Higher and Sustained Cell-Mediated Immune Responses After 3 Doses of mRNA COVID-19 Vaccine in Patients With Inflammatory Bowel Disease on Anti–Tumor Necrosis Factor Therapy

INTRODUCTION: Studies suggest that the generation of durable T-cell immunity following coronavirus disease 2019 (COVID-19) vaccination protects against severe disease. The aim of this study was to measure cell-mediated immune response (CMIR) 1–2 months and 6 months after a third dose of a COVID-19 mRNA vaccine. METHODS: This prospective study (HumoRal and CellULar initial and Sustained immunogenicity in patients with inflammatory bowel disease [IBD]) evaluated CMIR at 28–65 days (t1) after dose 2, 28–65 days (t2) (n = 183) and 6 months (±45 days) (t3) (n = 167) after a third dose of an mRNA COVID-19 vaccine. A small cohort had blood sample available 28–65 days (t4) (n = 55) after a fourth dose. Primary outcomes were CMIR at (t2) and (t3). Secondary outcomes included the effect of immunosuppressing IBD medications on CMIR and response at (t4). RESULTS: All patients had measurable CMIR at all time points. CMIR increased at t2 compared with that at t1 (median 1,467 responding cells per million (interquartile range [IQR] 410–5,971) vs 313 (94–960) P < 0.001). There was no significant waning in t2 vs t3 or significant boosting at t4. Those on anti–tumor necrosis factor monotherapy had a higher CMIR compared with those not on this therapy at t2 (4,132 [IQR 1,136–8,795] vs 869 [IQR 343–3,221] P < 0.001) and t3 (2,843 [IQR 596–6,459] vs 654 [IQR 143–2,067] P < 0.001). In univariable analysis, anti–tumor necrosis factor monotherapy was associated with a higher CMIR at t2 (P < 0.001) and t3 (P < 0.001) and confirmed in a multivariable model (P < 0.001). DISCUSSION: A third dose of a COVID-19 vaccine boosts CMIR, and the response is sustained in patients with IBD.


INTRODUCTION
Four coronavirus disease 2019 (COVID-19) vaccines, mRNA-1273 (Moderna), BNT162b2 (Pfizer-BioNTech), Ad.26.COV2.S (Johnson & Johnson; no longer available in the United States), and NVX-CoV2373 (Novavax) have been found to be highly effective in the general population (1,2).The original clinical trials evaluating the efficacy of these vaccines excluded patients with inflammatory bowel disease (IBD) and other immunosuppressed populations, who historically may have a lower immune response to non-COVID-19 vaccines (3,4).COVID-19 vaccines have been found to be safe in patients with IBD with similar rates of localized and systemic adverse events as found in the general population (5,6).In addition, COVID-19 vaccination is not associated with IBD flares with low rates reported in prospective studies (6).
Multiple studies have evaluated humoral immunogenicity of COVID-19 vaccines and found that most patients with IBD are able to mount measurable antibody responses after the 2-dose mRNA vaccine series (95%-99%) (7)(8)(9)(10)(11)(12).Robust humoral responses are seen after 3 doses with greater than 99% response rate, which are higher rates of seroconversion than that seen in other immunosuppressed populations such as solid organ transplant recipients (13)(14)(15).Furthermore, patients with IBD have a sustained humoral immune response to COVID-19 vaccines with most patients having measurable antibodies 6 months after a third dose of a COVID-19 vaccine (16).However, those on anti-tumor necrosis factor (TNF) therapy may have lower antibody concentrations after 2 or 3 doses of COVID-19 vaccines, with faster waning of antibodies (12).Those with waning humoral immunity may be more susceptible to breakthrough infections and boosting by additional doses are protective (17).While antibody concentrations wane with time after vaccination, cellular immunity may persist (18).SARS-CoV-2-specific cellular immune responses provide robust memory, mediate recognition of viral variants, and are important for viral clearance (18).In addition, many viral variants of concern may evade humoral immunity, but cellular responses induced by vaccines show strong protection against these variants (19).Thus, evaluating the COVID-19 vaccine-induced cell-mediated immune response (CMIR) in patients with IBD on immune-modifying therapy is of utmost importance.
In contrast to vaccine-induced humoral immunity, fewer studies have evaluated vaccine-induced CMIR in patients with IBD.Initial studies found that most patients develop CMIR after the mRNA primary series (9,20,21).Two studies found that those on anti-TNF therapy have a higher CMIR compared with those on non-TNF biologics or immunomodulators after 2 doses of vaccine (20,21).Two studies that evaluated CMIR immediately after a third dose of COVID-19 vaccine found that those on anti-TNF therapy mounted similar (10) or lower CMIR (22) compared with healthy controls.To further evaluate the CMIR after COVID-19 vaccination, the primary aim of this study was to measure CMIR at 1-2 months and approximately 6 months after a third dose of mRNA COVID-19 vaccine.Secondary aims were to evaluate the impact of a fourth dose; impact of immunemodifying therapies; effects of homologous or heterologous boosting; and prior COVID-19 infection on CMIR.We also aimed to examine the correlation between CMIR and humoral immune responses.We hypothesized that those on anti-TNF therapy would continue to demonstrate higher CMIR compared with patients with IBD on other treatments, as we have previously seen in our HumoRal and CellULar initial and Sustained immunogenicity in patients with IBD (HERCULES) cohort.

Study design
The HERCULES study is a prospective study evaluating humoral responses and CMIR of COVID-19 vaccines in patients with IBD.Participants with IBD were enrolled at the University of Wisconsin-Madison.Methods and inclusion criteria for HERCULES have been previously described (8).In brief, eligibility criteria were

INFLAMMATORY BOWEL DISEASE
Caldera et al a diagnosis of IBD, ages 18-85 years, on stable doses of maintenance therapy at least 2 months, and 3 mRNA vaccine doses.Patients were divided into the following groups: (i) The nonsystemic immunosuppressive group: on mesalamine monotherapy or no therapy for IBD or on vedolizumab monotherapy.Vedolizumab was considered in this group because previous studies have shown that it does not seem to affect vaccine response (4,23).(ii) The immunosuppressed group included the following: thiopurine therapy group: on azathioprine at least 2 mg/kg or 6 MP 1 mg/kg; anti-TNF therapy group: on maintenance therapy infliximab (at least every 8 weeks), golimumab (at least monthly), adalimumab (at least every 2 weeks), or certolizumab (at least monthly); anti-TNF combination therapy group: on anti-TNF therapy as described earlier along with either 15 mg of methotrexate or azathioprine at least 1 mg/kg or 6 MP 0.5 mg/kg; ustekinumab therapy group: on either ustekinumab monotherapy or combination therapy with methotrexate or azathioprine; tofacitinib therapy group: on tofacitinib at least 5 mg PO BID; and corticosteroid therapy group: on any one of the systemic immunosuppressive groups and any dose of corticosteroids.
We previously reported the CMIR at 28-65 days (t 1 ) after the 2 dose primary series (21).In this study, we analyzed CMIR in lymphocytes available at 28-65 days (t 2 ) and approximately 6 months (645 days) (t 3 ) after a third dose of an mRNA COVID-19 vaccine.A small cohort had blood sample available 28-65 days (t 4 ) after a fourth dose of an mRNA vaccine.Samples were analyzed from individuals who were not lost to follow-up from the original cohort, had not changed treatment since the initial 2 dose series, and had received a third or fourth dose of a COVID-19 vaccine.Change in treatment was defined as changing in treatment group (e.g., from anti-TNF therapy to ustekinumab therapy).A change in dose or addition of an immunomodulator to a participant on a biologic was not considered a change in therapy.Participants who changed therapy between t2 and t3 were excluded.
Completion of an mRNA vaccines series was confirmed by review of the Wisconsin Immunization Registry (WIR).A history of COVID-19 infection was defined by a molecular-confirmed diagnosis of SARS-CoV-2 infection through polymerase chain reaction (PCR) confirmed in chart review or nucleocapsid antibody seroconversion after 28-65 days (t 2 ) or approximately 6 months (645 days) (t 3 ) after a third dose.All patients signed informed consent.The study received instituational review board (IRB) approval at the University of Wisconsin-Madison School of Medicine and Public Health.

Wisconsin Immunization Registry
The WIR is a state-wide database maintained by the Department of Health and Family Services of the State of Wisconsin in which vaccine data for each Wisconsin resident is stored.The WIR captures 97% of vaccines administered in the state, and 98.5% of Wisconsin residents have an active WIR record.The WIR does not capture vaccines administered outside the state, and all Wisconsin vaccine providers are required to enter the record of COVID-19 vaccine administration into the registry (24).The WIR has been previously used to evaluate COVID-19 vaccine uptake in patients with IBD (25).

Outcomes
The primary outcomes of this study were CMIR at 1-2 months (28-65 days) (t 2 ) and approximately 6 months (645 days) (t 3 ) after a third dose of an mRNA COVID-19 vaccine.We chose these time points to evaluate CMIR immediately after immunization (t 2 ), the same time point used when we evaluated humoral and CMIR after a 2-dose series and sustained immunogenicity at 6 months because that is the typical time that has been used to evaluate sustained immunogenicity of vaccines in other studies (4,26).
Secondary outcomes included CMIR at 28-65 days (t 4 ) after a fourth dose, evaluating the effects of homologous boosting (receiving same mRNA vaccine as primary series) when compared with heterologous boosting (different from the primary vaccine), effects of immune-modifying therapies on CMIR, particularly anti-TNF monotherapy, anti-TNF combination therapy with an immunomodulator, and history of COVID-19 infection (as defined earlier), and correlations between CMIR and humoral immune responses.

Humoral immune response measurement
Nucleocapsid and spike protein S1 receptor binding domainspecific antibodies were measured in sera at 28-65 days (t 2 ) and approximately 6 months (645 days) (t 3 ) after a third dose of an mRNA COVID-19 vaccine and at 28-65 days (t 4 ) after a fourth dose.Specific antibodies measured in sera using electrochemiluminescent immunoassays performed at LabCorp were antinucleocapsid and anti-spike S1 receptor-binding domainspecific IgG antibodies reported as U/mL and mcg/mL respectively, as previously described at every time point (8).

CMIR measurements
Fluorospot assays were performed to quantitate antigenspecific T cells capable of secreting interferon (IFN)-g with use of the human IFN-g FluorospotPlus kit (Mabtech) from samples at (t 2 ), (t 3 ), and at (t 4 ), as previously described (21).Antigen-specific T cells were enumerated using spike protein peptide pools 1 1 2 (Stemcell Technologies, 1 mg/mL), which are pools of 158 peptides each consisting of 15-mer peptides with 11-amino acid overlaps that span amino acids 1-1,273 of the spike protein.

Data collection
The following clinical variables were collected for participants.At the baseline visit, demographic data including age, gender, race, ethnicity, IBD diagnosis and treatment, and COVID-19 vaccine history.At the subsequent visits, patient interview and medical record review were performed to look for medication changes and COVID-19 vaccine history.

Statistical analysis
CMIR between groups and time points were compared using Mann-Whitney U and Wilcoxon signed rank tests, respectively.Univariable and multivariable linear regression analyses were conducted to assess the association of CMIR at 28-65 days after dose 3 (t 2 ) and at approximately 6 months (645 days) after dose 3 (t 3 ) with age, sex, mRNA vaccine type, homologous or heterologous boosting, previous COVID infection, and IBD therapy (aminosalicylate monotherapy, thiopurine monotherapy, anti-TNF therapy, vedolizumab, ustekinumab, systemic corticosteroids, or tofacitinib).To evaluate the association between CMIR and these variables, we used a multivariable linear regression model with variable that had a P value , 0.05.The residuals had constant variance and were normally distributed.Multivariable linear regression American College of Gastroenterology

INFLAMMATORY BOWEL DISEASE
Cell-Mediated Immune Responses was performed to estimate the relationship among age, anti-TNF therapy, aminosalicylates, and CMIR.All tests were 2-sided with P value , 0.05 considered statistically significant.All analyses were performed using SPSS V27 or R studio 4.2.1.
Cell mediated immune response post third dose (t 2 and t 3 ) CMIR increased at (t 2 ) compared with after the primary 2-dose series (t 1 ) (t 2 median 1,444 responding cells/million [IQR]    2c).No correlation between antibody concentrations and CMIR was found at any time point (Figure 2d and see Supplementary Figure 1, http://links.lww.com/CTG/B89).The number of participants who received heterologous vaccine series was insufficient to test the hypothesis that CMIR  In univariable analysis evaluating CMIR at (t 2 ), increasing age (beta coefficient (b 5 251.16,P 5 0.009), being on aminosalicylate therapy (b 5 21,517.9,P 5 0.046), or being on systemic corticosteroids (b 5 22,805.9,P 5 0.035) was associated with a lower CMIR.Being male (b 5 1,199.4,P 5 0.037) and being on anti-TNF monotherapy were associated with a higher CMIR (b 5 2,627.4,P , 0.001).Those on anti-TNF combination therapy did not have higher or lower CMIR (Table 3).In our multivariable model, we confirmed that being on anti-TNF monotherapy was associated with a higher CMIR (b 5 2,056.9,P , 0.001).Being male was also associated with higher CMIR (b 5 1,152.4,P , 0.036).No other variables were associated higher or lower CMIR in the multivariable model (Table 2).
Univariable and multivariable analyses at approximately 6 month post third dose (t 3 ) In univariable analysis evaluating CMIR at (t 3 ), increasing age (beta coefficient (b 5 246.5, P 5 0.02) or being on aminosalicylate therapy (b 5 21,752.8,P 5 0.024) was associated with a lower CMIR, while being on anti-TNF monotherapy was associated with a higher CMIR (b 5 2,469.9,P , 0.001).Those on anti-TNF combination therapy did not have higher or lower CMIR (Table 3).
In our multivariable model, we confirmed that being on anti-TNF monotherapy was associated with a higher CMIR (b 5 2,089.3,P , 0.001).No other variables were associated with a lower or higher CMIR in the multivariable model (Table 3).

DISCUSSION
We evaluated CMIR at approximately 1-2 months and 6 months after a third mRNA COVID-19 vaccine and found that all patients with IBD regardless of their treatment regimen had measurable responses.We compared CMIR after 2 doses of vaccine (89% in HERCULES study) with CMIR after 3 doses of vaccine, after which 100% of study participants had measurable CMIR, demonstrating benefit of a third dose of mRNA vaccine (21).This has not been evaluated in previous studies in the population with IBD.In addition, our study did not show waning of CMIR at 6 months post 3-dose series.Those on anti-TNF monotherapy had higher CMIR at both time points after the third dose compared with those not on anti-TNF therapy or those on anti-TNF combination therapy with an immunomodulator, which is similar to what we saw after 2 doses of mRNA vaccine (21).Furthermore, those on anti-TNF therapy did not have significant waning of CMIR at 6 months.These findings can reassure patients and providers that even among patients with IBD on immune-modifying therapies such as anti-TNF therapy, immune responses to the vaccine are robust and durable.This is also, to our knowledge, the only study to examine CMIR after a fourth dose of vaccine.Although this subset of patients was smaller, we failed to show a boost in CMIR with a fourth dose of vaccine.Two previous studies have evaluated CMIR after a third dose of the COVID-19 vaccine (10,22).The COVID-19 vaccine-induced antibody and T-cell responses in immunosuppressed

INFLAMMATORY BOWEL DISEASE
Caldera et al patients with IBD after the third vaccine dose (VIP) study examined antigen-specific T-cell responses 28-49 days after the third vaccine dose (10).Most of the patients with IBD (191/202; 95%) mounted a CMIR.CMIR was similar among COVID-19 infection-naive patients with IBD on thiopurines, infliximab, thiopurine plus infliximab, ustekinumab, or vedolizumab, compared with healthy controls.Infection-naive patients receiving tofacitinib had lower CMIR compared with healthy controls.In those with previous SARS-CoV-2 infection, CMIR against SARS-CoV-2 nucleocapsid mapped epitope pool was significantly lower in patients treated with ustekinumab than in healthy controls.
The systemic and T cell-associated responses to SARS-CoV-2 immunization in gut inflammation (STAR SIGN) study evaluated humoral responses and CMIR 2-16 weeks after receiving the third mRNA vaccine dose among 139 patients on biologics and 110 healthy controls (22).In contrast to the VIP and this study, a significantly higher proportion of those on anti-TNF therapy (15/73, 20.5%) failed to mount an adequate CMIR to the third dose of vaccine, compared with healthy controls (2/100, 2%), as measured by an IFN-g release assay.There was no significant difference between CMIR in anti-TNF-treated and non-anti-TNF-treated patients with IBD.The VIP study and our study each used ELIspot or Fluorospot to measure CMIR, which has been shown to detect CMIR in immunosuppressed patients more often (27).In addition, these assays measure different aspects of T-cell response.ELISpot/ Fluorospot quantifies the number of T cells that respond, whereas the IFN-g release assay measures the amount of cytokine produced by T cells in response to antigen stimulation.
While the STAR SIGN study found decreased CMIR among those on anti-TNF therapy, the Impact of Biologic and Immunomodulatory Therapy on SARS-CoV-2 Infection and Immunity in Patients with IBD (CLARITY-IBD), Coronavirus Risk Associations and Longitudinal Evaluation-IBD, and HERCULES study groups found that patients with IBD on anti-TNF therapy have similar or augmented CMIR following vaccination with the primary series.The CLARITY-IBD study examined humoral and CMIR to the BNT162b2 and ChAdOx1 nCOV-19 vaccines among patients with IBD treated with infliximab or vedolizumab (9).Approximately one-fifth of patients failed to mount a CMIR in both groups after the primary series.The Coronavirus Risk Associations and Longitudinal Evaluation-IBD study examined CMIR among 303 patients with IBD and found all participants had a T-cell clonal response.Spike-specific T-cell response reduced significantly with age.There were no significant differences in response among those on ustekinumab, vedolizumab, tofacitinib, or steroids.CMIR was augmented among those on anti-TNF therapy after adjusting for age and sex.The HERCULES study found that most of the patients with IBD developed a CMIR after receiving 2 doses of mRNA COVID-19 vaccines.Those on anti-TNF therapy had an augmented CMIR (21).This finding was confirmed in this study where CMIR was augmented in those on anti-TNF monotherapy at both time points after a third dose of COVID-19 vaccine.Those on anti-TNF combination therapy with an immunomodulator did not have an augmented response showing that an augmented CMIR is potentially linked to anti-TNF monotherapy alone.The reason for the augmented response on anti-TNF therapy is unknown but is postulated due to the effects of TNF-a on vaccine-induced humoral and cellular responses (28).TNF alpha supports B-cell maturation and downregulates T-cell expansion.The lack of downregulation would result in an augmented CMIR, and the blunted B-cell maturation would cause lower antibodies in patients on anti-TNF therapy (29,30).These findings would suggest that patient on anti-TNF monotherapy may be at lower risk of severe disease from The CMIR is responsible for immunologic memory and recognition of viral variants and clearance, generation, and maintenance of antibody responses and is an important component of COVID-19 vaccine efficacy in addition to humoral immunity.The CMIR can protect against viral variants, which may evade humoral immunity (18).The VIP study, STAR SIGN study, and current study demonstrate discordant CMIR and humoral immune responses, suggesting that there is not a strong correlation between antibody and CMIR among patients with IBD (21).Previous studies in healthy controls have also demonstrated discordance in CMIR and antibody response (19).By contrast, the CLARITY-IBD study demonstrated a modest positive correlation between CMIR and antibody concentration among those receiving an mRNA vaccine (9).The discordance in these results demonstrate that both humoral and cell mediated immunity must be evaluated to determine the impact of different immune-modifying therapies on COVID-19 vaccine immunity.Each component only accounts for a part of vaccine-induced immunity.
Both mRNA-1273 (Moderna) and BNT162b2 (Pfizer-BioNTech) elicit a T-helper type1 immune response producing IFN-g shown in other studies in which only minimal T helper type 2 response was measured (31,32).The presence of CMIR has been shown to protect individuals from infection, even in the absence of neutralizing antibodies (33).SARS-CoV-2-specific cell-mediated immune responses seem to wane more slowly compared with antibody concentrations.In addition, these CMIR recognize conserved viral antigens with less than 15% of T cells recognizing viral mutations sites (19,34).Therefore, CMIR could provide protection from viral variants.All patients with IBD in this study mounted CMIR, which persisted over the vaccine dosing interval.Our study showed slower waning of CMIR compared with antibody concentrations.Clinicians can consider that other benefits of CMIR, such as recognition of at least some viral variants, which have been demonstrated in other studies may offer some additional vaccine protection in patients with IBD.
Our study has several strengths.We evaluated sustained CMIR over multiple time points among patients on stable treatment regimens for at least 2 months.This allowed us to assess the effect of different immunosuppressive medications on vaccine response.We assessed CMIR 6 months after the third dose of vaccine and were able to include a small cohort evaluating immunogenicity after a fourth dose of vaccine.Our study is limited by the lack of a reference healthy control population, inability to evaluate the impact of disease activity on vaccine response, and small representation of certain treatment regimens (tofacitinib, thiopurine monotherapy, and systemic steroids).We were also limited in being unable to evaluate the real-world effectiveness of vaccine in preventing breakthrough infections because the goal of HERCULES was evaluating humoral and cell-mediated immunogenicity of COVID-19 vaccine.Patients in the study were not followed up or asked about breakthrough infections during the study period.Thus, we were unable to evaluate the correlation of breakthrough infections with CMIR or IBD therapy.Evaluating breakthrough infections is a complex issue because the goal of vaccination is preventing severe disease that results in hospitalization.However, some variants are highly contagious and result in higher rates of infections.Furthermore, the impact of IBD therapies on breakthrough infections has been variable in multiple studies (35,36).Future studies should evaluate CMIR and humoral immune response in a larger population of patients in different treatment arms.
In summary, we found that patients with IBD should receive a third dose of an mRNA vaccine to maximize the CMIR and this response is sustained for approximately 6 months.We also confirmed our previous finding that patients on anti-TNF therapy have an augmented CMIR to mRNA COVID-19 vaccines.

Figure 1 .
Figure 1.Flowchart of study participants in HERCULES study.Flowchart of participants in the HERCULES study at the different time points t 1 -t 4 .Blood draw timings: (a) (t 1 ) 28-35 days after 2 dose mRNA vaccine series; (b) (t 2 ) 28-65 days after a third dose; (c) (t 3 ) approximately 6 months (645 days) after a third dose of an mRNA COVID-19 vaccine; (d) (t 4 ) 28-65 days after a fourth dose of an mRNA vaccine.HERCULES, HumoRal and CellULar initial and Sustained immunogenicity in patients with inflammatory bowel disease.

Table 3 .
Multivariable models evaluating cell-mediated immune response approximately 6 months (645 days) after third mRNA dose (t 3 ) (n 5 167) CoV2 given the augmented CMIR because T cells can help prevent onset of severe disease for most cases.