COVID-19 vaccination has significantly impacted the global pandemic by reducing the severity of infection, lowering rates of hospitalization, and reducing morbidity/mortality in healthy individuals. However, the degree of vaccine-induced protection afforded to renal transplant recipients who receive forms of maintenance immunosuppression remains poorly defined. This is particularly important when we factor in the emergence of SARS-CoV-2 variants of concern (VOCs) that have defined mutations that reduce the effectiveness of Ab responses targeting the Spike Ags from the ancestral Wuhan-Hu-1 variants employed in the most widely used vaccine formats. In this study, we describe a qualitative, longitudinal analysis of neutralizing Ab responses against multiple SARS-CoV-2 VOCs in 129 renal transplant recipients who have received three doses of the Pfizer–BioNTech COVID-19 vaccine (BNT162b2). Our results reveal a qualitative and quantitative reduction in the vaccine-induced serological response in transplant recipients versus healthy controls where only 51.9% (67 of 129) made a measurable vaccine-induced IgG response and 41.1% (53 of 129) exhibited a significant neutralizing Ab titer (based on a pseudovirus neutralization test value >50%). Analysis on the VOCs revealed strongest binding toward the wild-type Wuhan-Hu-1 and Delta variants but none with both of the Omicron variants tested (BA1 and BA2). Moreover, older transplant recipients and those who are on mycophenolic acid as part of their maintenance therapy exhibited a profound reduction in all of the analyzed vaccine-induced immune correlates. These data have important implications for how we monitor and manage transplant patients in the future as COVID-19 becomes endemic in our populations.

The COVID-19 pandemic remains a major global threat with >767,750,853 confirmed cases and 6,941,095 deaths (1). The mortality rate is reported to be >7-fold higher among kidney transplant recipients as compared with the general population, with younger recipients (age between 20 and 39 y old) reporting values of 23-fold higher versus healthy controls (2). SARS-CoV-2 vaccines have been proven effective for the general population of healthy individuals and are linked to reduced morbidity and mortality following exposure to the circulating SARS-CoV-2 pathogens (3). However, vaccine-induced protection in immunocompromised individuals remains controversial (4). This is particularly the case for transplant recipients where previous reports have indicated case fatality rates as high as 27.3% postinfection despite vaccination (2). Moreover, there remains a significant degree of contradiction in the published literature on the effectiveness of augmented boosting protocols where differences in the form of maintenance immunosuppression combined with underlying comorbidities confound the analysis of the patient cohorts (5–7). Thus, analyzing correlates of vaccine-induced immune protection in individual transplant patients should aid the identification and triage of individuals who remain at risk.

In this study, we assessed Ab responses against multiple SARS-CoV-2 VOCs following three doses of the Pfizer–BioNTech COVID-19 vaccine (BNT162b2) in renal transplant recipients. We measured Ab binding plus neutralizing activity at four time points, including prevaccination, postdose 2 (21–24 d after the second dose), predose 3 (60–67 d before dose 2), and peak response (21–24 d after dose 3). We also examined correlations in response between age, sex, lymphocyte count, estimated glomerular filtration rate (eGFR) status, diabetic status, length of transplant, and immunosuppressive drug regimen with neutralizing Abs. This represents a detailed analysis of vaccine-induced SARS-CoV-2 Ab responses in Asian transplant recipients.

We recruited transplant recipients >21 y of age who were eligible for SARS-CoV-2 vaccination. The patients included in this study were recruited from July 29, 2021 to October 13, 2021. Recent transplant recipients (<3 mo) and/or with ongoing rejection were excluded from this study. The study was approved by the National Healthcare Group Domain-Specific Review Board and was conducted in accordance with the Declaration of Helsinki for human research. The reference cohort was recruited under the COVID-19 PROTECT study (2012/00917), and the transplant cohort was recruited under Domain-Specific Review Board reference no. 2021/00630. Written informed consent was collected from all participants before study commencement. All patients used in this study were uninfected individuals. Patients who were infected with COVID were excluded from the study.

All kidney transplant recipients received three doses of the Pfizer–BioNTech BNT162b2 mRNA vaccine whereas the healthy volunteers received two doses. Plasma samples from all 129 patients of the transplant cohort were collected at visit 1 before vaccination (termed prevaccination), visit 4 at 21–24 d after dose 2 (termed postdose 2), visit 5 at 60–67 d after dose 2 (termed predose 3), and visit 6 at 21–24 d after dose 3 (termed peak response). Additional information including age, sex, ethnicity, and medical history were collected from all participants. As for the reference cohort, a total of 168 healthy volunteers were recruited and blood samples were taken at prevaccination, 21–24 d after the first dose, 21–24 d after the second dose (termed peak response), and 6 mo after the first dose 2 (termed prebooster). None of the reference cohort had a known history of SARS-CoV-2 infection or a history of immunosuppressive treatments. The reference cohort tested negative for antinucleocapsid IgG levels. Importantly, note that peak response for kidney transplant recipients and the healthy volunteers were at postdose 3 and postdose 2, respectively.

SARS-CoV-2 pseudotyped lentivirus that expresses SARS-CoV-2 Wuhan-Hu-1 Spike protein was made using a third-generation lentivirus system and quantified as described previously (8). Chinese hamster ovary (CHO) cells that expresses ACE2 were cultured in complete medium (DMEM/high glucose with sodium pyruvate [Life Technologies, no. 10569010), supplemented with 10% FBS [HyClone, no. SV301160.03], 10% MEM nonessential amino acids [Life Technologies, no. 1110050], 10% Geneticin (Life Technologies, no. 10131035], and 10% penicillin/streptomycin [Life Technologies, no. 15400054]) at 37°C with 5% CO2 at a density of 5 × 104 cells per well in a 96-well white/clear flat bottom plate (Corning, no. 353377). After 24 h, plasma samples that were diluted 80-fold were incubated with pseudovirus at 2 × 106 infectious units/ml at 37°C for 1 h, after which this pseudovirus/plasma mixture was added to the Chinese hamster ovary–ACE2 monolayer cells for another hour of incubation at 37°C. Complete medium was added subsequently to this mixture for further incubation. After 48 h, the cells were washed and a luciferase assay (Promega, no. E8130) was carried out per the manufacturer’s protocol. The percentage neutralization was calculated from luminescence values as follows: Neutralization % = {[readout (unknown) − readout (infected control)]/[readout (uninfected control) − readout (infected control)]} × 100.

Receptor binding domains (RBDs) for different SARS-CoV-2 variants (i.e., Wuhan-Hu-1, Delta, Omicron BA1, and Omicron BA2) as well as human ACE2 receptor conjugated with the human Ab Fc region were expressed and purified as described previously (9). ELISA was performed to detect serological IgG binding to RBD variants as described previously (8). Briefly, all RBD variants were coated at 1 µg/ml in a 96-well MaxiSorp plate for 1 h and then blocked for 2 h. Plasma samples that were diluted 100-fold were added to the plates for 1 h of incubation. ACE2–Fc was used as a positive control. Another hour of incubation was performed with 10,000-fold diluted goat anti-human IgG HRP (Invitrogen, no. 31413) for the detection of Abs. A plate wash was conducted after each incubation step. Signal was detected by tetramethylbenzidine substrate (Thermo Scientific, no. 34029), and the reaction was stopped with 1 M H2SO4. The reported OD450 was calculated as follows: Reported OD450 = OD450 of plasma binding to RBD − OD450 of plasma binding to blocking buffer.

All negative OD450 values after subtracting the background were arbitrarily assigned as 0.

The ACE2–RBD binding inhibition ELISA was carried out as described previously (8). The ACE2 inhibition ELISA assay has been validated using a World Health Organization reference reagent, which is an international standard comprised of pooled convalescent plasma samples. Briefly, RBDs of SARS-CoV-2 Wuhan-Hu-1 and Delta were coated at 2 µg/ml whereas RBDs of Omicron BA1 and BA2 variants were coated at 4 µg/ml in a 96-well MaxiSorp plate for 1 h. Plasma samples were diluted 5-fold, and ACE2–peroxidase (conjugated using a peroxidase-labeling kit–NH2, Abnova, no. KA0014), secondary Ab used was at 600 ng/ml for both Omicron BA1 and BA2, or 200 ng/ml for Wuhan-Hu-1 and Delta. All subsequent steps were performed the same as the RBD variant binding ELISA. Baseline ACE2–RBD binding was established for each RBD variant using 5-fold diluted heat-inactivated FBS. ACE2–Fc was used as a positive control to ensure a detectable inhibitory effect. The inhibition percentage was calculated as follows: Inhibition % = {[readout (negative control) − readout (sample)]/[readout (negative control)]} × 100.

All negative inhibition percent values were arbitrarily assigned as 0.

Continuous data were represented as medians with interquartile range (IQR) in case of nonnormal distributions. Categorical demographic data and medical history information are presented as absolute numbers with proportions (%). Neutralizing Ab levels and ACE2 inhibition data across time points were analyzed using a nonparametric Kruskal–Wallis test, followed by a Tukey’s test to correct for multiple comparisons. Binding and ACE2 inhibition data on SARS-CoV-2 VOCs RBD variants were compared with the wild-type Wuhan-Hu-1 strain using the a Kruskal–Wallis test followed by a Tukey test. Associations between inhibition from ACE2–RBD binding inhibition ELISA and categorical data were modeled using simple linear regression. A Pearson correlation coefficient, r2 value, and corresponding 95% confidence intervals (CIs) were reported. Chosen statistical tests were performed among prevaccine and postdose 2 datasets as well as predose 3 and peak response datasets for the study cohort. Reference cohort analysis was conducted between prevaccine and postdose 2 as well as peak response and postbooster. Statistical analysis results were chosen to show significant associations.

Missing data were handled on a listwise deletion basis (Supplemental Fig. 1). All statistical tests were two-tailed when applicable, and a p value of <0.05 was used to denote statistical significance. All experiments were performed with three technical repeats. Median and IQR are shown in all graphs unless stated otherwise. All analyses were performed using the statistical software GraphPad Prism 9.0.

A total of 129 kidney transplant recipients were enrolled into this study, with 44.9 and 55.1% female and male patients, respectively. The median age when first vaccine dose was administered was 52 y of age (IQR 43–59 y). This cohort included the following ethnic demographics: Chinese ethnicity (68.2%), followed by Malays (15.5%), Indians (11.6%), and other ethnicities (4.7%) (Fig. 1A, 1B). The main cause of kidney failure was reported to be glomerulonephritis (70.6%), followed by diabetic kidney disease (13.1%), autosomal dominant polycystic kidney disease (3.9%), and other causes (12.4%). Most participants were transplanted for a median time of 6 y (IQR 3–13 y). Demographic and other relevant clinical histories are listed in Fig. 1F.

FIGURE 1.

Definition of study cohort and analysis of serological responses against Wuhan-Hu-1.

(A and B) Demographic distribution of (A) males and (B) females in study cohort with values representing percentage. (C) Binding activity against the wild-type Wuhan-Hu-1 strain at four time points: prevaccine (prevaccination), postdose 2 (21–24 d after dose 2), predose 3 (60–67 d after dose 2), and peak response (21–24 d after dose 3). The dashed line represents the minimum threshold for optimum binding at OD450 of 0.5. (D) Neutralizing activity assessed by PVNT and (E) ACE2 Inhibition at four time points were measured by ELISA. Dashed line represents the minimum threshold for optimum neutralization at 50% for PVNT and 40% for ACE2 inhibition assays. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test. (F) Clinical characteristics in 129 infection-naive transplant recipients according to sex, age, ethnicity, cause of transplant, comorbidities, immunosuppressant drug regimen, and duration of immunosuppressants. Data represent median and interquartile range (IQR) unless stated otherwise. **p < 0.01, ****p < 0.0001.

FIGURE 1.

Definition of study cohort and analysis of serological responses against Wuhan-Hu-1.

(A and B) Demographic distribution of (A) males and (B) females in study cohort with values representing percentage. (C) Binding activity against the wild-type Wuhan-Hu-1 strain at four time points: prevaccine (prevaccination), postdose 2 (21–24 d after dose 2), predose 3 (60–67 d after dose 2), and peak response (21–24 d after dose 3). The dashed line represents the minimum threshold for optimum binding at OD450 of 0.5. (D) Neutralizing activity assessed by PVNT and (E) ACE2 Inhibition at four time points were measured by ELISA. Dashed line represents the minimum threshold for optimum neutralization at 50% for PVNT and 40% for ACE2 inhibition assays. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test. (F) Clinical characteristics in 129 infection-naive transplant recipients according to sex, age, ethnicity, cause of transplant, comorbidities, immunosuppressant drug regimen, and duration of immunosuppressants. Data represent median and interquartile range (IQR) unless stated otherwise. **p < 0.01, ****p < 0.0001.

Close modal

We measured serological levels of IgG by ELISA at four relevant time points to define the overall vaccine-induced serological response in transplant recipients. The four time points chosen were prevaccination, postdose 2, predose 3, and postdose 3, which we defined as peak response. We observed a significant increase in the level of Abs upon receiving two and three doses of vaccination, that is, postdose 2 and postdose 3 (peak response) with median OD value of 0.144 (IQR 0.025–0.621) and 0.578 (IQR 0.095–1.088), respectively (Fig. 1C). Pseudovirus neutralization test (PVNT) results indicated a gradual increase in neutralizing Abs between time points, with a median value of 2.9% (IQR 0–22.9%), 5.7% (IQR 0–19.6%), and 25.0% (IQR 0–91.8%) for postdose 2, predose 3, and peak response (Fig. 1D). We performed surrogate ACE2–RBD binding inhibition ELISA to measure cross-neutralizing Abs against VOCs; these also revealed a weak response in 8.3% of transplanted patients (IQR 0.4–62.5%) (Fig. 1E). Importantly, note that the cutoff for positive neutralization for PVNT and inhibition assay are 50 and 40%, respectively, whereas the cutoff value for positive binding was set at OD 0.5. Any value below that is considered insignificant. These thresholds are important, as they were obtained upon normalizing the values with World Health Organization international standards for a positive response. This also allows us to compare our data with data published in other studies.

To understand whether Ab response was age-dependent, we divided the participants into three subgroups, that is, 20–39, 40–59, and >60 y of age, and analyzed the level of Abs at four time points as described above. Results in Fig. 2A show that all three groups had significant Ab levels at peak response, with the younger participants having the best Ab response with a median OD value of 1.122 (IQR 0.146–1.340) among the three groups. For participants >40 y of age, despite having a significant difference in the Ab level at peak response, it was comparatively low, measuring only 0.528 (IQR 0.094–1.039) and 0.439 (IQR 0.047–0.957) for participants between 40 and 59 and >60 y of age. In addition, we found no significant difference in Ab response between these two groups as well. Fig. 2B shows the percentage of inhibition/neutralizing capacities between the three age groups. As expected, we found that participants from the younger age group (20–39 y old) measured significantly better neutralizing capacity than did the two older groups, with a median value of 47.6% (IQR 1.6–79.2%). The older participants from 40 to 59 and >60 y of age also exhibited a low percentage of neutralizing Abs at 8.0% (IQR 0.4–60.4%) and 4.7% (IQR 0.5–25.1%).

FIGURE 2.

Binding activity and ACE2 inhibitory response of serological IgG against RBD variants.

(A) Serological IgG binding activity against SARS-CoV-2 Wuhan-Hu-1 strain (wild-type [WT]) across three age groups (20–39, 40–59, and >60 y of age) at four time points, measured by ELISA. (B) Binding activity to four RBD variants: SARS-CoV-2 Wuhan-Hu-1 strain (WT), Delta, Omicron BA1 (O BA1), and Omicron BA2 (O BA2). (C) ACE2–RBD inhibitory response across three age groups and (D) inhibitory response to four different RBD variants, analyzed by ELISA. (E and F) Ab response to different variants and time points, in ascending age, represented by heatmaps: (E) binding activity and (F) ACE2–RBD binding inhibition. All time points are shown for WT and Delta subsets (n = 129), predose 3 (n = 14), and peak response (n = 129) O BA1 and O BA2 subsets, respectively, each. Age group subsets were consistent across all four time points: 20–39 (n = 20), 40–59 (n = 81), and >60 y of age (n = 28). Dashed line represents the minimum threshold for optimum absorbance (A and B) at 0.5 OD450 and ACE2 inhibition (C and D) at 40%. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test. Data are represented as median and interquartile range (A–D). Cell data are represented as mean (E and F). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 2.

Binding activity and ACE2 inhibitory response of serological IgG against RBD variants.

(A) Serological IgG binding activity against SARS-CoV-2 Wuhan-Hu-1 strain (wild-type [WT]) across three age groups (20–39, 40–59, and >60 y of age) at four time points, measured by ELISA. (B) Binding activity to four RBD variants: SARS-CoV-2 Wuhan-Hu-1 strain (WT), Delta, Omicron BA1 (O BA1), and Omicron BA2 (O BA2). (C) ACE2–RBD inhibitory response across three age groups and (D) inhibitory response to four different RBD variants, analyzed by ELISA. (E and F) Ab response to different variants and time points, in ascending age, represented by heatmaps: (E) binding activity and (F) ACE2–RBD binding inhibition. All time points are shown for WT and Delta subsets (n = 129), predose 3 (n = 14), and peak response (n = 129) O BA1 and O BA2 subsets, respectively, each. Age group subsets were consistent across all four time points: 20–39 (n = 20), 40–59 (n = 81), and >60 y of age (n = 28). Dashed line represents the minimum threshold for optimum absorbance (A and B) at 0.5 OD450 and ACE2 inhibition (C and D) at 40%. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test. Data are represented as median and interquartile range (A–D). Cell data are represented as mean (E and F). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Close modal

Next, we evaluated Ab responses against four different VOCs: wild-type Wuhan-Hu-1, Delta, and Omicron BA1 and BA2. Our results revealed that the IgG binding response was strongest toward the wild-type Wuhan-Hu-1 strain followed by the Delta variant with median OD450 values of 0.578 (IQR 0.095–1.088) and 0.470 (IQR 0.060–1.004), respectively, at peak response. The responses against Omicron variants were negligible, with values of 0.093 (IQR 0.020–0.338) and 0.120 (IQR 0.022–0.483) for BA1 and BA2, respectively (Fig. 2C). In terms of the neutralizing capacity, we found no significant difference between the four variants (Fig. 2D).

Fig. 2E and 2F show the heatmap of Ab response and percentage inhibition at different time points comparing age and different VOCs. Overall, the younger age groups showed good binding and neutralizing capacity for the both Wuhan and Omicron variants at peak response.

We investigated the relationship between possible risk factors and the vaccine-induced Ab response by performing a correlation analysis. Fig. 3A shows the correlation between OD value and the level of Ab with neutralization based on the percentage of ACE2–RBD binding inhibition. Our analysis revealed a positive relationship between the level of Abs and the neutralizing capacity at postdose 2 and peak response with an r value of 0.8023 (CI 0.7320–0.8556, p < 0.0001) and 0.7910 (CI 0.7172–0.8471, p < 0.0001), respectively. A weak negative correlation was observed when we examined age (r = 0.1999) (Fig. 3B) to neutralization activity but none for eGFR values (Fig. 3C). A weak correlation with an r value of 0.3075 (CI 0.144–0.450) was observed with lymphocyte counts (Fig. 3D). We also found no significant differences in Ab response in relationship to sex (Fig. 3E) or the presence of diabetes (Fig. 3F), regardless of time point analyzed.

FIGURE 3.

Correlation between common risk factors and vaccine-induced Ab responses.

(A) Correlation of ACE2–RBD binding inhibition (%) with IgG binding activity for viral Spike–RBD (absorbance OD450). (B) Correlation of percentage neutralization with age (years) at two time points, that is, postdose 2 (days after) and peak response (days after), including normalized data. (C) Correlation of neutralization with eGFR (ml/min/1.73 m2) values at two time points, that is, postdose 2 and peak response. (D) Correlation of neutralization with lymphocyte count × 109 at peak response. (E) Comparison of percentage of neutralization with diabetes status at four time points. (F) Comparison of percentage of neutralization between sexes at four time points. Data were modeled by a simple linear regression; Pearson correlation coefficients and p values are included in the legend. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test (C and E). Data are represented as median and interquartile range. **p < 0.01, ***p < 0.001 (E and F).

FIGURE 3.

Correlation between common risk factors and vaccine-induced Ab responses.

(A) Correlation of ACE2–RBD binding inhibition (%) with IgG binding activity for viral Spike–RBD (absorbance OD450). (B) Correlation of percentage neutralization with age (years) at two time points, that is, postdose 2 (days after) and peak response (days after), including normalized data. (C) Correlation of neutralization with eGFR (ml/min/1.73 m2) values at two time points, that is, postdose 2 and peak response. (D) Correlation of neutralization with lymphocyte count × 109 at peak response. (E) Comparison of percentage of neutralization with diabetes status at four time points. (F) Comparison of percentage of neutralization between sexes at four time points. Data were modeled by a simple linear regression; Pearson correlation coefficients and p values are included in the legend. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test (C and E). Data are represented as median and interquartile range. **p < 0.01, ***p < 0.001 (E and F).

Close modal

We examined the impact of different immunosuppressive drug regimens on vaccine-induced Ab responses. Most of the patients in this cohort (84%) were on three immunosuppressive drugs. Patients were pooled into four different groups based on the combination of drugs that they were on. The groupings were as follows: 1) CMP: calcineurin inhibitors (CNIs), mycophenolic acid (MPA), and prednisone; 2) CAP: CNIs, azathioprine, and prednisone; 3) M: mTOR inhibitors and prednisone with or without MPA; and 4) others (O): CNIs and prednisone or azathioprine and prednisone or CNIs, leflunomide, and prednisone (Supplemental Table I). The four groups were then compared for neutralizing capacities at different time points.

We found no differences between the different drug regimens (Fig. 4A) at prevaccination as expected. However, upon completing two doses of vaccination, individuals on the CAP regimen fared significantly better compared with the other groups based on a median value of 48.5% (IQR 17.4–65.8%) of neutralizing Abs compared with those on CMP, M, or O with a median of 16.9% (IQR 12.3–23.2%), 20% (IQR 15.9–30.2%), and 10.3% (IQR 6.3–75.5%), respectively (Fig. 4B). The Ab titer dropped two mo postdose 2 before increasing to a median value of 76.1% (IQR 37.2–90.5%) neutralization at peak response, with at least 75% (9 of 12) of individuals in this group having >50% neutralizing Abs. Patients from the other groups responded poorly even after three doses of vaccination with only 7.1% (IQR 0–51.8%), 6.2% (IQR 0.6–61.7%), and 39.6% (IQR 0–81.6%) neutralizing Abs in the CMP, M, and O regimens, respectively (Fig. 4C, 4D). Within these groups, 24.2% (23 of 95) of CMP, 33.3% (4 of 12) of M, and 44.4% (4 of 9) of O had a response of >50% inhibition/neutralization, suggesting an overall weak level of protection. In addition, we found no correlation between time of transplantation or the duration of immunosuppression on vaccine-induced Ab response (Fig. 4E).

FIGURE 4.

Impact of different immunosuppressive maintenance regimens on vaccine-induced Ab responses.

(AD) The ACE2–RBD inhibitory response against SARS-CoV-2 Wuhan-Hu-1 strain (WT) of different immunosuppressive drug regimens at (A) prevaccination, (B) postdose 2, (C) predose 3, and (D) peak response. (E) Correlation of percentage of inhibition with months posttransplant at four time points. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test (AD). Data are represented as median and interquartile range. Correlation data were modeled by a simple linear regression; Pearson correlation coefficients and p values are included in the figure. Data plotted as mean ± SEM. *p < 0.05, ***p < 0.001.

FIGURE 4.

Impact of different immunosuppressive maintenance regimens on vaccine-induced Ab responses.

(AD) The ACE2–RBD inhibitory response against SARS-CoV-2 Wuhan-Hu-1 strain (WT) of different immunosuppressive drug regimens at (A) prevaccination, (B) postdose 2, (C) predose 3, and (D) peak response. (E) Correlation of percentage of inhibition with months posttransplant at four time points. Data were analyzed by a Kruskal–Wallis test followed by a Tukey multiple comparison test (AD). Data are represented as median and interquartile range. Correlation data were modeled by a simple linear regression; Pearson correlation coefficients and p values are included in the figure. Data plotted as mean ± SEM. *p < 0.05, ***p < 0.001.

Close modal

The advent of COVID-19 vaccination has significantly improved outcomes following SARS-CoV-2 exposure in the general population. However, immunocompromised patients and, in particular, organ transplant recipients remain at significant risk as seroconversion after immunization remains low (10). Humoral immune responses upon vaccination differ significantly between individual patients within transplant cohorts. As such, it is important to monitor immune correlates of protection at the individual level. In Singapore, most of the population, including the transplant community, was vaccinated with the Pfizer–BioNTech COVID-19 vaccine. Blood samples were collected at seven different time points, but we only focused on four specific time points that are most relevant to this cohort, considering the overall poor Ab responses observed.

First, we measured the Ab level to understand the overall response in this population. SARS-CoV-2 Spike Ag binding assays revealed a median OD value of 0.5 at peak response, suggesting a weak Ab response after receiving three doses of vaccine. As Ab binding titer does not necessarily correlate with protection, we also measured the strength of neutralizing Abs in these individuals (11, 12). Our results indicated a weak neutralizing capacity in our transplant cohort with a median value of 8.3% inhibition. This is significantly lower compared with our reference cohort, which consisted of age- and sex-matched healthy volunteers who measured 100% neutralization (Supplemental Fig. 2), but is similar to other groups that reported poor responses in kidney transplant recipients (13). In addition, 51.9% (67 of 129) of study participants demonstrated optimal OD values >0.5 and 41.1% (53 of 129) demonstrated PVNT values of >50%, with most of these individuals from the younger age group. By implication, we estimate that less than half of the vaccinated transplant cohort was sufficiently protected.

Our evaluation of the response from the different age group revealed better responses from the younger age group: 20- to 39-y-old participants exhibited a better Ab response for both binding and neutralization as compared with participants >40 y of age. Participants from the older age group measured significantly lower responses. Analysis of Ab responses to different variants showed good binding to both the wild-type Wuhan-Hu-1 and Delta variants as expected from the nature of the vaccine Ag. We observed poor Ab binding responses in all cohorts to the Omicron variants.

We also performed a correlation analysis to understand whether common factors such as age, sex, interval posttransplant, diabetic status, eGFR values, and lymphocyte counts are risk factors that influence the differential SARS-CoV-2 Ab responses observed in our transplant cohort. We found no differences in responses between male and female participants, as opposed to previous reports on the healthy population, which indicated that females had better Ab responses compared with male recipients (14). However, weak correlations were observed when analyzed with age and lymphocyte count, resembling a number of previous reports on similar cohorts (15–17).

Analysis on the impact of different immunosuppressants revealed that patients who are on the CNIs, azathioprine, and prednisone combination fared significantly better compared with patients on the CNIs, MPA, and prednisone combination. MPA is an agent is known to inhibit Ab production by depleting B cells with persistent B cell exhaustion (18). A recent study by Kantauskaite et al. (19) suggested that dose reduction of MPA by 33% could help to improve vaccination outcomes in kidney transplant recipients. Thus, repeated doses of vaccine may not be efficient when immunosuppressant reduction is not implemented, bearing in mind the potential risk of rejection with immunosuppressant reduction. Whereas other studies demonstrated a negative correlation between interval posttransplant and neutralizing Ab activity (20), we found no relationship between interval posttransplant and neutralizing activity in our Asian patient cohort (r = 0.1930, IQR 0.0235–0.3517).

Previous studies on other (nonrenal) solid organ transplants have reported different outcomes. One study on liver transplant recipients reported that a third dose of BNT162b2 vaccine was able to elicit Ab response in almost 95% of the participants (21). Almost 67% of heart transplant patients were found to have positive Ab responses after the third dose, with corresponding neutralizing activity (22). However, recipients of lung transplant were reported to have worse outcomes, with only 16% of the patients exhibiting an Ab response after three doses of vaccine (23).

This study has a few limitations. First, we acknowledge the small number of participants, with unequally distributed demographics. Second, all of the patients in this study were vaccinated with a single type of vaccine, namely the Pfizer BNT162b2 formulation. Other studies analyzing seroconversion in patients receiving the mRNA-1273 (Moderna) vaccine or recipients receiving heterologous vaccine regimens reported better immune responses (6, 16). We also acknowledge that ∼85% of our study participants are >40 of age where age is an important risk factor that influences the vaccine response (24, 25). However, our study demonstrates several key strengths, including a well-curated population, with complete data on transplant and perivaccination variables, all receiving a consistent vaccination protocol permitting meaningful longitudinal analysis of Ab responses.

In conclusion, we found a significant degree of heterogeneity in vaccine-induced serological responses in our Asian vaccine cohort. Although previous reports have suggested that three doses of vaccine were sufficient to mount an adequate immune response in transplant recipients, this may not be applicable to all. In our study, we found that only 41% of the participants had Abs with >50% neutralizing activity and this is influenced by key factors such as the immunosuppressive drug maintenance regimen. As such, is it vital to monitor immune responses in a personalized, individual context in transplant recipients to identify individuals who remain at risk. This is particularly important for those who are older or on immunosuppressive drug regimens that include MPA.

The authors have no financial conflicts of interest.

The online version of this article contains supplemental material.

This work was supported by the National University Health System (NUHSRO/2021/048) research grant.

CI

confidence interval

CNI

calcineurin inhibitor

IQR

interquartile range

MPA

mycophenolic acid

PVNT

pseudovirus neutralization test

RBD

receptor-binding domain

VOC

variant of concern

1
World Health Organization
.
2023
.
Coronavirus disease (COVID-19) pandemic
. .
2
Medina-Pestana
,
J.
,
M. P.
Cristelli
,
R. D.
Foresto
,
H.
Tedesco-Silva
,
L. R.
Requião-Moura
.
2022
.
The higher COVID-19 fatality rate among kidney transplant recipients calls for further action
.
Transplantation
106
:
908
910
.
3
Chi
,
W. Y.
,
Y. D.
Li
,
H. C.
Huang
,
T. E. H.
Chan
,
S. Y.
Chow
,
J. H.
Su
,
L.
Ferrall
,
C. F.
Hung
,
T. C.
Wu
.
2022
.
COVID-19 vaccine update: vaccine effectiveness, SARS-CoV-2 variants, boosters, adverse effects, and immune correlates of protection
.
J. Biomed. Sci.
29
:
82
.
4
Reischig
,
T.
,
M.
Kacer
,
T.
Vlas
,
P.
Drenko
,
L.
Kielberger
,
J.
Machova
,
O.
Topolcan
,
R.
Kucera
,
S.
Kormunda
.
2022
.
Insufficient response to mRNA SARS-CoV-2 vaccine and high incidence of severe COVID-19 in kidney transplant recipients during pandemic
.
Am. J. Transplant.
22
:
801
812
.
5
Kamar
,
N.
,
F.
Abravanel
,
O.
Marion
,
C.
Couat
,
J.
Izopet
,
A.
Del Bello
.
2021
.
Three doses of an mRNA Covid-19 vaccine in solid-organ transplant recipients
.
N. Engl. J. Med.
385
:
661
662
.
6
Thomson
,
T.
,
M.
Prendecki
,
S.
Gleeson
,
P.
Martin
,
K.
Spensley
,
R. C.
De Aguiar
,
B.
Sandhu
,
C.
Seneschall
,
J.
Gan
,
C. L.
Clarke
, et al
.
2022
.
Immune responses following 3rd and 4th doses of heterologous and homologous COVID-19 vaccines in kidney transplant recipients
.
EClinicalMedicine
53
:
101642
.
7
Dib
,
M.
,
N.
Le Corre
,
C.
Ortiz
,
D.
García
,
M.
Ferrés
,
C.
Martinez-Valdebenito
,
C.
Ruiz-Tagle
,
M. J.
Ojeda
,
M. A.
Espinoza
,
A.
Jara
, et al
.
2022
.
SARS-CoV-2 vaccine booster in solid organ transplant recipients previously immunised with inactivated versus mRNA vaccines: a prospective cohort study
.
Lancet Reg Health Am
16
:
100371
.
8
Gu
,
Y.
,
J. M.
Low
,
J. S. Y.
Tan
,
M. S. F.
Ng
,
L. F. P.
Ng
,
B.
Shunmuganathan
,
R.
Gupta
,
P. A.
MacAry
,
Z.
Amin
,
L. Y.
Lee
, et al
.
2022
.
Immune and pathophysiologic profiling of antenatal coronavirus disease 2019 in the GIFT cohort: a Singaporean case-control study
.
Front Pediatr.
10
:
949756
.
9
Raghuvamsi
,
P. V.
,
N. K.
Tulsian
,
F.
Samsudin
,
X.
Qian
,
K.
Purushotorman
,
G.
Yue
,
M. M.
Kozma
,
W. Y.
Hwa
,
J.
Lescar
,
P. J.
Bond
, et al
.
2021
.
SARS-CoV-2 S protein:ACE2 interaction reveals novel allosteric targets
.
eLife
10
:
e63646
.
10
Lee
,
A. R. Y. B.
,
S. Y.
Wong
,
L. Y. A.
Chai
,
S. C.
Lee
,
M. X.
Lee
,
M. D.
Muthiah
,
S. H.
Tay
,
C. B.
Teo
,
B. K. J.
Tan
,
Y. H.
Chan
, et al
.
2022
.
Efficacy of covid-19 vaccines in immunocompromised patients: systematic review and meta-analysis
.
BMJ
376
:
e068632
.
11
Garcia-Beltran
,
W. F.
,
E. C.
Lam
,
M. G.
Astudillo
,
D.
Yang
,
T. E.
Miller
,
J.
Feldman
,
B. M.
Hauser
,
T. M.
Caradonna
,
K. L.
Clayton
,
A. D.
Nitido
, et al
.
2021
.
COVID-19-neutralizing antibodies predict disease severity and survival
.
Cell
184
:
476
488.e11
.
12
Khoury
,
D. S.
,
D.
Cromer
,
A.
Reynaldi
,
T. E.
Schlub
,
A. K.
Wheatley
,
J. A.
Juno
,
K.
Subbarao
,
S. J.
Kent
,
J. A.
Triccas
,
M. P.
Davenport
.
2021
.
Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection
.
Nat. Med.
27
:
1205
1211
.
13
Benotmane
,
I.
,
G.
Gautier
,
P.
Perrin
,
J.
Olagne
,
N.
Cognard
,
S.
Fafi-Kremer
,
S.
Caillard
.
2021
.
Antibody response after a third dose of the mRNA-1273 SARS-CoV-2 vaccine in kidney transplant recipients with minimal serologic response to 2 doses
.
JAMA
326
:
1063
1065
.
14
Jalkanen
,
P.
,
P.
Kolehmainen
,
H. K.
Häkkinen
,
M.
Huttunen
,
P. A.
Tähtinen
,
R.
Lundberg
,
S.
Maljanen
,
A.
Reinholm
,
S.
Tauriainen
,
S. H.
Pakkanen
, et al
.
2021
.
COVID-19 mRNA vaccine induced antibody responses against three SARS-CoV-2 variants
.
Nat. Commun.
12
:
3991
.
15
Vaiciuniene
,
R.
,
B.
Sitkauskiene
,
I. A.
Bumblyte
,
E.
Dalinkeviciene
,
E.
Ziginskiene
,
D.
Bagdonas
,
R.
Augliene
,
K.
Petruliene
,
I.
Bagdziuniene
,
I.
Skarupskiene
, et al
.
2021
.
Immune response after SARS-CoV-2 vaccination in kidney transplant patients
.
Medicina (Kaunas)
57
:
1327
.
16
Sanders
,
J. F.
,
F. J.
Bemelman
,
A. L.
Messchendorp
,
C. C.
Baan
,
D.
van Baarle
,
R.
van Binnendijk
,
D. A.
Diavatopoulos
,
S. C.
Frölke
,
D.
Geers
,
C. H.
GeurtsvanKessel
, et al
RECOVAC Collaborators
.
2022
.
The RECOVAC immune-response study: the immunogenicity, tolerability, and safety of COVID-19 vaccination in patients with chronic kidney disease, on dialysis, or living with a kidney transplant
.
Transplantation
106
:
821
834
.
17
Grupper
,
A.
,
L.
Rabinowich
,
M.
Ben-Yehoyada
,
E.
Katchman
,
R.
Baruch
,
T.
Freund
,
D.
Hagin
,
S. B.
Shlomo
,
D.
Schwartz
,
I. F.
Schwartz
, et al
.
2022
.
Humoral response to the third dose of Sars-Cov-2 vaccine in kidney transplant recipients
.
Transplant. Proc.
54
:
1439
1445
.
18
Eickenberg
,
S.
,
E.
Mickholz
,
E.
Jung
,
J. R.
Nofer
,
H. J.
Pavenstadt
,
A. M.
Jacobi
.
2012
.
Mycophenolic acid counteracts B cell proliferation and plasmablast formation in patients with systemic lupus erythematosus
.
Arthritis Res. Ther.
14
:
R110
.
19
Kantauskaite
,
M.
,
L.
Müller
,
J.
Hillebrandt
,
J.
Lamberti
,
S.
Fischer
,
T.
Kolb
,
K.
Ivens
,
M.
Koch
,
M.
Andree
,
N.
Lübke
, et al
.
2022
.
Immune response to third SARS-CoV-2 vaccination in seronegative kidney transplant recipients: Possible improvement by mycophenolate mofetil reduction
.
Clin. Transplant.
36
:
e14790
.
20
Hasan
,
I.
,
T.
Rashid
,
S.
Suliman
,
H.
Amer
,
R. M.
Chirila
,
M. L.
Mai
,
T.
Jarmi
,
S.
Khouzam
,
P. M.
Franco
,
C. W.
Heilig
,
H. M.
Wadei
.
2021
.
Predictors of disease severity and outcome of hospitalized renal transplant recipients with COVID-19 infection: a systematic review of a globally representative sample
.
Rom. J. Intern. Med.
59
:
10
42
.
21
Tomiyama
,
T.
,
R.
Suzuki
,
N.
Harada
,
T.
Tamura
,
K.
Toshida
,
Y.
Kosai-Fujimoto
,
T.
Tomino
,
S.
Yoshiya
,
Y.
Nagao
,
K.
Takeishi
, et al
.
2023
.
A third dose of the BNT162b2 mRNA vaccine sufficiently improves the neutralizing activity against SARS-CoV-2 variants in liver transplant recipients
.
Front. Cell. Infect. Microbiol.
13
:
1197349
.
22
Peled
,
Y.
,
E.
Ram
,
J.
Lavee
,
A.
Segev
,
S.
Matezki
,
A.
Wieder-Finesod
,
R.
Halperin
,
M.
Mandelboim
,
V.
Indenbaum
,
I.
Levy
, et al
.
2022
.
Third dose of the BNT162b2 vaccine in heart transplant recipients: immunogenicity and clinical experience
.
J. Heart Lung Transplant.
41
:
148
157
.
23
Dauriat
,
G.
,
L.
Beaumont
,
L. B.
Luong Nguyen
,
B.
Renaud Picard
,
M.
Penhouet
,
B.
Coiffard
,
M.
Salpin
,
X.
Demant
,
C.
Saint Raymond
,
N.
Carlier
, et al
.
2023
.
Efficacy of three COVID-19 vaccine doses in lung transplant recipients: a multicentre cohort study
.
Eur. Respir. J.
61
:
2200502
.
24
Falahi
,
S.
,
A.
Kenarkoohi
.
2022
.
Host factors and vaccine efficacy: implications for COVID-19 vaccines
.
J. Med. Virol.
94
:
1330
1335
.
25
Bajaj
,
V.
,
N.
Gadi
,
A. P.
Spihlman
,
S. C.
Wu
,
C. H.
Choi
,
V. R.
Moulton
.
2021
.
Aging, Immunity, and COVID-19: how age influences the host immune response to coronavirus infections?
Front. Physiol.
11
:
571416
.
This article is distributed under the terms of the CC BY-NC-ND 4.0 Unported license.

Supplementary data