2.1 Study design

A convenience sample of solid organ (liver, heart, kidney, lung, or liver-intestine) transplant recipients aged 5–17yo was recruited via clinic communications during May 2020–April 2022 after vaccination with ≥2 doses of mRNA vaccine. Recipients were excluded after receiving any SARS-CoV-2-specific convalescent plasma or monoclonal antibody. Clinical data, including age, sex, date of transplantation, dates of vaccine doses, and maintenance immunosuppressant medications, were collected. History of COVID-19 was established based on chart review and SARS-CoV-2 polymerase chain reaction (PCR) results available through Stanford's electronic medical records (EMR) system or through Care Everywhere, a regional shared EMR network. National Institutes of Health criteria were used to determine severity of disease.¹⁷ In subjects transplanted within the 12 months prior to vaccination, induction immunosuppressive agents were recorded. When available, immunosuppressant levels drawn within 2 weeks of vaccine-response testing were collected.

SARS-CoV-2-specific immunoglobulin G (IgG), RBD-ACE2 blocking Ab, and SARS-CoV-2-specific interferon (IFN)-gamma(𝛾) release assay (IGRA) were performed on subjects' serum samples ≥2 weeks after V2 and/or V3. For individual subjects with quantitative data after both V2 and V3, change in RBD ACE-2 blocking Ab and change in IFN-γ response were defined as the subject's absolute change in quantitative response between V2 and V3.

2.2 Laboratory testing

Anti-SARS-CoV-2 Spike S1 domain IgG ELISA was performed on lithium heparin plasma using the EUROIMMUN instrument and reagents (Lübeck, Germany) per the manufacturer's instructions. The anti-SARS-CoV-2 IgG antibody level was reported as the ratio of optical density (OD) of the sample over the OD of the calibrator. The ratio was defined as follows: <0.8 negative; ≥0.8 to <1.1 borderline; ≥1.1 positive. RBD-ACE2 blocking Ab was measured in subjects with positive IgG via an RBD-ACE2 competition ELISA assay.¹³ Plasma samples were incubated in wells coated with SARS-CoV-2 spike RBD protein prior to the addition of ACE2-mFc that binds to any open RBD sites. Horseradish peroxidase-conjugated goat anti-mouse IgG (Invitrogen) was then added with subsequent OD measurement (Molecular Devices), with higher measurements indicating more open RBD sites and reduced anti-RBD antibodies in the plasma. OD values were then converted to a range of percentages to reflect blocking activity.

Stanford's laboratory-developed IGRA¹⁸ was performed as follows: freshly collected blood (>3 ml) in a lithium heparin tube was (1) left unstimulated as negative control, stimulated with (2) a single SARS-CoV-2 peptide pool (Miltenyi Biotec, Bergisch Gladbach, Germany) for CD4+ T cells and (3) two SARS-CoV-2 peptide pools for CD8+ T cells, and (4) stimulated with mitogen as positive control. A single CD4+ T cell megapool (CD4+ pool) consisted of 221 predicted HLA class II CD4+ T-cell epitope peptides covering the entire viral proteome except for the spike protein which was covered with 253 15-mer peptides overlapping by 10 residues. Two CD8+ T cell megapools (CD8+ pools A and B) together consisted of 628 predicted HLA class I CD8+ T-cell epitopes from the entire SARS-CoV-2 proteome. After incubation for 24 hours, IFN-γ concentration in the plasma fraction was measured with an automated ELISA instrument in international units (IU) per ml. Positive IFN-γ response was defined as: Nil ≤8.0 and SARS-COV-2 Antigen minus Nil >0.35. Negative IFN- γ response was defined as: Nil≤8.0 and SARS-COV-2 Antigen minus Nil<0.35 and Mitogen-Nil ≥ 0.5.

2.3 Statistical analysis

Fisher's exact test was performed on categorical variables. Welch's T-test and the Mann–Whitney U-test were performed for comparison of normally distributed continuous variables and non-normally distributed continuous variables, respectively. Kruskal–Wallis one-way analysis of variance was used to compare IFN-γ response and RBD-ACE2 blocking Ab levels between groups or between time points. Paired testing was used for subgroup analyses.

The study was approved by Stanford University's institutional review board (IRB #59811). Informed consent was obtained in person, verbally, or via a RedCap electronic consent form; the Stanford IRB reviewed and approved the verbal consent process and electronic consent form.

3.1 Baseline characteristics

A total of 33 subjects were enrolled, including 17 liver, 7 kidney, 7 heart, 1 lung, and 1 liver/intestine transplant recipients. Median age at first dose was 15 years (IQR 12–16 years), and 4 subjects were under 12 years old. In all, 19 (57.6%) were males. All subjects received only BNT162b2 vaccine. Median time from transplantation to vaccination was 7.0 years (IQR 1.5–13.5 years), with 6 subjects (18.2%) vaccinated within the first year after transplantation. In all, 31 (93.9%) subjects were on tacrolimus, and 13 (39.4%) were also on mycophenolate. All six subjects vaccinated within the first year after transplantation received anti-thymocyte globulin as induction immunosuppression, and two of the six (33%) also received methylprednisolone. Two subjects were diagnosed with SARS-CoV-2 infection via nasopharyngeal PCR prior to vaccination.

3.2 Humoral and cell-mediated responses after vaccination

Of the 33 enrolled subjects, 25 underwent laboratory testing after V2, 21 underwent testing after V3, and 13 underwent testing after both V2 and V3 (Table 1). After V2, all subjects had IgG testing, and after V3, 19 subjects had IgG testing and 2 had IGRA testing only. Laboratory testing was performed an average of 9.24 weeks (SD 4.76) after V2 and 9.29 weeks (SD 6.12) after V3.

Overall, a majority of subjects had positive IgG and IGRA responses after V2 and V3, with significant increases in quantitative responses from V2 to V3. Positive antibody responses improved from 21/25 (84.0%) after V2 to 19/19 (100.0%) after V3, and RBD-ACE2 blocking Ab demonstrated a significant increase from a mean blocking activity of 52.8% after V2 to 80.5% after V3 (p = .007, Figure 1). The proportion of positive IGRA responses increased from 17/23 (73.9%) after V2 to 16/17 (94.1%) after V3, although the increase in SARS-CoV-2-specific IFN-γ response values from a mean of 2.26 IU/ml after V2 to 4.10 IU/ml after V3 was not significant (p = .11, Figure 2). All subjects showed adequate response to the nonspecific mitogen control built into the IGRA, confirming good overall T-cell function.

In the six subjects who were <1 year from transplant at time of first vaccine dose, RBD-ACE2 blocking Ab increased from a mean of 2.0% to 85.0% from V2 to V3 (p < .001). Similarly, IFN-γ response increased from a mean of 0.80 IU/ml to 7.81 IU/ml from V2 to V3 (p < .01). In this subgroup, the average time between V2 and laboratory testing was 7.83 weeks (SD 3.44), and the average time between V3 and testing was 11.71 weeks (SD 8.38).

3.3 Immune response by modifying factors

Positive SARS-CoV-2 IgG and IGRA after V2 were not associated with antimetabolite use, number of immunosuppressants, type of organ, or age at vaccination. Because all subjects had a positive SARS-CoV-2 IgG after V3, Fischer's exact test could not be performed to detect differences between groups.

Clinical factors associated with the degree of quantitative response after V2 or V3 were also assessed. Following V2, liver transplant (vs. kidney transplant), antimetabolite use, and being more than 1 year from transplant (vs. <1 year from transplant) were associated with higher quantitative RBD-ACE2 blocking Ab (p < .01, p < .05, p < .05, respectively). There were no significant differences in RBD-ACE2 blocking Ab by age at vaccination or serum tacrolimus level. There were also no significant differences in the quantitative IFN-γ response values by these modifying factors after V2. Following V3, there were no significant differences in the quantitative RBD-ACE2 blocking Ab or IFN-γ response values by organ type, immunosuppressant regimen, age at vaccination, time since transplant, or serum tacrolimus level.

3.4 Change in response by modifying factors

Of the 13 subjects who had testing after both V2 and V3, nine subjects had quantitative response measurements at both timepoints. In subgroup analyses for these nine subjects, the average change in RBD-ACE2 blocking Ab was greater in kidney transplant recipients (62.0%) than in liver transplant recipients (17.5%), though this benefit did not reach significance (p = .06). Neither immunosuppressant regimen nor tacrolimus level was associated with change in RBD-ACE2 blocking Ab or change in IFN-γ response. Following V3, subjects <1 year from transplant saw a greater change in RBD-ACE2 blocking Ab response (73.3%), compared to subjects ≥1 year from transplant (26.7%) (p = .05).

3.5 Clinical outcomes

Seven subjects developed SARS-CoV-2 infection after vaccination (five after V3 and two after V2). Of these breakthrough cases, two were asymptomatic, three had mild illness, one had severe illness, and one did not have enough clinical information available via EMR to determine disease severity. No subjects required hospitalization. All but one breakthrough case occurred in December 2021 or later, and median time after SOT was 6 years (IQR 1–14 years). One study subject died during the study period, though not due to SARS-CoV-2 infection; this subject had received two doses of vaccine.