A total of 326 patients with BA prior to liver transplantation were recruited from January 2018 to December 2018 at the Department of Liver Surgery of Renji Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China). The exclusion criteria were as follows: foreign patients (n = 15) and those with missing vaccination records (n = 23). A total of 288 patients with detailed vaccination records were assessed. Eleven patients never received the HBV vaccine before liver transplantation, and 57 patients had at least one of the following markers positive before cohort entry: HBsAg/HBeAg/HBeAb/HBcAb ( Figure 1 ). A total of 220 patients received at least one dose of the HBV vaccine. Their demographic data, clinical history, physical examination, and laboratory test results (i.e., HBsAb titers and bile acids levels) were all extracted from their medical records during the first visit.

For the flow cytometric analysis of CD4⁺ and B cell subsets in peripheral blood mononuclear cells (PBMCs), blood samples were collected from 49 BA children aged 4-21 months who met the inclusion criteria: (1) daily dose of glucocorticoid for children ≥ 10 kg not exceeding 20 mg/d; (2) daily dose of glucocorticoid for children weighing less than 10 kg not exceeding 2 mg/kg/d; and (3) the child had taken glucocorticoids exceeding the above dose for no more than two weeks.

HBV marker tests (HBsAg, HBsAb, HBeAg, HBeAb and HBcAb) in the serum of patients with BA before transplant were performed by ELISA (Genesis RMP150; Tecan Group, Zurich, Switzerland) following the manufacturer’s protocol. Children were classified as nonresponders or responders based on HBsAb levels: nonresponders (HBsAb levels < 10 mIU/mL), low/medium responders (HBsAb levels 10–200 mIU/mL), and high responders (HBsAb levels ≥ 200 mIU/mL).

Blood samples of 49 children with BA were collected in BD Vacutainer^® Blood Collection Tubes (BD). PBMCs were isolated by density gradient centrifugation with Ficoll-Paque (GE Healthcare) following the manufacturer’s protocol. In brief, blood samples were diluted with 0.9% NaCl (1:1) and gently loaded to the Ficoll layer, followed by density gradient centrifugation (500 g for 20 min at 20°C, acc/dec: 3/2). The mononuclear cell layer (intermediate white layer) was isolated, collected in a new tube, and rinsed with 0.9% NaCl. The cells were stored in liquid nitrogen or used for further assays.

Isolated PBMCs were incubated with the following fluorochrome-conjugated monoclonal antibodies: CD3 (HIT3a), CD4 (OKT4), CD19 (SJ25C1), CD20 (2H7), CD25 (BC96), CD27 (M-T271), CD38 (HIT2), CD45RA (HI100), CD127 (A019D5), CCR6 (G034E3), CXCR3 (G025H7), IgG (G18-145), IgM (CD38), and IgD (IA6-2). Dead cells were stained with Zombie Aqua (Biolegend). After staining at room temperature for 30 min, the cells were resuspended and examined with a FACS analyzer (Fortessa X-20, BD) and analyzed with FlowJo software (BD).

The flow cytometric gating strategy was slightly modified with respect to our previously reported protocol (36). CD4⁺ T cells (CD3⁺CD4⁺), CD19⁺ B cells (CD3^-CD19⁺), Th1 cells (CD3⁺CD4⁺CXCR3⁺CCR6^-), Th2 cells (CD3⁺CD4⁺CXCR3^-CCR6^-), Th17 cells (CD3⁺CD4⁺CXCR3^-CCR6⁺), Treg cells (CD3⁺CD4⁺CD25⁺CD127^-), CD19⁺CD27⁺IgG⁺ post-class-switched memory B cells, naive B cells (CD3^-CD19⁺IgD⁺CD27^-), class-switched memory B cells (CD3^-CD19⁺IgD^-CD27⁺), non-class-switched memory B cells (CD3^-CD19⁺IgD⁺CD27⁺), double-negative memory B cells (CD3^-CD19⁺IgD^-CD27^-) and plasma B cells (CD3^-CD19⁺IgD^-CD27⁺IgM^-CD20^-CD38⁺).

All data were analyzed using SPSS (version 22) or GraphPad Prism (version 8). Continuous variables are displayed as medians (IQRs) and analyzed using two-sample t tests or Wilcoxon rank-sum tests; categorical variables are determined as percentages and analyzed using Fisher’s exact test. Multiple logistic regression models were used to calculate ORs and corresponding 95% CIs to analyze the associations between patients with TBA ≥ 200 μmol/L and TBA < 200 μmol/L, adjusting for dose of vaccination, weight, height, Kasai surgery, TBIL and CHOL. All statistical tests were 2-tailed, and a p-value < 0.05 was considered statistically significant.

The study was approved by the Ethical Committee of Renji Hospital, School of Medicine, Shanghai Jiao Tong University [(2017)209]. All patients were provided with written informed consent.

A total of 326 patients with BA prior to liver transplantation were recruited from January 2018 to December 2018 at the Department of Liver Surgery of Renji Hospital, Shanghai Jiao Tong University School of Medicine (Shanghai, China). Fifteen foreign patients and 23 patients without vaccination records were excluded. A total of 288 patients with detailed vaccination records were assessed. The median age was 7 months (range: 4 months - 159 months). From birth to liver transplantation, 277 (96.2%) patients received at least one dose of the HBV vaccine. The vaccination rates for other vaccines included in the China National Immunization Program were also analyzed. The vaccination rates for the HBV (96.2%, 277/288) and bacillus Calmette-Guérin (BCG) (84.7%, 244/288) were much higher than the rates of meningococcal serogroup A/C (Men AC) (41.2%, 7/17), measles/mumps/rubella (MMR) vaccine (31.3%, 15/48), poliomyelitis (polio) (25.3%, 73/288), hepatitis A (HAV) (25.0%, 12/48), Japanese encephalitis (JE) (15.0%, 18/120), diphtheria/tetanus/pertussis (DTP) (14.2%, 41/288), meningococcal serogroup A (Men A) (13.5%, 29/215), and varicella (VAR) vaccines (10.8%; 7/65) ( Table 1 ). These results show that vaccination coverage in young patients with BA prior to liver transplantation are very low compare to the overall vaccination coverage of over 95% in China (37).

Given that one of the greatest risks for young children with BA is HBV infection, we focused on the analysis of HBV vaccination coverage in the BA pediatric population. Among 288 patients with vaccination records, 11 (3.8%) did not receive the timely birth-dose vaccination, and 57 (19.8%) patients tested positive for at least one of the following markers: hepatitis B surface antigen (HBsAg)/hepatitis B e antigen (HBeAg)/hepatitis B e antibody (HBeAb)/hepatitis B core antibody (HBcAb), revealing a very high HBV infection rate (19.8%). Even more importantly, the level of HBV was twofold higher than the one previously reported from the biopsy of tissues in infants with biliary atresia (HBV infection rate was 9%) (38).

The median age of the 220 patients who received at least one dose of the vaccine was 7 months (range: 4 months - 131 months); 84 (38.2%) were male, and 136 (61.8%) were female. A total of 113 (51.4%) patients received one dose of the HBV vaccine, 85 (38.6%) received two doses, and only 22 (10%) completed three doses of the HBV vaccine according to the China’s National Immunization Program ( Table 2 ), demonstrating that the full coverage of the HBV vaccine in young patients with BA is extremely low. Seventy-five (34.1%) patients were nonresponders (HBsAb titer lower than 10 mIU/mL), 94 (42.7%) patients were low/medium responders (HBsAb titer between 10 and 200 mIU/mL), and 51 (23.2%) patients were high responders (HBsAb higher than 200 mIU/mL). The proportion of high responders was significantly increased in patients received two or three doses of the vaccine ( Table 3 ). Taken together, these results demonstrate the low rate of full-dose HBV vaccine coverage and a low ratio of high responders in young children with BA.

Various factors modulate the immune response to vaccination (27). To investigate the potential factors that may impact the antibody responses to the HBV vaccine, the patients were divided into two groups based on HBsAb titers: non/low/medium responders (HBsAb < 200 mIU/mL, n = 51) and high responders (≥ 200 mIU/mL, n = 169) ( Figure 2A ). Fisher’s exact test assay was applied to compare parameters, including body weight, height, and clinical parameters. There were no differences between the two groups according to demographic parameters, such as age, sex, weight, height and BMI. Next, laboratory biochemical parameters were compared. Levels of albumin (ALB), alanine aminotransferase (ALT), aspartate transaminase (AST), gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP), total bilirubin (TBIL), cholesterol (CHOL) and white blood cell, neutrophil, lymphocyte and monocyte counts were similar between the non/low/medium and high responders ( Table 2 ). Surprisingly, total bile acid (TBA) levels were dramatically reduced in the high responders compared to the non/low/medium responders, with a nearly 25% reduction (median: 144.4 µmol/L vs 188.8 µmol/L) (p < 0.01) ( Figure 2B ). The history of the Kasai surgery, vaccination dosage and time since vaccination were also compared between the two groups. We found that vaccination dosage and time since vaccination were significantly different between the non/low/medium and high responders ( Table 2 ). Altogether, these assays reveal that serum bile acids are decreased in responders with high HBsAb titers.

To further analyze the association of serum bile acid levels with HBsAb titers in non/low/medium and high responders, the patients were subdivided into low (TBA < 200 µmol/L) and high total serum bile acids (TBA ≥ 200 µmol/L). There were 127 (57.7%) patients with serum TBA < 200 µmol/L, with a median value of 139.3 μmol/L. Ninety-three patients had TBA ≥ 200 µmol/L (median value: 279.3 μmol/L), which was 2-fold higher than the patients with low serum TBA. There were no significant differences between the high and low serum TBA levels according to demographic parameters, such as age, sex, weight, height and BMI. In addition, the laboratory biochemical parameters ALB, ALT, AST, GGT, ALP, TBIL, white blood cells, neutrophils, lymphocytes and monocytes in patients with low and high serum TBA were similar. Remarkably, only CHOL levels in the low TBA patients (median: 4.3 mmol/L) were significantly lower than those in the high TBA patients (median: 5.9 mmol/L) (p < 0.01) ( Table 4 ).

Among the 127 patients with low serum TBA, 88 were non/low/medium responders with HBsAb titers < 200 mIU/mL (median: 12.7 mIU/mL). Only 39 patients were high responders (median: 830.2 mIU/mL). Among the 93 patients with high serum TBA, 81 were non/low/medium responders with HBsAb titers < 200 mIU/mL (median: 16.67 mIU/mL), and only 12 patients were high responders (median: 333.2 mIU/mL). The ratio of high responders/non/low/medium responders (39/88) in low serum TBA patients was 3.0-fold higher than the ratio of high responders/non/low/medium responders (12/81) in high TBA patients ( Figure 2C ). Logistic regression analysis after adjustment for covariates (weight, height, vaccine dose, Kasai surgery, TBIL and CHOL) ( Table 4 ) showed that only high serum TBA levels were associated with non/low/medium responders to the HBV vaccine (OR = 0.394; 95% CI = 0.170-0.969, p = 0.03) ( Table 5 ). In summary, the patients with high serum bile acid levels were associated with poor HBV vaccine responses.

Bile acids play vital roles in regulating immune responses, such as natural killer cells, macrophages, dendritic cells, B cells and CD4⁺ T cells (31, 34, 35, 39). CD4⁺ T cells contribute to the antibody response during vaccination, and bile acids selectively promote regulatory T cells and inhibit Th17 cells and IL-22-producing cells (34, 40). To analyze CD4⁺ T cell profiling in young children with BA, 49 young patients with BA were recruited. Circulating CD4⁺ T cell subsets in human peripheral blood mononuclear cells (PBMCs) were analyzed by FACS. Contrary to previous findings that bile acids selectively modulate Treg and Th17 cell responses (34), there were no correlations between CD4⁺ T cell ratios (r² = 0.00, p = 0.93), Treg (r² = 0.01, p = 0.51), Th1 (r² = 0.01, p = 0.32), Th2 (r² = 0.01, p = 0.60), Th17 cells (r² = 0.03, p = 0.24) and serum total bile acid levels in young children with BA ( Figure 3 ).

B cells are antibody-producing cells and play crucial roles in vaccine efficacy (41, 42). There was no correlation between total CD3-CD19⁺ B cells and total BA (r² = 0.00, p = 0.75). TBA levels did not correlate with naive B cells (r² = 0.00, p = 0.75), class-switched memory B cells (r² = 0.02, p = 0.37), non-class-switched memory B cells (r² = 0.01, p = 0.61), double-negative memory B cells (r² = 0.01, p = 0.66) or plasma cells (r² = 0.00, p = 0.91) ( Figures 4A–G ). Notably, TBA levels were negatively correlated with CD19⁺CD27⁺IgG⁺ post-class-switched memory B cells (r² = 0.13, p = 0.01) ( Figure 4H ) and the median percentage of CD19⁺CD27⁺IgG⁺ cells in Low TBA group was 0.77%, which was much higher than 0.42% in High TBA group ( Figure 4I ), indicating that bile acids selectively regulate B cell subsets, specifically inhibiting post-class-switched memory B cells. These evidences show that serum BA levels are negatively correlated with CD19⁺CD27⁺IgG⁺ post-class-switched memory B cells.