Humoral and cellular immune responses to COVID-19 mRNA vaccines in immunosuppressed liver transplant … – Nature.com

Anti-RBD IgG titers and plasma neutralizing activity induced by COVID-19 mRNA vaccination in HDs and LTRs

We enrolled 44 HDs and 54 LTRs to comprehensively evaluate mRNA vaccine-induced antibodies and cellular immune responses (Table1). The mRNA vaccines, Pfizer BNT162b2 or Moderna mRNA-1273 were investigated. Blood samples were obtained at five-time points: before vaccination, 1, 3, and 6 months after the second vaccination, and 1 month after the third vaccination (Fig.1a).

a Schematic overview of the cohort. b Anti-RBD IgG endpoint titers in HDs (black) and LTRs (red) (sample size, pre: 25 vs 12, 1m after 2nd: 25 vs 54, 3m after 2nd: 24 vs 53, 6m after 2nd: 44 vs 54, 1m after 3rd: 44 vs 51). c Multivariable logistic regression model (OR and 95% CI) for predictors of weak and strong responders (lower and higher than median antibody titer in HDs at 1 month after third vaccination, respectively). d Anti-RBD antibody titers in HDs (black), LTRs taking only a calcineurin inhibitor (CNI group, red) and LTRs taking CNI and other medications (CNI+other drug(s) group, blue) (sample size, pre: 25 vs 1 vs 11, 1m after 2nd: 25 vs 20 vs 27, 3m after 2nd: 25 vs 23 vs 30, 6m after 2nd: 44 vs 23 vs 31, 1m after 3rd: 44 vs 21 vs 29). e Anti-RBD IgG endpoint titers in HDs 1 month after 2nd vaccination (black) and in CNI+other drug(s) group 1 month after 3rd vaccination (blue) (sample size, 25 vs 29). f Fold-induction in anti-RBD IgG endpoint titers after third vaccination (HDs: black, CNI: red, CNI+other drug(s): blue). Pie charts represent the proportion of individuals with fold-induction > 1, and gray slice shows frequency of negative responders. (sample size, 44 vs 21 vs 29). g pVNT50 against SARS-CoV-2 Wuhan-1 (HDs: black, CNI: red, CNI+other drug(s): blue). h pVNT50 in HDs 1 month after 2nd vaccination (black) and in CNI+other drug(s) group 1 month after 3rd vaccination (blue) (sample size, 25 vs 29). i Fold-induction in pVNT50 after third vaccination. Pie charts represent the proportion of individuals with fold-induction > 1, and gray slice shows frequency of negative responders (HDs: black, CNI: red, CNI+other drug(s): blue) (sample size, 44 vs 21 vs 29). P values (two-sided) were calculated using the MannWhitney U-test. All experiments were performed once. Error bars indicate the interquartile range.

All LTRs were administered CNIs, such as tacrolimus or cyclosporine. Some LTRs took additional medications, such as the metabolic antagonist MMF, a steroid, or the mTOR inhibitor everolimus. Specifically, 23, 12, 2, 11, 5, and 1 LTRs had taken only a CNI; CNI and MMF; CNI and everolimus; CNI and a steroid; CNI, MMF, and a steroid; and CNI, everolimus, and a steroid, respectively. Seven LTRs received entecavir, a drug used to treat hepatitis B, and immunosuppressive therapy.

Anti-RBD antibody titers in LTRs were significantly lower than those in HDs at all time points (Fig.1b) (p<0.0001 at 1 and 3 months, p=0.0005 at 6 months after the second vaccination, p=0.0002 after the third vaccination). Anti-RBD antibody titers in all HDs exceeded the WHO standard (dashed line, 1000U/mL); however, 53.2% of LTRs had anti-RBD antibody titers below the WHO standard at 1 month after the second vaccination. However, anti-RBD antibody titers in 92.2% of the LTRs after the third vaccination exceeded the WHO standard, suggesting that effective immune responses can be achieved in immunosuppressed LTRs by the third vaccination.

Interestingly, the variability in antibody levels among LTRs was wide compared with that in HDs. Therefore, we aimed to identify the factors that affect the variability in antibody production in LTRs. LTRs that obtained anti-RBD antibody levels higher and lower than the median value of antibody titers in HDs after the third vaccination were categorized as strong and weak responders, respectively. We conducted a multiple logistic regression analysis with clinical parameters (Fig.1c), suggesting that taking multiple drugs decreased antibody levels (p=0.0048, OR=0.0285).

We regrouped LTRs for comparison between LTRs taking only a CNI and taking a CNI and more drugs (CNI+other drug(s)) (Fig.1d). There was no difference in the antibody titers between the CNI group and HDs after the third vaccination. Contrarily, antibody titers were significantly lower in the CNI+other drug(s) group than in the HDs and the CNI group (p<0.0001 among HDs vs. CNI+other drug(s), p<0.0001 among CNI vs. CNI+other drug(s)). However, the anti-RBD antibody titers after the third vaccination in the CNI+other drug(s) group were the same as those in HDs 1 month after the second vaccination (Fig.1e; p=0.3255). After the second vaccination, anti-RBD antibodies in plasma were induced in 49 of 54 LTRs. The 5 LTRs in whom anti-RBD antibodies were not induced after the second vaccination all showed induction of the antibodies after the third vaccination. However, there was one individual who, despite having a positive plasma anti-RBD antibody titer after the second dose, did not benefit from the third booster dose and tested negative. This individual was taking three medications, namely CNI, MMF, and steroids (5mg/day), and had a low anti-RBD antibody titer even after the second vaccination.

Additionally, LTRs were regrouped based on clinical information apart from medication (Supplementary Fig.1). Antibody titers were considerably lower in deceased donor liver transplant (DDLT) than in living-donor liver transplant (LDLT) (Supplementary Fig.1a). Furthermore, antibody titers in LTRs less than 12 years after transplantation were lower (Supplementary Fig.1b). LTRs who experienced rejection reactions after transplantation also exhibited lower antibody titers than those who did not (Supplementary Fig.1c). LTRs who have taken MMF also exhibited lower antibody titers than those who have not (Supplementary Fig.1d). These factors are related to the regimen of immunosuppressive drugs, and the multivariate analysis suggested that the number of drugs has the most significant impact. Noteworthily, antibody titers of 89.7% in the CNI+other drug(s) group were increased by the third vaccination, and the fold induction of antibody titers in the CNI+other drug(s) group was similar to that in HDs (Fig.1f; p=0.7666).

Next, we measured the changes in the neutralizing activity of plasma from HDs and LTRs (Fig.1g). Neutralizing activity in most of the CNI+other drug(s) was below the detection limit after the second vaccination, and was significantly lower than that in HDs after the third vaccination (p=0.0001). Contrarily, the neutralizing activity in the CNI+other drug(s) group after the third vaccination was similar to that in HDs one month after the second vaccination (Fig.1h; p=0.2985). Furthermore, although the fold-induction of neutralizing activity in CNI+other drug(s) by the third vaccination was significantly lower than that of HDs, 82.8% of the CNI+other drug(s) group got a booster effect (Fig.1i) (p=0.0006 among HDs vs. CNI+other drug(s)). These results suggest that the third doses of mRNA vaccine are worthwhile for the induction of neutralizing activity in LTRs, but may not be sufficient compared to HDs.

Generally, immunosuppressive drugs, including CNIs, contribute to the suppression of T-cell responses. To investigate whether the reduction in antibody titers in LTRs is affected by changes in CD4 helper T-cell function, we performed flow cytometry analysis to evaluate the CD4+ T-cell responses. The frequency of total SARS-CoV-2 spike-specific CD4+ Tcells was measured using CD154 as an activation marker (Supplementary Fig.2a). The frequency of spike-specific CD4+ Tcells in CNI+other drug(s) at 1, 3, and 6 months after the second vaccination was significantly lower compared to HDs (Fig.2a; p=0.0117, p=0.0208, and p=0.0047 at 1, 3, and 6 months after the second vaccination, respectively). There was no significant difference between HDs and the CNI group at 1 month (p>0.9999), 3 months (p=0.6506), and 6 months (p=0.1379) after the second vaccination. Moreover, there were significant differences between the CNI and CNI+other drug(s) groups 3 months (p=0.024), and 6 months (p=0.0051) after the second vaccination (Fig.2a). However, there is no significant difference among HDs, the CNI group, and the CNI+other drug(s) group after the third vaccination. Regardless of HDs or LTRs, spike-specific CD4+ Tcells decreased over time after the second mRNA vaccination (Supplementary Fig.2b).

Frequency of spike-specific CD154+ (a), Th1 (b), and Th2 (c) CD4+ Tcells in total memory Tcells from HDs (black), CNI group (red), and CNI+other drug(s) group (blue). df Fold-induction of spike-specific CD154+, Th1, and Th2 CD4+ Tcells by the third vaccination. Pie charts represent the proportion of individuals with fold-induction higher than 1, and gray slice shows frequency of negative responders. (HDs: black, CNI: red, CNI+other drug(s): blue). g The ratio of spike-specific Th1 to Th2 CD4+ Tcells (HDs: black, CNI: red, CNI+other drug(s): blue). P values (two-sided) in (a) to (g) were calculated using the Mann-Whitney U-test. h Correlation matrix of antibody and CD4+ T-cell responses in HDs and LTRs. Shades of blue represent positive correlations approaching 1, while shades of red denote negative correlations nearing -1. P values (two-sided) were calculated using the Spearmans rank test. Sample size, 1m after 2nd: 23 vs 17 vs 26, 3m after 2nd: 22 vs 16 vs 22, 6m after 2nd: 43 vs 21 vs 29, 1m after 3rd: 43 vs 20 vs 29. All experiments were performed once. Error bars indicate the interquartile range.

Next, we measured the cytokine profiles of the total spike-specific CD4+ Tcells (Fig.2b, c, Supplementary Fig.2a). The frequency of Th1 cells in CNI+other drug(s) after the second vaccination was significantly lower compared to HDs (Fig.2b). On the contrary, the frequency of Th2 cells was higher in the CNI group than in HDs (Fig.2c). The frequency of total CD154+ spike-specific CD4+ Tcells and Th1 cells increased by the third mRNA vaccination in HDs and LTRs, and there was no significant difference between HDs and LTRs after the third vaccination (Fig.2a, b).

We next examined the effect of the third booster on memory CD4+ T cell responses by calculating the fold-induction of CD154+, Th1, and Th2 cell frequencies. We observed a boost effect in ~75% of individuals for CD154+ and Th1 cells in all groups, and in ~50% of individuals for Th2 cells (Fig.2df). Furthermore, Th1/Th2 ratio in LTRs was significantly lower compared to HDs (Fig.2g), suggesting that LTRs are more susceptible to the induction of Th2-biased CD4+ T-cell responses.

We next evaluated the correlation between CD4+ T-cell and antibody responses. One month after the second vaccination, the frequency of CD154+CD4+ T and Th1 cells was positively correlated with anti-RBD antibody titers in HDs and LTRs (Fig.2h). Moreover, CD4+ T-cell frequency before the third vaccination positively correlated with antibody titers after the third vaccination (HDs: r=0.299, p=0.049 for CD154+CD4+ Tcells vs. anti-RBD IgG; LTRs: r=0.483, p=0.0004 for CD154+CD4+ Tcells vs. anti-RBD IgG; r=0.433, p=0.0019 for Th1 CD4 Tcells vs. anti-RBD IgG). These results suggest that long-term CD4+ T-cell responses after the second vaccination contribute to the booster effect on antibody levels after the third vaccination.

In addition to antibodies and CD4+ Tcells, CD8+ T-cell responses also contribute to defense against SARS-CoV-2 infection22,23. However, COVID-19 mRNA vaccines reportedly have a lower ability to induce CD8+ Tcells than CD4+ Tcells24. Moreover, few reports demonstrate CD8+ T-cell responses to mRNA vaccines in LTRs. Therefore, we investigated whether spike-specific CD8+ Tcells were induced in LTRs and compared their frequency with HDs. We defined 4-1BB+CD69+CD8+ Tcells as spike-specific CD8+ Tcells in the PBMCs stimulated with spike peptides (Supplementary Fig.3a). Spike-specific CD8+ Tcells were detected in 100% of HDs and 93% of LTRs 1 month after the second vaccination (Fig.3a). However, the frequency of spike-specific CD8+ Tcells by the third vaccination did not increase in most HDs and LTRs (Fig.3b, HDs 55.8%, CNI 55%, and CNI+other drug(s) 42.9%). Compared to HDs, the frequency of LTRs was significantly lower at all time points, regardless of taking single or multiple drugs (Fig.3a). Furthermore, in contrast to antibody responses, there was no correlation between spike-specific CD8+ and CD4+ T cell responses (Fig.3c). These results suggest that the third boost effect on memory T-cell responses differs between CD4+ and CD8+ Tcells. We then checked the differentiation status of the spike-specific CD8+ T cells induced by vaccination using CD27, CD45RO, and CD57 markers to define central memory (CM; CD27+CD45RO+), effector memory (EM; CD27-CD57-), and effector (CD27-CD57+) subsets. As a result, the phenotypes of spike-specific CD8+ T cells were changed from CM to EM at 6 months after 2nd vaccination in both the HDs and LTRs who showed positive effects of boosting spike-specific CD8+ T-cell responses (Healthy boost+ and LTR boost+), although the phenotypes of total memory CD8+ T cells were not changed over time (Fig.3d, e). After 3rd mRNA vaccination, HDs and LTRs showed different phenotypes of spike-specific CD8+ T cells, with decreased CM and increased EM and Effector in HDs, but a trend toward increased CM in LTRs.

a Frequencies of spike-specific CD69+4-1BB+CD8+ Tcells in total memory Tcells from HDs (black), CNI group (red), and CNI+other drug(s) group (blue). b Fold-induction of spike-specific CD69+4-1BB+CD8+ Tcells after third vaccination. Pie charts represent the proportion of individuals with fold-induction > 1, and gray slice shows frequency of negative responders (HDs: black, CNI: red, CNI+other drug(s): blue). c Correlation matrix of CD4+ and CD8+ T-cell responses. Shades of blue represent positive correlations approaching 1, while shades of red denote negative correlations nearing -1. P values were calculated using the Spearmans rank test. Frequencies of CM, EM and effector within CD8+ total memory T cells (d) and spike-specific CD69+4-1BB+CD8+ Tcells (e) in individuals who did (boost+) or did not (boost-) receive boost effect from 3 doses of mRNA vaccine (HDs boost: gray, HDs boost+: black, LTRs boost: red, LTRs boost+: dark red). P values (two-sided) were calculated using the Wilcoxon matched-pairs signed rank test compared to 1 month after 2nd vaccination. Frequency of spike-specific CD69+4-1BB+CD8+ Tcells expressing GZMA (f), GZMB (g), and Perforin (h) (HDs: black, CNI: red, CNI+other drug(s): blue). i Expression of multiple cytotoxic molecules in spike-specific CD69+4-1BB+CD8+ Tcells. Each colors arc length and pie charts area represent the expression of each cytotoxic molecule (GZMA: red, GZMB: blue, Perforin: green) and cells expressing the indicated number of cytotoxic molecules (0: yellow, 1: green, 2: blue, 3: red), respectively. P values (two-sided) in (a), (b), (f), (g), and (h) were calculated using the MannWhitney U-test. Sample size, 1m after 2nd: 23 vs 17 vs 26, 3m after 2nd: 22 vs 16 vs 22, 6m after 2nd: 43 vs 21 vs 29, 1m after 3rd: 43 vs 20 vs 29. All experiments were performed once. Error bars indicate the interquartile range.

Furthermore, we previously reported that differences in the expression patterns of cytotoxic molecules could observe qualitative differences in mRNA vaccine-induced spike-specific CD8+ Tcells20. Therefore, we compared the expression of cytotoxic molecules in spike-specific CD8+ Tcells between HDs and LTRs. Supplementary Fig.3b shows the expression patterns of GZMA, GZMB, and Perforin, and gating. Regardless of HDs or LTRs, most spike-specific CD8+ Tcells expressed GZMA before and after the third vaccination (Fig.3f). The proportion of cells expressing GZMA in CNI+other drug(s) was significantly, but slightly, lower than that in HDs before the third vaccination (p=0.0237). However, the proportion of cells expressing GZMB and Perforin was not different between HDs and LTRs before and after the third boost (Fig.3g, h). Furthermore, the expression profiles of GZMA, GZMB, and Perforin were not significantly different between the groups (Supplementary Fig.4a, b). The proportion of subpopulations expressing GZMA, GZMB, and Perforin was approximately 20% in the spike-specific CD8+ Tcells of each group, and the proportion of subpopulations expressing only GZMA was over 50% (Fig.3i). However, we did not observe any qualitative differences in spike-specific CD8+ Tcells induced by the third boost.

HDs and LTRs were vaccinated with an mRNA vaccine based on the Wuhan-1 strain, and the induced antibodies potentially reduced the effectiveness against the recently emerged Omicron sublineages. Therefore, we measured the antibody titers before and after the third boost against RBD corresponding to the Omicron sublineages, and found that anti-RBD antibody titers before the third boost against all sublineages were significantly reduced compared to those against the Wuhan-1 (Fig.4a, b). Among sublineages, the anti-RBD antibody titers against BQ.1.1 and XBB were particularly reduced (HDs, 8.43-fold reduction; CNI, 5.23-fold reduction; CNI+other drug(s), 4.41-fold reduction against BQ.1.1, HDs, 11.9-fold reduction; CNI, 6.35-fold reduction; CNI+other drug(s), 4.41-fold reduction against XBB). Furthermore, the neutralizing activity before the third boost was below the detection limit for BA.5, BQ.1.1, and XBB in most individuals (Fig.4c). Furthermore, there was no change in the trend toward lower antibody titers for each Omicron sublineage (Fig.4d, e). In particular, the CNI+other drug(s) group showed significantly lower anti-RBD antibody levels against all sublineages than the HDs and CNI groups. Additionally, there was a slight improvement in neutralizing activity against the BA.5 strain, but not BQ.1.1 and XBB strains, by the third vaccination (Fig.4f).

a, d Anti-RBD antibody endpoint titers against indicated strains at (a) pre- and (d) post-third boost (HDs: black, CNI: red, CNI+other drug(s): blue). Fold-change of anti-RBD IgG against variants of concern endpoint titers at (b) pre- and (e) post-third boost relative to Wuhan-1. The minus symbol denotes increased resistance. Shades of red indicate a decrease in antibody titers, with darker shades signifying a larger negative fold change. pVNT50 against strains at (c) pre- and (f) post-third boost (HDs: black, CNI: red, CNI+other drug(s): blue). P values (two-sided) in (a), (c), (d), and (f) were calculated using the MannWhitney U-test. P values (two-sided) in (b) and (e) were calculated using the Wilcoxon matched-pairs signed rank test. Sample size, pre-3rd boost: 44 vs 23 vs 31, post-3rd boost: 44 vs 21 vs 30). All experiments were performed once. Error bars indicate the interquartile range.

Collectively, these results suggest that the third vaccination with the Wuhan-1 mRNA vaccine may not be sufficient to induce antibody responses against Omicron sublineages, particularly BQ.1.1 and XBB, in HDs and LTRs.

Finally, we investigated the differences in cellular immunity against Omicron sublineages between HDs and LTRs. The frequency of spike-specific CD154+CD4+ Tcells was evaluated in PBMCs before the third boost. There was no difference in response to the Wuhan-1 and mutant strains in all groups (Supplementary Fig.5a, b). The same trend was observed for spike-specific Th1 CD4+ Tcells (Fig.5a, b). However, the frequency of CD154+CD4+ Tcells and Th1 cells responding to mutant strains in HDs after the third boost was significantly and slightly lower than that of cells responding to Wuhan-1(Supplementary Fig.5c, d, Fig.5c, d). The same trend was observed in spike-specific Th2 CD4+ Tcells (Supplementary Fig.4eh). These results indicate that, unlike antibody responses, CD4+ T-cell responses induced by mRNA vaccines can react to Omicron sublineages. Moreover, LTRs resulted in CD4+ T-cell responses to Omicron sublineages with comparable reactivity to those in HDs.

a Comparison of spike-specific Th1 CD4+ T-cell frequency against spike peptides in CD4+ total memory Tcells at pre-third boost (HDs: black, CNI: red, CNI+other drug(s): blue). b Fold-change of spike-specific Th1 CD4+ T-cell frequency against variants of concern at pre-third boost relative to Wuhan-1. The minus symbol denotes increased resistance. Shades of blue represent an increase in fold change, with darker shades indicating a larger positive fold change. Conversely, shades of red denote a decrease, with darker shades signifying a larger negative fold change. c Comparison of spike-specific Th1 CD4+ T-cell frequency against spike peptides in CD4+ total memory Tcells at post-third boost (HDs: black, CNI: red, CNI+other drug(s): blue). d Fold-change of spike-specific Th1 CD4+ T-cell frequency against variants of concern at post-third boost relative to Wuhan-1. The minus symbol denotes increased resistance. Shades of blue represent an increase in fold change, with darker shades indicating a larger positive fold change. Conversely, shades of red denote a decrease, with darker shades signifying a larger negative fold change. P values (two-sided) were calculated using the Wilcoxon matched-pairs signed rank test. Sample size, 1m after 2nd: 23 vs 17 vs 26, 3m after 2nd: 22 vs 16 vs 22, 6m after 2nd: 43 vs 21 vs 29, 1m after 3rd: 43 vs 20 vs 29. All experiments were performed once. Error bars indicate the interquartile range.

Next, we investigated CD8+ T cell responses to Omicron sublineages. Interestingly, the frequency of spike-specific CD8+ T-cell responses to mutant strains was not significantly decreased, regardless of the pre- and post-third boost (Fig.6a, b). The fold-changes in the frequency of CD8+ T-cell responses to mutant strains relative to Wuhan-1 are shown (Fig.6c, d). Collectively, these results demonstrate that mRNA vaccines induce CD8+ T-cell responses reactive to BA.5, BQ.1.1, and XBB mutant strains and that these responses are maintained in LTRs.

Comparison of spike-specific CD69+4-1BB+ CD8+ T-cell frequency against spike peptides in CD8+ total memory Tcells at (a) pre- and (b) post-third boost (HDs: black, CNI: red, CNI+other drug(s): blue). Fold-change of spike-specific CD8+ T-cell frequency against variants of concern at (c) pre- and (d) post-third boost relative to Wuhan-1. The minus symbol denotes increased resistance. Shades of blue represent an increase in fold change, with darker shades indicating a larger positive fold change. Conversely, shades of red denote a decrease, with darker shades signifying a larger negative fold change. P values (two-sided) were calculated using the Wilcoxon matched-pairs signed rank test. Sample size, 1m after 2nd: 23 vs 17 vs 26, 3m after 2nd: 22 vs 16 vs 22, 6m after 2nd: 43 vs 21 vs 29, 1m after 3rd: 43 vs 20 vs 29. All experiments were performed once. Error bars indicate the interquartile range.

Original post:

Humoral and cellular immune responses to COVID-19 mRNA vaccines in immunosuppressed liver transplant ... - Nature.com