Factors associated with immune responses to SARS-CoV-2 vaccination in individuals with autoimmune diseases

Patients with autoimmune diseases are at higher risk for severe infection due to their underlying disease and immunosuppressive treatments. In this real-world observational study of 463 patients with autoimmune diseases, we examined risk factors for poor B and T cell responses to SARS-CoV-2 vaccination. We show a high frequency of inadequate anti–spike IgG responses to vaccination and boosting in the autoimmune population but minimal suppression of T cell responses. Low IgG responses in B cell–depleted patients with multiple sclerosis (MS) were associated with higher CD8 T cell responses. By contrast, patients taking mycophenolate mofetil (MMF) exhibited concordant suppression of B and T cell responses. Treatments with highest risk for low anti–spike IgG response included B cell depletion within the last year, fingolimod, and combination treatment with MMF and belimumab. Our data show that the mRNA-1273 (Moderna) vaccine is the most effective vaccine in the autoimmune population. There was minimal induction of either disease flares or autoantibodies by vaccination and no significant effect of preexisting anti–type I IFN antibodies on either vaccine response or breakthrough infections. The low frequency of breakthrough infections and lack of SARS-CoV-2–related deaths suggest that T cell immunity contributes to protection in autoimmune disease.


Clinical assessments
Demographic data, concomitant medications, comorbid diseases, and vaccine type were entered into a central Health Insurance Portability and Accountability Act (HIPAA)-compliant REDCap database.Disease activity was categorized as either inactive, mild, moderate, or severe using the Clinical Disease Activity Index (CDAI) for RA (1), SLEDAI-2K for SLE (2), Pemphigus Disease Area Index (PDAI) for pemphigus (3), or physician determination for other diagnoses.Disease flares were defined as an increase in disease activity prompting changes in immunosuppressive medication.Treatment details-including current use of immunosuppressive, immunomodulatory, and biologic medications; history of B cell depletion or cyclophosphamide within the last 12 months; current oral corticosteroid dose; and infusions of > 100 mg of methylprednisolone within the last 6 months-were recorded at the Pre V visit.Changes in medications were recorded at all subsequent visits.Additional medications recorded at postvaccine visits must have been taken for > 4 weeks, except for cyclophosphamide and B celldepleting agents, which could be taken for shorter periods.Temporary suspension of immunosuppressives or biologics at the time of vaccination according to or close to American College of Rheumatology (ACR) guidelines was also recorded (4).
After incubation, PBMCs were centrifuged and washed with fluorescence-activated cell sorting (FACS) buffer (PBS/2mM EDTA/1% BSA).Cells were stained with the fluorophore conjugated antibodies and viability dye (Supplementary Table 10), for 30 min at 4°C.Stained PBMCs were washed twice with FACS buffer and transferred to 5 mL round bottom flow tubes.A Becton Dickerson (BD) LSR Fortessa X-20 cell analyzer was used to acquire data on all samples.Data were analyzed using BD FACS Diva (v9.0) and FlowJo (v10.8.1).

Autoantibody arrays
Antigens (Supplementary Table 9) were conjugated to uniquely barcoded, carboxylated magnetic beads (MagPlex-C, Luminex Corp.) as previously described (29).Five µl of bead array was added to each well of a 384-well plate (Greiner BioOne).Forty-five µl of diluted serum or plasma per well was transferred into the 384-well plate containing the bead array.Samples were incubated for 60 minutes on a shaker at room temperature and then left overnight at 4°C.Beads were then washed with 3 × 60 µl PBS-Tween on a plate washer (EL406, Biotek) and incubated with 50 µl of 1:1000 diluted R-phycoerythrin (R-PE)-conjugated, Fc-γ-specific goat anti-human IgG F(ab')2 fragment (Jackson ImmunoResearch, 106-116-098) for 30 minutes.The beads were then washed with 3 × 60 µl PBS-Tween and re-suspended in 50 µl PBS-Tween, and the plates were then analyzed with a FlexMap3D TM instrument (Luminex Corp.).Beads were validated on positive control plasma or serum samples with known reactivity patterns derived from subjects with confirmed prior SARS-CoV-2 infection and/or autoimmune diseases (29).Healthy control sera were obtained by the Stanford Biobank prior to the pandemic.pp.subject had also been on belimumab that was stopped at Post V2 1. 1 subject was MMF non-adherent and on an equivalent dose of prednisone ³ 30 mg; 1 subject in which MMF was held (F694) 2. 1 subject on an equivalent dose of prednisone ³ 30 mg, and was non-adherent to MMF (F323) 3. 1 subject in which MMF and hydroxychloroquine were stopped, and methotrexate continued, also on an equivalent dose of prednisone ³ 30 mg (500497) 4. subject on an equivalent dose of prednisone ³ 30 mg 5. 1 subject on an equivalent dose of prednisone ³ 30 mg 6. azathioprine held for vaccine dosing 7. 3 subjects in which methotrexate was held (R118, F677, R038) 8. 1 subject in which methotrexate was held and hydroxychloroquine was continued (R105) 9. 1 subject in which methotrexate was held (F717) 10. 2 subjects in which MMF was held (R148, ACE-P-012) 11. 1 subject in which belimumab was held, and methotrexate and hydroxychloroquine were continued (F158) 12. belimumab was stopped and adalimumab started and taken between the 1 st and 2 nd Pfizer vaccines, while methotrexate was continued (500940) 13. 1 subject in which methotrexate was held and hydroxychloroquine was continued (F355) 14. 1 subject in which MMF was held and hydroxychloroquine was continued (F692) 15. 5 subjects were non-adherent to MMF 16. 3

Supplementary Figure 2. AIM Assay Quality Controls. A-B: Dot plot colored by treatment group comparing total CD4 (A) and CD8 (B) counts vs % AIM+ after SARS-CoV-2 peptide pool stimulation. To limit false positives/negatives, samples with fewer than 1,000 CD4 (2.85%) and 750 CD8 T (6.55%) cells were removed from subsequent analysis. C-F: CD4 (C, D) and CD8 (E, F) AIM+ frequency after SARS-CoV2 (C, D) or CMV (E, F) peptide pool stimulation of
Linear regression modeling was used to determine correlations of anti-spike IgG values, CD4 T cell % AIM, or CD8 T cell % AIM at thePost V1 visit with other serological and clinical variables.Preprocessing: The variables considered for inclusion in the models were: ethnicity, race, age, sex, autoimmune diagnosis, comorbid disease (diabetes, hypertension, coronary artery disease, multiple comorbidities, other, or none), BMI, disease activity severity (inactive, mild, moderate or severe), medications, months since last B cell depleting agent, total number of doses of B cell depleting agent, COVID exposure, number of weeks since full COVID-19 vaccination, vaccine type, MMF/MPA level, flare in disease activity requiring an increase or change in medication, mean corticosteroid dose since last visit, cytokine levels (IFNa1,2,6,7, IFNg, IL-17A, IL-17F, GM-CSF, TNFa, IL-6, CXCL10), anti-NC, anti-spike IgG, CD4 T cell level, and CD8 T cell level.Certain medications were considered together: B cell inhibition (ocrelizumab, obinutuzimab, ofatumumab, rituximab), TNF inhibitors (adalimumab, certolizumab, etanercept, golimumab, infliximab), IL-12/23 and IL-17A inhibitors (ustekinumab, secukinumab), Jak inhibitors and IL-6 receptor antagonist (tocilizumab), and MMF/MPA.Variables with over 50% missing data were excluded from the dataset.The log2 of the anti-spike IgG and T cell values were used.KNN imputation was performed on all non-one-hot-encoded data to increase the amount of data points available for training and testing.Model building: A linear regression model was used to determine correlations between the serological or clinical variables and the three outcomes of interest (anti-spike IgG values, CD4 T cell % AIM, or CD8 T cell % AIM).70% of the data was used as a training set, of which 30% was used as a validation set.The remaining 30% of the data was used as a test set.The variables were added into the model sequentially.If including a new variable improved the performance of the linear regression model, as measured by the R 2 value, the new variable was kept in the final model.If including a new variable did not improve the model's performance, it was removed from the dataset.The final performance of the model was evaluated on a validation data set that included thirty percent of the data the testing data that was set aside prior to building the model.The model's accuracy was measured through evaluating the R 2 value.The coefficients of every variable included in the model were also reported.

4 :
Pre V and Post V1 samples from anti-NC-healthy donors post vaccination.Positive SARS-CoV2 response thresholds, represented as dotted lines, were calculated using the uninfected Pre V group (n=11 for CD4 and n=10 for CD8) as follows: mean + 2*SD and mean + 10*SD.Statistical significance was determined by unpaired two-tailed t test.** p < 0.01, **** p < 0.0001.G, H: CD4 (G) and CD8 (H) AIM frequencies after the indicated stimulation showing consistency of a single deidentified healthy donor included in each run of the AIM assays.Data represent mean ± SEM (n = 37).I, J: No differences in CD4 (I) and CD8 (J) AIM frequencies in response to CMV peptide pool stimulation at sequential visits according to SARS-CoV2 exposure and medication.Each data point represents an individual subject.Supplementary Figure 3. Breakthrough infection frequency and severity at each visit.Frequency and severity of infections acquired since the previous visit are shown in each bar as a percentage of subjects with available data at each visit.The subject number for each visit is shown above each bar.Supplementary Figure Autoantibodies to selected analytes A-E: Line plots showing changes in MFI levels for 5 analytes in matched samples from 241 vaccinated subjects with a range of autoimmune diseases.B-D: Autoantibodies against Ro60/SSA, La/SSB, and Smith demonstrate stable MFI levels throughout the vaccine series.E: Newly detected anti-TPO autoantibodies in a small subset of subjects.F-I: Anti-cardiolipin IgG (F, G) and IgM (H, I) autoantibodies in matched samples from 186 anti-NC-(green) and 63 anti-NC+ (black) subjects.Values >10 are considered positive.J, K: Anti-desmoglein1 (J) and 3 (K) autoantibodies in matched samples from 23 subjects with pemphigus.Each connected set of data points represents 1 subject.P = NS for all comparisons performed usingKruskal-Wallis ANOVA.Supplementary Figure 5. Heatmap representing serum IgG detected with an 83-plex array of cytokines and chemokines, traditional autoimmune-associated antigens, and viral antigens.239 vaccinated autoimmune subjects are represented, grouped by whether they were infected with SARS-CoV-2 by the Post V1 visit.Representative data from 16 prototype samples and 8 healthy control subjects are included.Within each group, samples are clustered and annotated by the visit at which the sample was taken (Pre V, yellow; Post V1, blue; Post B1, red).Analytes that were not cross-reactive and had a value above 5000 MFI are shown.Analytes in each group of antigens are color coded and individual antigens in each group of antigens are shown in B.

Supplementary Table 1: Response rates to vaccination and booster in autoimmune subjects and healthy controls
AI: autoimmune subjects; HC: healthy controls *Numbers of autoimmune subjects in each groupSupplementary

Table 2 . Medication Use Associated with Anti-Spike Antibody and Anti-NC Status
abatacept was stopped prior to booster dosing cc. 1 subject in which mycophenolic acid was started at Post V1 and then held for booster dosing dd. 1 subject in which both methotrexate and HCQ were held for booster dosing ee. 1 subject non-adherent to MMF, and clinical trial drug started at Post V3; 1 subject non-adherent to MMF and on tacrolimus ff. 1 subject non-adherent to MMF, and received a mean steroid dose of 400 mg prednisone equivalent between Pre (1)fingolimod(1), glatiramer acetate (1), interferon beta (1), secukinumab (1), tocilizumab (1), cladribine (1), JAK inhibitor (1), other (1) l. dimethyl fumarate (1), glatiramer acetate (3), interferon beta (5), natalizumab (1), other (9) 1 subject who was ANA positive, without a clear autoimmune diagnosis, and was started on hydroxychloroquine before the Post V1 visit ** 9 subjects who received B cell depletion alone, with the last dose > 60 months prior to vaccination, were excluded from this table subjects non-adherent to MMF at the Post V1 visit, 1 of these subjects also had a breakthrough infection !! belimumab held for booster dosing; HCQ and azathioprine continued; also had breakthrough infection @@ MTX held at Post V1 and at booster; HCQ continued at Post V1 and held at booster && MTX held at Post V1; subject also with breakthrough infection ^^ on adalimumab only at the Post V1 visit &&& clinical trial drug + belimumab stopped at Post V2 ++ 1 subject with HCQ held at booster visit *** 3 subjects non-adherent to MMF at the Post V1 visit !!! 1 subject non-adherent to MMF at the Post V1 visit; also received a mean of 400 mg prednisone between the Pre V and Post V1 visits ^^ 1 subject on HCQ alone at the Post V1 visit, and was on HCQ and methotrexate at the Post B1 visit 1. 1 subject with breakthrough infection, anti-NC negative $$ also, with breakthrough infection *