Cytokine autoantibodies in SARS-CoV-2 prepandemic and intrapandemic samples from an SLE cohort

Introduction

Autoantibodies directed to cytokines, including type I interferon (T1IFN), have been reported in a wide variety of infectious and autoinflammatory diseases including in up to 27% of patients with SLE.1–3 A remarkable finding reported during the SARS-CoV-2 pandemic was that pre-existing autoantibodies directed to T1IFN were associated with a higher risk of developing severe COVID-19.4–6 Given the importance of IFN in host viral defences,7 these observations raised the question of whether anti-T1IFN autoantibodies are associated with an increased risk of severe COVID-19 in SLE, a disease in which T1IFN plays a key role.8

Methods

We examined the prevalence of cytokine autoantibodies in prepandemic and intrapandemic sera from a SLE cohort using a commercially available 15-PLEX anticytokine array (Luminex Corporation, Austin, Texas, USA), which included detection of IgG autoantibodies directed to IFN-gamma (γ), IFN-beta (β), B cell activating factor (BAFF)/B lymphocyte stimulator (Blys), tumour necrosis factor alpha (α), granulocyte colony stimulating factor, interleukin (IL)-1α, IL-6, IL-8, IL-10, IL-12 (p40), IL-15, IL-17A, IL-17F, IL-18 and IL-22. In addition, autoantibodies to IFN-α (IFN-α2b) were detected by an addressable laser bead immunoassay (ALBIA) using the full-length purified human protein (Abcam, Cambridge, UK) following protocols as previously published.9 A cut-off of 500 median fluorescence intensity was established at two standard deviations above the mean of age-matched healthy controls, following the requirements for validation of laboratory tests in the accredited laboratory (MitogenDx) that performed these tests. SLE-related autoantibodies (anti-double stranded DNA (dsDNA), anti-Smith (Sm), anti-U1-ribonucleoprotein (RNP), anti-ribosomal P, anti-SSA/Ro60, anti-Ro52/(tripartite motif containing-21) TRIM21 and antichromatin) were detected in the prepandemic and intrapandemic sera by commercially available multianalyte immunoassays as previously published.9 10 Prepandemic serum samples from the SLE cohort were biobanked prior to 01 January 2020 and intrapandemic samples were collected from 15 March 2020 to 31 January 2021; all patients were unvaccinated at the time of serum collection.11 All samples were tested for SARS-CoV-2 antibodies using an ELISA measuring IgA and IgG antispike 1 (S1) protein (Euroimmun AG, Lübeck, Germany) and an assay detecting IgG antibodies to nucleocapsid (N), S1 receptor binding domain (RBD) and S1 (XMAP: Luminex Corporation, Austin, Texas, USA).11 RT-PCR tests were performed if clinically indicated and results collected retrospectively through to 31 January 2021.

Cytokine autoantibodies were compared between patients with SLE with SARS-CoV-2-positive serology and/or RT-PCR confirmed SARS-CoV-2 infection (SARS-CoV-2 serology methods previously published)10 vs the remaining cohort using a χ² test. The Bonferroni method was used to correct for multiple comparisons (17), with a corrected α of 0.0029 being considered statistically significant. Potential associations between each cytokine autoantibody and each SLE-related autoantibody in the prepandemic and intrapandemic periods were assessed using a χ² test, with a Bonferroni corrected α of 0.00045 being considered significant to account for the 112 comparisons.

Patient and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.

Results

The SLE cohort consisted of 173 patients (94.8% female, mean age 48.5 years, mean disease duration 11.7 years, 42.8% non-White race/ethnicity, 83.2% prescribed hydroxychloroquine, 28.9% corticosteroids and 43.9% other immunomodulators).11 Intrapandemic samples from nine patients had SARS-CoV-2-positive serology and/or RT-PCR confirmed SARS-CoV-2 infection (SARS-CoV-2 serology + and RT-PCR+ (n=3); SARS-CoV-2 serology + and RT-PCR not done (n=3); SARS-CoV-2 serology− and RT-PCR+ (n=3)), yielding 164 patients with no laboratory evidence of SARS-CoV-2 exposure. The frequency of cytokine autoantibodies in the prepandemic and intrapandemic samples stratified by SARS-CoV-2 exposure is shown in table 1, the patient characteristics and cytokine autoantibodies for the nine exposed patients are shown in table 2, and a graphical representation of the distribution of cytokine autoantibodies and their titers is shown in figure 1 (patients are sorted from left to right based on descending T1IFN autoantibody titres separately for prepandemic and intrapandemic samples).

Request permission

In the prepandemic samples, 33.3% (3/9) of patients with serological and/or RT-PCR evidence of SARS-CoV-2 exposure versus 28.0% (46/164) without exposure had at least one cytokine antibody (p=1.00). In the intrapandemic samples, 22.2% (2/9) of exposed versus 23.1% (38/164) without exposure had at least one cytokine antibody (p=1.00). Overall, there was no difference in the frequency of cytokine autoantibodies in prepandemic compared with intrapandemic samples (28.3% (49/173) vs 23.1% (40/173), respectively; p=0.33). In the entire cohort of 173 patients, the most common cytokine autoantibodies in prepandemic samples were directed against IFN-γ (9.8%), followed by IL-6 (8.7%), IL-10 (5.8%) and BAFF (4.6%) (table 1 and figure 1) compared with 5.2% for IFN-γ and anti-IL-6, 5.8% for IL-10% and 6.4% for BAFF in intrapandemic samples. In the prepandemic samples, none of the exposed patients had autoantibodies to IFN-γ versus 10.4% (17/164) without exposure (p=0.66). In the intrapandemic samples, 11.1% (1/9) of the exposed patients had autoantibodies to IFN-γ versus 4.9% (8/164) without exposure (p=0.96). In the prepandemic and intrapandemic samples, none of the exposed patients had autoantibodies to T1IFN (IFN-α or IFN-β) versus 3.7% (6/164) without exposure had autoantibodies to IFN-α and 0% without exposure had autoantibodies to IFN-β (p=1.00 and not applicable, respectively).

When cytokine autoantibodies were analysed for associations with SLE-related autoantibodies, only in intrapandemic samples were there associations with p values less than 0.05: anti-IFN-α and anti-dsDNA (intrapandemic p=0.015 vs prepandemic p=0.7035); IL-12 and anti-dsDNA (intrapandemic p=0.043 vs prepandemic p=0.816); IL-12 and anti-ribosomal P (intrapandemic p=0.009 vs prepandemic p=1.000); and IL-12 and anti-Sm (intrapandemic p=0.005 vs prepandemic p=1.000). However, after correcting for multiple comparisons none of these associations met the reduced α (α=0.00045) to be considered statistically significant.

Discussion

The frequency of cytokine autoantibodies observed in our study is in keeping with previous publications that used similar assay techniques reporting that up to 27% of patients with SLE had cytokine autoantibodies.1 3 In the present study, none of the patients with SLE with confirmed exposure to SARS-CoV-2 had autoantibodies to T1IFN and, as we previously reported,11 none of the 173 patients in our cohort, including the 9 patients with confirmed exposure, had severe COVID-19 symptoms necessitating hospitalisation. Hence, we could not confirm that T1IFN autoantibodies increase the risk for severe SARS-CoV-2. The overall observed frequency of anti-IFNγ autoantibodies in our study is consistent with a previous report of 11% in patients with SLE.1 In an earlier communication, we reported that this SLE cohort had a lower rate of prepandemic seropositivity and a slightly lower to similar rate of intrapandemic seropositivity compared with contemporaneous controls.11 Granted, in the current report we did not study the neutralising effect of cytokine autoantibodies, but it would be difficult to explain T1IFN neutralising activity in the absence of detectable T1IFN cytokine autoantibodies. In addition, we found a non-significant association between IFNγ autoantibodies and dsDNA autoantibodies in intrapandemic samples, which is similar to a previous report that anti-IFNγ autoantibodies were associated with anti-dsDNA.1 Further, due to the size of our cohort and its geographical context, our results may not be generalisable to a broader population of patients with SLE. Because of interest in vaccine responses in patients with SLE,12–15 current efforts are focusing on a possible relationship of cytokine autoantibodies to adverse or unusual reactions.