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Zhang Y, Zhang XD, Liu X, Li YS, Ding JP, Zhang XR, et al. Reference gene screening for analyzing gene expression across goat tissue. Asian-Australas J Anim Sci. 2013;26(12):1665–71. doi: 10.5713/ajas.2013.13199 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Koramutla MK, Aminedi R, Bhattacharya R. Comprehensive evaluation of candidate reference genes for qRT-PCR studies of gene expression in mustard aphid, Lipaphis erysimi (Kalt). Sci Rep. 2016;6:25883. doi: 10.1038/srep25883 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Palombella S, Pirrone C, Cherubino M, Valdatta L, Bernardini G, Gornati R. Identification of reference genes for qPCR analysis during hASC long culture maintenance. PLoS One. 2017;12(2):e0170918. doi: 10.1371/journal.pone.0170918 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Tricarico C, Pinzani P, Bianchi S, Paglierani M, Distante V, Pazzagli M, et al. Quantitative real-time reverse transcription polymerase chain reaction: normalization to rRNA or single housekeeping genes is inappropriate for human tissue biopsies. Anal Biochem. 2002;309(2):293–300. doi: 10.1016/s0003-2697(02)00311-1 [DOI] [PubMed] [Google Scholar]
- 51.Scicluna BP, Uhel F, van Vught LA, Wiewel MA, Hoogendijk AJ, Baessman I, et al. The leukocyte non-coding RNA landscape in critically ill patients with sepsis. Elife. 2020;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Li C, Hu X, Li L, Li JH. Differential microRNA expression in the peripheral blood from human patients with COVID-19. J Clin Lab Anal. 2020;34(10):e23590. doi: 10.1002/jcla.23590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Wu J, Liu X, Shao J, Zhang Y, Lu R, Xue H, et al. Expression of plasma IFN signaling-related miRNAs during acute SARS-CoV-2 infection and its association with RBD-IgG antibody response. Virol J. 2021;18(1):244. doi: 10.1186/s12985-021-01717-7 [DOI] [PMC free article] [PubMed] [Google Scholar]