Role of senescent CD4⁺ T cells in breakthrough infection of the new variant strain of SARS-CoV-2 in elderly patients

Abstract

Background

In late 2022, Beijing, China saw a large-scale BF.7 Omicron variant breakthrough infection. However, the impact of COVID-19 vaccines on elderly breakthrough-infected patients’ antibodies and immune cells response was unclear.

Methods

We recruited 329 inpatients over 65 with BF.7 breakthrough infections. We analyzed the link between vaccination and survival in 67 sampled patients, investigating changes in antibody levels, cytokine profiles, as well as immune phenotypes.

Results

Experiments revealed that while vaccination could raise antibody levels in the elderly, it showed no significant neutralizing activity against the emerging COVID-19 variant XBB. Flow cytometry showed vaccination increased the proportion of CD4⁺ senescent T cells. Moreover, we found that elevated frequencies of CD4⁺ Tsens cells were associated with reduced antigen-specific CD4⁺ T cell activation, diminished IL-2 production, and lower proportions of Tfh cells, which ultimately leading to impaired neutralizing antibody production, particularly against new emerging variants.

Conclusions

We found the immunological efficacy of inactivated vaccines in elderly patients is influenced by the proportion of CD4⁺ senescent T cells. Future elderly vaccination strategies need optimization in dose number and timing, and future vaccine design should aim to minimize the generation of CD4⁺ senescent T cells.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-025-06756-0.

Introduction

Since the end of 2019, the novel coronavirus, referred to as SARS-CoV-2, has quickly spread around the world [1]. After 2021, the Omicron variant B.1.1.529 featuring numerous mutations in the spike protein has spread across various countries. The original BA.1 was rapidly replaced by BA.2, which then evolved into a wide range of subvariants, including BA.5, BA.5.2.1, BF.7, BA.2.75, BA.2.75.2, BA.2.10.1, etc. as well as the XBB.1.5 in 2023 [2, 3]. After XBB.1.5, new variants continue to emerge, BA.2.86, JN.1and EG.5 have started to spread globally (Fig. 1) [4, 5]. Evidence indicated that alterations in their antigenic properties have allowed these variants to evade serum neutralization in vaccinated individuals [6]. This decline in neutralization capacity underscores the challenges posed by the evolving virus, as newer variants demonstrate increasing immune evasion although previous vaccination efforts has been made [3].

Fig. 1.

The schematic representation of the SARS-CoV-2 Omicron BA.2 subvariants and strains used in this experiment

During December 2022 and January 2023, most residents in China experienced breakthrough infections with the BA.5/BF.7 strain. During the wave of infections in Beijing, China, the BF.7 subtype was reported to be the dominant circulating strain at the time [7]. Prior to the breakthrough infections, at the start of 2021, the Chinese government began promoting COVID-19 vaccinations among the public. As of January 13, 2023, a total of 3.488 billion doses of COVID-19 vaccines have been administered in China [3]. The most commonly used COVID-19 vaccines in mainland China are inactivated vaccines, which include CoronaVac (Sinovac Life Sciences Co., Ltd.) and BBIBP (Beijing Institute of Biological Products Co., Ltd.) [8]. Reports indicated that these two vaccines can reduce symptoms in patients after injection and it was reported the administration of two doses of vaccines can reduce mortality rates among elderly patients [9, 10].

As mentioned previously, despite the majority of the elderly population has been vaccinated with inactivated WT subvariant vaccines, BA.5/BF.7 breakthrough infections have seriously impacted the lives of many elderly individuals, raising concerns regarding vaccine effectiveness among older adults. Recent studies indicate that individuals vaccinated with wild-type vaccines demonstrate limited cross-neutralization capabilities against emerging Omicron subvariants, a phenomenon attributed to original antigenic sin [11, 12]. Original antigenic sin refers to the immune system’s tendency to respond to subsequent infections or vaccinations based primarily on the memory of the first encountered antigen, which may impede effective responses to new variants. However, the presence of this phenomenon in aging populations remains unvalidated. In addition to that, the increased senescent T cells have also been observed to impair the B cell immune response in COVID-19 patients [13]. The influence of vaccination doses on the proportion of senescent T cells, as well as the impact of these senescent T cells on antigen-specific antibody production, has not been thoroughly elucidated in elderly patients aged 65 and older. Aging is a critical risk factor associated with increased severity of COVID-19, adverse outcomes, and potentially diminished vaccine efficacy [14]. Timothy et al. suggested a negative correlation between age and antibody response following vaccination [15]. Another study revealed that the reinfection rate among elderly individuals stands at 15.86%, which is approximately twice as high as the 6.46% rate observed in younger populations [16]. Furthermore, elderly patients exhibit significantly greater vulnerability to reinfection [3, 17, 18]. To protect the vulnerable population from recurrent infections and disease progression, it is essential to evaluate the immune response of older adults to new subvariants following BA.5/BF.7 breakthrough infections.

In this study, we recruited 329 elderly participants aged over 65 from Beijing, China, who were infected with the BF.7 variant between December 2022 and January 2023. These patients received varying doses of inactivated vaccines. We collected serum samples during their recovery period to analyze neutralizing antibody titers. Additionally, assays for various cytokines in serum were conducted. Furthermore, we obtained patients’ peripheral blood mononuclear cells (PBMCs) for immunological profiling and assessments of cellular function. The objective of this study is to investigate whether elderly individuals exhibit the phenomenon of original antigenic sin following vaccination and to evaluate the effects of different vaccination doses on both the quantity and functionality of immune cells in elderly patients experiencing BF.7 breakthrough infections. Ultimately, this research aims to provide guidance for vaccination approaches specifically designed for the elderly population.

Materials and methods

Study design and sample collection

We conducted this study at Peking University’s Third Hospital (Beijing, China). From December 23, 2022, to January 19, 2023, the inpatients were admitted to the hospital. And all of them were infected with the SARS-CoV-2 mutant BF.7 during the BF.7 infection wave. In advance of enrollment the patients had been confirmed with SARS-CoV-2 (BF.7) infection by a nucleic acid–positive test. All patients were confirmed as SARS-CoV-2-naïve individuals with no prior history of infection. Patients were divided into two groups depending on whether they get vaccinated or not. And additional subgroup analyses were also performed with stratification by the number of vaccine injection doses that patients received. And all patients had an interval more than 6 months between final vaccine dose and confirmed SARS-CoV-2 infection. Peripheral venous blood samples were collected from enrolled patients firstly. And then, the samples were centrifuged for 10 min at 2000 g. Plasma was then aliquoted from the samples after centrifugation and stored at -80 °C until used. The residual red blood cells were lysed with 1×RBC Lysis Buffer (BD, NJ, USA) for white blood cells (WBCs) collection. And the immunophenotyping assay of white blood cells was conducted using flow cytometry. After density gradient centrifugation in Ficoll, peripheral blood mononuclear cells (PBMCs) were obtained from white blood cells and we used PBMCs for intracellular cytokine staining and ELISPOT assay. The clinical analysis and subsequent experimental sample selection process in this article was shown in Fig. 2.

Fig. 2.

The clinical analysis and subsequent experimental sample selection process

Immunophenotyping assay

For flow cytometry assay, three panels were divided. The first panel (Panel S1) were designed for phenotype identification including innate leukocytes, T cells and B cells. The second panel (Panel S2) was designed for memory T cells and senescent T cells phenotyping. The third panel (Panel S3) was designed for the cytokines of T cells assay while the fourth panel (Panel S4) was designed for analyzed T-cell activation following in vitro restimulation with WT, BA.5, BF.7, XBB pseudovirus or WT COVID-19 Spike peptides in 10 vaccinated elderly patients. PBMCs were co-cultured with pseudovirus or peptides for 36 h prior to Panel S4 flow cytometry staining. For cell surface makers staining, 1 × 10⁶ white blood cells were stained with the antibodies for 20 min in the dark at room temperature. For assessment of cytokines production in T cells, the PBMCs were cultured for 5 h at 37 ℃ with PMA (Biolegend, CA, USA) and Brefeldin A (Biolegend, CA, USA). The surface makers of T cells were stained firstly in Panel S3 and Panel S4. PBMCs were then fixed and permeabilized using the Staining Buffer Kit (BioLegend, CA, USA). Finally, intracellular proteins were stained. Flow cytometric analyses were conducted with a Cytoflex cytometer (Beckman, CA, USA), an ID7000 cytometer (Sony, TYO, JP) and Kaluza analysis software v. 2.1.1. Supplementary Table 1 listed all antibodies in the four panels. Supplementary Figs. 1, 2, 3 and 4 showed the gating strategy.

Evaluation of total IgG and IgM titers specific to SARS-CoV-2

The titer of antigen-specific total IgG and IgM were measured by ELISA. In brief, 1 µg/ml of WT spike protein, or WT and Omicron BF.7 RBD protein (Sino Biological, Beijing, China) were coated in 96-well plate at 4 °C overnight. And then the plates were blocked by ELISA Assay Diluent (Biolegend, CA, USA) at room temperature for 2 h. After blocking, serial dilutions of serum samples were added to blocked plates. Antibodies bound to the plates were either detected with goat anti-human IgM-HRP antibody (Sigma, SL, USA) for IgM measurements or goat anti-human IgG-HRP antibody (Invitrogen, CA, USA) for IgG measurements. After 45 min of culture, TMB substrate (Biolegend, CA, USA) were added. ELISA stop solution was added to stop the rection 10 min later. Spark reader (Tecan, Männedorf, Switzerland) was used to measure absorbance at 450 nm and 630 nm. With a cutoff of 0.15 OD 450–630 nm, the endpoint dilution titer was calculated in GraphPad Prism.

Pseudo-virus neutralization assay

Neutralization assays were conducted with pseudovirus. Briefly, serum samples were diluted by 1:100 for the initial dilution and followed by 3-fold serial dilutions until 1:24300. The stock solutions of pseudo-virus were diluted to 1.5 × 10⁴ TCID₅₀/mL by culture medium. And then, the diluted serum samples were added with 50 all of pseudo-virus (WT, BA.5, BF.7, XBB) dilution and incubated at 37 °C incubator for 1 h. Then, HEK-293T cells which express human ACE2 receptor were seeded into the 96-well containing the virus-serum mixture and incubated at 37 ℃ for 48 h. After an incubation, 100µL of the culture medium was removed and the same volume of Bio-Lite luciferase reagent (Vazyme, NJ, China) was added. 3 min later, the luminescence of medium was read by Spark microplate reader (Tecan, ZRH, Switzerland). The medium which only contained cells and virus was the positive group (VC) and the medium which contained only cells was the negative group (CC). The endpoint dilution titer for plasma was calculated using the following formula:

Enzyme-linked immunospot assay (ELISPOT)

After PBMCs was collected, 100 µL of PBMCs (3 × 10⁵/well) were added to the precoated ELISPOT plate (MabTech, STO, Sweden) and stimulated overnight with 2 mg/ml peptide pool for SARS-CoV-2 S protein (MabTech, STO, Sweden) in a 37 °C incubator. PBS was used as negative control and PMA (BioLegend CA, USA) was used as positive control. After incubation the plate was washed with wash buffer and then 100 µL biotinylated anti-human IFN-γ antibody was added. The plate was incubated for 2 h at room temperature. Washing 3 times with PBS, AEC substrate solution was added to the plate. Once noticeable spot emerged in the positive control, sterile water was used to stop the development. Lastly, using ELISPOT reader AID ELISPOT to count the numbers of spot-forming cells (AID, BB, Germany).

Assay for multiplexing plasma soluble factors

The kit of human CD8/NK panel, human proinflammation chemokine panel 1 and human chemokine panel 2 and from LEGEND PlexTM (Biolegend, CA, USA) was used to analyze 44 plasma samples in the vaccinated group or unvaccinated group. According to the manufacturer’s instructions, the assay was performed. After the sample preparation, analyses were performed using CytoFLEX S flow cytometry (Beckman Coulter, CA, USA) and the final data were analyzed using online software from Biolegend.

Statistical analysis

Continuous data were presented as mean values ± SD. Categorical data were expressed as number (%). And the P values were determined using two-tailed t test or One-way ANOVA, as each figure’s caption indicates. Comparisons between continuous variables were performed using two-tailed Student’s t test or one-way analysis of variance. Comparisons between categorical variables were performed using Pearson’s chi-squared test. Logistic regression model was used to evaluate the associations of administration of COVID-19 vaccines and survival during and after hospitalization. Three models with different covariates were adopted. Model 1 was adjusted by sex and age. Model 2 was further adjusted by smoking, alcohol consumption, days of hospital admission, aspirin use, and surgery. Model 3 was further adjusted by symptoms during COVID-19 infection (Phlegm, dyspnea, cough, and fever). Statistical significance was defined as a P value less than 0.05. GraphPad Prism 9.0 and STATA 18.0 were used for statistical analyses.

Results

Patient clinical characteristics

All patient’s survival status followed at 1 month, 3 months, 6 months, and 1 year after discharge (Table 1). According to model 3 of logistic regression, among those vaccinated, the odd ratios (ORs) were 0.51 (0.16 to 1.63), 0.73 (0.27 to 2.00), 0.75 (0.27 to 2.05), 0.80 (0.31 to 2.10), and 0.76 (0.28 to 2.09) for death during hospitalization, 1 month after discharge, 3 months after discharge, 6 months after discharge, and 1 year after discharge, respectively (Supplementary Table 2). Stratified by doses, there is still no statistically significant results.

Table 1.

Characteristics of the participants in this study

Characteristic	All (n = 329)	Doses of COVID-19 vaccines			P
		0 dose (n = 259)	2 doses (n = 23)	3 doses (n = 47)
Age (mean, SD)	80.44 (8.25)	81.03 (8.27)	78.70 (7.88)	78.02 (7.90)	0.040
Sex (n, %)					0.082
Male	149 (61.32)	111 (57.81)	12 (80.00)	26 (72.22)
Female	94 (38.68)	81 (42.19)	3 (20.00)	10 (27.78)
Smoking (n, %)	67 (20.36)	50 (19.31)	6 (26.09)	11 (23.40)	0.634
Alcohol consumption (n, %)	45 (13.68)	33 (12.74)	3 (13.04)	9 (19.15)	0.499
Death during hospitalization (n, %)	50 (15.20)	43 (16.60)	2 (8.70)	5 (10.64)	0.385
Death within 1 month after discharge (n, %)	63 (23.16)	53 (24.65)	4 (22.22)	6 (15.38)	0.449
Death within 3 months after discharge (n, %)	63 (23.25)	53 (24.65)	4 (23.53)	6 (15.38)	0.452
Death within 6 months after discharge (n, %)	69 (26.85)	58 (28.86)	5 (27.78)	6 (15.79)	0.248
Death within 1 year after discharge (n, %)	69 (30.26)	58 (32.58)	5 (29.41)	6 (18.18)	0.254

One month after discharge, 57 participants were lost to follow-up (0 doses: 44 participants; 2 doses: 5 participants; 3 doses: 8 participants); Three months after discharge, 58 participants were lost to follow-up (0 doses: 44 participants; 2 doses: 6 participants; 3 doses: 8 participants); Six months after discharge 72 participants were lost to follow-up (0 doses: 58 participants; 2 doses: 5 participants; 3 doses: 9 participants); One year after discharge, 101 participants were lost to follow-up (0 doses: 81 participants; 2 doses: 6 participants; 3 doses: 14 participants)

To evaluate the neutralizing antibody titers against pre-Omicron variants and Omicron subvariants from half to one month after the onset of COVID-19, plasma or serum samples (plasma samples take priority) of 67 inpatients with severe COVID-19 from Peking University Third Hospital were collected. The 67 patients (48 males and 19 females) were aged from 65 to 93 years old. Among them, 20 and 22 participants were administered 2 and 3 doses of inactivated COVID-19 vaccines, respectively. The detailed demographic characteristics of the 67 participants can be seen in Supplementary Table 3. Since Omicron BF.7 infection was the most prevalent in Beijing from December 2022 to January 2023, we assumed that almost all the patients in this study were probably infected with BF.7 subvariants. Written informed consent was obtained from all patients included in this study.

The level of spike-specific neutralizing antibodies positively correlated with the dosage of inactivated wild-type vaccines

Humoral immunity, mediated by B cell and antibodies, represents a crucial component of antiviral defense mechanisms [19, 20]. We collected serum samples from 50 elderly inpatients (aged over 70 years) who experienced breakthrough infections. The participants were divided into two groups: unvaccinated (n = 25) and vaccinated (n = 25). Subsequently, we assessed the levels of spike-specific neutralizing antibodies (IgG or IgM), targeting both S1 and RBD, in the serum samples from these two groups. The results indicate that the antibody titers against both the wild-type SARS-CoV-2 strain and the BF.7 Omicron variants in the vaccinated group were significantly higher than those observed in the unvaccinated group. Furthermore, the trend of IgM titers was found to be consistent with that of IgG (Fig. 3a and e). Studies have demonstrated that the immunogenicity of inactivated whole-virus particle vaccines increases with the dosage administered [10, 20]. Additionally, we conducted a comparison of spike-specific neutralizing antibody levels among inpatients who received 0 doses (n = 25), 2doses (n = 11) and 3 doses (n = 14) of inactivated virus vaccines. We observed that the IgG titers against the wild-type SARS-CoV-2 strain exhibited an increase corresponding to the number of vaccine doses administered (Fig. 3b and c). However, the dose-dependent response in the titers of IgG against the BF.7 SARS-CoV-2 strain has diminished (Fig. 3d). For the titers of IgM in the wild-type (WT) strain, we observed that the titers in the group receiving three doses were significantly higher compared to those in the group receiving no doses (Fig. 3f and g). And IgM antibody titers against BF.7 SARS-CoV-2 strain amongst different injection times were no significant different (Fig. 3h).

Fig. 3.

The spike-specific antibody level after the vaccine injection in different immunization manner of vaccinations. (a) The comparison of the titers of IgG against S1 and RBD peptides of WT or BF.7 subvariants between unvaccinated group (n = 25) and vaccinated group (n = 25) (b-d) The titers of IgG against S1 peptides of WT subvariants, RBD peptides of WT subvariants or S1 peptides of BF.7 subvariants after 0–3 doses of the vaccine (e) The comparison of the titers of IgM against S1 and RBD peptides of WT or BF.7 subvariants between unvaccinated group (n = 25) and vaccinated group (n = 25) (f-h) The titers of IgM against S1 peptides of WT subvariants, RBD peptides of WT subvariants or S1 peptides of BF.7 subvariants after 0–3 doses of the vaccine. n = 25 in the group injected 0 dose of vaccine, n = 11 in the group injected 2 doses of vaccines, n = 14 in the group injected 3 doses of vaccines. *, p < 0.05; **, p < 0.01; ***, p < 0.001

Pseudovirus neutralizing titers against early COVID-19 variants increased after three vaccine doses except the XBB strain

After breakthrough infections with the BA.5/BF.7 variants, it has been demonstrated that the Omicron subvariant XBB possesses the ability to evade neutralization, even in individuals who have received inactivated vaccines [3]. We focused on the antibody levels against four distinct SARS-CoV-2 pseudovirus strains (WT, BA.5, BF.7, and XBB) following vaccination in elderly individuals. Specifically, we assessed the 50% neutralization geometric mean titers (GMTs) for these four subvariants in serum samples from elderly participants who received 0 dose, 2 doses, and 3 doses of the vaccine. Our findings indicated that the GMTs among elderly inpatients who had received three doses of the vaccine exhibited a significant increase compared to those who did not receive any vaccination (Fig. 4a-c). Furthermore, we found that despite elderly patients having received the inactivated vaccine, there were no vaccine-boosted responses observed in the geometric mean titers (GMTs) against the novel XBB subvariant (Fig. 4d) [11, 21].

Fig. 4.

(a-d) The titer of neutralizing antibodies against WT, BA.5, BF.7, and XBB pseudovirus was measured in convalescent sera by pseudovirus neutralization tests. n=25 in the group injected 0 dose of vaccine, n=19 in the group injected 2 doses of vaccines, n=20 in the group injected 3 doses of vaccines. *, p< 0.05; **, p <0.01.

Vaccines slightly enhance the ability of T cells to recognize and respond to BF.7 and XBB

It was reported that there has been a relative underestimation of the importance of T cell immunity in the control of SARS-CoV-2 [22]. It is optimal for vaccines to elicit both humoral and cellular immunity. In addition to the analysis of humoral immunity, cellular immunity-related cells were also examined in both vaccinated and unvaccinated elderly inpatients. We found that the percentage both of myeloid DCs (mDCs) and plasmacytoid DCs (pDCs) in the vaccinated group tended to increase and the increase was more significant in mDCs (Fig. 5a-b). Furthermore, the trend in the percentage of CD8⁺ T cells was found to be consistent with that of DCs. The patients received vaccines exhibited elevated proportions of CD8⁺ T cells (Fig. 5d-e). To further assess the functionality of T cells, we conducted an ELISPOT assay in which peripheral blood mononuclear cells (PBMCs) were stimulated with four types of pseudoviruses. The results from the ELISPOT assay indicated that T cells in the vaccinated group exhibited sensitivity to both wild-type (WT) and BF.7 subvariants, while showing insensitivity to BF.7 and XBB subvariants (Fig. 5f).

Fig. 5.

The proportion and function of immune cells in elderly inpatients with BF.7 breakthrough infection. (a-b) The percentage of mDCs and pDCs in WBCs from elderly inpatients with or without vaccination, n = 25 in unvaccinated group and n = 25 in vaccinated group. (c-d) A comparison of CD4⁺ and CD8⁺ T cells proportions in WBCs from elderly inpatients received vaccines or not, n = 25 in both groups. (e) The ratio of the percentage of CD8⁺ T cells to that of CD4⁺ T cells. (f) ELISPOT assay enumerated the number of IFN-γ secreting T cells following stimulation with four distinct types of pseudovirus, n = 6 in unvaccinated group and n = 6 in vaccinated group. *, p < 0.05; **, p < 0.01; ***, p < 0.001

Patients with low neutralizing antibodies titers have a higher proportion of CD4⁺ Tsens

After the vaccination, the populations of memory CD4⁺ and CD8⁺ T cells are usually concerned [23]. We further analyzed the subsets of CD4⁺ and CD8⁺ T cells. T cells were divided in 4 subsets which are naive T cells (Tn), central memory T cells (Tcm), effecter memory T cells (Tem) and terminally-differentiated effector memory T cells (Temra). From the results of the experiments, we observed that the proportion of memory T cells did not show a significant increase in elderly patients with BF.7 breakthrough infections who had previously received inactivated virus vaccines (Fig. 6a-b).

Fig. 6.

The influence of vaccine injection doses on the proportion of CD4⁺ T cell senescence. (a-b) Comparison of the CD8⁺ or CD4⁺ T cells subsets between elderly inpatients who have not received injection of vaccine or that received at least two doses of inactivated vaccines (c) The percentage of Tsens from elderly inpatients with or without vaccination. (d) The ratio of Tsen/Tn cells from elderly inpatients with or without vaccination (e) The percentage of Tsens from elderly inpatients with 0 dose, 2 doses or 3 doses vaccination. (f) The ratio of Tsen/Tn cells from elderly inpatients with 0 dose, 2 doses or 3 doses vaccination. n = 25 in unvaccinated group or unvaccinated group, n = 11 in 2 doses group and n = 14 in 3 doses group

Previously study show that the quantity of senescent T cells (Tsens) has a certain impact on the production of antibodies in the population after vaccination [13]. The absence of CD28 and the presence of CD57 are key indicators of senescent T cells [24]. So we also attempted to preliminarily elucidate the impact of senescent T cells on antibody production in elderly patients. The experimental results indicated that there is an increasing trend of CD4⁺ Tsens in elderly individuals after vaccination (Fig. 6c-f).

Elevated CD4⁺ Tsens impair neutralizing antibody production through attenuated CD4⁺ T cell activation and reduced T follicular helper cell proportions

To further investigate the relationship between CD4⁺ Tsens and the production of specific antibodies in elderly patients following vaccination. We categorized elderly patients into two groups -low senescent T cell group (the cell proportion ≤ 4.815%) and high senescent T cell group (the cell proportion > 4.815%) based on the median percentage of senescent T cells. We compared the antibody titers between groups with high and low levels of CD4⁺ Tsens within the same vaccination dose. We found that in the group receiving three doses with high levels of CD4⁺ Tsens, the neutralizing antibody titers were significantly lower than those in the group with low levels of senescent CD4⁺ T cells (Fig. 7b and c). Among the patients who received three doses of vaccine, the difference in antibody titers against the BF.7 RBD peptides between the high and low CD4⁺ Tsens groups was more significant compared to the difference in antibody titers against the WT RBD peptides (Fig. 7b and c). Moreover, in the high Tsens group, we observed a reduced proportion of activated CD4⁺ T cells (CD4⁺ CD69⁺) coupled with an elevated frequency of exhausted CD4⁺ T (CD4⁺ PD-1⁺) cells (Fig. 7d and e). Concurrently, elderly patients in this group exhibited a marked decrease in Tfh cell populations (Fig. 7f). And our findings also revealed significantly reduced proportions of IL-2-secreting CD4⁺ and CD8⁺ T cells in the high CD4⁺ Tsens group (Fig. S7 a-b).

Fig. 7.

The impact of CD4⁺ Tsens on immune activation in vaccinated elderly patients. (a-c) Neutralizing antibody titers against WT S1, WT RBD, and BF.7 RBD peptides in elderly patients with CD4⁺ Tsens ≤ 4.815 (low CD4⁺ Tsens) or CD4⁺ Tsens > 4.815 (high CD4⁺ Tsens) after vaccination with different doses in the group with the percentage of low CD4⁺ Tsens, the numbers of patients receiving 0 dose, 2 doses, and 3 doses were n = 13, n = 5, and n = 7, respectively. In the group with the percentage of high CD4⁺ Tsens, the numbers of patients receiving 0 dose, 2 doses, and 3 doses were n = 12, n = 6, and n = 7, respectively. (d-f) Proportional changes in CD4⁺CD69⁺, CD4⁺PD-1⁺, and Tfh cells in high and low CD4⁺ Tsens groups after co-culture with pseudovirus variants or WT COVID-19 Spike peptides, n = 5 in high or low CD4⁺ Tsens groups *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001

Vaccination can reduce the level of inflammatory cytokines and cytokine storms risk in the elderly

Finally, we conducted a comparative analysis of the 44 plasma cytokines between the two groups using Cytometric Bead Array (CBA) technology. From the CBA results, we can see that there were two aspects decrease in cytokine levels between two groups. On one hand, the levels of T cell toxicity-related cytokines, including IFN-γ and perforin, as well as the T cell chemotaxis-related cytokine I-TAC, exhibited a significant decrease following vaccine administration in elderly individuals [25, 26] (Fig. 8a). And there was also a certain decreased level of IL-2 and Granzyme A in the vaccinated group (Fig. 8a). The above results showed that the function of T cells may be affected in elderly after vaccines injection. On the other hand, the inflammatory cytokines associated with disease severity, including IL-10, MCP-1, sFasL, GROα and IL-6 exhibited a certain decrease trend (Fig. 8a). These findings suggest that the administration of inactivated vaccines may mitigate the severity of COVID-19 symptoms in elderly individuals experiencing BF.7 breakthrough infections. On the other hand, there was an increased trend in cytokines such as IL-1β and IL-12p70 after vaccination in elderly patients, but the differences were not statistically significant (Fig. 8b).

Fig. 8.

The plasma level of soluble cytokines in unvaccinated group or vaccinated group. (a) The cytokines with decreased concentrations (b) The cytokines with increased concentrations n = 25 in unvaccinated group and n = 25 in vaccinated group

Discussion

Elderly people were more likely to contract SARS-CoV-2 and develop illness [27]. It is crucial to find a best vaccination strategy for protecting the elderly from COVID-19 infections. To determine factors influencing neutralizing antibody production in vaccinated ≥ 65-year-old elderly patients, we deeply assessed immune cell subsets and cytokines in breakthrough-infected elderly patients post-inactivated COVID-19 vaccination. This study confirms that the proportion of senescent CD4⁺ T cells in elderly patients affects the immune efficacy of inactivated vaccines. Specifically, senescent CD4⁺ T cells impair T cell proliferation and activation by reducing IL-2 secretion. Additionally, the decreased proportion of T follicular helper (Tfh) cells may contribute to the poor humoral immune response and reduced neutralizing antibody production caused by senescent CD4⁺ T cells.

Consistent with published findings, elderly recipients of inactivated WT vaccines generated robust neutralizing antibodies against ancestral (WT), BA.5, and BF.7 strains, yet exhibited markedly reduced titers against emerging XBB subvariants [28, 29]. This progressive decline in cross-variant neutralization indicates waning effectiveness of inactivated vaccines against evolving SARS-CoV-2 lineages. Notably, this limitation extends to mRNA platforms: older adults show substantially impaired neutralizing capacity against Omicron variants after mRNA vaccination [30], with persistent deficiencies against contemporary BA.2.86 and JN.1 sublineages [31]. These convergent findings demonstrate that age-related impairment of vaccine-induced variant protection constitutes a platform-independent limitation for COVID-19 vaccines.

mRNA vaccines (e.g., BNT162b2) elicit robust dual humoral and cellular immunity, while inactivated vaccines (e.g., Sinovac, Sinopharm) primarily depend on humoral responses. Although mRNA platforms activate CD8⁺ T cells independently of CD4⁺ T cell help, both mRNA (BNT162b2) and adenovirus-vectored (AZD1222) vaccines critically require CD4⁺ T cells for protective antibody generation, establishing CD4⁺ T cells as central mediators of COVID-19 vaccine efficacy [32]. Crucially, this CD4⁺ T cell-dependent protection faces cross-platform erosion with aging, evidenced by universal declines in antibody titers and antigen-specific T-cell responses [30, 31]. In older adults, reduced mRNA vaccine efficacy correlates with diminished CXCR3 + T follicular helper (Tfh) and CD4⁺ Th1 cells, linking impaired CD4⁺ T cell function to weaker immunogenicity [33]. Similarly, our study of inactivated vaccines in elderly patients reveals that elevated CD4⁺ Tsens directly compromise CD4⁺ T cell activation (Fig. 7d) and reduce Tfh levels (Fig. 7e), limiting antibody breadth against emerging variants. These findings demonstrate that CD4⁺ T cell senescence is a shared mechanism undermining vaccine efficacy in aging. Supporting this universality, murine studies show reversing cellular senescence restores aged vaccine responses (e.g., enhanced IFN-γ, IgG⁺ B cells) [34]. Translationally, the senomorphic agent rapamycin improves influenza vaccine responses in older adults [35, 36]. Therefore, T cell senescence is pivotal to impaired vaccine efficacy in aging—a consistent effect across mRNA and inactivated COVID-19 vaccines. Targeting senescence represents a promising strategy to enhance immunogenicity in elderly populations, potentially extending to other viral vaccines.

It is worth noting that a recent study has shown that individuals who received three doses of the inactivated vaccine exhibited weaker immune responses to the new Omicron subvariants, as compared to those who had only two doses and experienced a BA.2 breakthrough infection [12]. While our serological analysis similarly showed elevated antigen-specific IgG/IgM by ELISA after three doses, pseudovirus neutralization assays demonstrated no significant inter-dose differences in elderly patients. This functional antibody deficit aligns with our finding that original antigenic sin is minimally operative in elderly cohorts (Fig. 4)—a distinction from conventional immune imprinting patterns. Notably, over 80% of vaccine effectiveness against COVID-19 mortality derives from the primary dose [10]. In consideration of our previous findings that vaccination may increase the proportion of CD4⁺ Tsens, which may lead to poor immune response against emerging of new subvariants, the number of COVID-19 vaccine doses for elderly individuals may need to take multiple factors into account.

Our comparative analysis of breakthrough infections in elderly inpatients revealed that vaccination confers protection against cytokine storms through dual cellular and molecular mechanisms. At the cellular level, vaccinated individuals exhibited elevated dendritic cell counts—particularly enriched in myeloid DCs (mDCs) over plasmacytoid DCs (pDCs)—alongside significant monocyte reduction (Fig. 5a-b). This monocyte decline is critically important given their established role in antigen presentation to CD8⁺ T cells, which drives IFN-γ production [37]. Concomitantly at the cytokine level, vaccination substantially downregulated proinflammatory mediators including IFN-γ, I-TAC, MCP-1, MIP-1α, IL-6, perforin, and IL-10 (Fig. 8). Collectively, these coordinated adaptations—enhanced mDC activity, disruption of monocyte-dependent T cell stimulation, and broad-spectrum cytokine suppression—demonstrate how vaccines rebalance immune responses to attenuate inflammatory cascades, thereby mitigating cytokine storm risk and reducing severe disease outcomes in the elderly [10, 38]. These findings underscore the importance of targeting mDC pathways in future vaccine design for aging populations.

Our study has several limitations. First, sample collection was restricted to a single timepoint (0.5-1month post-infection), necessitating future longitudinal monitoring of breakthrough infections in elderly cohorts. Second, due to participant frailty constraints, ELISPOT analyses were limited to PBMCs from only six patients per group across four viral variants. Third, exclusive enrollment of Chinese inactivated vaccine recipients raises the question of whether CD4⁺ Tsens-mediated impairment of neutralizing antibody production generalizes to mRNA-vaccinated elderly populations.

In summary, we found that increasing CD4⁺ Tsens cells may contribute to reduced production of neutralizing antibodies in old patients with inactivated COVID-19 vaccines. Critically, this Tsens-driven antibody deficiency correlates with substantially reduced protection against breakthrough infections by emerging variants like XBB. Elevated frequencies of CD4⁺ Tsens cells were associated with reduced antigen-specific CD4 + T cell activation, diminished IL-2 production, and lower proportions of Tfh cells. Most importantly, reducing CD4⁺ Tsens cells may be one of the key factors to consider in the design of COVID-19 vaccines targeting new variants.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Manuscript writing: Zhaoyuan Liang, Hua Zhang, Lu Xu, Jianling Yang. Experimental Design: Zhaoyuan Liang, Lu Xu, Jianling Yang, Jie zhang, Lixiang Xue. Data Analysis: Zhaoyuan Liang, Hua Zhang, Lu Xu, Jianling Yang, Lin Zeng. Patient Recruitment: Yongchang Sun, Ning Shen. Ethical Compliance: Yongchang Sun, Ning Shen. Manuscript Revision: Lixiang Xue, Siyan Zhan, Jianling Yang, Lin Zeng, Jie Zhang. Review and editing: All authors. Acquisition: Siyan Zhan, Lixiang Xue, Yongchang Sun. Supervision: Ning Shen.

Data availability

All data presented in this article have been included in the main manuscript and supplementary materials, both of which have been uploaded.

Declarations

Ethics approval and consent to participate

This study’s sampling and experimental steps were approved by the Ethics Committee of Peking University Third Hospital (License No. IRB00006761-M2022865).

Consent for publication

All participants (or their legal guardians) provided explicit consent for the anonymized publication of research data. And written informed consent for research participation was obtained from all participants.

Competing interests

The authors state no conflict of interest.