Distinct features of humoral and cellular immunity against Omicron breakthrough infection among chronic liver disease patients: An experience from a follow-up cohort

ABSTRACT

Immunity response following vaccination and infection toward the SARS-CoV-2 Omicron variant remains limited in chronic liver disease (CLD) population. This study aims to investigate humoral and cell-mediate immunity to Omicron BA.4/5 among CLD patients with/without breakthrough infections (BI). This longitudinal study characterized immune responses to Omicron infection across three CLD patient groups: vaccinated without BI (group 1, n = 10), vaccinated with BI (group 2, n = 43), and unvaccinated infected (group 3, n = 7). SARS-CoV-2-specific antibodies/T-cell responses and lymphocyte phenotypes were quantified by ELISA and flow cytometry. Logistic regression identified immunological correlates of BI event, viral clearance and liver disease progression. Dynamic immunity changes after infection analyzed using Spearman’s correlation. Patients in group 2 exhibited the highest levels of neutralizing antibodies (NAbs) against wildtype (WT) and Omicron BA.4.5, anti-RBD IgG, and total anti-SARS-CoV-2 antibodies compared to groups 1 and 3. Comparable T cell responses were observed between two vaccinated groups, stronger than unvaccinated individuals. Hybrid immune (vaccination plus infection) elicited stronger NAbs-WT, total anti-SARS-CoV-2 antibodies, anti-RBD IgG, MBCs, and Omicron-specific IFN-γ⁺/IL-4⁺CD4⁺T cells compared to primary Omicron infection. Besides, severe liver disease was not only a risk factor for BI, but also impaired Omicron-specific IFN-γ⁺CD8⁺T, CD38⁺CD4⁺T, and CD45RA⁺CD8⁺T cells responses after infection. Importantly, IgG⁺/IgM⁺MBCs and NAbs showed positive associations with central memory (CD45RA⁻CD62L⁺CD4/8) or naive (CD45RA⁺CD62L⁺CD4) T cells. Low levels of CD45RA⁺CD62L⁺CD8⁺T cells and resting MBCs were identified as the independent risk factors for viral clearance and disease progression, respectively. BI predominantly stimulates humoral immunity without enhancing cellular responses. Weaker immune response induced by primary Omicron infection underscores the need for additional vaccinations in vaccine-naïve CLD patients. The positive correlations between humoral/cellular immunity and memory/activated T cells reveal their coordinated role in adaptive immune defense against SARS-CoV-2. These findings emphasize the importance of regular immune monitoring and timely interventions to optimize protection for CLD patients against emerging Omicron subvariants.

Introduction

The Omicron lineage of SARS-CoV-2 harbors the highest number of mutations in the spike protein, conferring a notable capacity for antibody escape.^1,2 Patients with chronic liver disease (CLD) exhibit diminished immunogenic responses to SARS-CoV-2 vaccines and reduced vaccine efficacy compared to the general population.^3-,6 Consequently, this population is at increased risk of breakthrough infections (BI) post-vaccination and advanced clinical outcomes following infection, particularly among individuals with severe liver disease (SLD) such as cirrhosis.⁷

Several studies have demonstrated that SARS-CoV-2-specific antibody and T cell responses are associated with a reduced risk of BI or symptomatic infection post-vaccination in healthy individuals.^8-,11 Notably, among solid organ transplant recipients, a medium titer of anti-receptor binding domain (RBD) antibodies has been identified as a risk factor for BI compared to high titers.¹² Furthermore, liver transplant recipients and patients with cirrhosis who developed hybrid immunity (vaccination combined with infection) exhibited significantly higher anti-spike antibody titers than uninfected individuals with an equivalent number of vaccine-induced spike antigen exposures.¹³ Convalescent individuals also mount more robust and broader immune responses after hybrid immunity compared to naïve-vaccinated individuals.¹⁴ However, a comprehensive longitudinal analysis of vaccine‐induced humoral and cellular responses in CLD patients, with or without BI, remains lacking.

To address this gap, we conducted a follow-up study on a CLD cohort with pre-infection immunological baseline data. The study aimed to: (i) compare SARS-CoV-2-specific humoral and cellular immune responses in vaccinated and unvaccinated CLD patients with or without Omicron variant infection, (ii) analyze the interaction between humoral and cellular immunity, and (iii) determine independent immunological factors associated with BI, viral clearance and disease progression.

Methods

Patients enrolled and study design

This study was conducted at the Fifth Hospital of Shijiazhuang between December 2022 and May 2023, as part of a follow-up cohort for vaccine immunogenicity in CLD patients.³ The inclusion criteria were: (1) adults ( >18 years old) with clinically or pathologically confirmed CLD; (2) receipt of inactivated SARS-CoV-2 vaccines or no vaccination; (3) first-time SARS-CoV-2 infection confirmed by RT-PCR or rapid antigen testing of nasopharyngeal swabs; (4) availability of quantitative data on SARS-CoV-2-specific antibodies prior to infection; (5) completion of a background questionnaire documenting SARS-CoV-2 infection history, associated symptoms, vaccination status, and negative conversion time (NCT) of nucleic acids (Appendix 1 in Supplementary Data); and (6) voluntary participation in the study. Exclusion criteria included: (1) unclear test results or uncertain timing of infection; (2) inability to adhere to follow-up procedures or undergo standard blood collection; and (3) a history of repeated SARS-CoV-2 infection.

BI was defined as the first occurrence of SARS-CoV-2 infection in CLD patients after receiving the COVID-19 vaccination. Based on official reports from the Chinese Centers for Disease Control (https://www.chinacdc.cn/jkzt/crb/zl/szkb_11803/jszl_13141/202302/t20230208_263674.html), all convalescent individuals in this study were infected with the Omicron variant.

CLD patients were categorized into three groups based on their vaccination status and SARS-CoV-2 infection history: vaccinated SARS-CoV-2-naïve individuals (group 1, n = 10), vaccinated SARS-CoV-2 convalescent individuals (group 2, n = 43), and unvaccinated convalescent individuals (group 3, n = 7). Whole blood samples were collected from all participants to isolate plasma and peripheral blood mononuclear cells (PBMCs) to assess humoral and cellular immune responses (Figure 1).

Figure 1.

Study design.

Clinical characteristics and laboratory data were obtained from the Hospital Information Systems (Table S1).

SARS-CoV-2-specific antibody detection

In terms of humoral immunity, total anti-SARS-CoV-2 antibodies (including IgM, IgG, and IgA; Wantai Biological Pharmacy Enterprise Co.,Ltd., China), neutralizing antibodies (NAbs) against SARS-CoV-2 wild-type (WT) and Omicron BA.4/5 subvariant (GenScript cPass™, USA), together with anti-RBD IgG (Proprium Biotech Co, Ltd., China) levels in plasma were measured using commercial ELISA kits, according to the manufacturer’s instructions, as previously described.³ Antibody responses were evaluated in ten CLD patients of group 1, 43 of group 2, and seven of group 3. The detailed protocol of SARS-CoV-2-specific antibody arrays was provided in Supplementary Methods.

Cell-mediated immune responses

SARS-CoV-2-specific T cell responses were assessed using an intracellular cytokine staining (ICS) assay, with the following subsets analyzed: T helper (Th) cells (IFN-γ⁺CD4⁺CD3⁺[Th1] and IL-4⁺CD4⁺CD3⁺[Th2]) and cytotoxic T (Tc) cells (IFN-γ⁺CD8⁺CD3⁺[Tc1] and IL-4⁺CD8⁺CD3⁺[Tc2]). Specifically, PBMCs were stimulated with SARS-CoV-2 spike glycoprotein peptide (SPs, 2 μg/mL) or B.1.1.529-Omicron SPs (detailed in Supplementary Data) in RPMI 1640 medium at 37°C in a 5% CO₂ incubator for 18 hours. Brefeldin A (10 μg/mL, Sigma-Aldrich) was concurrently added to the cultures. After stimulation and washing, anti-human CD45-FITC, CD3-APC-Cy^TM7, CD4-APC, and CD8-PerCP-CyTM5.5 antibodies were added and incubated in the dark for 15 minutes. After fixation and permeabilization, PBMCs were stained with anti-human IFN-γ-PE and IL-4-PE/Cy7 antibodies at 4°C for 1 hour in the dark. Finally, cells were washed, resuspended in 300 μL of PBS, and analyzed using flow cytometer (BD FACSCanto II).

Memory B cells (MBCs), as well as memory, activated, and functional T cells, and natural killer (NK) cells, were analyzed using fresh EDTA anticoagulated blood. For each patient, four aliquots of 50 μL blood were stained with fluorochrome-conjugated monoclonal antibodies (Table S2) and incubated in the dark for 15 minutes. Hemolysin (Beckman, USA) was added to each tube and incubated for 10 minutes. Subsequently, samples were washed and resuspended in 200 μL of PBS and detected by FACSCanto II.

SARS-CoV-2-specific T cell responses to WT and Omicron SPs, and lymphocyte phenotypes were evaluated in eight CLD patients of group 1, 21 of group 2, and six of group 3. Flow cytometry data were processed by FlowJo software (version 10). The gating parameters and representative cell population distributions are presented in Figures S1-S5.

Statistical analysis

Statistical analyses were conducted using R version 4.1.3, SPSS Statistics 26.0 (IBM Corp, USA), and Prism 8.0 (San Diego, USA). The Shapiro-Wilk test was applied to determine the normality of quantitative data. T-test or ANOVA were respectively applied to compare two groups or multiple groups for data with normal distribution; otherwise, the Wilcoxon test or Kruskal-Wallis test was performed. The χ² test was used to compare categorical variables.

Univariable analyses employed binary logistic regression to assess potential risk factors, and calculating odds ratios (ORs) with 95% confidence intervals (CIs) for both BI occurrence, prolonged NCT ( > 7 days), and liver disease progression. Disease severity (SLD vs NSLD) and baseline humoral immunity profiles (total anti-SARS-CoV-2 antibodies, NAbs-WT/-BA.4/5, and anti-RBD IgG) were evaluated for BI risk. Prolonged NCT analysis incorporated disease severity, gender (male vs female), and proportions of CD62L⁺CD45RA⁺CD8⁺T and CD4⁺T cells. Disease progression analysis included vaccination status (vaccinated vs unvaccinated), total anti-SARS-CoV-2 antibodies, anti-RBD IgG, Omicron-specific IFN-γ⁺CD4⁺T cells, CD28⁺CD8⁺T cells, total MBCs, and resting MBCs. Variables demonstrating significant associations in univariable analyses were included in subsequent multivariable logistic regression models. All models satisfied multicollinearity assumptions (variance inflation factor < 3; tolerance > 0.3). Analyses were conducted using IBM SPSS Statistics (v26.0).

Correlation analyses of non-normally distributed data were performed using Spearman’s correlation coefficients, and visualized with the “ggcorrplot” and “corrplot” packages in R. P < .05 was considered statistically significant.

Results

BI of the Omicron variant in CLD patients

Among the 53 participants who were followed up and had received at least two doses of the inactivated vaccine, 81.13% experienced a BI event, occurring at a median of 350 days after the last vaccine dose. Notably, SLD was a risk factor for BI (Table S3).

No significant intergroup differences were observed in gender, CLD types, treatment regimen, or liver indicators. Liver function parameters were also comparable in CLD patients between baseline and follow-up assessments in three groups (Table S4). COVID-19 symptoms were reported 39 out of 43 participants in group 2 and all seven in group 3. Fever and fatigue were the most commonly observed clinical manifestations. Besides, vaccinated individuals with BI were more likely to present with cough and sore throat, whereas unvaccinated individuals reported headache, myalgia, and arthralgia (Table 1).

Table 1.

Demographic, clinical, and laboratory characteristics of CLD patients at follow-up.

Characteristic	Group 1 (n = 10)	Group 2 (n = 43)	Group 3 (n = 7)	P value
Age (year)	57.9 ± 15.98	48.12 ± 10.07	56.29 ± 10.44	.0228
Male/female, n	5/5	30/13	6/1	.2765
Type of CLD, n (%)				0.0553
Hepatitis Ba	2 (20%)	25 (58.14%)	1 (14.29%)
Cirrhosisb	6 (60%)	10 (23.26%)	4 (57.14%)
Liver cancer	2 (20%)	8 (18.60%)	2 (28.57%)
Vaccination status, n (%)				<0.0001
2nd dose of inactivated vaccine	0 (0%)	7 (16.28%)	0 (0%)
3rd dose of inactivated vaccine	9 (90%)	36 (83.72%)	0 (0%)
4th dose of inactivated vaccine	1 (10%)	0 (0%)	0 (0%)
Unvaccinated	0 (0%)	0 (0%)	7 (100%)
Period of last vaccination at the time of sampling (day)	462.2 ± 163.8	503 [379, 503]	/	.3965
Period of last vaccination at the time of BI (day)	/	350 [326, 411]	/
Period of first BI at the time of sampling (day)	/	146.5 [107, 164]	97.83 [23.57]	.0077
Negative conversion time of SARS-CoV-2 (day)	/	7 [4, 9]	6.5 [6, 9]	.6577
COVID-19 symptoms, n (%)
Fever	/	32 (74.42%)	7 (100%)	.3237
Fatigue	/	14 (32.56%)	4 (57.14%)	.2345
Headache	/	8 (18.60%)	3 (42.86%)	.1697
Muscle and joint pain	/	11 (25.58%)	3 (42.86%)	.3835
Sore throat	/	15 (34.88%)	2 (28.57%)	> 0.9999
Nasal obstruction	/	9 (20.93%)	1 (14.29%)	> 0.9999
cough	/	18 (41.86%)	2 (28.57%)	.6872
Altered taste and smell	/	9 (20.93%)	2 (28.57%)	.6407
Treatments, n (%)	8 (80%)	42 (97.67%)	6 (85.71%)	.0902
Antiviral treatment c, n (%)	8 (80%)	41 (95.35%)	6 (85.71%)	.0758
Chinese patent medicines d, n (%)	2 (20%)	13 (30.23%)	3 (42.86%)	.5980
Hepatitis B parameters
Hepatitis B surface antigen (IU/mL)	329.2 ± 367.4	935.7 [190.5, 3594]	62.65 [51.4, 1255]	.0537
Hepatitis B surface antigen (+), n (%)	5 (50%)	33(76.74%)	5(71.43%)	.2395
Hepatitis B surface antibody (mIU/mL)	0.5 [0.01, 2.805]	0.5 [0.5, 1.078]	0.5 [0.5, 1.505]	.8049
Hepatitis B surface antibody (+), n (%)	0(0%)	1(2.33%)	0(0%)	.8179
Hepatitis B virus DNA (+), n (%)	1(10%)	8(18.60%)	1(14.29%)	.7926
Liver function parameters
Alanine aminotransferase (U/L)	26.78 ± 14.43	22.9 [16.9, 34.1]	32.61 ± 7.996	.2883
Aspartate aminotransferase (U/L)	29.33 ± 13.13	20 [16, 25.4]	41.69 ± 21.6	.0021
γ-glutamyl transferas (U/L)	18.8 [12.55, 66.73]	19.7 [14.4, 30.6]	31.6 [17.8, 155.4]	.3483
Alkaline phosphatase (U/L)	104.3 ± 27.5	73 [62, 99]	131.6 ± 58.2	.0032
Total bilirubin (μmol/L)	22.67 ± 14.08	12.7 [9,9, 17]	21.4 [14.6, 30.2]	.0310
Direct bilirubin (μmol/L)	7.3 [4.025, 12.13]	4.4 [3.6, 6.2]	10.36 ± 6.526	.0154
Total protein (g/L)	74.68 ± 4.939	75.5 [72.6, 78.9]	78.9 [77, 88.3]	.0647
Albumin (g/L)	41.37 ± 6.306	45.9 [44.1, 48]	38.91 ± 5.704	.0046
Lactate dehydrogenase (U/L)	192.2 ± 43.27	184.9 ± 39.75	226.1 ± 47.39	.0563
Blood cell counts
White blood cells (×109/L)	4.44 ± 1.387	5.53 [4.545, 6.66]	4.443 ± 1.304	.0376
Lymphocytes (×109/L)	1.741 ± 0.7735	1.755 [1.45, 2.155]	1.406 ± 0.6491	.2304

^aViral hepatitis has not yet progress to cirrhosis.

^bCirrhosis has not yet progress to liver cancer.

^cAntiviral drugs include entecavir, tenofovir disoproxil fumarate, tenofovir alafenamide fumarate, sofosbuvir and velpatasvir.

^dChinese patent medicines include Ruangan Huajian Granule, Fuzheng Huayu Preparation, Dahuang Zhechong Pill, Chaihuangyigan Granule, compound Yiganling, Biejiajianwan.

Normally distributed data are shown as mean±standard deviation, non-normally distributed data are shown as median [Q1, Q3]; Categorical variables are shown as number (percentages).

BI augments the antibody response in vaccinated CLD individuals

Following Omicron infection in the CLD population, levels of NAbs-WT and -BA.4/5, total anti-SARS-CoV-2 antibodies, and anti-RBD IgG were markedly elevated compared to baseline (Figure 2(A–D) and Table S5).

Figure 2.

The comparisons of four SARS-CoV-2 specific antibodies among CLD patients before and after breakthrough infection (BI). A-D: the inhibition rates (%) of neutralizing antibodies (NAbs) against Omicron subvariant BA.4/5 (A) and wild type (WT) (B), the concentrations (BAU/ml) of anti-RBD IgG (C), and the levels of total antibodies against SARS-CoV-2 (D) among vaccinated CLD patients without BI (group 1) or with BI (group 2), and unvaccinated CLD patients with Omicron infection (group 3). The seroconversion rates of SARS-CoV-2 specific antibodies were shown in the upper part of the figure, positive rate was presented as white. According to manufacturer’s instructions, the inhibition rate of NAbs above 30%, the concentrations of IgG anti-RBD antibodies ≥10 BAU/ml, and the optical density (OD) values of total anti-SARS-CoV-2 antibodies above 0.19 were regarded as positive.

Specifically, the inhibition rate of NAbs-BA.4/5 was significantly higher in group 2 than in group 1 (P = .0320). Meanwhile, seropositivity rates of NAbs-BA.4/5 and -WT were increased in group 2 compared with both baseline levels and those in group 1. Notably, detectable levels of NAbs-BA.4/5 were also observed in group 1, suggesting the likelihood of previous, most likely asymptomatic, infection.

Full-course inactivated SARS-CoV-2 vaccination was pivotal in inducing antibody production in CLD patients. In addition to elevated NAbs-BA.4/5, group 2 exhibited higher inhibition rates of NAbs-WT (P = .0007), higher concentrations of anti-RBD IgG (P = .0010), and increased levels of total anti-SARS-CoV-2 antibodies (P = .0003) compared to group 3. Similarly, seropositivity for NAbs-WT was greater in group 2 than in group 3 (P = .0005). In addition, total anti-SARS-CoV-2 antibody levels in group 3 remained lower than in group 1 (P = .0097).

While previous studies reported that SLD patients developed inferior antibody titers³ and impaired RBD-specific MBC responses post vaccination compared to those with general disease,¹⁵ our data revealed comparable SARS-CoV-2-specific antibody levels across patients with varying disease severity and etiologies following BI (Figure S6A-S6C).

CLD patients exhibit robust and durable cellular immunity following post-vaccine infection

The proportions of IFN-γ⁺CD4⁺T, IL-4⁺CD4⁺T, IFN-γ⁺CD8⁺T, and IL-4⁺CD8⁺T cells were comparable between groups 1 and 2, but were higher than those in group 3 (Figure 3(A–B), Table 2, Figure S1, and Table S6). Following Omicron infection, patients in group 2 exhibited a significant increase in the frequency of Omicron- or WT-specific IFN-γ⁺CD4⁺T cells (P = .0403 and P = .0988, respectively) and IL-4⁺CD4⁺T cells (P = .0216 and P = .0826, respectively) compared to those in group 3. Moreover, the frequency of Omicron spike-specific IL-4⁺CD4⁺T cells (P = .0193) was higher among CLD patients in group 1 than in group 3. Similar levels of Omicron-specific-IFN-γ⁺ and IL-4⁺CD8⁺T cells were observed in groups 3 and 1, indicating a comparable cellular response between unvaccinated individuals infected with Omicron and vaccinated Omicron-naïve patients. These results suggest that vaccine-induced SARS-CoV-2-specific T cell responses are robust and durable.

Figure 3.

Cellular immune responses among CLD patients without and with breakthrough BI. A-B: the differences of SARS-CoV-2 specific IFN-γ⁺CD4⁺T (Th1) and IL-4⁺CD4⁺T (Th2) in vaccinated CLD patients experienced Omicron BI (group 2) in contrast with those did not experience BI (group 1) and unvaccinated CLD patients with infection (group 3). Omicron (B.1.1.529): the SARS-CoV-2 Omicron (B.1.1.529) variant of concern (VOC) and its sub-lineages, PBMCs with stimulation of Omicron specific peptide. WT: wild type, PBMCs with stimulation of SARS-CoV-2 specific peptide (ancestral strain). C: percentages of memory B cell (MBC) subsets in total MBC. D: percentages of memory T cell subsets in total memory CD4⁺ (left) and CD8⁺T (right) cells. E: percentages of functional subsets (CD28⁺) in total CD4⁺ cells. F-G: percentages of activated subsets (CD38⁺/HLA-DR⁺) in total CD4⁺ cells. H: percentages of functional subsets (CD28⁺) in total CD8⁺ cells. I-J: percentages of activated subsets (CD38⁺/HLA-DR⁺) in total CD8⁺ cells.

Table 2.

SARS-CoV-2 specific cellular responses in CLD patients with/without breakthrough infection.

Group	Group 1		Group 2		Group 3
	Omicron (B.1.1.529)	WT	Omicron (B.1.1.529)	WT	Omicron (B.1.1.529)	WT
IFN-γ+CD4+/CD4+T cell (%)	4.135 ± 0.8589	4.05 ± 1.489	4.374 ± 1.49	4.126 ± 2.097	1.89[0, 4.315]	1.785[0, 4.265]
IL-4+CD4+/CD4+T cell (%)	1.98 ± 0.8983	2.045 ± 1.198	1.64[0.78, 2.87]	1.695[0.8575, 3.53]	0.595[0, 1.348]	0.96 ± 1.129
IFN-γ+CD8+/CD8+T cell (%)	3.278 ± 1.418	2.633 ± 0.8735	3.552 ± 1.192	3.301 ± 1.466	3.977 ± 2.907	2.935 ± 1.694
IL-4+CD8+/CD8+T cell (%)	3.03 ± 1.545	3.085 ± 0.951	3.351 ± 1.761	3.344 ± 1.66	3.382 ± 2.477	2.023 ± 1.631

Normally distributed data are shown as mean±standard deviation, non-normally distributed data are shown as median [Q1, Q3].

Comparable proportions of memory T/B cells, as well as activated and functional T cells, were observed in patients from groups 1 and 2 (Figure 3(C–J), Figure S2-S4, and Table S6). Notably, vaccination appeared to modulate post-infectious cellular immune responses. When comparing CLD patients in groups 2 and 3, significantly higher percentages of total MBCs (P = .0226), resting MBCs (P = .0185), IgM⁺MBCs (P = .0089), CD45RA⁺CD62L⁺CD8⁺T cells (P = .0822), and CD28⁺CD8⁺T cells (P = .0229) were observed in vaccinated individuals. In contrast, decreasing trends were observed for atypical (P = .0098) and intermediate (P = .0445) MBCs, as well as activated T cells, including HLA-DR⁺CD4⁺T cells (P = .0772), CD38⁺CD8⁺T cells (P = .0286), and HLA-DR⁺CD8⁺T cells (P = .0226).

In contrast to humoral immunity, disease severity significantly impacted cellular immune responses. Specifically, SLD patients exhibited reduced proportions of Omicron-specific IFN-γ⁺CD8⁺T cells, CD38⁺CD4⁺T cells, and CD45RA⁺CD8⁺T cells. Cellular immunity profiles did not differ significantly between patients with cirrhosis versus liver cancer, except for decreased IgG⁺MBC subsets observed in SLD patients classified as Child-Pugh Class B (Figure S6D-F).

Correlation of humoral and cellular immune responses in CLD patients with BI

The interplay between humoral and cellular immune responses was assessed after BI. As depicted in Figure 4(A–G), markers of humoral immunity, including IgG⁺/IgM⁺MBCs, NAbs-WT, and -BA.4/5, were positively associated with central memory T cells (TCM, CD45RA⁻CD62L⁺CD4/CD8⁺T cell) and naive T cells (CD45RA⁺CD62L⁺CD4⁺T cells). In addition, SPs-specific T cell responses were correlated with memory and activated T cell populations (Figure 4(H–K)). Specifically, Omicron/WT-specific IL4⁺ responses were positively associated with CD45RA⁺CD62L⁻CD4⁺T cells, which are late effector memory cells re-expressing CD45RA. In contrast, IFN-γ⁺ responses were positively correlated with the activation marker CD38 on T cells.

Figure 4.

The correlation analysis of humoral and cellular immunity among CLD patients experienced BI. A: correlation heatmap displays the associations between SARS-CoV-2 specific antibodies and cellular immune responses. B-E: the humoral immunity (including IgG⁺ or IgM⁺ MBCs and NAbs) showed a positive association with memory T cells. H-K: associations between SARS-CoV-2-specific cell response and memory (H and I) or activated T cell (J and K). L-M: the proportion of CD38⁺CD8⁺T cell and wild type-specific IFN-γ⁺CD8⁺T cell response waned with the days from BI. The confidence interval (95%) is represented by the shadowed areas.

Humoral and cellular immune responses remained largely stable over time following BI, with the exception of CD38⁺CD8⁺T cells (Figure 4(L)). Notably, WT-specific IFN-γ⁺CD8⁺T cells were positively correlated with CD38⁺CD8⁺T cells, suggesting that the waning IFN-γ⁺CD8⁺T cells frequency over time since BI may be attributed to the decline in CD38⁺CD8⁺T cells (Figure 4(M)).

Interestingly, in group 3, a gradual decline in NAbs and anti-RBD IgG was observed over time following infection (Figure 5(A–C)). In contrast, patients in group 2 exhibited sustained antibody responses after hybrid immunity, indicating that vaccination enhances the durability of antibody responses in convalescent individuals (Figure 5).

Figure 5.

SRAS-CoV-2 specific-antibody levels dynamic over time in convalescent individuals. A-D: distribution of (A) NAbs against Omicron BA.4/5 levels, represented in inhibition rate (%), (B) NAbs toward wild type levels, represented in inhibition rate (%), (C) anti-RBD IgG levels, represented in log10(BAU/mL); and (D) total antibodies against SARS-CoV-2 levels, represented in the OD value at 450 nm over time (days from positive Omcrion variant dection result). Blue and orange colors represent vaccinated convalescent CLD patients (group 2) and unvaccinated convalescent individuals (group 3), respectively.

Risk factors associated with negative conversion time after BI

The nucleic acid NCT in CLD patients after BI was seven days. Patients with prolonged NCT ( >7 days) exhibited a higher proportion of CD4⁺T cells (P = .0308), whereas the proportion of CD45RA⁺CD62L⁺CD8⁺T cells was lower (P = .0561) (Figure 6(A, B)). No difference in antibody levels was observed between the groups at baseline or after BI.

Figure 6.

Risk factors associated with long negative conversion time in CLD patients after BI. A-B: the changes of total CD4⁺T cell (A) and CD45RA⁺CD62L⁺CD8⁺T cell (B) in patients with long NCT ( >7 day) compared with those with short NCT (≤7 day). C: risk factors associated with NCT of SARS-CoV-2 BI are shown in forest plot where odd ratios (OR) with 95% confidence interval (CI) and P values were represented at right panel. Factors with OR >1 were risk factors in contributing prolonged NCT >7 among patients with BI, while factors with OR < 1 were considered as protective factors.

Multivariate logistic regression analysis identified CD4⁺T cells as a risk factor for prolonged NCT, while CD45RA⁺CD62L⁺CD8⁺T cells were found to be protective, contributing to more rapid clearance of SARS-CoV-2 (Figure 6(C)).

Post-infection immune responses correlated to CLD progression

Liver disease exacerbation was significantly more frequent in unvaccinated (4/7, 57.1%) than vaccinated (6/43, 14.0%) CLD patients following Omicron infection (P = .023), characterized by diminished SARS-CoV-2-specific antibodies (total anti-SARS-CoV-2, anti-RBD IgG), T cell responses (Omicron-specific Th1, functional CD8⁺T cells), and total MBCs (especially resting MBCs) (Figure 7). Multivariate logistic regression analysis revealed low resting MBC levels as an independent predictor of disease progression (Table S7), supporting its potential prognostic utility.

Figure 7.

Levels of SARS-CoV-2-specific antibodies and lymphocyte subsets in CLD patients with liver disease progression or without after Omicron infection. A. The comparison of total anti-SARS-CoV-2 antibodies, anti-RBD IgG, and NAbs toward BA.4/5 and WT in CLD individuals with different prognosis. B. The percentages of Omicron-specific Th1, functional CD8⁺T cell, total MBCs, and resting MBCs in CLD patients with/without liver disease progression.

Discussion

Following China’s reopening policy in December 2022, a substantial number of individuals, especially vulnerable populations, were infected with the Omicron variant. In this context, we conducted a follow-up study among CLD patients to evaluate humoral and cellular immune responses before and after Omicron infection, with the aim of informing post-reopening public health strategies.

Our data revealed a significant increase in all antibody levels after Omicron infection compared to baseline. Notably, both the neutralization potency and seroconversion rate of NAbs-BA.4/5 were higher in CLD patients with BI than in those without, reflecting an antigen-induced booster effect. Consistent with reports in healthy population,^14,16,17 vaccinated individuals maintained superior magnitude and durability of SARS-CoV-2-specific antibody responses compared to their unvaccinated counterparts post-infection. The maintenance of neutralizing activity may be attributed to the expanded total MBCs, including resting MBCs and IgM⁺MBCs, observed in CLD patients with BI. This finding aligns with established evidence that MBCs provide long-term protective immunity against severe SARS-CoV-2 infection.^18,19 Collectively, these results suggest that convalescent CLD patients could derive substantial benefit from prior vaccination, highlighting the need for additional vaccines in unvaccinated CLD patients to achieve broad and effective protection.

Vaccination, BI, and natural infection all contributed to the induction of SARS-CoV-2-specific T cell responses. Nearly all vaccinated CLD patients exhibited robust IFN-γ⁺CD4⁺T, IL-4⁺CD4⁺T, IFN-γ⁺CD8⁺T, and IL-4⁺CD8⁺T cells. This observation aligns with previous studies reporting that T cell responses are not significantly altered by additional antigen exposure through BI.^20-22 The absence of differences may reflect the main function of T cells in facilitating early viral clearance and mitigating disease severity, rather than preventing initial infection.²³ Notably, after infection, Omicron-specific IFN-γ⁺CD4⁺T and IL-4⁺CD4⁺T cells increased in vaccinated CLD patients rather than the unvaccinated. Memory T cells induced by vaccination generate antigen-specific helper cells and cytotoxic T lymphocytes, which mediate directly antiviral effect by promoting antibody production and secreting antiviral cytokines.²⁴ In contrast, unvaccinated patients had lower levels of naive, functional, and activated CD8⁺T cells, suggesting T‐cell exhaustion following natural SARS‐CoV‐2 infection.^24-26 Altogether, these findings suggest that in vaccinated CLD patients, the immune evasion of the Omicron variant impair humoral responses more than cellular responses.

In CLD patients with BI, IgG⁺/IgM⁺MBCs, and NAbs-WT were positively associated with CD45RA⁻CD62L⁺CD4/8⁺T cells, and NAbs-BA.4/5 positively linked to CD45RA⁺CD62L⁺CD4⁺T cells. These highlight the potential role of memory and activated T cells in adaptive immune defense against SARS-CoV-2. CD45RA⁻CD62L⁺T cells, classified as TCM, localize primarily in secondary lymphoid organs and exhibit rapid proliferation in response to antigenic stimulation.²⁷ High antibody responders exhibited increased S1-specific CD4 TCM following vaccination, suggesting that long-term humoral responses link to the individuals’ ability to produce and maintain memory T cell populations.²⁸ The co-expression of CD45RA and CD62L is a characteristic of naive T cells, which can proliferate and differentiate into memory cells upon antigenic stimulation.²⁹ Within the first post-BI week, vaccine-primed antibodies and memory T cell activation constrained viral replication, driving MBCs activation and neutralizing antibody production that promoted residual viral clearance by week two.³⁰ CD45RA⁺CD62L⁺CD8⁺T cells was identified as the protective factor for NCT among CLD patients, underscoring the pivotal role of vaccine-induced memory T cell responses in virus clearance. Additionally, IFN-γ⁺CD8⁺T cells specific to Omicron or WT were positively correlated with CD38⁺T cells. CD38 is an activation marker of T lymphocytes during chronic viral infections.³¹ In HIV patients, CD38 effector T cells enhanced immunopathology and end-organ disease by increasing proinflammatory cytokines production,³² suggesting that CD38⁺T cells may be potential targets for attenuating ongoing immune activation in CLD patients.

This study has several strengths and limitations. One strength is the inclusion of unvaccinated individuals, which allowed for direct comparison of antibodies and cellular immune responses between vaccinated and unvaccinated convalescent CLD patients. The difference between the two subsets emphasizes the importance of vaccination, particularly for vaccine-naïve patients. Another strength is the follow-up CLD patients with immunological assessment. Although pre-infection immunological parameters did not predict the occurrence of BI, this finding suggests that other factors (like disease severity), may play a more prominent role in risk stratification for BI among CLD patients. However, several limitations should be acknowledged. First, the relatively small sample sizes in groups 1 and 3 reflect recruitment challenges posed by high vaccination coverage (81.4%)³³ and population infection rates (83.3% in our study). Second, stimulation-based ICS assays are limited in detecting nonfunctional virus-specific T cells and phenotypically altered T cell subsets during in vitro stimulation. These technical constraints highlight the need for future work incorporating MHC multimer staining to complement ICS for more comprehensive evaluation of SARS-CoV-2-specific T cells.

In summary, our findings highlight three key insights in CLD patients. Immune evasion of Omicron variant primarily affects humoral rather than cellular immunity. Hybrid immunity outperforms infection-alone immunity, supporting the need for additional vaccination in vaccine-naïve CLD patients to elicit stronger and broader immunity against future infections. SARS-CoV-2 specific-NAbs/T cell responses correlated with memory/activated T cell, with CD45RA⁺CD62L⁺CD8⁺T cells and resting MBC independently predicting NCT and disease progression, respectively, broadens our knowledge of cellular mechanisms of viral clearance and clinical outcomes. Continuous monitoring of immune parameters is warranted for the effective management of CLD patients and to improve their prognosis during the ongoing Omicron wave. Future studies should characterize nonfunctional virus-specific T cells and phenotypically altered T cell subsets during antigenic stimulation, and investigate their correlation with SARS-CoV-2 infection severity.

Biographies

Yongzhe Li, Professor and PhD/Postdoctoral Supervisor at Peking Union Medical College Hospital, holds esteemed positions including membership in the Laboratory Medicine Branch of the Chinese Medical Association, and Vice Chairmanship of the Laboratory Professional Committee of the Chinese Research Hospital Society. Professor Li has led eight projects funded by the National Natural Science Foundation and has been granted nine invention patents, published over 100 papers in prominent international academic journals, including the New England Journal of Medicine, Nature Genetics, Arthritis & Rheumatology, ect. His contributions have been recognized with awards such as the Chinese Medical Science and Technology Award from the Chinese Medical Association, the Beijing Science and Technology Award, and the Huaxia Medical Science and Technology Award. Professor Li mainly focuses on the pathogenesis of autoimmune diseases and the clinical application of laboratory diagnostic techniques. He first reported the clinical implications of multiple high-titer antiphospholipid antibodies in COVID-19 patients in the NEJM, highlighting the potential increased risk of long-term autoimmune diseases following immune damage induced by COVID-19.

Erhei Dai, Dean, Chief Physician, Professor, and Master’s Supervisor at the Fifth Hospital of Shijiazhuang, specializes in the diagnosis and research of pathogenic microorganisms. He has made significant contributions to understanding the natural history of chronic hepatitis B and mother-to-child transmission prevention, the pathogenesis of liver fibrosis and its treatment with traditional Chinese medicine, as well as the molecular epidemiology of novel coronavirus pneumonia. His research endeavors have been recognized with five provincial and ministerial science and technology progress awards. Additionally, he has served as the chief editor or editor of 14 academic works and holds four national invention patents. Dai has received numerous honorary titles, including Provincial Management Excellent Expert, Provincial Excellent Science and Technology Worker, Municipal High-Level Talent, Municipal Management Professional and Technical Top Talent, and Outstanding Contributions to Young and Middle-Aged Experts in the City.

Funding Statement

This work was supported by the Beijing Natural Science Foundation [M23008], the National High Level Hospital Clinical Research Funding [2022-PUMCH-B-124] for YZL and Key R&D project of Hebei Province [22377744D] for HXG.

Disclosure statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Ethical approval statement

This study was performed in accordance with the Declaration of Helsinki, approved by the Medical Ethics Committee of the Fifth Hospital of Shijiazhuang (2022–017–1). The informed consent was obtained from all CLD patients.

Abbreviations

BI:

Breakthrough infection

CLD:

chronic liver disease

ELISA:

enzyme-immunosorbent assay

EDTA:

ethylene diamine tetraacetic acid

MBC:

memory B cells

NAb:

neutralization antibody

NCT:

negative conversion time

PBMCs:

peripheral blood mononuclear cells

PBS:

phosphate buffered saline

RBD:

receptor binding domain

RT-PCR:

reverse-transcription polymerase chain reaction

SARS-CoV-2:

Severe acute respiratory syndrome coronavirus 2

SLD:

severe liver disease

sVNT:

Surrogate Virus Neutralization Test

SPs:

Spike glycoprotein peptide pools

VOC:

Variants of Concern

WT:

wildtype.

Supplementary Information

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2025.2544466