We included all kTx recipients aged over 12 years old followed in one of the three Pediatric Nephrology Departments in Paris (Robert Debré Hospital, Necker Hospital and Trousseau Hospital) who were vaccinated against SARS-CoV-2 with the Pfizer SARS-CoV-2 mRNA BNT162b2 vaccine between 30 January 2021 and 21 December 2021. French health authorities approved vaccination in children with comorbidities more than 16 years old on 20 January 2021 and extended it to children aged 12–15 years old on 01 June 2021. Specific guidelines in adult patients with SOT recommended three injections of mRNA vaccine but no specific pediatric guidelines were available. Therefore, the vaccination strategy was left to the treating physician’s decision with some performing three injections systematically and others only in patients with low anti-S IgG 1 month after the second injection. Patients with a proven (positive SARS-COV2 PCR or home-antigen test) natural infection prior to vaccination only received 2 doses of vaccine (Figure 1A). All centers evaluated patients’ humoral and cellular responses. Blood samples were collected between 21 and 90 days after vaccine injection and processed immediately in a centralized laboratory (Immunology department, Robert Debré Hospital). Clinical and biological data were collected retrospectively. In order to analyze the effect of COVID-19 infection on vaccination, patients were considered as having had an infection if they had a positive PCR, a positive anti-N serology or a positive anti-N T-cell response. The study was approved by Robert Debré Hospital Ethics committee.

Anti-Spike SARS-CoV-2 antibody levels were evaluated by chemiluminescent immunoassay in plasma using the LIAISON^® SARS-CoV-2 TrimericS IgG kit according to manufacturer recommendations (Diasorin^R). A serological positive response was defined as anti-S IgG response >33.8 BAU/mL. We also present results based on a higher cut-off 264 BAU/mL. This Ab level was found associated with 80% of vaccine efficacy against primary symptomatic COVID-19 (264 BAU/mL) in previous studies [14]. In some patients, serological response after the 2nd vaccine injection was tested in an outside laboratory. These results were collected and used to classify patients in responders and non-responders, but these patients were excluded from the comparison of Ab titers. Anti-Nucleocapsid SARS-CoV-2 IgG levels were evaluated by chemiluminescent immunoassay in plasma using the Alinity I^R anti-N SARS-CoV-2 IgG kit according to manufacturer recommendations (Abbott^R). A serological positive response was defined as anti-N IgG index response >1.4 (Figure 1B).

Neutralizing antibodies were quantified using the GenScript^R SARS-CoV-2 surrogate Virus Neutralization Test (sVNT). Briefly, the kit detects the ability of antibodies in the plasma of patients to block the interaction between the HRP-conjugated SARS-CoV-2 RBD fragment (HRP-RBD) and the human ACE2 protein (hACE2). Circulating neutralizing antibodies form with HRP-RBD a complex that get removed during washing. Unbound HRP-RBD is captured on a hACE2 pre-coated plate and reacts with the added TMB by changing the color of the solution. The absorbance is inversely dependent on the level of cirulating neutralizing antibodies. HRB-RBD protein used in the kit is selected based on the strain of SARS-CoV-2 tested (wild-type, Delta or Omicron). A positive neutralizing antibody activity is defined as more than 30% according to manufacturer recommendations. Serum from one of the 32 kTx children followed after two vaccine doses was not available for measuring the neutralizing antibody response against SARS-CoV-2 strains (Figure 1B).

Human peripheral blood mono-nucleated cells (PBMCs) were isolated from fresh blood samples by density gradient centrifugation (Leucosep) to obtain a final cell concentration of 2.5 10⁶ PBMCs/mL in AIM-V-Medium (Ficher Scientific, Suisse). Anti-Spike and anti-Nucleocapsid cellular responses were evaluated on fresh cells using the T-spot COVID ELIspot system from Oxford Immunotec^R (United Kingdom). Briefly, 50 µL of PHA (positive control), AIM-V-Medium (negative control), Spike-antigen mix and Nucleocapsid-antigen mix were added in wells of anti-IFNγ Abs pre-coated plate. 100 μL of the diluted cell suspension were added in each well and the plate was incubated for 16–20 h at 37°C, 5% CO₂. After discarding supernatant, wells were washed three times with PBS. Afterwards, 50 µL of conjugate reagent (anti-IFNγ Abs, conjugated to alkaline phosphatase) were added to each well and incubated for 1 hour at 4°C in the dark. After washing, 50 µL of substrate solution was added to each well at room temperature for 7 min leading to form insoluble precipitate at the site of reaction. Then, the plate was washed and dried. T-spots were counted by an Elispot-reader (Bioreader^R 6000-E Biosys, Germany) and results expressed as SFC (Spot Forming Cell)/250,000 PBMCs. Results were positive for a specific antigen if negative control was ≤10 spots, positive control ≥20 spots and the Antigen-Mix >4 spots per well (according to manufacturer recommendations) (Figure 1B).

Patient characteristics and immunological response to vaccine are presented as medians and interquartile ranges for continuous variables, and counts and percent for categorical variables. Chi-square test and Mann-Whitney U test were used to compare characteristics between the groups for categorical and continuous variables, respectively. A significant statistical difference was assumed when the p-value was <0.05. All analyses were conducted using GraphPad PRISM version 5.00.288 (GraphPad Software Inc., San Diego, CA, United States) and SAS 9.1. The reporting of the data followed the STROBE statement.

Table 1 describes patients’ and transplants’ characteristics at the time of the first vaccine injection. 64 patients aged 16.9 years (14.9; 17.6) were included, 49 of whom received their transplant from a deceased donor (76.5%). Patients were 56% male; the first causes of ESKD were urological abnormalities in 31.2%, hereditary nephropathies in 37% and glomerular diseases in 19% of the patients. Fifteen patients were transplanted preemptively and 49 patients were transplanted after a median time on dialysis of 1.56 years (1.11; 2.69) (37 in hemodialysis and 12 in peritoneal dialysis). In maintenance oral immunosuppressive treatment, 36 (56%) patients received an association of tacrolimus with mycophenolate mofetil (MMF) or mycophenolic acid (MPA), 18 (28.1%) tacrolimus with azathioprine and 35 (54.7%) with low doses steroids (median 5 mg) as a third immunosuppressive treatment. At first vaccine, 8 patients had lymphopenia <1,500/mm³, 3 had hypogammaglobulinemia < 5 g/L and 26 (40%) had eGFR <60 mL/min.

All patients were vaccinated with the Pfizer SARS-CoV-2 mRNA BNT162b2 vaccine. Twenty patients received only 2 doses of vaccine including 11 with a known previous natural infection with SARS-CoV-2 (median time from infection to first vaccine injection 132 days IQR [81; 291], and 44 (62.5%) patients without known history of natural infection received three doses (Figure 1B).

Anti-spike IgG antibodies (anti-S Abs) were quantified in the plasma from 32 kTx children after two doses of vaccine (Figure 2A). 87.5% (28/32) had a positive response (defined as ≥33.8 BAU/mL according to the manufacturer) with a median antibody titer in responders at 1825 (IQR, 637–4,883) BAU/mL. The majority (81.3%, 26/32) had an antibody titer above 264 BAU/mL, Ab levels associated with 80% of vaccine efficacy [14]. KTx children after two vaccine doses were then classified according to their history of natural infection; based either on positive SARS-COV2 PCR or home-antigen test, or on positive anti-N humoral or cellular response. Fourteen of the 32 kTx children evaluated had a history of natural SARS CoV-2 infection before their 2nd vaccine dose and all of them (14/14) had a positive humoral response above the 264 BAU/mL cut-off after vaccination. In comparison, only 77.8% (14/18) of children without previous natural infection had a positive humoral response (p = 0.059) and 66.7% (12/18) reaching the 264 BAU/mL cut-off (p = 0.017). Among responders, anti-S antibody titers were significantly higher in children with natural infection [median: 3,265 (IQR, 1,492–8,178) BAU/mL] than in children without previous natural infection [median: 1,107 (IQR, 593–2,658) BAU/mL] (p = 0.007).

Anti-S Abs were quantified in the plasma for 40 kTx children after three vaccine doses (median time between 3rd dose and serology 39 days IQR (28; 72) (Figure 1A). Serological responses after the second dose of vaccine were determined only for 20 children (for 12 children in a laboratory that provides only qualitative results and for 8 in our central laboratory). Twenty children were only tested after their third dose (Figure 1A). Children were classified in three groups according to their humoral response to the 2nd vaccine dose: Responder for positive response, Non-responder for negative response and Unclassified (NC) for children with missing data. As expected, all (100%, 8/8) responders presented a positive humoral response after their 3rd vaccine dose, with seven (87.5%) achieving Ab levels above the 264 BAU/mL cut-off. Median antibody titer after the 3rd dose in these patients was 1805 (IQR, 783–2,485) BAU/mL. In comparison, the 3rd vaccine dose led to a positive humoral response in 75% (9/12) of non-responders with a lower median Ab titer in responders at 355 (IQR, 140–3,865 BAU/mL) (p = 0.028) and only 41.7% of them reaching the 264 BAU/mL cut-off (Figure 2B).

As mentioned above, eight kTx children were tested centrally after the second and the third doses of vaccine including 3 patients with no response after the 2nd dose (Figure 2C). The median antibody titers increased from 159 (IQR, 5–1,458) BAU/mL after 2 vaccine doses to 1,150 (IQR, 201–2,108) BAU/mL after 3 vaccine doses (p = 0.085). Ab level above the 264 BAU/mL cut-off were achieved respectively for 5/8 (62.5%) of children after the 2nd dose and 6/8 (75%) children after the 3rd dose. Interestingly, two of the 3 non responders after the 2nd dose developed a positive humoral response after their 3rd vaccine dose, while responders did not show a significant increase in their anti-S IgG titers after the 3rd injection.

Neutralizing antibodies against wild-type, Delta and Omicron strains were quantified using a SARS-CoV-2 pseudo-neutralization Antibody Detection kit (ELISA) [15] in the serum from 31 kTx children after two doses of vaccine (Figure 3A). The positive response (defined as ≥30% according to the manufacturer) against wild-type, Delta and Omicron strains was respectively achieved for 25/31 (80.6%), 24/31 (77.4%) and 18/31 (58.0%) of kTx children. The median neutralizing antibody activity decreased from 95.3% (IQR, 91–96.9) against the wild-type strain to 69.2% (IQR, 53.8–93.5) against Omicron (p = 0.003). Among children with a positive neutralizing antibody response, 72% of them had antibodies able to neutralize the three SARS-CoV-2 strains, 24% both the wild-type strain and Delta, and 4% the wild-type strain only (data not shown). All the children with a history of natural SARS-CoV-2 infection before their 2nd vaccine dose (14/14) had a positive neutralizing antibody response against the wild-type strain (Figure 3A, right panel). In comparison, only 64.7% (11/17) of children without previous natural infection had a positive neutralizing response (p = 0.008). In agreement, median of neutralizing antibody activity against wild-type strain was significantly stronger in children with natural infection [95.9% (IQR, 93.5–97.3)] than in children without previous natural infection [90.2% (IQR, −0.4–95.9)] (p = 0.008). Similar significant differences were observed for neutralizing antibody activity against Delta (p = 0.0026) and Omicron strains (p = 0.0002). It is however interesting to note that 12/14 (85.7%) of children with natural infection maintained a neutralizing response against Omicron instead of 6/17 (35.3%) of children without previous natural infection.

Neutralizing antibodies against wild-type, Delta and Omicron strains were quantified in the serum for 40 kTx children after three vaccine doses (Figure 3B). 75% (30/40) had a positive response against the wild-type strain. The percentage of children with positive response decreased with variants, reaching 40% (16/40) of children with Omicron (p = 0.0015). Similarly, the median neutralizing antibody activity decreased form 93.4% (60.9–96.7) against the wild-type strain to 69.6% (41.8–85.5) against Omicron (p = 0.0003). All (100%, 8/8) children with humoral response to the 2nd vaccine dose presented a positive neutralizing response after their 3rd vaccine dose against the wild-type strain with the median of neutralizing activity at 94.5% (IQR, 75.3–96.9) (Figure 3A, right panel). In comparison, the 3rd vaccine dose led to a neutralizing activity in 67.7% (8/12) of children classified non-responders after the 2nd dose, with a lower median at 73.7% (IQR, 42.7–95.4) (p = 0.018). Frequency of positive neutralizing response and median neutralizing activity levels decreased from the wild-type strain to Delta, then to Omicron. Only 4/8 (50%) of responders and 3/12 (25%) of non-responders presented a positive neutralizing response against Omicron after their third dose.

As mentioned above, only eight kTx children were centrally evaluated after both the second and the third doses. Among five patients with no neutralizing activity against the wild-type strain after the 2nd dose (Figure 3C), two developed a positive response after their 3rd vaccine dose. The median antibody titers increased from 3.2 (IQR, −5.8–85.8) % after 2 vaccine doses to 89.5 (IQR, 21.6–96.7) %L after 3 vaccine doses (p = 0.085). Interestingly, the neutralizing activity against Omicron was not improved after the 3rd injection.

Anti-S antibody levels strongly correlated with neutralizing antibody activity for the wild-type strain (Figure 4A), Delta (Figure 4B) and Omicron (Figure 4C). However, the percentage of children with a positive anti-S antibody response (≥33.8 BAU/mL) without a positive neutralizing activity (cut off 30%) increased from 6/71 for wild-type strain, to 11/71 for Delta, then 27/71 for Omicron. Interestingly, level of anti-S antibody in children without neutralizing activity against the wild type strain were all below the 264 BAU/mL cut-off associated with 80% of vaccine efficacy [14]. Conversely, some patients with high anti-S antibody titers showed no neutralizing activity against the VOCs despite antibody titers up to 736 and 1,690 BAU/mL for delta and omicron variants, respectively.

Spike-specific T cells were quantified in 28 kTx children after two vaccine doses and in 23 kTx children after three vaccine doses. Specific spike memory T cell were observed with a median of 6.5 (IQR, 2.0–32) SFC in children after two doses of vaccine. Medians of Spike-specific T cell response were, respectively, 13.5 (IQR, 2.3–51.8) SFC in the 12 children with previous natural infection, and 5 (IQR, 1–8.8) SFC in the 16 children without previous natural infection. Specific memory T cell were observed with a median of 7 (IQR, 1.0–23) SFC in children after three doses of vaccine. Median of Spike-specific T cell response tended to be higher in children with a humoral response after 2 doses (10 (IQR, 1–36) SFC) as compared to non-responders (3 (IQR, 1.5–145.5) SFC). Interestingly, children with the higher level of Spike-specific T cell response in the non-responder group had a natural infection between the 2nd and the third dose (Figures 5A,B).

Among 51 patients tested for both anti-Spike humoral and Spike-specific T cell responses, none had a spike-specific cellular response without a positive humoral response, whereas 9/51 had a positive humoral response without spike-specific cellular response detected (Figure 5C).