Expression levels of SARS-CoV-2 IgM, IgG, and neutralizing antibodies in a Chinese cohort and detection of peptide-specific antibodies in COVID-19 patients Authors: Suli Yang, Xiaosha Wen, Yi Lin, Ranrong Zhang, Dixian Luo Published: 2026 ABSTRACT AbstractBackgroundLong COVID is a complex multisystem disorder affecting approximately 65 million individuals worldwide, with autoimmune mechanisms implicated in its pathogenesis. This study aimed to elucidate the role of SARS-CoV-2-induced autoantibodies and immune dysregulation in the development of Long COVID, focusing specifically on autoimmunity and aberrant immune activation. The investigation sought to explore the underlying mechanisms through serological and peptide-based analyses.MethodsSerum samples from 315 healthcare workers, serving as a control cohort, were analyzed for the presence of SARS-CoV-2 IgM, IgG, and Neutralizing Antibodies (NAbs), with stratification based on age, gender, vaccination timing, and symptom clusters. Bioinformatic approaches were employed to identify SARS-CoV-2 epitopes with homology to human proteins, screening for epitopes of the SARS-CoV-2 Spike (S) and Nucleocapsid (N) proteins and their human homologous peptides through analyses of hydrophilicity, antigenicity, and homology prediction. An ELISA-based method was utilized to detect antibody responses to these peptides in both COVID-19 infected groups (categorized as mild, moderate, or severe) and an uninfected group.ResultsThe primary findings indicated a high seropositivity rate for IgG (97%) and Neutralizing Antibodies (94%) among the control group, with notable age-related trends: IgM levels increased with age, whereas IgG and Neutralizing Antibodies showed a decline. Female participants demonstrated higher IgG levels compared to their male counterparts. Antibody levels did not exhibit significant differences across acute or persistent symptom clusters (≥ 6 months post-infection). However, neurological symptoms, whether acute or chronic, were associated with elevated IgG and Neutralizing Antibody titers, suggesting that neurotropic infections may elicit a more robust humoral immune response. Bioinformatic analyses identified 29 potential epitopes within the S protein and 10 within the N protein, with 91 and 24 human homologous peptides, respectively. The detection of peptide-specific antibodies in the serum of the COVID-19 infected group revealed that antibodies against eight peptides displayed a greater than 2.5-fold difference between the COVID-19-infected and uninfected groups.ConclusionThese findings underscore age- and gender-related variations in antibody responses, identify immunodominant peptides with potential implications for COVID-19 symptoms, and suggest that cross-reactive antibodies targeting viral-human homologous peptides (e.g., S68-CHL1) may play a role in autoimmune mechanisms associated with COVID-19.Trial registrationClinical trial number: Not applicable.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12879-026-13532-y.Keywords: Long COVID, Autoantibodies, Molecular mimicry, SARS-CoV-2, Immune dysregulation Background Long COVID, also known as Post-Acute Sequelae of SARS-CoV-2 Infection (PASC), represents a complex multisystem disorder that emerges following COVID-19 and is frequently associated with severe symptomatology. The SARS-CoV-2 virus has the potential to inflict systemic damage across various organ systems, including the respiratory, cardiovascular, central nervous, digestive, endocrine, urogenital, and rheumatologic/immune systems, leading to long-term sequelae. It is estimated that approximately 65 million individuals worldwide are currently afflicted with Long COVID [1]. Extensive research has demonstrated that the majority of Long COVID patients exhibit a multitude of heterogeneous symptoms affecting multiple organs, including cardiovascular diseases, thromboembolic disorders, cerebrovascular incidents, type 2 diabetes, chronic fatigue syndrome, and autonomic dysfunction [2]. Certain symptoms may persist for extended durations, potentially lasting years or even a lifetime [3]. A significant number of patients experience debilitation to the point of work incapacity, resulting in considerable socioeconomic and psychological burdens. In light of the incomplete understanding of its pathogenesis and the absence of effective treatments, elucidating the underlying mechanisms and developing targeted interventions are pressing priorities. Current hypotheses regarding the pathogenesis of Long COVID encompass several mechanisms: ① the presence of persistent viral reservoirs within tissues [4]; ② immune dysregulation [5], potentially accompanied by the reactivation of latent pathogens such as herpesviruses [6]; ③ disruptions in the microbiome, including alterations in the virome [7, 8]; ④ autoimmunity induced by molecular mimicry [9–11]; ⑤ microvascular thrombosis and endothelial dysfunction [12, 13]; and ⑥ dysfunctional signaling in the brainstem and vagus nerve [14]. Given the clinical similarities between Long COVID and autoimmune diseases, this study concentrates on the role of autoimmunity and aberrant immune activation in Long COVID. Current research on long COVID SARS-CoV-2 infection induces immune dysregulation Individuals with moderate COVID-19 cases exhibit prolonged T-cell abnormalities lasting over 13 months, characterized by T-cell exhaustion and increased expression of PD-1 on central memory T cells [15, 16]. Sustained activation of the innate immune system, depletion of naïve T and B cells, and elevated responses of type I and III interferons persist for up to 8 months following infection [5]. Comparative analyses among Long COVID patients, healthy controls, and COVID-19-recovered individuals without Long COVID indicate that Long COVID patients, even 14 months post-infection, continue to demonstrate increased immune dysregulation Elevated autoantibodies in long COVID Patients with Long COVID may exhibit elevated levels of autoantibodies that target angiotensin-converting enzyme 2 (ACE2) [11], β2-adrenergic receptors, muscarinic M2 receptors, angiotensin II AT1 receptors, and MAS receptors [9]. These autoantibodies can disrupt cytokine and chemokine activity as well as Fc-receptor-mediated immune cell function [17]. Studies utilizing animal models have demonstrated that the presence of autoantibodies can exacerbate SARS-CoV-2 infection and inhibit the activation of myeloid and lymphoid cells [17]. High-throughput sequencing has identified 2,770 autoantibodies against various tissues, organ systems, and immunomodulatory proteins in COVID-19 patients, a number significantly greater than that observed in uninfected controls. Low titers of anti-SARS-CoV-2 antibodies and inadequate immune responses have been identified as predictors of Long COVID risk [18, 19]. Severe cases of Long COVID are associated with deficient CD4+/CD8 + T-cell responses [20], and elevated levels of autoantibodies are inversely related to viral antibody titers, increasing susceptibility to breakthrough infections [10]. The presence of viral proteins and RNA in reproductive, cardiovascular, cerebral, muscular, hepatic, pulmonary, and lymphoid tissues may play a role in initiating the pathogenesis of Long COVID [21, 22]. Molecular mimicry plays a pivotal role in the pathogenesis of autoimmunity observed in long COVID The production of autoantibodies frequently arises from molecular mimicry, a process in which proteins from pathogens (including viruses, bacteria, and fungi) exhibit structural similarities to human proteins. This structural homology leads to the generation of cross-reactive antibodies that target these shared epitopes, resulting in tissue damage, inflammation, and the production of autoantibodies. Notably, SARS-CoV-2-neutralizing antibodies have been shown to cross-react with tissues in the intestines, kidneys, lungs, heart, and brain. Furthermore, antinuclear antibodies (ANAs), typically associated with systemic lupus erythematosus (SLE), have been observed acutely in COVID-19 cases, and there is documented comorbidity between SLE and Long COVID [23, 24]. Atypical memory B cells (AtMs), which are characteristic of SLE, may be expanded in certain COVID-19 cases [25] and are associated with chronic inflammation [26]. Organ-specific autoantibodies have been linked to distinct phenotypes of Long COVID: anti-nucleocapsid IgG is associated with neurological symptoms, multiple autoantibodies with gastrointestinal manifestations, and anti-IFN-α2 with respiratory sequelae [10]. These associations appear to be independent of demographic factors or the severity of the initial disease. Consequently, profiling autoantibodies may serve as a predictive tool for identifying clinical subtypes of Long COVID. Alongside specialized diagnostic tools like MRI and tilt-table testing, new diagnostic methods for Long COVID are being developed, such as microclot imaging, corneal confocal microscopy, and changes in the ECG QRS complex. Proposed biomarkers for Long COVID include elevated cytotoxic exosomes [27] and immune markers [28]. However, treatments for autoimmune or immune-dysregulated Long COVID are still limited. BC 007, which neutralizes G-protein-coupled receptor autoantibodies, is effective against capillary damage, loss of taste, and fatigue [29]. Therapeutic apheresis has shown effectiveness in reducing Long COVID symptoms, but its use is limited due to high costs and variable outcomes. This study uses high-titer antibody screening against SARS-CoV-2 peptide libraries to find cross-reactive human tissue antigens, with the goal of identifying key autoantigens that drive Long COVID and establishing predictive biomarkers for monitoring the disease. Methods Assessment of SARS-CoV-2 antibody expression levels in the control cohort Between February 1 and February 15, 2023, a total of 315 serum samples were obtained from hospital staff during routine health examinations at Shenzhen Nanshan People’s Hospital. These samples were subsequently analyzed for the presence of SARS-CoV-2-specific IgM, IgG, and neutralizing antibodies. In addition, participants were required to complete two structured questionnaires: the *COVID-19 Infection and Vaccination Survey*, which documented their history of SARS-CoV-2 infection, the number of vaccine doses received, the interval since the last vaccination, and any pre-existing medical conditions; and the *Post-COVID-19 Long COVID Symptom Survey*, which evaluated the frequency of infections and the persistence of symptoms following infection.For the purpose of analysis, participants were stratified into vaccination cohorts based on the duration since their last vaccination (< 3 months, 3–6 months, 6–12 months, or > 12 months). Additionally, symptomatology was categorized into acute and chronic manifestations (≥ 6 months post-infection), encompassing respiratory symptoms (such as fever, cough, headache, sore throat, myalgia, conjunctivitis, and nasal congestion), gastrointestinal symptoms (including diarrhea and abdominal pain), neurological symptoms (such as sleep disturbances, memory/attention deficits, dizziness, anosmia/ageusia, anxiety, and depression), and cardiovascular symptoms (including reduced cardiorespiratory endurance, palpitations, hypertension, coronary artery disease, and arrhythmias). The presence of SARS-CoV-2 IgM/IgG and neutralizing antibodies was assessed utilizing the commercial “2019-nCoV IgM/IgG/Neutralizing Antibody Detection Kit (Magnetic Particle Chemiluminescence Method)” from Shenzhen YHLO Biotech. The detection of IgM and IgG antibodies employed the fusion antigen of the 2019-nCoV nucleocapsid and spike S1 proteins, while the detection of neutralizing antibodies utilized the 2019-nCoV RBD protein. Measurements were conducted using the iFlash 3000 chemiluminescence immunoassay analyzer (YHLO). Screening and synthesis of SARS-CoV-2 epitopes and human homologous peptides for the identification of cross-reactive neo-antigens The amino acid sequences of the SARS-CoV-2 Spike glycoprotein (S) and Nucleoprotein (N) were retrieved from the UniProt database, with accession numbers P0DTC2 and P0DTC9, respectively. The integrity of these sequences was confirmed by verifying structural domains using the NCBI Conserved Domain Database. Predictions of protein secondary structures were conducted via the ExPASy proteomics server, employing a combination of the Chou-Fasman and GOR algorithms. These analyses identified β-turns as the predominant structural features, attributed to their flexibility and surface exposure. A comprehensive physicochemical characterization was undertaken, which included hydrophilicity profiling using the Kyte-Doolittle scale with a window size of 9, calculation of the antigenic index via the Jameson-Wolf method, and prediction of surface accessibility through Emini’s implementation with a threshold of ≥ 1.0. Regions with high scores were indicative of potential surface exposure and were identified as probable B-cell epitopes, ranging from 6 to 12 amino acids in length. Any discrepancies between the algorithms were addressed through consensus scoring.Subsequent homology screening was conducted utilizing NCBI BLAST with rigorous parameters, including a bitscore for similarity assessment and an E-value threshold of < 10^-5 for statistical significance, to identify human proteins exhibiting high sequence similarity to the predicted SARS-CoV-2 epitopes. Ultimately, the selected viral epitopes and their homologous human peptide counterparts were chemically synthesized through solid-phase peptide synthesis for further examination. ELISA detection of peptide-specific antibodies in serum from COVID-19 infected patients The study involved the detection of peptide-specific antibodies in the serum of patients infected with COVID-19. The cohort consisted of 31 hospitalized patients diagnosed with COVID-19 (ICD-10 code U07.1), who were selected through systematic random sampling from Shenzhen Nanshan People’s Hospital between February 1 and May 6, 2023. Patients with co-infections from other respiratory pathogens, those with immunocompromised conditions, and individuals with incomplete follow-up data were excluded from the study. The patients were stratified based on disease severity into mild/moderate (n = 16) and severe (n = 15) categories. A uninfected group comprised 15 healthy individuals without a history of COVID-19 infection. For the detection of autoantibodies, a standard curve was established using a 10-point serial dilution (ranging from 60 to 5 µg/mL) of a 90 µg/mL standard. Samples were processed by combining 40 µL of dilution buffer with 10 µL of serum to achieve a final 5× dilution. After a 30-minute incubation at 37 °C, followed by five washes with Wash Solution, detection was carried out using an HRP-conjugate and TMB substrate. The reactions were terminated, and the optical density (OD) was measured at 450 nm within 15 min. The autoantibody titers obtained were subsequently correlated with clinical symptom profiles for further analysis. Statistical analysis In this study, categorical variables are presented as absolute frequencies alongside their respective proportions (n [%]), whereas continuous variables are described using medians in the descriptive analyses. Nonparametric tests were utilized, specifically the Mann-Whitney U test for comparisons between two independent sample groups and the Kruskal-Wallis H test for comparisons among multiple independent sample groups. All statistical analyses were conducted using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA), with statistical significance determined by a two-tailed p-value threshold of < 0.05. Results The demographic information of the study subjects is presented in Table 1. Table 1Demographic table of the study subjectsGroupsControl CohortInfected groupUninfected groupMild/moderateSevereAge(y)18-44y227(72.1%)11(68.7%)4(26.7%)7(46.6%)45-65y80(25.4%)5(31.3%)7(46.6%)7(46.6%)> 65y8(2.5%)04(26.7%)1(6.8%)Genderfemale231(73.3%)9(56.3%)4(26.7%)8(53.4%)male84(26.7%)7(43.7%)11(73.3%)7(46.6%)Vaccination(months)< 317///3–619///6–1279///> 12(n = 171)171///symptom onsetFebruary,2023/16(100%)15(100%)/ Demographic table of the study subjects Analysis of SARS-CoV-2 IgM, IgG, and neutralizing antibodies in the control cohort The investigation into SARS-CoV-2 IgM, IgG, and neutralizing antibodies among the 315 control subjects revealed distinct demographic and serological characteristics. The age distribution was as follows: 72% of participants were aged 18–44 years, 25.4% were aged 45–65 years, and 2.6% were over 65 years. The gender composition comprised 26.7% male and 73.3% female participants. Serological assessments indicated that 1% of the cohort tested positive for IgM (with a cutoff value exceeding 7 AU/mL), 97% tested positive for IgG (with a cutoff value exceeding 10 AU/mL), and 94% tested positive for neutralizing antibodies (with a cutoff value exceeding 10 AU/mL). Age-stratified analysis revealed significant differences in IgG levels among different age groups (P = 0.0113), whereas IgM and neutralizing antibody levels did not exhibit significant variation across age groups (IgM: P = 0.44, Neutralizing Antibodies: P = 0.245) (Fig. 1). Furthermore, female participants demonstrated significantly higher IgG titers compared to their male counterparts (P = 0.0085), although no sex-based differences were observed for IgM or neutralizing antibodies (IgM: P = 0.77, Neutralizing Antibodies: P = 0.377) (Fig. 1).Among the 286 participants who completed vaccination surveys, all of whom had received at least two doses, stratification by time since the last vaccination (> 12, 6–12, 3–6, and < 3 months) revealed no significant variations in IgM, IgG, or neutralizing antibody levels across these temporal groups (IgM: P = 0.7123, IgG: P = 0.8866, Neutralizing Antibodies: P = 0.9346) (see Supplementary Materials, Fig. 1). A: SARS-CoV-2 IgM, IgG, and neutralizing antibody levels in serum by SARS-CoV-2 infection Symptom, B: SARS-CoV-2 IgM, IgG, and neutralizing antibody levels in serum by Persistent Symptom Clusters at 6 Months Post-Infection The median levels of the three antibody types (IgM, IgG, and neutralizing antibodies) indicated that IgM antibodies gradually increased with advancing age (median: 18–44 years: 0.5 AU/mL to > 65 years: 1.46 AU/mL), whereas IgG and neutralizing antibodies progressively declined (IgG: 18–44 years: 360.7 AU/mL to > 65 years: 285.8 AU/mL; Neutralizing Antibodies: 18–44 years: 652 AU/mL to > 65 years: 144.5 AU/mL). Additionally, IgM antibody levels were elevated in the > 12-month group (0.53 AU/mL) compared to the < 3-month group (0.33 AU/mL). IgG and neutralizing antibodies reached peak levels 3–6 months post-vaccination (IgG: 395.9 AU/mL; Neutralizing Antibodies: 854 AU/mL), followed by a subsequent decline. Analysis of antibody levels by SARS-CoV-2 infection symptoms Within the control cohort, 51 participants completed the “Post-COVID-19 Long COVID Symptom Survey.” The acute phase symptoms were predominantly characterized by viral respiratory manifestations (96.1%), followed by neurological symptoms (33.3%). In terms of symptom combination distribution, 32 participants (62.7%) exhibited solely respiratory symptoms, which included typical upper respiratory syndrome manifestations such as cough, sore throat, and fatigue. Meanwhile, 15 participants (29.4%) experienced both respiratory and neurological symptoms, indicating potential viral neurotropism. Additionally, 4 participants (7.8%) presented with concurrent respiratory and gastrointestinal symptoms, such as diarrhea and vomiting. Notably, 2 participants (3.9%) exhibited symptoms across all three systems—respiratory, neurological, and gastrointestinal—indicating a severe phenotype. Analysis of antibody levels by persistent symptom clusters at six months post-infection Among the 51 control subjects who completed the “Post-COVID-19 Long COVID Symptom Survey,” the chronic phase sequelae exhibited diverse characteristics affecting multiple systems, including the respiratory, neurological, and cardiovascular systems. Of the 51 individuals who had recovered, 21 cases (41.2%) reported no sequelae, while 30 cases (58.8%) experienced at least one type of sequela. Specifically, single-system sequelae were observed in 10 cases (19.6%), predominantly affecting the respiratory or neurological systems. Dual-system sequelae were identified in 13 cases (25.5%), with common combinations being respiratory and neurological or neurological and cardiovascular systems. Triple-system sequelae were present in 7 cases (13.7%), involving simultaneous impact on the respiratory, neurological, and cardiovascular systems, which is defined as the multi-system involvement phenotype of long COVID. Symptom Outcome Characteristics: Respiratory symptoms: Of the 49 cases presenting with acute-phase respiratory symptoms, 21 cases (42.9%) exhibited persistent chronic respiratory symptoms, whereas 28 cases (57.1%) achieved complete remission. Neurological symptoms: Seventeen cases exhibited acute-phase neurological symptoms, while 21 cases presented with chronic-phase neurological symptoms, indicating that neurological symptoms in some patients were either delayed-onset or persistent. Gastrointestinal symptoms: All six cases with acute-phase gastrointestinal symptoms achieved complete remission in the chronic phase, resulting in a 0% persistence rate and suggesting a favorable prognosis for digestive system symptoms. Cardiovascular symptoms: Among individuals without acute-phase cardiovascular symptoms, 14 cases (27.5%) developed new cardiovascular symptoms during the chronic phase, characterized by palpitations, chest tightness, and reduced cardiopulmonary endurance, thereby representing a newly emerging category of sequelae. The antibody levels associated with different symptom phenotypes are presented in Table 2. We conducted an analysis by categorizing participants into two groups based on the presence of neurological versus non-neurological symptoms (Fig. 1). The median levels of three antibody types-IgM, IgG, and neutralizing antibodies-indicated that individuals with neurological involvement exhibited the highest levels of IgG and neutralizing antibodies in both the acute phase of SARS-CoV-2 infection (IgG: 399.8 AU/mL, NAb: 860.1 AU/mL) and at six months post-infection (IgG: 397.8 AU/mL, NAb: 847.5 AU/mL). Table 2Comparison of antibody levels among different symptom phenotypes (Mean ± SD)Symptom group n IgM (AU/mL)IgG (AU/mL)overall510.80 ± 0.80344.31 ± 87.20Acute phase–Respiratory Symptoms490.81 ± 0.81342.04 ± 88.24Acute Phase-Neurological Symptoms170.75 ± 0.76346.98 ± 66.60Acute Phase-Gastrointestinal Symptoms60.79 ± 0.96317.06 ± 158.52Chronic Phase -With Sequelae300.85 ± 0.77353.80 ± 54.08Chronic Phase -Without Sequelae210.73 ± 0.84330.75 ± 120.17 Comparison of antibody levels among different symptom phenotypes (Mean ± SD) Screening for peptides with high homology to SARS-CoV-2 peptides Our study commenced with the retrieval of SARS-CoV-2 Spike glycoprotein (S) and Nucleoprotein (N) sequences from the UniProt database. Through comprehensive bioinformatic analyses, which included secondary structure prediction and assessments of hydrophilicity, antigenicity, and surface accessibility, we identified multiple potential antigenic epitopes within both proteins. Specifically, the S protein was found to contain 29 candidate epitopes (refer to Supplementary Materials, Fig. 2; Table 1), while the N protein exhibited 10 candidate epitopes (refer to Supplementary Materials, Fig. 2; Table 2). Notably, BLAST analysis against human proteomes available in the NCBI database revealed 91 peptides in the S protein and 24 peptides in the N protein that exhibited homology to human sequences (see Supplementary Tables 3 and Table 2, respectively). We used the EMBOSS Needle tool to search for homology between the SARS-CoV-2-derived peptides in Supplementary Tables 1 and 2 and the human homologous peptides in Supplementary Tables 3 and 4. Supplementary materials Tables 5 and Table 6 lists, for each human homologous peptide, the corresponding viral peptide ID, the sequence region used for homology determination(length), and the homology percentage(Identity). Levels of the 8 Peptide-Specific Antibodies in the uninfected, mild/moderate group, and severe group Based on these results, we synthesized selected peptide fragments to evaluate antibody responses in COVID-19 convalescent sera using peptide-based enzyme-linked immunosorbent assay (ELISA). Analysis of peptide-specific antibody detection results in serum from COVID-19 infected groups Utilizing 154 protein-derived peptides obtained through the aforementioned methods-comprising 29 peptides from the SARS-CoV-2 S protein, 10 from the SARS-CoV-2 N protein, 91 from human homologous S proteins, and 24 from human homologous N proteins-as solid-phase peptides, a peptide ELISA was conducted to screen serum samples from the three study groups. The distribution of antibody titers for all peptide segments is depicted in Supplementary Materials Fig. 3. Antibodies against seven peptides demonstrated a greater than threefold difference (Table 3), and antibodies against 57 peptides exhibited a greater than twofold difference (Supplementary Materials Table 7) between the severe group and the uninfected group. Additionally, antibodies against 22 peptides showed a greater than twofold difference (Supplementary Materials Table 8) between the mild/moderate group and the uninfected group. Antibodies against eight peptides exhibited a greater than 2.5-fold difference (Supplementary Materials Table 9) between the COVID-19-infected group and the uninfected group. Statistical significance was evaluated using GraphPad Prism v8.0, employing an independent samples t-test. Antibodies against these eight peptides showed statistically significant differences across the three groups (P < 0.05)(Fig. 2). Table 3Comparison of Peptide-Specific Antibodies levels between the severe group and the uninfected group (Peptide-Specific Antibodies with threefold differences)Peptide-Specific Antibodiesuninfected groupSevere groupFold ChangeS670.2438970.9617333.943202S700.2941421.0920643.712706XGN60.2591600.9296273.587072S710.2900370.9357953.226472S850.3385391.0759113.178101S900.2581190.7868283.048318S370.2270940.6872033.026066 Comparison of Peptide-Specific Antibodies levels between the severe group and the uninfected group (Peptide-Specific Antibodies with threefold differences) Antibodies against eight peptides exhibited a greater than twofold difference (Supplementary Materials Table 10) between the COVID-19-infected group and the uninfected group.The tissue localization and expression profiles of target antigens and their homologous peptides were examined utilizing the Protein Atlas database (refer to Supplementary Materials, Table 11). Our serological investigation of COVID-19 convalescent patients demonstrated significantly elevated antibody titers against peptides associated with reproductive organs, as well as pulmonary and neurological tissues. Discussion The humoral immune response to SARS-CoV-2 infection is initially characterized by the production of IgM antibodies, which is subsequently followed by the generation of IgG and neutralizing antibodies (NAbs) that provide prolonged protection. In our control cohort of 315 healthcare workers, seropositivity rates were observed at 1% for IgM, 97% for IgG, and 94% for NAbs. These findings are consistent with previous studies. For instance, a study conducted by Iyer et al. [30] involving 343 patients reported that IgM antibodies disappear by 49 days post-infection, whereas IgG antibodies persist for up to three months following symptom onset. Similarly, research by Najlaa Assaid et al. [31] indicated that the IgM positivity rate decreased to 19% one month after symptom onset, while the IgG positivity rate increased to 80% three months post-symptom onset. Additionally, Fengling Chen et al. [32] investigated the dynamics of IgM and IgG antibody levels in 1,493 subjects, finding that following complete vaccination, IgM antibodies peaked in the second week and became undetectable by the third week, whereas IgG antibodies remained at a relatively high level throughout a 25-week follow-up period.In our study, 286 individuals, representing 90.7% of the sample, had been infected with COVID-19 within the preceding three months. The levels of IgM antibodies exhibited a downward trend, while the levels of IgG and neutralizing antibodies remained elevated. Consequently, the positive rate of IgM antibodies was relatively low, in contrast to the high positivity rates observed for IgG and neutralizing antibodies. IgM antibodies are typically the first to be secreted during the early stages of pathogen infection but possess a short half-life, rendering them useful indicators of acute infections. In contrast, IgG antibodies emerge later than IgM antibodies and have a longer persistence, serving as markers for mid- to late-stage or past infections. Both IgG and neutralizing antibodies are integral to the humoral immune response against SARS-CoV-2, providing some degree of protection against viral infections. Our study also found that IgM antibody levels gradually increased with advancing age, while IgG and neutralizing antibody levels exhibited a progressive decline. In a related study, Najlaa Assaid et al. [31] conducted a three-month longitudinal follow-up of 169 individuals infected with COVID-19, finding that the median IgM antibody levels in individuals over 50 years of age were slightly higher than those in the cohort under 50, which aligns with the findings of our study.Their research demonstrated that IgG levels in patients over the age of 50 were significantly higher than those observed in younger cohorts at all follow-up intervals, a finding that contradicts our results. This discrepancy may be attributable to the limited sample size of participants over 65 years old in this study, as well as the possibility that the antibodies detected were a result of COVID-19 vaccination, whereas the study by Najlaa Assaid concentrated on antibody levels following natural SARS-CoV-2 infection. It is important to note that antibodies elicited by vaccination differ from those generated through natural immunity. Fengling Chen et al. [32] observed that the younger demographic (18–45 years old) exhibited higher IgG peaks and a longer persistence (approximately two weeks) post-vaccination, whereas the older demographic (> 60 years old) displayed lower IgG peaks with a shorter duration (approximately one week), findings that are consistent with the results of our study. Similarly, Jo-Lewis BANGA NDZOUBOUKOU et al. [33] reported that the mean anti-IgG levels in older age groups (42–52 years old and > 53 years old) were significantly lower than those in the younger group, which also corroborates our findings.Research has demonstrated that age is a significant determinant of vaccine-induced humoral immune responses, independent of the presence of chronic diseases [34]. As individuals age, their immune response to viral infections diminishes. In this study, the median levels of IgM, IgG, and Neutralizing Antibodies were observed to be lower in males compared to females; however, this difference did not reach statistical significance. Contradictorily, Fengling Chen et al. [32] reported that male participants exhibited significantly higher IgM levels than females within the first two weeks post-vaccination, with no significant gender difference by the third week. This inconsistency may be attributed to variations in the intensity of immune responses between sexes. Furthermore, numerous studies have indicated that females generally mount stronger immune responses to viruses and vaccines, potentially due to sex hormones, genetic factors, or environmental influences [35]. Although the present study did not explicitly identify a correlation between gender and the three antibody types, the limited sample size constrains our ability to definitively exclude such an association.The association between antibody levels following SARS-CoV-2 infection and variables such as age and gender necessitates further exploration through extensive cohort studies. In examining the correlation between serum antibody levels and the timing of vaccination, the study observed no statistically significant differences in the levels of the three antibodies at various vaccination time points. Nonetheless, as the duration post-vaccination extended, the median level of IgM antibodies exhibited an upward trend, whereas IgG and neutralizing antibodies (NAb) initially increased and subsequently decreased, with the highest levels of IgG and NAb observed in the 3–6 months post-vaccination group. Hage NOBIN [36] conducted an investigation into the serum levels of SARS-CoV-2 IgM and IgG antibodies among 1,031 residents aged over five years, revealing that the positive rate of serum IgG antibodies was 97.68% in participants who received two vaccine doses, 97.22% in those who received a single dose, and 80% in unvaccinated individuals. Numerous studies have demonstrated that following vaccination, serum IgM levels peak approximately two weeks post-vaccination, while IgG antibody levels peak around three months.While serum IgG antibodies may decline over time, they persist at elevated levels [32, 33], aligning with the findings of this study. This suggests that vaccination enhances the production of IgG and neutralizing antibodies, thereby offering some protection against SARS-CoV-2 infection and contributing to the formation of an immune barrier against COVID-19. However, this study did not evaluate the levels of memory B cells and T cells, limiting its ability to connect the results to acquired immune memory. Concerning the relationship between antibody levels and symptoms at the time of infection, the median IgM antibody level was highest in the respiratory/gastrointestinal/nervous system group and lowest in the nervous system group. Conversely, the nervous system group exhibited the highest median levels of IgG antibodies and neutralizing antibodies. Regarding long-term sequelae six months post-infection, the respiratory system group demonstrated the highest median IgM antibody level, whereas the nervous system group maintained the highest median levels of IgG and neutralizing antibodies.The findings of this study indicate that individuals within the nervous system group exhibited the highest median levels of IgG antibodies and neutralizing antibodies both at the time of infection and during the manifestation of sequelae symptoms. These antibodies are instrumental in sustaining prolonged immune responses. SARS-CoV-2 itself facilitates the production of cross-reactive antibodies, and specific autoantibodies are strongly correlated with long COVID symptoms across various systems. Su et al. (2021) observed that long COVID patients with predominant nervous system symptoms had elevated levels of anti-SARS-CoV-2 nucleocapsid protein IgG, whereas gastrointestinal-related long COVID was associated with multiple autoantibodies during the convalescent phase or even at diagnosis. Notably, IFN-α2 autoantibodies were exclusively linked to respiratory-related long COVID. These findings suggest a connection between SARS-CoV-2 antibodies and the manifestation of COVID-19 symptoms. Nevertheless, the limited number of questionnaires collected in this study constrains the reliability of the results, underscoring the need for further research with larger sample sizes to facilitate a more comprehensive exploration. Peptide-based detection represents a precise and efficient strategy for the identification of SARS-CoV-2 infections [37–39]. This approach utilizes synthetic peptides that replicate key antigenic epitopes of viral surface proteins, such as the spike protein S and nucleocapsid protein N, thereby facilitating the capture of specific antibodies present in the serum of individuals infected with SARS-CoV-2. Employing phage display technology, Li et al. constructed an extensive peptide library and identified several high-affinity epitopes capable of binding antibodies from COVID-19 patients. Notably, the S430-451, S470-481, and S501-515 epitopes within the receptor-binding motif (RBM) demonstrated strong immunogenicity, with the S470-486 epitope proving particularly effective in eliciting neutralizing antibodies that inhibit viral entry [40]. These findings have not only advanced diagnostic assays but have also informed the development of peptide-based vaccines. Repeated administration of selected epitopes can simulate natural T-cell immune responses, thereby offering enhanced protection for immunocompromised individuals.In a parallel investigation, Herrscher et al. [41] utilized proteomics and serological profiling to identify antibody epitopes linked to the severity of COVID-19, specifically NC2, NC3’, and NC5’ within the N protein, M1 within the M protein, and SK27/SK28 within the S protein. The variations in antibody levels targeting these epitopes were found to correlate with disease progression, indicating their potential utility as biomarkers for clinical stratification and prognosis. Similarly, our study has identified differentially expressed peptide antigens across mild, moderate, and severe cases. Although the peak of the pandemic has subsided, the post-acute sequelae of COVID-19 (Long COVID) continue to pose significant challenges. Looking ahead, the development of highly specific peptide-based diagnostic kits and therapeutic vaccines may present promising strategies to mitigate these persistent health issues. In the serum of individuals infected with SARS-CoV-2, the detection of peptide-specific antibodies revealed that antibodies against eight peptides exhibited a greater than 2.5-fold difference when compared to the uninfected group. The tissue localization and expression patterns of the target antigens and their homologous peptides were analyzed using the Protein Atlas database. Our serological analysis of convalescent COVID-19 patients demonstrated significantly elevated antibody titers against peptides associated with reproductive organs, pulmonary, and neurological tissues. Notably, the S68 peptide displayed particularly strong immunoreactivity, a finding of potential clinical significance due to its high homology (82%) with human CHL1, a neural adhesion molecule predominantly expressed in brain tissue. Importantly, CHL1 is abundantly expressed in the CCF-STTG1 glioblastoma cell line, indicating its potential as a model system for studying these autoantibody responses. Conclusion The results underscore the differences in antibody responses associated with age and gender, identify immunodominant peptides that may be pertinent to COVID-19 symptomatology, and propose that cross-reactive antibodies targeting viral-human homologous peptides (such as S68-CHL1) could play a role in autoimmune mechanisms related to COVID-19. These insights hold potential for the development of biomarkers and the formulation of targeted therapeutic interventions. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Abbreviations Angiotensin converting enzyme 2‌ Antinuclear antibodies ANalysis Of VAriance Angiotensin I Atypical memory B cells BerlinCures 007 Cell adhesion molecule with homology to L1CAM CoronaVirus Disease 2019 Electrocardiogram Enzyme linked immunosorbent assay Horseradish peroxidase Interferon Interleukin-4 Interleukin-6 interquartile ranges Magnetic Resonance Imaging Neutralizing Antibodies National Centre for Biotechnology Information programmed cell death protein 1 Post-Acute Sequelae of SARS-CoV-2 Infection Receptor-binding motif Severe Acute Respiratory Syndrome Coronavirus 2 Systemic lupus erythematosus 3,3′,5,5′-Tetramethylbenzidine Shenzhen YHLO Biotech Co., Ltd Author contributions S.Y: Conceptualization, Investigation, Writing – Original Draft, Writing – Review & Editing. Xiaosha Wen: Methodology, Formal Analysis, Visualization, Writing – Original Draft. Yi Lin: Investigation, Data Curation. Ranrong Zhang: Resources, Validation. Dixian Luo: Conceptualization, Supervision, Project Administration, Funding Acquisition. All authors have read and agreed to the published version of the manuscript. Funding This work was financially supported by Municipal Financial Subsidy of Nanshan District Medical Key Discipline Construction; National Key R&D Program of China (2023YFA0915602); Shenzhen Natural Science Foundation Project (JCYJ20240813114502004); General Projects of the National Natural Science Foundation of China (82273236); Newly introduced discipline leader fund project in Nanshan District of Shenzhen City (NSZD2023020). Data availability Data and Material Availability Statement. The datasets analyzed during the current study are not publicly available due to patient privacy concerns, but anonymized data may be made available from the corresponding author upon reasonable request and with permission from the institutional ethics committee. Declarations Ethics approval and consent to participate All participants provided informed consent.This research is guided by the ethical principles of the Declaration of Helsinki. Ethical approval for the study was granted by the Research Ethic Committee of Shenzhen Nanshan People’s Hospital(committee’s reference number:EC-20240202-1014). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Suli Yang and Xiaosha Wen contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Data and Material Availability Statement. The datasets analyzed during the current study are not publicly available due to patient privacy concerns, but anonymized data may be made available from the corresponding author upon reasonable request and with permission from the institutional ethics committee.