This study is the first to demonstrate the immunogenicity of heterologous prime-boost vaccination with an inactivated vaccine followed by the ChAdOx1 nCoV-19 viral vector vaccine in patients with SARDs receiving immunosuppressive drugs. Furthermore, we demonstrated the magnitude of humoral and cellular immunogenicity and evaluated neutralizing activity against emerging Omicron BA.2 VOC. We found an impaired humoral response and a preserved cellular immune response. Surprisingly, the levels of IFN-γ+ CD4+ T cells, IFN-γ+ CD8+ T cells, and TNFα+ CD4+ T cells were higher in the SARDs group compared to the healthy group. GC, MMF, and AZA were associated with a diminished humoral immune response. There was no difference in the AEFI between the SARDs and healthy groups.

A heterologous prime-boost or “mix and match” strategy was employed to maximize vaccine immunogenicity, and mitigate the risk for adverse reactions and vaccine shortage14. Using an adenovirus vector as a prime followed by other platforms as a booster improved neutralizing antibodies and Th1 T cell responses in a specific pathogen-free BALB/c mouse model15. Proof-of-concept heterologous vaccine studies, using the ChAdOx1-nCoV-19 vaccine boosted by BNT162b2 vaccine, revealed superior anti-RBD IgG, neutralization titer and T cell reactivity compared to homologous regimens9,16. Preliminary studies of heterologous prime-boost vaccination regimens with the inactivated SARS-CoV-2 vaccine followed by the ChAdOx1-nCoV-19 vaccine found that this strategy resulted in an excellent humoral response compared with homologous regimens with inactivated vaccine or adenoviral vector vaccines13. However, no heterologous prime-boost studies have reported outcomes for patients with SARDs. Results of our first study demonstrated the immunogenicity of a heterologous prime-boost regimen with an inactivated vaccine followed by the ChAdOx1 SARS-CoV-19 vaccine. This approach provides another option for patients with SARDs for whom homologous mRNA or adenoviral vaccines are contraindicated.

Our study demonstrated 93.3% seropositivity, which is higher than the 70.4% seropositivity obtained with a homologous inactivated vaccination regimen in 910 patients with SARDs2. Likewise, a 56.3% seroconversion rate was reported by Seree-aphinan et al.17. Data following two doses of homologous ChAdOx1 SARS-CoV-2 vaccine in patients with SARDs are lacking; however, Shenoy et al. revealed that 90.2% patients had detectable antibodies after vaccination with the ChAdOx1 SARS-CoV-19 vaccine18. A large case-control study of mRNA-based COVID-19 vaccines by Furer et al. reported 86% vaccine responders in the autoimmune inflammatory rheumatic diseases cohort, similar to results of Braun-Moscovici et al. and Ferri et al.3,19. The high seropositivity rate herein may be explained by the beneficial effect of the heterologous strategy. Nevertheless, the results may be impacted by the intensity of immunosuppressive drugs and the small number of patients. Our results revealed a seropositivity rate at least equal to homologous mRNA, or adenovirus vectors, and homologous inactivated vaccines.

The methods to evaluate the anti-SARS-CoV-2 S-protein assays are diversified, leading to difficulty comparing the results between studies. The WHO International Standard for COVID-19 serological tests aims to harmonize humoral immune response assessments using binding antibody units (BAU) as the universal reporting system. However, Infantino M et al. revealed uninterchangeable differences in commercial quantitative anti-SARS-CoV-2 S-protein assays, even with conversion to BAU/ml20. Thus, we could not directly compare our quantitative anti-SARS-CoV-2 S-protein assay results to those of other studies. Nevertheless, our study demonstrated a significant reduction in antibody levels compared to those of healthy groups, similar to the studies by Furer et al. and Ferri et al., which involved either the BNT162b2 or mRNA-1273 vaccine3,19.

Neutralizing activities were evaluated to predict vaccine efficacy; different SARS-CoV-2 vaccines utilize a wide range of Nabs. Nonetheless, reduced Nabs to the Omicron VOC is a global health problem due to its ability to escape host immunity21. A previous study confirmed that this variant is resistant to therapeutic antibodies and reduced neuralization capacity to double BNT162b2 vaccination22. No previous studies have investigated heterologous CoronaVac/ChAdOx1. Our study is the first to evaluate the neutralization of Omicron BA.2 VOC in patients with SARDs administered heterologous vaccine regimens. The results revealed significantly lower neutralizing titers in patients with SARDS compared with healthy group, and nearly half of these patients were negative in the PRNT. This finding indicated the diminished humoral vaccine immunogenicity to the Omicron VOC in patients receiving immunosuppressive drugs. These data support current recommendations for additional booster doses for patients with SARDs.

Evidence supporting vaccine-induced T cell responses is sparse and controversial. Prendecki et al. demonstrated a preserved T cell response following the completion of a second primary series of BNT162b2 mRNA or ChAdOx1-nCoV-19 vaccines in patients with SARDs. The T cell response was detected in 81.8% of patients who receiving immunosuppressive drugs, despite B-cell depletion4. Bitoun et al. reported a similar finding; there was no difference in CD4 T cell secreting IFN-γ and TNF levels between patients receiving immunosuppressive drugs, rituximab, and healthy group23. In addition, the CD8+-induced TNF response against spike peptides tended to be reduced in patients with a defective humoral response23,24. Conversely, Miyara et al. revealed a cellular response of only 57% using IFN-γ secretion levels in patients with SLE with a neutralizing antibody response after two doses of BNT162b2 vaccine25. Our study demonstrated that the vaccine induced a more significant polyfunctional Th1 cytokine response in both CD4 + and CD8 + T cells. These findings can be explained by the effects of ChAdOx1-nCoV-19 vaccines, which predominately induce the T cell response, and the lack of a calcineurin inhibitor. Greater Th1 cytokine secretion compared with healthy group may be explained by the baseline cellular subsets in patients with autoimmune disease26,27.

The use of immunosuppressive drugs contributes to impaired vaccine immunogenicity. We found that GC, MMF, and AZA were correlated with reduced anti-RBD IgG levels. The effects of GC and MMF were concordant with the previous studies1,2,3,4. However, MTX did not diminish immunogenicity, which contrasts to the previous findings2,25,28,29. This may be explained by a lower dose of MTX (10–15 mg per week) and the fact that patients temporarily withheld MTX for 1 week after each vaccine dose. Impaired humoral immunogenicity with AZA was a notable finding of our study. This effect was similar to immunogenicity reported following administration of an influenza vaccine in patients with SLE30. Recently, a study of the mRNA vaccine demonstrated significantly reduced antibody titers in patients with SARDs who received AZA (p = 0.01)31. Rituximab is associated with humoral impairment and strongly correlated with vaccine non-response;3,4,5 however, this study did not evaluate the impact of rituximab and instead aimed to focus on conventional immunosuppressive drugs.

Although half the patients in our cohort had SLE, we did not expect different SARD subtypes to significantly impact vaccine immunogenicity. Prior studies, which included all SARD subtypes, demonstrated that the reduction of vaccine immunogenicity depends on the intensity of immunosuppressive drugs and the vaccine platform2,3,19. These findings are consistent with guidelines for post-vaccination immunosuppressive drug discontinuation, in which recommendations depend on the type of immunosuppressive drug32. Recent studies in SLE populations revealed that MMF and GC dosage were significantly associated with impaired humoral immunogenicity after a primary series of SARS-CoV-2 vaccination33,34. SARS-CoV-2 vaccine immunogenicity in SLE and other SARDs were compared by Ammitzbøll et al., who reported similar seropositivity rates between 61 SLE patients and 73 rheumatoid arthritis patients after administration of an mRNA vaccine35. Furer et al. also demonstrated acceptable seroconversion rates in patients with SLE, RA, psoriatic arthritis and ankylosing spondylitis, while antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis and idiopathic inflammatory myositis had the lowest seroconversion rates. This finding was explained by the intensity of immunosuppressive drugs, especially rituximab3. In summary, although the SARD subtypes in our study varied, treatment with immunosuppressive agents was found to be the major risk factor for reduced immunogenicity.

To our knowledge, this is the first prospective cohort study to compare the immunogenicity of heterologous prime-boost vaccination with inactivated vaccine followed by ChAdOx1-nCoV-19 in patients with SARDs compared with age- and sex-matched healthy controls. A strength of this study was the use of PRNT to evaluate Omicron VOC and T cell immunogenicity. Validation in larger studies analyzing the effects of each SARD subtype and immunosuppressant is required. The immunogenicity of a booster dose in this strategy is worthy of follow-up owing to the intact T-cell responses. Furthermore, these findings may support practical vaccination recommendations for patients with autoimmune rheumatic diseases who cannot be vaccinated with homologous mRNA-based vaccines.

Our study had several limitations. First, the study included a small number of patients. Second, we did not evaluate the pre-vaccination antibody status of participants; however, all patients declared no prior SARS-COV-2 infection. Third, we did not include biological or small-molecule drugs which may impact cellular immunogenicity; however, this reflects a case of real-world immunosuppressive use in populations with limited access to biological drugs. Last, because the patients did not visit the clinic simultaneously, cryopreserved cells were used for the cellular analysis. Peripheral blood mononuclear cell (PBMCs) obtained from patients treated with immunosuppressive drugs provided very limited yields. Furthermore, we did not include unstimulated controls. The responses observed could be a result of auto-reactivity as well as S1 reactivity.

Patients with SARDs who received immunosuppressive drugs, mainly GC, MMF, and AZA, had a lesser humoral vaccine response compared with healthy controls. The finding of a preserved T cell response is a highlight of this study. T cells may be essential in the prevention of severe COVID-19 disease, which is important for this valuable group. This vaccine regimen may be an option for patients with SARDs who are hesitant or have contraindications to the mRNA vaccine. However, neutralizing titer to the Omicron strain was reduced, supporting the necessity for a third booster dose.