Determinants of Antibody Response to SARS-CoV-2 Vaccines in Liver Transplant Recipients: The Role of Immunosuppression Reduction

Liver transplant recipients on chronic immunosuppression show an attenuated antibody response after SARS-CoV-2 vaccination. Adjusting immunosuppressants during vaccination remains debated. We enrolled 380 liver transplant recipients receiving 2 doses of a protein subunit, mRNA, or a vector vaccine. The patients were informed to temporarily suspend immunosuppression for 2 weeks for both vaccination doses. We measured anti-live-SARS-CoV-2 spike neutralizing antibody levels at 1–2 months after the second vaccination; 83.9% of patients had humoral responses (SARS-CoV-2 NT50 ≥ 9.62 IU/mL) to 2 doses of vaccines. The mRNA (86.7%) and protein subunit vaccines (85%) yielded higher response rates than the vector vaccines (40.9%). Immunosuppression suspension during the two vaccinations yielded a higher response rate (91.5% vs. 57.7%). Only eight patients (2.1%) experienced transaminase level elevation of thrice the normal value (>110 IU/L) after the second vaccination. Most recovered spontaneously after resuming immunosuppression. Multivariate analysis revealed ABO incompatibility, white blood cell count <4000, lymphocyte count <20%, tacrolimus trough level >6.5 ng/mL, and no immunosuppression adjustment as independent risk factors to nonresponse. The mRNA and protein subunit vaccines yielded a higher response rate. Immunosuppression suspension for 2 weeks enhanced the antibody response. ABO incompatibility, leukopenia, lymphopenia, a high tacrolimus trough level, and no immunosuppression adjustment are associated with nonresponse.

The severity and mortality of SARS-CoV-2 infection is higher among transplant recipients on chronic immunosuppression [4,5]. Although SARS-CoV-2 vaccination is generally recommended [6], the efficacy of these vaccines among immunocompromised patients is unknown given their exclusion from all vaccine trials. Studies among transplant recipients have demonstrated a weaker humoral response [7][8][9][10]; however, the efficacy of

Serological Testing
We isolated the plasma from the obtained blood samples within 4 h of collection. Anti-live-SARS-CoV-2 spike neutralizing antibody levels were measured as described previously [16]. We conducted the live viral culture procedures in a biosafety level 3 facility regulated by the Taiwan Center for Disease Control. Briefly, Vero E6 cells were seeded in 96-well plates and incubated overnight. Tested sera were diluted in modified Eagle's medium (Thermo Fisher Scientific, Waltham, MA, USA) at an initial dilution factor of 20, and then further 2-fold serial dilutions were performed to a final dilution of 1:5120. SARS-CoV-2 viruses at 100 TCID50/50 µL (Wuhan wildtype, hCoV-19/Taiwan/CGMH-CGU-01/2020, GenBank accession MT192759) were mixed with sera in an equal volume and incubated at 37 • C for 1 h before adding to Vero E6 cells. The mixture was incubated at 37 • C for 5 days and then fixed using 4% formalin for 1 h and stained with 0.1% crystal violet for visualization.
A quantitative assay of the SARS-CoV-2 50% neutralization titer (NT50) was determined based on the median tissue culture infectious dose (the dilution of a virus required to infect 50% of given cell culture) and converted into an international unit (IU/mL, WHO Standardized) [17]. NT50 < 9.62 IU/mL and ≥9.62 IU/mL was considered a negative and a positive humoral response, respectively.
The immunosuppression regimen was not adjusted if it consisted of tacrolimus alone. In some patients, transient immunosuppression adjustment was performed if tacrolimus was combined with MMF/EV. The patients were asked to pause MMF or EV for 2 weeks from the date of vaccination. ABOi LDLTs underwent a desensitization protocol that consisted of rituximab and bortezomib in urgent acute liver failure cases. In the routine protocol, if the IgG and IgM antibody titers were ≤1:64, the patient underwent ABOi LDLT directly [18,19]. Rituximab (375 mg/m 2 ) was administered after surgery on postoperative day (POD) 1. If IgG and IgM anti-ABO titers were >1:64, rituximab (375 mg/m 2 ) was administered intravenously 3 weeks before LT. If the pre-transplant isoagglutinin titers remained at >1:64, plasmapheresis or plasma exchange was performed and repeated as needed to decrease the antibodies to ≤1:64 at surgery. Additional rituximab (187.5 mg/m 2 ) was administered on POD 1. In the urgent protocol for acute liver failure recipients, bortezomib (3.5 mg) was administered, and if the anti-blood type isoagglutinin titers were ≤1:64, LT was performed immediately. If the titers were >1:64, plasma exchange was arranged. Rituximab (375 mg/m 2 ) was administered on POD 1 [20].

Clinical Follow-Up
All patients were regularly followed up at the outpatient clinic every 1-2 months. Liver function, renal function, and tacrolimus trough levels were measured at every visit.

Safety
Participants were asked to report the adverse effects after each vaccination. Of the recorded adverse events, the systemic ones included dizziness, fever, chills, headache, fatigue, myalgia, dyspnea, and diarrhea; the local ones included pain and pruritus.

Statistical Analysis
We collected data including age, sex, timing and type of vaccination, underlying diseases, laboratory tests, and immunosuppression therapy before each vaccination.
Continuous data are expressed as medians and interquartile ranges. Differences in continuous variables were assessed using the Mann-Whitney U or Kruskal-Wallis test. Categorical variables were expressed as proportions and analyzed using the chi-square or Fisher exact test. We used a binary logistic regression model to identify independent factors predicting the serologic response after vaccination. Variables with p < 0.05 were entered in the multivariate analysis. Statistical analysis was performed with SPSS 26 statistical software (IBM-SPSS Statistics; IBM Corporation, Chicago, IL, USA), and GraphPad Prism 9.0 software (GraphPad, San Diego, CA, USA) was used for illustration. A p < 0.05 was considered statistically significant.
No donor livers were procured from executed prisoners. The institutional review board of Chang Gung Memorial Hospital (IRB No. 202102089B0) approved the study. All patients provided written informed consent.

Immunosuppression Regimens
At vaccination, 84 patients (22.5%) were receiving calcineurin inhibitor (CNI)-based monotherapy (mostly tacrolimus, 99.4%) and received no immunosuppression adjustment during the vaccinations; an additional 32 patients (8.6%) also had no immunosuppression adjustment. Of the six patients receiving triple immunosuppression (tacrolimus, MMF, and steroids), two stopped MMF during the second vaccination. Of the patients who had immunosuppression adjustments, 31 patients (8.3%) stopped MMF/EV at the second vaccination, and 227 patients (60.7%) temporarily suspended MMF/EV during both vaccinations.

Humoral Responses
Overall, 314 of 374 patients (83.9%) had humoral responses (SARS-CoV-2 NT 50 ≥ 9.62 IU/mL) to both doses of vaccines (Table 1). Among the nonresponsive patients, the proportions of those with alcoholic liver cirrhosis, ABOi, leukopenia, lymphopenia, hemodialysis or immunosuppression adjustment were greater, as were those with high tacrolimus levels. The age, sex, LT-to-vaccination interval, interval between the two vaccinations, history of cancer and renal function (assessed using EGFR) did not differ between the two groups.

Humoral Response by Type of Vaccination
The mRNA vaccines yielded a significantly higher median anti-SARS-CoV-2 antibody titer than vector or protein subunit vaccines ( Figure 2). Both the MRNA and the KT groups showed higher positive response rates (86.7% and 85%, respectively) than the AZ group (40.9%). Patient characteristics differed across the three groups ( Table 3). The MRNA group had a greater proportion of older patients. The AZ and KT groups had higher proportions of patients with alcoholic liver cirrhosis. Patients with ABOi were more common in the AZ group. The proportion of patients with a shorter vaccination interval was greater in the KT group. The MRNA and KT groups had a greater proportion of patients with immunosuppression adjustment.

Effect of Immunosuppression Adjustment
Temporary suspension of MMF/EV during the two vaccinations yielded the highest serologic response (91.5% vs. 57.7% in the group without modification; Figure 3). Of note, the antibody titers and response rates remained high (83.1%) among patients with CNI monotherapy, even without immunosuppression suspension. We observed no significant differences between MMF or EV (Supplementary Figure S1). Regarding graft safety after transient immunosuppression suspension during vaccination, serum aminotransferase levels did not differ significantly (Supplementary Figure S2). Only eight patients had a 3X elevation in transaminase levels than the normal (>110 IU/L) during follow-up after the second vaccination. Six patients recovered spontaneously after resuming immunosuppression: one patient in the no-adjustment group, three patients in the CNI monotherapy group, and two patients in the two rounds of immunosuppression adjustment group. One patient in the two rounds of immunosuppression adjustment group experienced a simultaneous biliary tract infection, which resolved with antibiotic treatment. Only one patient on the temporary 2-week MMF/EV suspension experienced acute rejection, but recovered uneventfully after pulse steroid therapy.

ABOi Patients
In our series, 12 patients underwent ABOi LDLT. One patient underwent deceased donor LT but was included in this group because of initial preparation for ABOi LDLT, including rituximab/bortezomib for B cell and plasma cell depletion. The median time from transplantation to vaccination in ABOi patients was 40 months (IQR, 24-81.5 months). Total bilirubin and tacrolimus trough levels were significantly higher among ABOi patients (Table 4).
Data on the percentage of CD19+ cells in the total peripheral blood lymphocyte population (%CD19) were available for only three patients and ranged between 6.7% and 14.5%. In general, ABOi patients had an impaired serological response. The median antibody titers were significantly lower, and the overall response rates were 53.8% (vs. 85% in compatible patients; Figure 4). Among patients who developed a humoral response, patients with CNI monotherapy without immunosuppression adjustment showed a 50% response rate, whereas those with two rounds of MMF/EV suspension had a response rate of 71.4%. Regarding the type of vaccines, most patients received mRNA vaccines (61.5%), and the positive immune response rate in this group was 75%.

ABOi Patients
In our series, 12 patients underwent ABOi LDLT. One patient underwent deceased donor LT but was included in this group because of initial preparation for ABOi LDLT, including rituximab/bortezomib for B cell and plasma cell depletion. The median time from transplantation to vaccination in ABOi patients was 40 months (IQR, 24-81.5 months). Total bilirubin and tacrolimus trough levels were significantly higher among ABOi patients (Table 4).  Data on the percentage of CD19+ cells in the total peripheral blood lymphocyte population (%CD19) were available for only three patients and ranged between 6.7% and 14.5%. In general, ABOi patients had an impaired serological response. The median antibody titers were significantly lower, and the overall response rates were 53.8% (vs. 85% in compatible patients; Figure 4). Among patients who developed a humoral response, patients with CNI monotherapy without immunosuppression adjustment showed a 50% response rate, whereas those with two rounds of MMF/EV suspension had a response rate of 71.4%. Regarding the type of vaccines, most patients received mRNA vaccines (61.5%), and the positive immune response rate in this group was 75%.

Vaccination Safety
No patient who received 2 doses of any vaccine experienced any serious adverse events (Supplementary Table S1).

Discussion
Studies have shown an attenuated immune response after SARS-CoV-2 vaccination among solid organ transplant recipients [8,21,22]. However, vaccination in this population is still recommended. Understanding the factors affecting an immune response after vaccination will provide further insights for clinical practice during booster vaccinations. In this regard, this study prospectively recruited patients without prior infection undergoing LT and evaluated mRNA, vector-primed, and protein subunit SARS-CoV-2 vaccines.
The mRNA and protein subunit vaccines yielded a superior neutralizing antibody response, with response rates >85%. After 2 doses of vector vaccines, only 40.9% showed a positive response. This finding is consistent with that of previous studies showing only a 34% response rate with pure vector vaccines [21], which increased to 70% after a heterologous regimen with booster mRNA vaccines [22]. Inactivated and protein subunit vaccines remain understudied. One study reported a response rate of 17% after administering an inactivated vaccine [23]. Our study included 20 patients receiving 2 doses of a protein subunit vaccine with similar response rates to mRNA vaccines.
Similar to some other studies, our study identified ABOi leukopenia, lymphopenia, a high tacrolimus trough level, and no immunosuppression adjustment as risk factors for a failure to develop humoral response. Other factors identified were advanced age, vaccination during the first year after transplantation, hypogammaglobulinemia, a lower estimated glomerular filtration rate, the use of high-dose steroids before vaccination, and maintenance with a triple immunosuppressive regimen [8,22,24]. Therapeutic approaches that may increase the immune response to vaccination include heterologous vaccination [22], supplementation with micronutrients [25], vitamin D [26], or immunoregulatory molecules like IL-31/IL-33 [27]. However, in the short term, the most important modifiable risk factor is the immunosuppression adjustment during vaccination. In our study, we asked some patients to stop MMF or EV for 2 weeks during the two vaccinations. The response rate among patients with tacrolimus monotherapy was 81%, whereas it was 91.2% among those who suspended MMF/EV twice; this is consistent with previous reports showing decreased humoral responses with antimetabolite treatment [12,24]. MMF or mycophenolic acid therapy inhibits B-cell proliferation and plasma cell formation, which explains the reduced antibody production [28]. With EV, however, the results have been inconsistent, with studies reporting increased [29,30], attenuated [31], and unaffected [15] humoral responses after vaccination. In our series, we found no differences between suspending MMF or EV or between suspending immunosuppression once or twice during both vaccinations. Furthermore, as the liver is an immunologically tolerant organ after transplantation, the temporary suspension of MMF/EV was safe, with only 8 of 213 patients showing abnormal liver function. Most patients recovered after resuming the immunosuppression regimen after vaccination, and only one patient was admitted for steroid recycle treatment.
Regele et al. stopped mycophenolic acid or azathioprine in 18 renal transplant patients 1 week before until 1 week after the 4th/5th dose of the BNT162b2 vaccine. Although renal function remained stable and there were no new donor-specific antibodies, there were no differences in the humoral response rate vs. control or the absolute antibody levels [15]. In another cohort of 29 renal transplant patients who had not mounted a humoral response to three previous vaccine doses, Schrezenmeier et al. reported that a temporary antimetabolite treatment hold for 5 weeks increased neutralizing antibodies, vaccine-specific B cells and the ex vivo activation of spike-specific T cells. Seroconversion was already observed on day 7 in 34.4% of patients without antimetabolite treatment, as compared with only 12% of patients in the control group [14]. Data on liver transplantations are scarce, but in general, there is no consensus on immunosuppression modulation at time of vaccination in transplant patients. The effectiveness of MMF or azathioprine dose reduction before and after receiving SARS-CoV-2 vaccination is being evaluated in prospective randomized trials [32,33], which may provide further evidence to guide clinical practice.
Patients undergoing ABOi LT showed a generally inferior immunogenicity. B cell and plasma cell depletion is the primary aim of the desensitization protocols for ABOi cases, with expected recovery after 9-12 months of treatment [34]. Of interest, the only patient who underwent LT within 12 months before vaccination showed a positive immune response, highlighting the predominant effect of the increased immunosuppression and higher tacrolimus trough levels in these patients. A previous study also reported that tacrolimus trough levels in patients with a negative response was 6.6 ± 2.2 ng/mL compared with 5.4 ± 2.0 ng/mL in patients with a positive antibody response [35]. When we compared the levels of neutralizing antibodies among patients who had temporarily suspended MMF/EV, antibody levels were significantly lower when the tacrolimus trough level was >6.5 ng/mL. Therefore, when the effects of B cell depletion, MMF, or EV are minimized, adjusting the CNI trough levels is critical to achieving a humoral response.
This study has several limitations. First, we did not include a control group. However, substantial data from large clinical trials are available regarding the efficacy of vaccines in healthy individuals. Second, we did not compare the humoral responses before and during vaccinations. Although the prevalence of SARS-CoV-2 infection during the study period was very low in Taiwan, we may have missed patients with asymptomatic SARS-CoV-2 infection that likely affected the serological response. Third, we did not evaluate cellular immune responses, which is an important determinant of a durable immune response after vaccination. Fourth, we did not evaluate the longitudinal immune response, after 3-5 booster doses, or the rate of SARS-CoV-2 infection post vaccination. Sterilizing immunity is the ideal scenario, but we were unable to make any conclusions regarding the antibody response efficacy against infection or severe disease. Finally, sample sizes were very small in some groups (KT and ABOi LTs), which limited the statistical power for subgroup analysis.
In conclusion, mRNA and protein subunit vaccines yield a higher rate of neutralizing antibodies, particularly when the immunosuppression therapy is adjusted and MMF or EV is temporarily suspended. Factors associated with vaccine nonresponse included ABOi, leukopenia, lymphopenia, a high tacrolimus trough level, and no immunosuppression adjustment. These factors may provide a platform for future strategies to enhance immune responses in additional booster doses. As a proof of concept, our results highlight the effectiveness and safety of antimetabolite suspension or immunosuppression minimization during SARS-CoV-2 vaccinations.
Supplementary Materials: The following supporting information can be downloaded at https:// www.mdpi.com/article/10.3390/vaccines10111827/s1: Table S1: Comparison of side effects after first and second SARS-CoV-2 vaccination in all patients; Figure S1: Comparison of the humoral responses (SARS-CoV-2 NT50 antibodies) under MMF or EV in patients with immunosuppression adjustment. (A) Adjustment once. (B) Adjustment twice. Data are expressed as median and interquartile range. n.s., not significant (Mann-Whitney U test). Abbreviations: MMF, mycophenolate mofetil; EV, everolimus; Figure S2: Serum aminotransferase levels of before (1) and after (2)   Informed Consent Statement: Written informed consent has been obtained from the patient(s) to publish this paper.