At the onset of the COVID-19 pandemic, despite limited genetic variation in the SARS-CoV-2 genome, some individuals developed respiratory failure, while others remained asymptomatic or experienced only mild symptoms (1), even before treatments or prophylactic vaccines became widely available. Older age and presence of clinical comorbidities such as obesity are associated with an increased risk of severe COVID-19 (2), particularly in the pre-vaccine era, possibly mediated through an impaired cellular environment resulting in distorted expression of interferons (IFNs) and an impaired antiviral immunity (3).

IFNs, a principal group of antiviral cytokines, are broadly classified into three main types based on their specific receptors and structural homologies, including IFN types I, II, and III (4). Type I IFNs include IFNα, β, ϵ, k, ω, and δ, type II comprises IFNγ (5), while type III IFNs comprise the interferon lambda (IFNλ) family and were initially termed IL-29, IL-28A, and IL-28B, but later renamed IFNλ1, 2, and 3, respectively (6).

Genome-wide association studies (GWAS) have identified single nucleotide polymorphisms (SNPs) that are associated with expression of a fourth type III IFN; IFNλ4. Together, the IFNλ family have been shown to impact on antiviral immune responses and correlate to infectious disease outcomes in HCV, HIV, CMV, and influenza viral infection (7, 8).

IFNλ4 expression is determined by two major SNPs: rs368234815 and rs117648444. The former harbors a frameshift mutation that changes the genotype from ΔG-expressing to TT-non-expressing. The ΔG genotype expression allows for the production of a functional IFNλ4 open reading frame for a functional IFNλ4 protein, while the TT allele results in a truncated non-functional protein which abolishes the expression of IFNλ4 (9). Importantly, IFNλ4 expression is found in only 50% of the world population with striking ethnic differences; 50% of Europeans, 90% of Africans, and 10% of Asian individuals (10).

The genotype of a neighboring SNP, rs117648444 (G/A: IFNλ4 P70/S70), can also modify the biological activity of rs368234815-ΔG (10). The rs117648444 activity variant SNP results in a missense mutation that gives rise to a non-synonymous mutation P70S (11). This hypomorphic rs117648444-A allele only occurs when the ΔG is present in rs368234815. The two SNPs can thus differentiate individuals into three genotypic groups: IFNλ4-null (rs368234815-TT/TT, rs117648444-G/G), IFNλ4-strong (rs368234815-TT/ΔG and ΔG/ΔG, rs117648444-G/A), and IFNλ4-weak (rs368234815-TT/ΔG and ΔG/ΔG, rs117648444-A/A) (10).

Obesity and a high body mass index (BMI) were classified as significant predictors of COVID-19 outcomes and severity (12). The relationship between inflammation and metabolism has recently become evident, as changes in metabolism, such as those seen in obesity, can lead to inflammation. Conversely, inflammation resulting from viral infections can also induce metabolic changes (13). Teran-Cabanillas et al. have demonstrated impaired interferon responses in individuals with obesity, specifically IFNα, IFNβ and IFNλ, affecting their immune function (14). This indicates the role that obesity plays in modulating immune responses to viral infections.

Dysregulated IFN responses are deemed a key factor in COVID-19 pathogenesis, but conflicting results are observed in different studies examining COVID-19 disease severity and IFN responses. Early induction of type I and III IFNs may produce a favorable outcome of COVID-19 disease progression (15). While type I IFNs act rapidly to induce an immune response via the expression of chemokines and cytokines (16), type III are considered to be tissue protective and usually lack the accompanying pro-inflammatory response induced by type I IFNs (17). However, a number of studies have also associated type I IFN responses with an increased expression of interferon stimulated genes (ISGs), proinflammatory genes, and cytokines in individuals with severe COVID-19 (18). Kwon et al, have shown inflated type I/II IFNs (IFNα and IFNγ) in severe COVID-19 within 5-10 days of symptom onset, notably IFNα tended to correlate with the viral load in their cohort (19). Conversely, other studies have shown a greatly weakened type I and III IFN response in early COVID-19 infection (20). The timing of IFN induction whether type I or III, may also be an integral part in regulating the outcome of COVID-19 disease.

Given this lack of clarity on the role of IFNs in COVID-19 infection, we sought to investigate what role, if any, expression of IFNλ4 plays in modifying disease severity in COVID-19 and additionally to explore relationships between early infection levels of other circulating type I, II, and III IFNs and COVID-19 disease severity, considering important clinical predictors such as age and presence of comorbidities such as obesity.

This study investigated the association of early host IFN responses with COVID-19 disease severity. We demonstrate that type III IFNs, and particularly IFNλ2, may play a role in COVID-19 disease outcome. Additionally, we found higher IFNλ2 levels in those with obesity and a relationship between IFNλ2 and COVID-19 disease severity that was only observed in those without obesity, suggesting an interaction between obesity and IFNλ2 responses that could help explain differing clinical outcomes to COVID-19 in people with obesity. However, when investigating the expression of two common IFNλ4 SNP genotypes of rs368234815 and rs117648444, we found that the expression of IFNλ4 had no impact on clinical disease severity in COVID-19 in this study.

Type I and III IFNs demonstrate a paradoxical role in mediating host immune responses to SARS-CoV-2 infection, as their role in early infection is not well elucidated, with early increased levels shown to be protective against severe disease in some studies, while others have shown the opposite (24, 25). Zaleska et al. found lower levels of IFNλ2 in severe COVID-19 compared to the moderate group, where the majority of the severe group had undetectable levels of IFNλ2 (26). In contrast, our unadjusted analyses showed higher IFNλ1 and IFNλ2 levels in those with moderate and severe COVID-19, possibly driven by the high prevalence of comorbidities in our cohort, which may have amplified the type III IFN response. Nonetheless, our findings align with Ruytinx et al. who demonstrated higher IFNλ1 in severe and critically ill individuals with COVID-19, with increased IFNλ1 also associated with a higher odds ratio of ICU mortality (27). However, after adjustment, only higher IFNλ2 remained associated with more severe COVID-19 in our cohort, with the opposite observed in the cohort of Rutinx et al. As we collected our samples within a range of 10 days of symptom onset, they collected their samples within 8 days of a positive SARS-CoV-2 PCR test. Additionally, there could be intrinsic differences of the analytical methods between both studies, although this may be a small difference between methods of time of sampling or cytokine measurement, it could still expound the differences in cytokine levels between our cohort and theirs.

The relative roles of IFNs in mediating host immune responses to respiratory viruses are complex. Our data supports a role for heightened type III IFNs, precisely IFNλ1 and IFNλ2, in driving more severe COVID-19. Type III IFNs are known to be potent cytokines provoked by ISGs (28), several cohorts have found increased transcriptional ISG signatures in individuals with severe COVID-19, where these transcriptional signatures have been linked back to higher SARS-CoV-2 viral loads which may also trigger a more intense cytokine storm (29). Taken together, these data support a scenario where increased expression of ISGs drive expression of type III IFNs which then contribute to the cytokine storm characteristic of severe COVID-19.

Interestingly we also observed differential associations of IFN levels and disease severity based on subjects’ BMI. Although we observed significantly higher IFNλ2 in subjects with obesity, the relationship between elevated IFNλ2 and severe disease was only observed in people without obesity. Various cohorts have shown that obesity is a major predictor of severe COVID-19 disease (30, 31). These studies have focused largely on epidemiological data without the consideration of inflammatory markers in their analytical models. As obesity is considered to be low-grade inflammation, it is not surprising that people with obesity experience an altered expression of pro-inflammatory cytokines, with the adipose tissue being an important producer of TNFα, IL1β, and IL6 (32). Our data would support a scenario whereby individuals with obesity have pre-existing elevations of inflammatory cytokines and interferons, including type III IFNλ2 placing them at higher risk of severe disease, whilst individuals without obesity who do not mount excessive IFNλ2 responses are relatively protected against severe COVID-19. This is confirmed by several studies that found elevated levels of type III IFNs, particularly IFNλ1 in people with obesity (33, 34).

Although, the different subtypes of the IFNλ family share highly homologous amino acid (aa) sequences which likely arose through gene duplication giving rise to paralogous sequences. IFNλ1 and IFNλ2 share 81% aa, IFNλ2 and IFNλ3 share 96% aa, while IFNλ4 may be more distantly related and possesses only 28% aa identity with other IFNλ types (7). Our outcomes showed IFNλ1 and IFNλ2 were significantly associated with severe COVID-19, whereas IFNλ4 expression was not. Our findings on IFNλ4 are consistent with previous data form a smaller Spanish cohort (N=177) of mainly Caucasian subjects (35). However, our outcomes contrast with another study that analyzed IFNλ4 expression between survivors and non-survivors in an Iranian population (N=750), which observed associations between IFNλ4-ΔG/ΔG genotype and COVID-19 mortality (36). This discrepancy between these outcomes could partially be explained by IFNλ4 allele frequencies within different ethnicities, as our cohort, although larger, consisted of a primarily Caucasian population, similar to the Spanish cohort, with lower frequencies of the ΔG allele in a European population.

These data build on the growing evidence surrounding the role of IFNλ4 in modifying host responses to viral infections, some beneficial and some less so. In Hepatitis C (HCV), the rs368234815 SNP is associated with differing responses to treatment depending on the HCV genotype (improved responses in HCV genotype 1 and 4 but not genotypes 2 and 3) (37), and the spontaneous clearance of HCV in the absence of treatment (38). In people with HIV, we previously reported on associations between rs368234815-ΔG/ΔG and increasing likelihood of normalizing CD4⁺:CD8⁺ ratio in response to antiretroviral therapy (39). In contrast, presence of rs368234815-ΔG/ΔG linked to IFNλ3-rs12979860 was associated with reduced clearance of respiratory RNA viruses in Rwandan children (40). These results suggest that the impact of host expression of IFNλ4 may have differing impacts on clinical outcomes depending on the virus and type of infection. Whether IFNλ4 polymorphisms impact upon COVID-19 severity and, most importantly, the ultimate outcome following SARS-CoV-2 infection requires further investigation.

Our study does have limitations. The high proportion of Caucasians likely resulted in an under ascertainment of the ΔG allele. A larger, more diverse cohort could provide additional insights into the impact of IFNλ4 expression and COVID-19 severity and outcome. As our samples were collected between March 2020 and June 2021, we were unable to explore the impact of COVID-19 vaccination on the interaction between type III IFN expression and COVID-19 disease severity, which may change following SARS-CoV-2 vaccination. Additionally, we did not measure neutralizing type I IFN autoantibodies nor inborn errors of type I IFN, which have also been implicated in contributing to COVID-19 disease severity (41). Furthermore, during this time period, a number of distinct SARS-CoV-2 variants of concern were in circulation with differing impacts on clinical severity which may also have impacted the analyses (42). Additionally, obesity was defined as BMI ≥ 30 kg/m², which may not fully capture other measures such as percentiles or obesity grades. However, given that most participants were Caucasian, this is unlikely to have significantly impacted our findings. Our ventilation data was limited to a binary variable, with only 15 of 321 participants requiring ventilation. Similarly, as most participants were either discharged or treated as outpatients, further analysis of ventilation duration and interferon levels between survivors and non-survivors was limited. Lastly, we did not assess the predictive value of IFN levels and IFNλ4 expression. Future studies with longitudinal data are needed.

In conclusion, common IFNλ4 genotypes were not associated with COVID-19 disease severity or expression of circulating IFNs. However, both higher IFNλ1 and IFNλ2 were associated with more severe COVID-19 disease. IFNλ2 was also higher in individuals with obesity, although associations between IFNλ2 and COVID-19 disease severity were only observed in individuals without obesity, suggesting that obesity may contribute to increased risk of severe COVID-19 through increased expression of IFNλ2, potentially impacting disease progression. These findings suggest that assessing IFNλ2 could help to identify those at a greater risk of severe disease, enabling earlier interventions and improved management strategies. Future work is required to validate these results and to study the impact of vaccination on the relationships between type III IFNs and COVID-19 severity.