Comparison by Age of the Local Interferon Response to SARS-CoV-2 Suggests a Role for IFN-ε and -ω.

Children generally develop a mild disease after SARS-CoV-2 infection whereas older adults are at risk of developing severe COVID-19. Recent transcriptomic analysis showed pre-activated innate immunity in children, resulting in a more effective anti-SARS-CoV-2 response upon infection. To further characterize age-related differences, we studied type I and III interferon (IFN) response in SARS-CoV-2 infected and non-infected individuals of different ages. Specifically, levels of expression of type I (IFN-α, -β, -ε and -ω), type III (IFN-λ1, -λ2 and -λ3) IFNs and of the IFN-stimulated genes, ISG15 and ISG56 were quantified in nasopharyngeal cells from diagnostic swabs. Basal transcription of type I/III IFN genes was highest among children and decreased with age. Among SARS-CoV-2-infected individuals, only IFN-ε and -ω levels were significantly higher in children and young adults whereas ISGs were overexpressed in infected adults. The occurrence of symptoms in children and the need for hospitalization in adults were associated to higher transcription of several IFN genes. Starting from a pre-activated transcription level, the expression of type I and III IFNs was not highly up-regulated in children upon SARS-CoV-2 infection; young adults activated IFNs' transcription at intermediate levels whereas older adults were characterized by higher ISGs and lower IFN-ε and -ω relative expression levels. Overall, our findings contribute to recognize components of a protective IFN response as a function of age, in the context of SARS-CoV-2 infection.

Introduction

From December 2019 up to now, the novel coronavirus named SARS-CoV-2 has caused a huge impact on health worldwide due to life-threatening COVID-19 disease. Risk factors predisposing to severe COVID-19 are male sex, the presence of comorbidities and an older age, whereas children commonly develop a mild disease after SARS-CoV-2 infection (1). The mechanisms that protect the younger from severe COVID-19 disease are yet to be defined; reasons may be a low expression of the ACE2 receptor, trained immunity, a higher capacity to regenerate cells after viral damage but also a strong and effective innate and/or adaptive immune response to SARS-CoV-2 (1–3). SARS-CoV-2 replication starts in nasopharyngeal mucosal cells and recognition of viral molecular patterns leads to a rapid activation of antiviral and inflammatory responses (4, 5). The antiviral response is largely mediated by the transcription of Interferon (IFN) genes, in particular the type I (IFN-α, -β) and type III (IFN-λ1-3) that are the most important IFNs produced by many cell types. Once the IFNs are secreted from infected cells and bind to specific receptors, their signaling upregulates the expression of numerous IFN-stimulated genes (ISGs). In infected cells, SARS-CoV-2 non-structural proteins are able to inhibit the innate immune receptor signaling and IFN production (6); on the other hand, SARS-CoV-2 is highly sensitive to type I/III IFNs when exogenously administered (7, 8). In patients, the key protective role of the IFN-response was confirmed by genetic studies showing that mutations in genes of the type I IFN pathway are far more frequent in patients with severe COVID-19 than in those with a mild disease (9). Moreover, circulating auto-antibodies (Abs) against type 1 IFNs were detected in around 10% of critically ill COVID-19 patients (10). While a fraction of severely COVID-19 patients is highly suppressed in their IFN production and activity due to genetic defects, the presence of anti-IFN auto-Abs or other causes (9–12), on the other hand, delayed and disproportionate IFN responses are involved in the progression to severe disease (4, 13, 14). Thus, to achieve control of SARS-CoV-2 infection, a fast and effective mucosal innate response appears crucial (15), as seen in adults that developed a mild COVID-19 (14–17). Consistent with this notion, children, which usually develop asymptomatic or very mild COVID-19, present a more robust early innate immune response in comparison with adults (18, 19). In particular, a higher expression of genes associated with IFN signaling was found in SARS-CoV-2 infected children with respect to adults, not stratified by age (18). Loske et al. found that basal levels of several Pattern Recognition Receptors and Interferon-stimulated genes (ISG) were pre-activated in children and that, upon SARS-CoV-2 infection, their expression reached higher levels in children compared to adults (19); in this study, the expression of the IFN coding genes could not be detected. Contrastingly, Gilbert et al. measured the mRNA levels of type I (IFN-α and -β) and type III (IFN-λ1-3) IFNs, reporting that IFN α and -β transcript levels were induced by SARS-CoV-2 infection, reaching more elevated levels in the adult/elderly with respect to the pediatric age in which only IFN- λ1 was comparatively higher (20).

In order to determine which IFNs are activated during an effective anti-SARS-CoV-2 immune response in the upper respiratory tract, we sought to compare this response among different age groups, with respect to negative controls of the same age. Specifically, we measured mRNA expression levels of genes encoding type I (including the less studied IFN-ε and -ω) and type III IFNs, and IFN-stimulated genes in nasopharyngeal cells from subjects of ages ranging from 3 to 90 years, that were evaluated also for symptoms and the need for hospitalization.

Materials and Methods

Study Subjects

Children/adolescents attending a pediatric ambulatory clinic of the “Umberto I”, Sapienza University Hospital in Rome, to performSARS-CoV-2 molecular tests, were enrolled (September-December 2020) after parent’s written consent (N=47). Children/adolescents were tested because of a previous contact with a SARS-CoV-2 positive case within their family or at school, or for the occurrence of respiratory symptoms. Moreover, medicine students and post-graduates practicing in the pediatric clinic were tested in the same period. Adults attending the “Umberto I” Emergency Departments (ED) were recruited during two periods (March-May 2020 and September-December 2020); they were health workers attending routinely testing for SARS-CoV-2 infection or subjects presenting to the ED with fever and/or respiratory symptoms. A total of 145 adults (>16 years of age) were enrolled following informed consent. SARS-CoV-2 detection was performed in the same virology diagnostic laboratory, from nasopharyngeal (NP) swabs, using RealStar® SARS-CoV-2 RT-PCR Kit 1.0 (Altona Diagnostics, Germany), targeting E and S viral genes. All SARS-CoV-2 positive patients were infected with the original D614G SARS-CoV-2 strain that was the only one circulating until late December 2020 when the Alpha variant was firstly detected in Italy. For subject tested at the ED, the threshold cycle (Ct) value from the diagnostic RT-PCR was obtained. Residual NP samples were stored for up to 3 hours at 4°C, then centrifuged, pellets added with guanidine isothiocyanate reagents and stored at -80 C; they were included in the study as cases if tested positive for SARS-CoV-2 or controls if negative. The occurrence of symptoms was reported by children’s parents and self-reported by adolescents and adults. Children/adolescents were categorized in symptomatic (those presenting one or more of the following symptoms: fever, cough, sore throat, dyspnea, ageusia, anosmia, vomiting, diarrhea) or asymptomatic. Contrastingly, nearly all SARS-CoV-2 positive adults presented with COVID-19 related symptoms and were retrospectively categorized on the base of the need for hospitalization. Demographic data, SARS-CoV-2 Ct values, the occurrence of symptoms, and the need for hospitalization were recorded de-identified. The study was approved by the Institutional Review board and the Ethics Committee (Policlinico Umberto I Hospital, Sapienza, University of Rome, Rif. 5836, Prot. 0690/2021).

Gene Expression Measurement

De-frozen residual NP cells were centrifuged and pellets used for total RNA extraction using RNA-extraction kits (Norgen Biotek Corporation, Canada), 2 µl of purified RNA was quantified on a NanoDrop Spectrophotometer (Thermo Fisher Scientific, Rome, Italy) to determine concentration and purity. Reverse transcription was performed on 300 ng of purified RNA using the High Capacity cDNA Archive Kit (Applied Biosystems, Monza, Italy). From cDNA, the expression levels of genes coding for IFNs type I (IFN-α, -β, -ε and -ω) and type III (IFN-λ1, -λ2 and -λ3) and for ISG15 and ISG56, well-known markers of type I/III IFNs’ activation, previously analysed in SARS-CoV-2 infections (11, 17, 21), were measured by quantitative Reverse Transcription-Real time PCR assays. Briefly, IFN genes transcripts were detected in co-amplification with the beta-glucuronidase (GUS) invariant gene, using the target specific primers and hydrolysis probes listed in our previous study (21), with the addition of primers and probes targeting IFN-ε (Hs.PT.58.4812867.g) and IFN-ω (Hs.PT.20160308.g) purchased from IDT (Integrated DNA Technologies, Coralville, Iowa, US). Transcript level of each gene was calculated using the threshold cycle (Ct) relative quantification to the GUS Ct of the same sample (the 2−ΔCt method). A gene transcript was considered not detectable when the Ct value was > 40 and was assigned a value of one tenth of the limit of detection. Residual cDNAs from SARS-CoV-2 negative samples were tested for the presence of 13 other respiratory viruses (influenza viruses A and B, respiratory syncytial virus, the human CoVs: OC43, 229E, NL63 and HKU1, adenovirus, rhinovirus, parainfluenza viruses 1-3, metapneumovirus), using a panel of home-made PCRs (22).

Statistical Analysis

For categorical variables (sex, age group, occurrence of symptoms, need for hospitalization), Pearson’s Chi square was used to test the statistical difference in proportion among independent groups. Age in years, mRNA expression values (non-normally distributed data) were compared using the non-parametric Kruskal–Wallis (KW) test among three or more independent groups or using the non-parametric Mann–Whitney test between two groups with Bonferroni correction for multiple comparisons. The Jonckheere-Terpstra (JT) test, a rank-based nonparametric test, was also used for determine significant trends in gene expression levels among the ordered age groups. Spearman’s rho coefficient was calculated to assess the correlation between age and expression level of the studied genes. The significance was fixed at the 5% level; the analysis was performed with SPSS v.27.0 for Windows.

Results

Out of 192 individuals (114 males and 78 females) enrolled in this study, 106 tested positive for SARS-CoV-2 RNA and 86 resulted negative to SARS-CoV-2 and to the other respiratory viruses. Levels of expression of genes coding for type I (IFN-α, -β, -ε and -ω), type III (IFNs λ1-3) and for ISG15 and ISG56 were measured in 192 NP samples. Transcript levels of all study genes were significantly higher levels in the SARS-CoV-2 positive- than in the negative- samples (p<0.001 for comparisons of all genes between the two groups; ). Males and females had comparable gene expression levels when considering all samples together or when comparing by sex SARS-CoV-2 negative and positive-samples separately (data not shown). To understand the impact of age on IFN genes expression during SARS-CoV-2 infections, samples were stratified in age groups: children up to 16 years of age (y), young adults (17-40 y), middle-aged adults (41-60 y), older adults (>60 y) and on the base of SARS-CoV-2 negative- or -positive test. Demographic and other data of enrolled subjects stratified by age groups are reported in . Median and interquartile range values for each study gene in the 8 groups are shown in .

Figure 1

SARS-CoV-2 infection induced activation of IFN genes in nasopharyngeal cells. Relative mRNA expression values (y-axis), calculated using the threshold cycle relative quantification (2-ΔCt), are reported for each gene; values are depicted as open circles for SARS-CoV-2 negative- and as filled circles for SARS-CoV-2 positive samples, triangles indicate samples with Ct values >40. Horizontal lines indicate the median value of the group indicated below the X-axis. All p-values, calculated using Mann-Whitney tests for comparison of each gene between SARS-CoV-2 negative and positive samples, are <0.001.

Table 1

Demographic and virologic data of study subjects (N=192) stratified by age groups.

	Female/Male(Female %)	Age a	SARS-CoV-2Ct value b	Symptoms	Hospitalized
SARS-CoV-2 Neg <16y (n=21)	8/13 (38.1%)	8.9 ± 3.4	NA c	NA	NA
SARS-CoV-2 Pos <16y (n=26)	9/17 (34.6%)	10.5 ± 3.4	NR d	15/26 (57.7%)**	0/26 (0%)**
SARS-CoV-2 Neg 16-40y (n=21)	13/8 (61.9%)	28.9 ± 6.9	NA	NA	NA
SARS-CoV-2 Pos 16-40y (n=24)	12/12 (50.0%)	27.1 ± 7.4	27.6 ± 6.2	21/24 (87.5%)**	7/24 (29.2%)**
SARS-CoV-2 Neg 41-60y (n=25)	11/14 (44.0%)	50.5 ± 5.2	NA	NA	NA
SARS-CoV-2 Pos 41-60y (n=28)	8/20 (28.6%)	53.1 ± 4.4	27.9 ± 6.6	28/28 (100%)**	24/28 (85.7%)**
SARS-CoV-2 Neg >60y (n=19)	8/11 (42.1%)	64.5 ± 2.5*	NA	NA	NA
SARS-CoV-2 Pos >60y (n=28)	9/19 (32.1%)	73.3 ± 8.5*	26.7 ± 5.3	28/28 (100%)**	22/28 (78.6%)**

Age in years (y) is expressed as mean ± standard deviation of each group.

The diagnostic test Ct value is expressed as mean ± standard deviation in adults’ groups.

NA, not applicable.

NR, not reported. *p < 0.01, for age comparison between SARS-CoV-2 negative vs positive older adults; **p < 0.001, for comparisons among the SARS-CoV-2 positive age groups.

Table 2

Gene expression results of study samples (N=192) stratified by SARS-CoV-2 infection and age group.

	SARS-CoV-2 negative samples by age group				p value b	SARS-CoV-2 positive samples by age group				p value b
Gene a	<16y	16-40y	41-60y	>60y		<16y	16-40y	41-60y	>60y
IFN-α	2.03 (0.58)	1.37 (1.87)	0.35 (1.58)	0.07 (0.54)	<0.001	2.39 (0.79)	1.94 (1.80)	2.30 (2.28)	1.75 (2.00)	0.076
IFN-β	1.40 (0.49)	0.46 (1.43)	0.33 (1.05)	0.11 (0.35)	<0.001	1.43 (0.52)	1.65 (1.67)	1.60 (1.73)	1.35 (1.53)	0.821
IFN-ε	2.30 (0.43)	1.38 (1.85)	0.26 (1.72)	0.20 (0.70)	<0.001	2.43 (0.72)	2.39 (2.60)	0.90 (2.64)	1.11 (2.45)	0.037
IFN-ω	2.43 (1.04)	1.08 (2.42)	0.23 (1.81)	0.10 (0.75)	<0.001	2.75 (0.66)	2.87 (2.60)	1.72 (2.80)	1.80 (2.60)	0.067
IFN-λ1	0.06 (0.03)	0.06 (0.05)	0.02 (0.03)	0.01 (0.03)	<0.001	0.06 (0.02)	0.07 (0.04)	0.06 (0.05)	0.05 (0.05)	0.415
IFN-λ2	2.22 (0.52)	2.06 (2.82)	0.17 (1.52)	0.10 (0.29)	<0.001	2.16 (0.87)	2.50 (2.80)	1.09 (2.76)	1.32 (2.53)	0.170
IFN-λ3	4.38 (1.70)	2.32 (4.31)	0.90 (2.68)	0.52 (1.60)	<0.001	4.32 (0.90)	4.84 (1.84)	3.93 (2.47)	3.77 (1.45)	0.100
ISG15	1.44 (0.54)	2.06 (4.88)	1.92 (2.57)	1.20 (2.33)	0.049	1.87 (0.92)	3.42 (4.08)	5.30 (36.51)	7.70 (20.41)	<0.001
ISG56	0.51 (0.23)	0.52 (0.33)	0.36 (0.23)	0.23 (0.33)	<0.001	0.66 (0.23)	0.75 (0.95)	0.96 (2.68)	0.82 (3.20)	0.046

Expression levels, calculated using 2-ΔCt, are indicated as median (interquartile range) of the group;

p-values for comparisons among the age groups are calculated using Kruskal-Wallis tests.

Basal Transcriptional Level of IFN Genes Is Elevated in Children and Decrease With Age

We first compared results from SARS-CoV-2 negative samples stratified by age groups ( ). All study genes, with the exception of ISG15, were far more expressed in children with progressively lower values found in young- and middle-aged adults and in the elderly (p<0.001). Uninfected young adults had levels of IFN-α, IFN-ε, IFN-ω, and IFN-λ3 transcripts at intermediate values between children and older adults; the level of IFN-β transcript was comparable to those found in the older adults’ groups, whereas IFN-λ1 and –λ2 were comparable to those in the children group ( ). These results were strengthened by the significant negative correlation between age and IFN-α, -β, -ε, -ω; the IFNs λ1-3; ISG56 ( ). Contrastingly, ISG15 levels showed little difference among groups (p=0.049), with no significant correlation with age.

Figure 2

Correlation between gene expression and age, in SARS-CoV-2 negative and positive subjects. Relative mRNA expression values (y-axis) are calculated using the threshold cycle relative quantification (2-ΔCt); Spearman’s rho coefficient and p-values are reported in the upper right corner. Panel A: correlation of gene expression with age in years of SARS-CoV-2 negative subjects. Panel B: correlation of gene expression with age in years of SARS-CoV-2 positive subjects.

IFN-ε and -ω are defective and ISGs Overexpressed in SARS-CoV-2 Infected Adults

The comparison of expression levels reached by IFN genes among SARS-CoV-2 positive age groups showed interesting differences. Regarding type I IFNs, transcripts of IFN-α and IFN-β did not differ among age groups ( ) and their values were not significantly correlated to age ( ). Contrastingly, levels of IFN-ε and -ω were similarly high in children and young adults and much lower in the older age groups (p=0.007; p=0.021 in JT test for trends, respectively; ) and were negatively correlated with age ( ). IFNs λ1-3 transcripts were comparable among the four age groups with no significant relationship with age ( , ). Notably, ISG56 and ISG15 were expressed at lower levels in the younger, and progressively increased with age (p=0.020; p<0.0001, in JT test for trends, respectively; ); a moderate significant positive correlation between age and ISG56 and ISG15 was found ( ).

IFN Genes Are Differentially Upregulated During SARS-CoV-2 Infection Depending on Age

To dissect the age specific extent of activation of the type I/III IFN response to SARS-CoV-2 infection, a comparison between SARS-CoV-2 negative and -positive samples was performed within each age-group ( ). In children, IFN-α (p=0.008) and ISG56 (p=0.012) were the only transcripts to be significantly higher in the SARS-CoV-2 positive with respect to the negative; there was a tendency to higher levels of IFN-ω (p=0.079) and ISG15 (p=0.057) whereas the other transcripts were comparable in SARS-CoV-2-positive and -negative children ( ). In young adults, comparisons between SARS-CoV-2 negative and -positive samples showed a significant activation of IFN-β, IFN-ε and ISG56 (p=0.024; p=0.017; p=0.018, respectively) and a trend toward higher levels of IFN-α, IFN-ω, IFN-λ2, IFN-λ3 and ISG15 (p=0.078; p=0.100; p=0.078; p=0.084; p=0.064, respectively) in the SARS-CoV-2-positive with respect to the -negative whereas IFN-λ1 levels were comparable between groups ( ). As regards the groups of middle-aged adults and older adults, all study genes were significantly activated in the SARS-CoV-2 positive subjects with respect to the negative of the same age group (p<0.005 for all comparisons, with the exception of ISG15 in the middle-aged adults, p=0.015; ).

Figure 3

Age-specific activation of IFN genes. Relative mRNA expression values (y-axis), calculated using the threshold cycle relative quantification (2-ΔCt), are depicted as open circles for SARS-CoV-2 negative and as filled circles for SARS-CoV-2 positive samples, triangles indicate samples with Ct values >40. Horizontal lines represent the median value of the group indicated below the X-axis. P-values, calculated using Mann-Whitney tests for comparison between SARS-CoV-2 negative and positive samples of the same age group, are shown with symbols representing: °0.1

0.05, *0.05

Age-specific activation of IFN genes. Relative mRNA expression values (y-axis), calculated using the threshold cycle relative quantification (2-ΔCt), are depicted as open circles for SARS-CoV-2 negative and as filled circles for SARS-CoV-2 positive samples, triangles indicate samples with Ct values >40. Horizontal lines represent the median value of the group indicated below the X-axis. P-values, calculated using Mann-Whitney tests for comparison between SARS-CoV-2 negative and positive samples of the same age group, are shown with symbols representing: °0.1