Patients

The present study utilizes data and biological samples from the Report-Age COVID project, an observational study conducted at the Italian National Center on Aging (IRCCS INRCA), Italy. The aim of this study is to provide a deeper understanding of COVID-19 disease in older hospitalized patients (age>65 years). All the selected subjects were cases of COVID-19 as confirmed by the positive detection of SARS-CoV-2 RNA in nasal/oro-pharyngeal swabs using real-time reverse transcriptase-polymerase chain reaction assay. The study protocol has been approved by the Ethics Committee of the IRCCS INRCA hospital, Ancona, Italy (reference number CE-INRCA-20008) and registered under the ClinicalTrials.gov database (reference number NCT04348396). Clinical and epidemiological data of hospitalization were gathered in a retrospective manner and anonymized prior to release. All the patients enrolled in the Report-Age COVID received treatment from INRCA hospital from March 1st 2020 to date. Among these patients, we selected 156 subjects from the database based on the availability of biological samples (serum and plasma at hospital admission). These selected patients were admitted to INRCA hospital between October 11th and December 31st 2020. None of them received the anti-COVID-19 vaccine. Most of the patients received corticosteroids during their hospital stays, without significant differences with respect to the outcome. Standard oxygen or nonivasive ventilation (NIV) were used for most of the patients, while few patients were transferred to an intensive care unit (ICU). The initial decision to admit or not a COVID 19 older patient to the ICU was taken after a comprehensive evaluation by a multidisciplinary team, composed by the anesthesiologist, the geriatrician, the cardiologist and the palliativist, as appropriate, which also involved the patient and the family, whenever possible.

In addition to the cohort of COVID-19 patients, and in order to compare values for n-cfDNA parameters in hospitalized older patients with and without SARS-CoV-2 infection, a group of 36 older adults was selected from the Report-Age project, a large-scale ongoing observational study on the health conditions of hospitalized older adults at INRCA Hospital (Trial Registration no. NCT01397682) [42]. The study protocol of the Report-Age project has been approved by the Ethics Committee of the IRCCS INRCA hospital, Ancona, Italy [42]. The control group was composed of older adults who accessed the INRCA hospital for common geriatric disorders in the period 2013-2017 (before the COVID-19 pandemic outbreak), with available biological (plasma) samples and clinical and follow-up data. Patients with evidence of infectious or acute respiratory diseases or with intra-hospital mortality, or with less than two-years of follow-up survival after hospital discharge, were excluded. The patients of this control non-COVID-19 group had comparable sex ratio, median age and prevalence of most of the common age-related diseases as the analysed COVID-19 cohort.

Plasma/serum collection and Nucleic acid extraction

EDTA plasma tubes were gently inverted 8 times and centrifuged. Tubes were centrifuged at 2500 x g at 4°C for 15 minutes. After the centrifugation of blood EDTA tube, three layers were obtained (plasma, buffy coat and erythrocytes). About 2.5 ml of plasma were collected from the upper part of the plasma layer, and placed in a new vial (avoiding collecting plasma from the lower part of the plasma layer near the buffy coat). The collected plasma EDTA volume was then subjected to a second centrifugation, conducted for 8 minutes at 10.000 RCF. After the centrifugation, the upper 80% of the volume was carefully collected (avoiding collecting of the lower 20% volume, enriched in cell debris and platelets) and stored in aliquots of 400-500 microliters, to be immediately frozen at -80°C. Immediately after collection, serum tubes were gently inverted 8 times and left at room temperature for 60 minutes to allow clotting, and then centrifuged at 2500 x g at 4°C for 15 minutes. After centrifugation, the top of the supernatant was carefully aspirated and stored in aliquots of 500-1000 microliters, to be immediately frozen at -80°C. Nucleic acids were extracted from 140 μl of plasma using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany). The kit does not discriminate between RNA and DNA, and it was used to extract plasma cell free DNA (of cellular origin) and viral RNA (for detection of SARS-CoV-2 RNAeamia) from the same plasma samples. The purified nucleic acid sample was conserved at -80°C before the analysis.

mt-cfDNA quantification

PCR amplification of mitochondrial cfDNA (mt-cfDNA) was measured using a real-time quantitative assay for Human cytochrome C oxidase subunit III (MT-CO3) gene. All assays were performed on a Rotor-Gene Q detection system (Qiagen) using a 72-well carousel.

The reaction mixture consisted of 2 μl of nucleic acids from plasma samples and 18 μl master mix, which was composed of 7 μl H2O, 10 μl SYBR qPCR Master Mix (Vazyme, Nanjing, China) and 1 μl of 5 μM forward and reverse primers, respectively. The primers used were: forward 5’-ATGACCCACCAATCACATGC-3’, reverse 5’- ATCACATGGCTAGGCCGGAG-3’ (IDT, Coralville, IA).

PCR conditions were set to: 95 °C for 3 min, followed by 40 cycles of 10 s denaturation at 95 °C, 30 s annealing at 55 °C. Absolute quantification of the target sequence in each biological sample was estimated by comparison to a RT-qPCR standard amplification curve. To generate the standard MT-DNA, a selected region of human purified genomic DNA containing the target sequence for MT-CO3 (forward 5’-ATGACCCACCAATCACATGC-3’, reverse 5’-ATCAATAGATGGAGACATAC-3’) was amplified by PCR under the following conditions: initial denaturation at 95° C for 1 min, then 35 cycles of denaturation at 95° C for 30 s, annealing at 50° C for 30 s and extension at 72°C for 1 min. The generated amplicon had a length of 775 nucleotides and a molecular weight of 478879.57 Da. After amplification, the MT-DNA standard was purified by using MinElute columns (Qiagen) and quantified using a NanoDrop spectrophotometer (ThermoScientific). Serial dilutions were then used to calibrate the RT-qPCR standard curves. Each sample was quantified in duplicate, and triplicates of the standard curve were included in each run.

Determination of abundance and integrity of n-cfDNA

The polymerase chain reaction protocol used for the quantification of nuclear cfDNA (n-cfDNA) and for the evaluation of n-cfDNA integrity was derived from a previous study [26]. Two sets of primers complementary to the consensus sequence of human Alu repeats were used: ALU115 primer forward, 5’-CCTGAGGTCAGGAGTTCGAG-3’; ALU115 primer reverse, 5’-CCCGAGTAGCTGGGATTACA-3’; ALU247 primer forward, 5’-GTGGCTCACGCCTGTAATC-3’; ALU247 primer reverse, 5’-CAGGCTGGAGTGCAGTGG-3’. The first set was used to amplify the 115-bp amplicon (ALU 115), the second for the 247-bp amplicon (ALU 247). The reaction was conducted using 0.2 μM each of forward primer and reverse primer (ALU 115 or ALU 247), 1x Ssofast mix containing enzyme and Syber Green dye (Bio-Rad Laboratories), in a reaction volume of 15 μl. Thermal cycles were conducted on a Qiagen Rotor-Gene instrument, as follows: pre-denaturation 2.5 min at 98°C, followed by 35 cycles of denaturation 15 sec at 95°, annealing/extension 60 sec at 64°C. A melting curve was conducted at the end of each run, to check for the presence of unspecific amplification. A calibration curve constructed by amplifying serial dilutions of a genomic DNA standard sample (0.02 to 200 pg/μl) was present in each assay run, and was used to assess the absolute equivalent concentration of genomic DNA in each plasma sample. Each sample was run in duplicate.

All assays were conducted in blind without knowing the sample identity. The ratio of concentrations calculated for each sample with the two set of primers, i.e. (concentration of Alu 247) / (concentration of Alu 115), hereafter indicated as Alu247/115, was used as index of n-cfDNA integrity.

Measurement of soluble biomarkers

Serum concentration of IL-6, IFN-α, TNF-α and IL-10 were assessed by using four Pro Quantum Immunoassays (Thermo Fisher Scientific) specific for each of the four biomolecules. Each reaction was conducted in duplicate, using 2 μl of serum, and run on an Agilent Aria Mx Real Time PCR instrument at the conditions indicated by the producer.

The plasma levels of NE and LL37 were measured using commercial ELISA kits (ABCAM, AB119553 PMN Elastase Human ELISA Kit and Hycult Biotech, HK321 Human LL-37). The serum levels of CD163 were measured using a commercial ELISA kit (R&DSystems, Human CD163 Immunoassay). All the serum/plasma samples were diluted 1:20 (for sCD163 and LL37 assayes) or 1:50 (for neutrophile elastase assay). All samples were ran in duplicate and samples with an intra-assay coefficient of variation below 10.0% were included in this study.

Assessment of RNAemia

The presence of SARS-CoV-2 RNA in plasma was assessed by analyzing 10 μl of nucleic acids extracted from plasma in a Real Quality RQ-SARS-CoV-2 real time PCR assay (AB Analitica). The assay was specific for the genes RdRp (encoding the RNA-dependent RNA polymerase) gene and S gene (encoding the spike protein). The presence of an endogenous control (targeting the human RNAse P gene) in the same assay was used as internal control to ensure proper sampling and nucleic acids extraction. The assay was used following the conditions indicated by the producer, on an Agilent Aria Mx Real Time PCR instrument. The plasma sample was considered positive for the SARS-CoV-2 RNAemia if both viral genes were detected with a cycle threshold (CT) < 40.

Statistical analysis

Continuous variables were reported as median and interquartile range after their non-normality have been assessed using Shapiro-Wilk test and comparison of variables between groups was performed by unpaired Mann-Whitney U test. Tertiles of NE, sCD163 and n-cfDNA integrity (Alu247/115) were calculated and, as other categorical variables, were expressed as absolute number.

Spearman rank correlation was conducted to check associations among different biomarkers and between these biomarkers and other parameters.

The Kaplan–Meier survival curves were used to estimate the association of in-hospital mortality risk with different levels of the analysed biomarkers. Cox proportional hazards analysis were used to derive age- and gender-adjusted (Model 1); age- gender- comorbidities- and Clinical Frailty Scale (CFS)-adjusted (Model 2); age- gender- comorbidities- Clinical Frailty Scale (CFS)- and SARS-CoV-2 RNAemia-adjusted (Model 3) and age- gender- comorbidities- Clinical Frailty Scale (CFS)- SARS-CoV-2 RNAemia- and treatments-adjusted (Model 4 and Model 5) hazard ratios (HR) and 95% confidence intervals (95% CI) of the association between all independent variables and study outcome. The choice to consider different Cox proportional hazards models, each containing the same tested variables (biomarkers), but adjusted for an increasing number of confounders, was aimed to verify if a statistically significant relationship between a biomarker and survival maintained its significance after the addition of confounders, and which (if any) confounders could determine a loss of significance in the tested variables. The length of hospital stay was used as the time to failure variable for the model. A two-tailed P value < .05 was considered significant.

The statistical analyses were performed using the IBM SPSS Statistics program (version 27) and the STATA version15.1 Statistical Software Package for Windows (Stata Corp, College Station, TX).