Omicron variant neutralizing antibodies following BNT162b2 BA.4/5 versus mRNA-1273 BA.1 bivalent vaccination in patients with end-stage kidney disease

Neutralization of Omicron subvariants by different bivalent vaccines has not been well evaluated. This study characterizes neutralization against Omicron subvariants in 98 individuals on dialysis or with a kidney transplant receiving the BNT162b2 (BA.4/BA.5) or mRNA-1273 (BA.1) bivalent COVID-19 vaccine. Neutralization against Omicron BA.1, BA.5, BQ.1.1, and XBB.1.5 increased by 8-fold one month following bivalent vaccination. In comparison to wild-type (D614G), neutralizing antibodies against Omicron-specific variants were 7.3-fold lower against BA.1, 8.3-fold lower against BA.5, 45.8-fold lower against BQ.1.1, and 48.2-fold lower against XBB.1.5. Viral neutralization was not significantly different by bivalent vaccine type for wild-type (D614G) (P = 0.48), BA.1 (P = 0.21), BA.5 (P = 0.07), BQ.1.1 (P = 0.10), nor XBB.1.5 (P = 0.10). Hybrid immunity conferred higher neutralizing antibodies against all Omicron subvariants. This study provides evidence that BNT162b2 (BA.4/BA.5) and mRNA-1273 (BA.1) induce similar neutralization against Omicron subvariants, even when antigenically divergent from the circulating variant.

Chronic kidney disease is a major risk factor for severe COVID-19 4 and those receiving hemodialysis are at increased risk for exposure to SARS-CoV-2 5 .Although patients with chronic kidney disease exhibit a robust antibody response to third-dose vaccination, new Omicron subvariants including BQ.1.1 and XBB.1.5have immune-evasive potential 6,7 .Therefore, we evaluated the neutralizing antibody response to Omicron subvariants following bivalent COVID-19 vaccination in 98 patients receiving hemodialysis and kidney transplant recipients.Given that the mRNA-1273 bivalent vaccine targets BA.1 while BNT162b2 targets BA.4/5, we examined differences in neutralizing antibody levels against BA.1, BA.5, BQ.1.1,and XBB.1.5induced by the different bivalent vaccines using a spiked-pseudotyped lentiviral neutralization assay.
In samples taken prior to receipt of the bivalent vaccine, the median time from the most recent COVID-19 vaccine dose was 236 days (interquartile [IQR] 189-290).Following receipt of the bivalent vaccine dose, serum samples were taken at a median of 25 days (IQR 24-27).Among participants, 26% (25/98) had prior confirmed COVID-19, as determined by RT-PCR or rapid antigen testing, while 41% (40/98) had a positive anti-nucleocapsid antibody before bivalent vaccination.At 1 month follow-up, 37% (36/98) had a positive anti-nucleocapsid antibody with one new seroconversion while 5 individuals who were initially seropositive became seronegative.No clinical COVID-19 infections were reported during the study period (Supplementary Table 1).

Discussion
Concerns have emerged regarding immune escape with newer Omicron subvariants 8,9 .Consistent with this, we observed a 46-fold decrease in viral neutralization against BQ.1.1 and 48-fold decrease against XBB.1.5 in comparison to wild-type.Consistent with other reports, we found that hybrid immunity enhanced neutralizing antibody response 10 .In addition, while BNT162b2 elicited higher absolute neutralizing antibodies in comparison to mRNA-1273 against all Omicron subvariants, these differences were not statistically significant after accounting for confounders including prior COVID-19 infection, number of vaccine doses, and hemodialysis versus kidney transplant recipients.Given that XBB.1.5 is presently the dominant circulating Omicron subvariant worldwide, this suggests that boosters targeting earlier subvariants may provide similar protection.Our results are consistent with a case series of dialysis patients, where bivalent vaccination increased anti-spike IgG 2.5 fold which correlated with BA.4/5 neutralization 11 .Our study findings conducted in a real-world setting are similar to the findings of a randomized controlled trial in 202 individuals, which compared the Pfizer/BioNTech BA.1 versus BA.4/5 vaccine and found similar neutralizing antibodies against BQ.1.1 and  XBB.1.5 with either bivalent vaccine 12 .However, this randomized trial differed from our study as it only included healthy individuals and did not evaluate the mRNA-1273 BA.1 vaccine.Other studies have not found a higher peak neutralizing antibody response with bivalent vaccines in comparison to monovalent vaccines, and suggested a role for immunologic imprinting wherein prior antigenic exposures prime B cell memory and limit the development of memory B cells and neutralizing antibodies against newer subvariants 8,13 .Kidney transplant recipients generally had both lower increases and absolute lower neutralizing antibodies levels post-vaccination, which is unsurprising given that transplant recipients are known to have attenuated humoral response to COVID-19 vaccination 14 .The number of prior vaccine doses (4 versus 5) did not meaningfully influence neutralizing antibody levels both prior to and following bivalent vaccination.
Strengths of this study include that it is the first to compare neutralizing antibody response against current circulating Omicron subvariants with bivalent vaccines targeting BA.1 versus BA.4/5.We had anti-nucleocapsid serology and clinical information available on all participants, allowing us to account for multiple factors including prior COVID-19 infection in our analysis.This study also focuses on a vulnerable patient population as both kidney transplant recipients and individuals receiving hemodialysis have lower humoral responses to COVID-19 vaccination and are at higher risk for severe COVID-19 outcomes 15,16 .Therefore, characterizing the dynamics of humoral response to bivalent vaccination will help to inform future vaccination strategies in vulnerable populations, including the timing of additional boosters.Similarly, a recent study in patients with hematologic malignancy demonstrated the utility of serologic testing in immunocompromised populations for predicting future risk of COVID-19 infection 17 .
This study does have limitations.Due to the observational nature of this study, there may have been residual confounding.Individuals who chose the mRNA-1273 BA.1 vaccine, which was available earlier in Canada, were more likely to have received prior mRNA-1273 vaccination in comparison to those who received the BNT162b2 BA.4/5 vaccine.However, in a sensitivity analysis wherein we adjusted for the initial three-dose vaccine types, the underlying conclusions were unchanged.Due to temporal differences in mRNA-1273 BA.1 versus BNT162b2 BA.4/5 vaccine availability, it is possible that differences in COVID-19 infections could have occurred based upon epidemiologic and participant factors.However, only one individual contracted COVID-19 (as determined by anti-nucleocapsid seroconversion) during follow-up.As such, COVID-19 infections were unlikely to have affected the underlying study conclusions.We did not evaluate cellular immunity, although prior studies have suggested that spike-specific CD8+ and CD4 + T-cell response increased only modestly after bivalent mRNA boosting 18 .In addition, while neutralizing antibody levels correlate with vaccine effectiveness 19,20 , the clinical implications of these findings on preventing COVID-19 infection require confirmation at a population-level, as there is evidence that bivalent vaccine effectiveness declines after one month 21 .Finally, we were not able to evaluate the mRNA-1273 BA.4/5 bivalent vaccine as it was not approved in our jurisdiction during the study period.
In conclusion, we found that bivalent vaccines increased neutralizing antibodies against Omicron subvariants, although neutralizing antibody titers were significantly attenuated against BQ.1.1 and XBB.1.5.We found that both BNT162b2 and mRNA-1273 bivalent vaccines elicited a similar neutralizing antibody response against Omicron subvariants including BQ.1.1 and XBB.1.5,suggesting that variant modified bivalent vaccines may confer protection against emerging circulating Omicron subvariants despite antigenic divergence.

Methods
This study protocol was approved by the respective Institutional Review Boards at Sunnybrook Health Sciences Centre and Unity Health Network (CTO #3604) as well as Michael Garron Hospital (REB # 856-2201-Inf-066).All procedures were in accordance with the ethical standards of the 1964 Declaration of Helsinki.This was a prospective observational cohort study involving individuals age ≥18 receiving hemodialysis or with a kidney transplant receiving the mRNA-1273 bivalent vaccine (Original and Omicron BA.1) or BNT162b2 bivalent vaccine (Original and Omicron BA.4/BA.5).The study was conducted between July 25 and November 30, 2022.The protocol for this study has been published previously 22 .Only participants unable to provide informed consent due to cognitive impairment or a language barrier if a translator was unavailable were excluded.Demographics, vaccination status, comorbidities, and medications were obtained from electronic patient records.Serum samples were taken prior to and one month following bivalent vaccination to evaluate neutralizing antibody levels against Omicron subvariants.Serum samples were collected in a serum separator tube and allowed to clot for 30 minutes and centrifuged for 10 minutes at 3000 × g-force at room temperature.Levels of binding IgG antibodies to each of the antigens (produced by the National Research Council of Canada) were normalized to reference standards included on each plate and expressed as relative ratios or World Health Organization International Standard units (BAU/mL).

Spiked-Pseudotyped lentiviral neutralization
The spiked-pseudotyped lentiviral neutralization assay was performed in HEK293T-ACE2/TMPRSS2 cells 14 .Lentiviral virus-like particles were generated from co-transfection of the viral packaging (psPAX2, Addgene, #12260), the ZsGreen and luciferase reporter construct (pHAGE-CMV-Luc2-IRES-ZsGreen-W, provided by Jesse Bloom), and the spike protein constructs: wild-type (D614G), Omicron BA.1, BA.5, BQ.1.1,XBB.1.5subvariants available at http://nbcc.lunenfeld.ca/resources), into HEK293TN cells (System Biosciences [LV900A-1]).To construct HDM_XBB.1.5,a human codon-optimized cDNA encoding the XBB.1 spike variant (Twist Bioscience) was first inserted into the mammalian expression plasmid HDM (a gift from Jesse Bloom) and modified to include the F486P mutation starting from the cDNA fragment encoding the XBB.1.5RBD.Viral supernatants were harvested, clarified, and filtered through 0.45-μm filters prior to storage at −80 °C.A viral titer assay was performed by infecting HEK293T-ACE2/ TMPRSS2 cells, followed by a luciferase assay to determine the relative luciferase unit (1:10 to 1:250 dilution of virus stock, depending on the virus titers of each variant).For the neutralization assay, diluted patient sera samples (1:22.5)were prepared and serially diluted by 3-fold over 7 dilutions, followed by incubation with diluted pseudovirus for 1 h at 37 °C prior to addition to HEK293T-ACE2/TMPRSS2 cells.Cells were lysed 48-hours post infection using the Bright-Glo Luciferase Assay System (Promega), and the luminescence signals were detected using a PerkinElmer EnVision instrument.HEK293TN and HEK293T-ACE2/TMPRSS2 cells are maintained at 85% confluency for no more than 25 passages.The inhibitory dilution with 50% virus neutralization (ID50) was calculated in GraphPad Prism version 9.5.0 using a nonlinear regression algorithm (log[inhibitor] versus normalized response -variable slope).In patients with an absence of 50% neutralization with undiluted serum, a log 10 ID 50 of zero was considered the threshold for detectable viral neutralization.

SARS-CoV-2 binding IgG antibodies
Binding IgG antibodies against the ancestral (D614G) full-length spike protein (anti-spike), its receptor binding domain (anti-RBD) (Identifier: VHH72-hFc1X7), and anti-nucleocapsid antibodies (HC2003; Source: Genscript; Identifier: Cat#A02039) were validated and measured on an automated enzyme-linked immunosorbent assay platform 23 .Binding antibody levels were reported as relative ratios to a synthetic standard included as a calibration curve on each assay plate.Thresholds for binding antibody seroconversion were determined by aggregating data from negative controls and calculating the mean +3 standard deviations (SD) and were measured at a 1:160 (0.0625 µL/well) dilution.Negative controls were obtained from pre-COVID-19 pandemic sera, blanks, and commercially purified IgG.The threshold for seropositivity was obtained from the mean of the log distributions of the controls which was found to result in a specificity for RBD of 100%, a sensitivity of 89%, while spike had a sensitivity of 99% and specificity of 94%.Seroconversion thresholds were 0.186, 0.19, and 0.164 for anti-RBD, anti-spike, and anti-nucleocapsid antibodies, respectively.

Statistical analysis
Descriptive statistics were used to describe baseline characteristics.Neutralizing antibodies were analyzed through linear mixed-effects models with random intercepts and fixed covariates adjusting for vaccine time point, bivalent vaccine type, number of COVID-19 vaccine doses (four versus five), kidney transplant recipients versus dialysis patients, and anti-nucleocapsid antibody status.In a sensitivity analysis, we additionally adjusted for initial two-dose vaccine and third-dose vaccine type to account for differences in immunogenicity between monovalent COVID-19 vaccines.With two-sided testing, P = 0.05 was considered statistically significant.We performed all analyses using R version 4.0.5 (R Project for Statistical Computing).

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
The serologic data generated in this study are openly available through the publicly accessible Borealis repository (https://doi.org/10.5683/SP3/9XUY6O).The patient-specific clinical data are available under restricted access through the publicly accessible CITF Databank.The source data generated in this study for figures are provided in the Supplementary Information/Source Data file provided with this paper.

Table 1 |
Baseline characteristics by bivalent vaccine type