Abstract
COVID-19 pandemic continues to challenge the world with a major public health problem, long COVID (LC), which is estimated to affect over 400 million people worldwide. Many unknowns remain regarding the mechanisms involved in LC. We investigated the impact of anti-SARS-CoV-2 antibody and IFN-γ responses on the development of LC and its various phenotypes. We studied a cohort of 137 convalescents following predominantly mild COVID-19 during the first pandemic wave (2020) and up to one-year post-infection. We found 45% of LC cases that were associated with a greater number and duration of acute-phase symptoms. Cardiovascular and/or gastrointestinal symptoms in the acute phase were associated to protection against LC development, while pulmonary, otorhinolaryngological, musculoskeletal and other symptoms were associated with increased risk of LC development. Regarding LC phenotypes, we observed risk associations and potentially deleterious effects of anti-SARS-CoV-2 antibodies for LC symptoms classified as general or other. In contrast, for vital organ-related LC symptoms, we found only protective associations, particularly for cardiovascular symptoms, which indeed had a low prevalence in LC (16%). Collectively, our data suggest that anti-SARS-CoV-2 antibodies play a protective role against vital organ-related LC symptoms, especially cardiovascular symptoms, but are insufficient in preventing or limiting other highly prevalent LC symptoms, such as neurological, psychiatric and pulmonary.
Similar content being viewed by others
Introduction
Despite the predominance of mild COVID-19 cases, it became clear by the end of the first year of the pandemic that a significant proportion of individuals had not fully recovered and developed a condition known as long COVID (LC)1. Different nomenclatures and definitions have been proposed for LC2, which limits comparisons across studies and highlights the need for a consensus1. Depending on the cohort—whether it comprises predominantly mild or hospitalized COVID-19 cases—the prevalence of LC varies from approximately 25%3 to over 70%4. Additionally, the incidence of LC has varied throughout the pandemic, influenced by factors such as circulating variants, symptom severity, and vaccination status5. Currently, the WHO defines LC as the persistence of symptoms or the development of new symptoms three months after SARS-CoV-2 infection onset, unexplained by other causes1. Different LC phenotypes have been identified, as reported in large US health datasets, which indicate that certain LC symptoms tend to co-occur6. Given the clinical heterogeneity and mechanistic complexity of LC, we propose that diverse phenotypes may be identified if additional biological parameters, such as immunological features and reactivation of other pathogens, are examined. These diverse biological features could provide novel insights into potential interconnected underlying mechanisms in LC.
Several non-mutually exclusive mechanisms have been proposed for LC7, including persistent viral reservoirs in various tissues8,9,10, reactivation of latent viral infections such as Epstein–Barr virus (EBV) and human herpes type 6 (HHV-6)11,12, immunological perturbations13, continuous low-grade inflammation14, endothelial cell dysfunction15, microthromboembolism16, dysbiosis17, and immune tolerance breakdown leading to pathological autoimmunity and tissue damage13. Many of these perturbations also occur during acute COVID-1918, suggesting that many SARS-CoV-2-infected individuals are either unable to reestablish homeostasis or take longer to do so. A similar phenomenon is observed in other infections, such as myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)19, which often follows EBV and/or HHV-6 infections20.
Higher symptom numbers during the COVID-19 are reported to be a risk factor for LC21. However, it remains unclear which specific pathophysiological features—individually or in combination—and how the intensity of the biological burden of COVID-19 play significant roles in LC development or in distinct LC phenotypes. Additionally, the extent to which disturbed adaptive immune responses contribute to LC pathophysiology is still debated. Some adaptive immune response perturbations involve inadequate antiviral responses11 and excessive immune responses that cause inflammatory damage22,23. Accordingly, the persistence of high levels of anti-SARS-CoV-2 antibodies following natural infection has been associated with LC24. It is hypothesized that infection may stimulate excessive or dysregulated immune responses, including autoimmune reactions and uncontrolled inflammation25.
We hypothesized that, in addition to the well-recognized protective effects of anti-pathogen antibody and cellular immune responses, potentially aberrant or insufficient immune responses to SARS-CoV-2 could contribute to the development of LC and/or specific LC phenotypes. This hypothesis is supported by robust data in the literature indicating that dysregulated immune responses to pathogens, particularly viruses26, may trigger immune tolerance breakdown and/or persistent inflammation, leading to tissue damage (reviewed in27). Bystander immune activation within the inflammatory microenvironment, tissue damage leading to “hidden” self-antigen exposure, and epitope spreading28 are important mechanisms (reviewed in27). The factors determining beneficial versus deleterious anti-pathogen immune responses have been investigated for decades and may vary in different settings. Key findings highlight the relevance of specific pathogen antigens, which may favor cross-reactivity with self-antigens and trigger inflammatory immune responses29,30,31, as well as the organism´s ability to fine-tune effector versus immunoregulatory immune responses and control inflammation. The capacity to restore homeostasis following infection involves multiple immunoregulatory mechanisms, mediated by immune cells, like Tregs32,33 and antibodies, which may exhibit immunoregulatory functions depending also on their isotype30,32.
To test this hypothesis, we studied the occurrence of LC in a convenience cohort of 137 individuals34 who predominantly experienced mild COVID-19 during the first wave of the pandemic, in 2020, prior to vaccination. We evaluated potential interconnections between antibodies and cellular immune responses to different SARS-CoV-2 antigens, both in the early convalescence period and during ongoing LC, while considering comorbidities and symptoms during the acute and LC phases. We found a 45% prevalence of LC and a significant association with a greater number and duration of acute-phase symptoms. We also detected associations between anti-SAR-CoV-2 antibody responses and the presence or absence of specific LC phenotypes, suggesting both potential protective and deleterious antibody effects. Among the significant associations, we highlight the potentially protective effects of IgA and IgM anti-NP antibodies against cardiovascular and psychiatric LC symptoms and of IgM anti-RBD antibodies against cardiovascular LC symptoms. Our findings indicate that antibody responses against different SARS-CoV-2 antigens likely play dual roles in long-term clinical outcomes and the development of distinct LC phenotypes. We conclude that following predominantly mild COVID-19, diverse anti-SARS-CoV-2 antibody responses may play an important role in protecting against vital organ-related LC symptoms, particularly cardiovascular symptoms, but offer limited protection against neuropsychiatric LC symptoms.
Materials and methods
Cohort
Our convenience cohort initially comprised 209 convalescent individuals infected with SARS-CoV-2 during the early phase of the pandemic, from February to March 2020, who were invited via social media. All participants had medium to high levels of education and socioeconomic status and most experienced mild COVID-19. Blood samples were collected at three timepoints: approximately 40 days after infection onset (T1: April–May 2020), 6 months after infection (T2: October–November 2020), and approximately 1 year after infection (T3: between April and May 2021). Participants who did not need medical assistance were classified as having mild symptoms, while those who were hospitalized were considered to have severe symptoms. At T1, participants completed a questionnaire on demographic data and COVID-19 symptoms. Between the T2 and T3 timepoints, in January 2021, structured electronic questionnaire focusing on long COVID directed was sent for participants for self-reporting the persistence and/or emergence of symptoms. We received responses from 137 participants. LC was defined as the persistence of COVID-19 symptoms at least 3 months following SARS-CoV-2 infection onset, or development of new ones for at least 3 months, with no alternative explanation1. Thirty-two samples collected before 2019 were used as non-COVID-19 controls. The study was approved by CAPPesq (Comissão de \(\acute{E}\)tica Para Análise de Projetos de Pesquisa do HC-FMUSP) and CONEP (Comissão Nacional de Ética em Pesquisa) (CAAE: 30155220.3.0000.0068). All study participants provided written informed consent and the study was conducted in accordance with the Declaration of Helsinki.
ELISA
An in-house enzyme-linked immunosorbent assay (ELISA) was used to quantify antibodies against three SARS-CoV-2 antigens35. Samples were diluted 1:100.
The results are expressed as the ELISA index, calculated by dividing the OD value obtained in the test sample minus the blank condition OD, by the cutoff, which was determined by using the mean OD in pre-pandemic samples (sera from 32 healthy individuals, collected before 2019) + 3 standard deviations (mean pre-pan OD + 3*SD) (index: OD test-OD blank/mean pre-pan OD + 3SD). Values greater than or equal to 1.2 were considered positive.
Virus neutralization assay
SARS‐CoV‐2 (GenBank: MT MT350282) was used for a cytopathic effect (CPE)-based virus neutralization test (VNT) in a Biosafety Level 3 laboratory, as previously described36. Vero cells (ATCC CCL‐81) (5 × 104 cells/mL) were used with diluted inactivated sera (1:20–1:5120) mixed in equal volumes with the virus (100 tissue culture infectious doses, 100% endpoint per well—VNT100). The serum + virus mixtures were transferred to the confluent cell monolayer, incubated at 5% CO2, 37 °C, for 3 days. After 72 h, the plates were analyzed by light microscopy. Gross CPE was observed in Vero cells, distinguishing the presence or absence of CPE‐VNT after fixing and staining plates with amido black. The neutralizing antibody titer was considered as the highest serum dilution capable of neutralizing virus growth. As a positive control, we used serum from a SARS-CoV-2 RT‐qPCR-positive individual displaying plaque reduction in the neutralization test > 1:640.
SARS-CoV-2–specific IFN-γ–producing cells by ELISPOT assay
PBMCs were stimulated with SARS-CoV-2 peptides (5 μg/mL), as previously described34. The cutoff was defined as the mean spot-forming unit (SFU) plus 3 standard deviations (SD) obtained in controls (18 serum samples collected before 2019). The cutoff values were 130 SFU/106 cells for CD4 (n = 20 peptides) and CD8 (n = 26 peptides) megapools34.
Statistical analysis
We used the Kolmogorov‐Smirnov and Shapiro‐Wilk tests to check the variables’ distribution patterns. T tests, Pearson’s correlations or analyses of variance were used for parametric data, and two‐tailed Mann‐Whitney U tests, two‐tailed Spearman’s correlations, Wilcoxon matched‐pairs signed‐rank tests or Kruskal‒Wallis tests for nonparametric variables. Multiple logistic regression models for developing LC or specific LC symptoms were used with the logit link method. For statistical analyses we used R statistical computing language version 4.3.1, considering significant p values < 0.05.
Results
Long COVID (LC) vs nonlong COVID (nLC): demographic and acute phase differences
We compared demographic and clinical features of 137 participants between those who developed long COVID (LC) (45%) and those who did not (nLC) (55%) (Table 1).
We found no significant differences in age (p = 0.05) or sex distribution (p = 0.36) between individuals who developed LC and those who did not. The overall comorbidity frequency was similar between LC (37.1%) and nLC (40%; p = 0.86). We observed no significant differences in hospitalization frequency during COVID-19 (p = 0.62) or hospitalization duration (p = 0.39), between the two groups. Only four individuals required intensive care unit (ICU) admission.
We found a greater number of acute-phase symptoms in the LC group (p < 0.001). Individuals with 4 or more acute phase symptoms had a greater risk of developing LC (OR = 2.74 [1.31–5.73], CI = 95%, p < 0.01). Those who developed LC also had longer acute-phase symptoms duration (median of 15 days vs. 12 days, p = 0.013) and a higher frequency of pulmonary (p = 0.03), neurological (p ≤ 0.0001), psychiatric (p = 0.009), otorhinolaryngological (p < 0.0001) and other symptoms (p = 0.0017) during COVID-19. Specific acute symptoms more frequent in the LC group included shortness of breath (p = 0.0026), fatigue (p = 0.0335), muscle pain (p = 0.0147), anxiety (p = 0.0147), depression (p = 0.0043) and distorted vision (p = 0.0269) (Table 2).
Long COVID: no differential anti-SARS-CoV-2 antibody levels or detection frequency
We found no significant differences in antibody detection frequency or levels of antibodies targeting SARS-CoV-2 antigens (NP, RBD or spike) between LC and nLC groups at any timepoint (p > 0.05). Levels of IgM, IgG or IgA anti-SARS-CoV-2 antibodies (against NP, RBD and spike) were not significantly different at any timepoint between individuals who developed LC and those who did not (all p > 0.05). Overall, the neutralizing antibody detection frequency was high in both groups but did not differ significantly at any time point (p > 0.05) (data not shown). Similarly, neutralizing antibody levels did not significantly differ between the LC and nLC groups (data not shown).
Additionally, we found no differences in antibody levels or detection frequency against other beta coronaviruses (HKU1, NL63, OC43 and 229E) measured at T1 between the groups (p > 0.05). All raw data on antibody levels, cellular responses and symptoms are available as a supplementary file.
Multivariate analysis: cardiovascular and gastrointestinal symptoms in the acute phase are associated with a decreased risk of long COVID
We performed multivariate analyses to determine a valid model for predicting the development of LC, considering sex, age, comorbidities, acute phase symptoms, and anti-SARS-CoV-2 antibody and IFN-γ responses at T1. The results did not show adequate accuracy, reaching only 58.9% accuracy with cross-validation.
A logistic regression model of LC development, considering only grouped acute-phase symptoms, achieved an accuracy of 79.58% in discriminating LC development, with cross-validation. In this model, cardiovascular and gastrointestinal symptoms during the acute phase negatively contributed to the risk of LC development (coefficient values [cv]: −0.239 and −0.758, respectively). In contrast, otorhinolaryngological (cv = 18.761), pulmonary (cv = 0.56), musculoskeletal (cv = 0.325) and other symptoms (cv = 0.582) contributed to an increased risk. When sex and age were included in the model, the accuracy slightly improved to 80.23%. For all other tested models, the accuracy was very low.
Sensitivity analysis using R2 decomposition revealed all significant associations remained robust against potential unmeasured confounding, with robustness values (RVs) consistently exceeding original R2 contributions (Supplementary Table S1). The otorhinolaryngological symptoms (ORLAg) showed the highest robustness (RV = 18.1%, original R2 = 17.8%), indicating this association would require an exceptionally strong confounder (> 18% residual variance explained) to be nullified. Gastrointestinal symptoms (TGIAg) and other predictors demonstrated intermediate robustness (RVs = 2.6–11.4%), all above conventional thresholds for meaningful confounding in observational studies.
The long COVID group
Demographic and clinical characteristics: greater number of long COVID symptoms in previously hospitalized individuals
Symptoms manifested during LC were grouped into system categories, according to NHS criteria, as previously reported37 (Table 3). Neurologic and psychiatric symptoms were the most common.
Within the LC group, only 10 individuals (16.13%) were hospitalized during the acute phase. The sex distribution did not differ significantly between previously hospitalized and non-hospitalized individuals (p = 0.45). However, the median number of symptoms during COVID-19 was higher in previously hospitalized individuals (p = 0.011).
In the LC group, 23 of 62 individuals (37%) had at least one comorbidity, namely, hypothyroidism/Hashimoto’s hypothyroidism (14.5%), hypertension (4.84%), asthma (8.06%), dyslipidemia (1.61%), diabetes (3.23%), depression (6.45%), arrhythmia (3.23%), obesity (5%), arrhythmia (3.23%), breast cancer (1.61%), fibromyalgia (3.23%), osteoporosis (1.61%), rhinitis (1.61%), prostate hyperplasia (1.61%), uterus cancer (1.61%), rheumatoid arthritis (1.61%), ulcerative rectocolitis (1.61%), anxiety (1.61%) and aneurism (1.61%).
Regarding the number of LC symptoms and anti-SARS-CoV-2 antibody features, the only observed difference was lower IgM anti-spike levels at T2 in individuals with ≥ 4 LC symptoms (ELISA index median = 0.087; 0.0–1.285) compared to those with < 4 symptoms (ELISA index median = 0.947; 1.017–3.739, p = 0.049).
Differential correlations between anti-SARS-CoV-2 antibody levels and neutralizing capacity in long COVID vs. non-long COVID: potential protective effect for IgA and deleterious for IgG anti-NP
Some anti-SARS-CoV-2 antibody levels were differentially correlated with neutralization intensity in LC and nLC. Both groups exhibited several correlations at different timepoints, with some shared and others exclusive (Table 4). We considered only strong correlations with r values ≥ 0.6 and p values < 0.05. Following these criteria, the only two correlations found exclusively in nLC were positive correlations between IgA anti-NP levels and neutralizing capacity at T1 (p < 0.001, r = 0.62) and between IgG anti-RBD levels and neutralization at T2 (p < 0.001; r = 0.76). In contrast, in LC, we found exclusive strong correlations at T1 between neutralization and IgM anti-spike (p < 0.001, r = 0.62), IgG anti-NP (p < 0.00, r = 0.60) and IgA anti-RBD (p < 0.001, r = 0.70) antibody levels. Additionally, at T3, we identified exclusive strong positive correlations between neutralization and IgG (p < 0.001, r = 0.67) and IgA (p < 0.001, r = 0.66) anti-spike levels, as well as IgG (p < 0.001, r = 0.72) and IgA (p < 0.001, r = 0.78) anti-RBD levels. Both groups shared strong positive correlations between IgM and IgG anti-RBD levels and neutralization at T1.
Associations involving anti-SARS-CoV-2 antibody isotypes and either the risk of developing LC or specific LC symptoms are summarized in Fig. 1. Overall, associations with anti-NP antibodies suggested more beneficial effects, indicating protection against the development of LC symptoms (risk/protection ratio: 0.62). In contrast, associations with anti-Spike (risk/protection ratio: 1.12) and anti-RBD (risk/protection ratio: 1.5) primarily indicated potential deleterious effects, increasing the risk of specific LC symptoms. Associations involving IgM antibodies against any antigen were less frequent but mostly suggested protective effect (78%) against specific LC symptoms. For IgG and IgA antibodies against any antigen, the associations predominantly indicated deleterious effects, with increased risk (IgG: 59%, IgA: 64%) for specific LC symptoms. A global analysis, considering all antigens and all antibody isotypes, revealed a balanced distribution of associations between risk and protection for developing LC or specific LC symptoms.
Associations of anti-SARS-CoV-2 IgM, IgG, and IgA antibodies (NP, RBD, Spike) with risk or protection for Long COVID (LC) symptoms. This graphical summary illustrates statistically significant protective (green) or risk-associated (red) relationships between anti-SARS-CoV-2 antibody responses and LC development or specific LC symptoms. Dotted-line empty rectangles indicate no significant association. Associations are categorized by antigen (all isotypes) or by isotype (all antigens), with each unique association counted once. Cardio: cardiological LC symptoms, ORL: Otorhinolaryngological LC symptoms, Gastro: gastrointestinal LC symptoms, Neuro: neurological LC symptoms, T1: the first timepoint in our study, early convalescence, approximately 40 days following COVID-19 onset. T2: the second timepoint in our study, approximately 6 months following COVID-19 onset; T3: the third timepoint in our study, approximately 1 year following COVID-19 onset. RBD: receptor-binding domain from the spike protein, NP: nucleocapsid protein, VNT: virus neutralization titer. VNT titers ≥ 20 were considered positive for neutralizing antibodies. All identified differential antibody profiles presented p values < 0.05. Statistical tests used: (i) for frequency detection, chi-square or Fisher tests were used; (ii) for antibody levels, Mann‒Whitney tests, and (iii) for correlation analyses, Spearman correlation.
Associations of anti-SARS-CoV-2 antibody detection with risk or protection for long COVID symptoms
Within the long COVID group, we evaluated whether the detection of specific anti-SARS-CoV-2 antibodies or neutralizing antibodies (VNT) at any timepoint was associated with LC symptoms, either individually or grouped by system category (NHS classification: gastrointestinal, otorhinolaryngological, pulmonary, musculoskeletal, cardiovascular, neuromuscular, psychiatric, general [fever, fatigue/weakness] and other [red eyes, vision alteration, dry eyes, dry mouth, runny nose and phlegm])37. For VNT evaluation, we considered two thresholds: (i) VNT ≥ 20) or (ii) VNT ≥ 160. Within the LC group, we detected significant associations between anti-SARS-CoV-2 antibody and/or VNT detection frequency, suggesting either risk or protection against specific LC symptoms. Anti-RBD and anti-NP antibody detections were associated with both risk and protection for distinct LC symptoms, while anti-Spike antibody detection was linked only to increased risk. Detection of any neutralizing antibody level (VNT ≥ 20), suggested protection against cardiovascular symptoms during LC. Conversely, detection of higher neutralizing capacity (≥ 160) was exclusively associated with risk of general symptoms (including fatigue) or symptoms classified as other during LC (Table 5).
We highlight some associations indicating protection against some specific LC symptoms: (i) protection against psychiatric symptoms was associated with IgM anti-NP detection at T3; (ii) protection against depression was linked to IgG anti-RBD detection at T3; (iii) protection against cardiovascular symptoms was associated with IgA anti-NP detection at T1; and (iv) protection against tachycardia and chest pain was linked to the detection of any neutralization intensity (VNT > 20) at T1 and T2 (for tachycardia) and at T1 (for chest pain). Additionally, we identified some associations indicating a higher risk of specific LC symptoms: (i) general symptoms were associated with IgA anti-NP detection at T1, IgG anti-RBD detection at T2; and the detection of exclusively high neutralizing capacity (VNT > 160) at T1, T2 and T3; (ii) fatigue was associated with IgG anti-NP detection at T1, IgG anti-RBD detection at T2; and the detection of exclusively high neutralizing capacity (VNT > 160) at T1, T2 and T3; and (iii) symptoms classified as other were associated with anti-spike IgM and IgA detection at T1.
Higher levels of anti-SARS-CoV-2 antibodies are associated with protection against vital organ-related long COVID symptoms but not against general symptoms
Several LC symptoms were associated with higher or lower levels of anti-SARS-CoV-2 antibodies at different timepoints (Fig. 2). Higher antibody levels were associated with protection against LC symptoms: (i) cardiovascular: IgM anti-NP (median ELISA index = 2.25 vs 0.833, p = 0.0435) and IgM anti-RBD (ELISA index = 2.06 vs 0.689, p = 0.0155) at T1, IgM anti-Spike, at T2 (ELISA index = 0.575 vs 0.0, p = 0.0150); (ii) pulmonary: IgG anti-Spike, at T2 (ELISA index = 1.78 vs 10.2, p = 0.045); (iii) gastrointestinal (GI): IgG anti-Spike, at T2 (ELISA index = 0.06 vs 9.735, p = 0.0163); (iv) psychiatric: IgM anti-NP, at T2 (ELISA index = 0.375 vs 0.08, p = 0.018); and (v) otorhinolaryngological (ORL): IgG anti-Spike, at T1 (ELISA index = 5.88 vs 10.06, p = 0.0332); and (vi) musculoskeletal: IgA anti-RBD, at T2 (ELISA index = 0.73 vs 0.085, p < 0.001). Individuals displaying the above LC symptoms had significantly lower levels of the above-cited antibodies. These protection associations showing higher antibody levels in individuals who did not present vital organ-related LC symptoms were observed only at T1 and T2 timepoints, mostly against Spike (4 of 8 associations), anti-RBD (2 of 8) and anti-NP (2 of 8) antibodies.
Differential antibody levels at different timepoints and long COVID symptoms. Anti-SARS-CoV-2 antibody levels were significantly different between the LC and nLC groups. (−): specified symptom not present; ( +): presence of specified symptom; NP: nucleocapsid protein; RBD: receptor-binding domain protein; S: spike protein; ORL: otorhinolaryngological LC symptoms; ELISA index: OD/cutoff. n = 137 (*p < 0.05; **p < 0.001, Mann‒Whitney).
In contrast, for some LC symptoms involving nonvital organs—general or other—we found higher levels of certain anti-SARS-CoV-2 antibodies. Higher levels of these antibodies were associated with LC symptoms: (i) general: IgA anti-NP (T1) (median ELISA index = 2.964 vs 2.007, p = 0.0349), IgG anti-RBD (T2) (median ELISA index = 3.403 vs. 0.381, p = 0.0081) (T3) (median ELISA index = 2.68 vs 1.259, p = 0.0261), IgA anti-Spike, (T3) (median ELISA index = 3.45 vs. 1.465, p = 0.0174), and (ii) other: IgA (median ELISA index = 4.485 vs 2.017, p = 0.044) and IgG (median ELISA index = 4.273 vs 2.842, p = 0.048) anti-NP, IgA anti-Spike (median ELISA index = 6.8 vs 2.095, p = 0.0131), all in T1.
All associations between anti-SARS-CoV-2 antibody responses and LC development or specific LC symptoms are summarized in Fig. 3. In multivariate analyses, the levels of all antibodies measured at any timepoint did not discriminate the occurrence of any LC symptom, whether analyzed by system category or individually.
Associations between anti-SARS-CoV-2 antibody responses and system-specific long COVID symptoms. Protection associations are represented in green balloons and risk associations in red balloons. A) Protection associations found exclusively for vital organ-related symptoms. Differential anti-SARS-CoV-2 antibody profiles and neutralizing antibody detection (VNT ≥ 20) were associated with protection against several system-stratified LC symptoms. B) Risk associations were observed for LC symptoms classified as general or other. Overall, risk and protection associations were balanced for LC development or specific LC symptoms. ORL: Otorhinolaryngological LC symptoms, Gastro: gastrointestinal LC symptoms, Neuro: neurological LC symptoms, T1: the first timepoint in our study, during early convalescence, approximately 40 days following COVID-19 onset. T2: the second timepoint in our study, at approximately 6 months following COVID-19 onset; T3: the third timepoint in our study, at approximately 1 year following COVID-19 onset. RBD: receptor-binding domain from the spike protein, NP: nucleocapsid protein, VNT: virus neutralization titer. VNT titers ≥ 20 were considered positive for neutralizing antibodies. All the identified differential antibody profiles showed p values < 0.05. The statistical tests used were: (i) for detection frequency, chi-square or Fisher tests; (ii) for antibody level analyses, Mann‒Whitney tests; and (iii) for correlation analyses, Spearman correlation test.
Long COVID: no differential IFN-γ-cellular response to SARS-CoV-2 peptides in the acute phase.
We assessed the T-cell SARS-CoV-2-specific IFN-γ response by EliSpot, using individual peptides and 2 peptide pools for CD4 and CD8 T cells. The frequency and magnitude of the IFN-γ T-cell response to SARS-CoV-2 in silico-predicted peptides, measured at T1, did not significantly differ between individuals who developed LC compared to those who did not (p > 0.05).
Discussion
We investigated whether specific clinical and immune response features triggered during COVID-19 impact the development of long COVID and whether anti-SARS-CoV-2 immune responses could be involved in the risk of or protection against the occurrence of specific LC symptoms.
Overall, we found no striking associations between anti-SARS-CoV-2 antibody or IFN-γ responses and LC development. However, we identified several associations suggesting risk or protection for different LC clinical phenotypes based on anti-SARS-CoV-2 antibody profiles. All associations indicating potential deleterious effects of anti-SARS-CoV-2 antibodies involved LC symptoms classified as general or other. In contrast, for vital organ-related LC symptoms, we found exclusively protective associations, particularly against cardiovascular, pulmonary, psychiatric, gastrointestinal, musculoskeletal and otorhinolaryngological symptoms.
Collectively, our data suggest that, following predominantly mild COVID-19, antibodies targeting diverse SARS-CoV-2 antigens play a protective role against vital organ-related LC symptoms, especially cardiovascular (16%) and gastrointestinal symptoms (9.6%), both of which were infrequent in our study.
In our cohort of primarily mild COVID-19 convalescents from the first wave of the pandemic, the prevalence of LC was 45%, with no association found with hospitalization during the acute phase. Although most reports indicate a higher prevalence following severe COVID-1938, our findings, along with those of others24, emphasize that LC can develop even after mild COVID-19. This highlights the mechanistic complexity involved in LC development, as well as the biological burden experienced during the acute phase of COVID-19. Since our study’s focused the recruitment during Brazil’s first wave (February–March 2020), early cases included primarily travelers and their contacts, resulting in a higher socioeconomic status (SES) cohort, also due to limited testing availability at that time. While this homogeneity strengthens internal validity by reducing acute-phase care disparities, it may limit generalizability to lower-SES populations who face greater comorbidity burdens and healthcare barriers39. Socioeconomic factors influence both immune responses and post-acute outcomes40, suggesting potential variability in LC manifestations across populations. Future studies should utilize public health systems and community-based recruitment for broader representation. Still, our findings raise important hypothesis baseline LC mechanisms in a controlled early-pandemic cohort.
We recognize the use of self-reported symptoms is a limitation of the study, which may be subject to recall bias. While common in pandemic research and necessary for remote data collection, it is true this approach lacks clinical validation. Notably, our finding that neurological symptoms predominated aligns with objective measures in other studies41,43, supporting our observations. Future work should combine patient-reported outcomes with clinical assessments and biomarker data to strengthen phenotype classification.
In concordance with other studies41, we found a greater number and duration of symptoms during acute COVID-19 associated with the development of LC. This likely reflects an impaired capacity to control viral replication and/or persistent inflammation affecting several organs and systems, particularly the lungs and the nervous system (CNS) in our study. In the LC group, we observed a higher frequency of pulmonary, neurologic and psychiatric symptoms during COVID-19. Indeed, both the lungs and the CNS suffer extensive damage during SARS-CoV-2 infection40,41, resulting in persistent inflammation and dysregulated immunological activation and repair mechanisms42,43. Moreover, SARS-CoV-2 detection in postmortem tissues indicates viral persistence, mostly reported in the lungs44, the CNS45 and the heart45. Notably, these organs are severely and persistently affected by COVID-1946, even following apparent clinical recovery44,47.
The contribution of biological burden intensity during COVID-19 to LC development involves immunological perturbations, including potentially deleterious anti-SARS-CoV-2 immune responses48. We found no differential anti-SARS-CoV-2 humoral or IFN-γ responses in the LC group. Despite limited and conflicting literature, LC has been associated with a decline in NP-specific interferon-γ-producing CD8 + T cells, suggesting that a sustained IFN-γ response may be protective49. In contrast, LC has also been associated with persistently elevated levels of type I and III interferons50, TNF, IP10, IL-1β, and IL-651. Our findings suggest that the anti-SARS-CoV-2 IFN-γ response, in early convalescence is unlikely to play a major role in promoting or preventing LC following mild COVID-19. However, we only studied IFN-γ, other cytokines may be relevant.
Nonetheless, we found differential correlations between the neutralizing capacity and anti-SARS-CoV-2 antibody levels when comparing LC and nLC participants. The exclusive strong correlations in nLC suggest that time synchronization between certain antigen-specific antibody levels and neutralization may play a role in limiting infection-induced biological distress, contributing to reestablishing homeostasis and preventing LC development52. These potentially protective correlations were noted only at T1 (IgA anti-NP) and T2 (IgG anti-RBD). Synchrony between neutralization and the levels of some anti-SARS-CoV-2 antibodies may be a relevant LC-preventive factor early in convalescence and during the six months following COVID-19. Considering that the NP is located outside the viral/host molecular binding region, in addition to the ability to neutralize viruses, other anti-NP antibody-mediated functions are likely relevant in preventing LC. Indeed, diverse antibody effector functions have been highlighted in COVID-1953 but not in LC. Importantly, SARS-CoV-2 NPs may accumulate in some cell types, as myenteric plexus neurons and megakaryocytes54, and bind to proinflammatory receptors, promoting inflammation and tissue damage55. Accordingly, anti-NP antibodies could play a protective role by promoting NP protein clearance from tissues, avoiding/limiting damage, in synchronic action with antibody neutralization.
Intriguingly, a strong positive correlation between neutralization and anti-NP was also associated with the risk for LC, also at T1, specifically for IgG rather than IgA. Notably, IgG anti-NP and the NP itself can induce IL-6 in vitro56. Thus, IgG but not IgA anti-NPs could favor sustained inflammation, contributing to LC development. Furthermore, high IgG anti-NP levels are correlated with increased nerve growth factor (NGF) levels57, and increased NGF is associated with sustained inflammation58, suggesting that this correlation may reflect a biological setting prone to chronic inflammation.
Both potentially protective and deleterious correlations between neutralizing capacity and anti-SARS-CoV-2 antibody levels/ involved all three antigens and all immunoglobulin isotypes. Our data indicate that in addition to their protective role in controlling infection, anti-SARS-CoV-2 antibodies may also have deleterious effects, as reported for anti-spike antibodies and the spike protein itself59, anti-S2 spike antibodies60 displaying particular glycosylation61, and for IgG anti-NP62. Certain IgG subclasses may mediate antibody-dependent enhancement (ADE), where antibodies facilitate viral entry into cells, potentially exacerbating disease severity64. Pro-inflammatory IgG responses, such as those involving afucosylated IgG, have also been associated with severe COVID-19 and cytokine storms48.
Our best multivariate analysis did not include antibody levels and revealed a model with approximately 80% accuracy in predicting LC development, based solely on grouped acute-phase symptoms. This accuracy is very similar to the value found for the model that included sex and age. Although significant sex and age differences have been reported regarding symptoms and severity of COVID-19 (reviewed by65 we did not observe a substantial impact. Our results are consistent with meta-nalysis data, indicating no significant impact of age, but they are discordant regarding the impact of sex66. Additionally, long COVID-19 has been linked to certain pre-existing medical comorbidities, such as diabetes and hypertension66 but only few participants of our cohort presented these conditions.
We found a protective effect for acute-phase gastrointestinal and cardiovascular symptoms, against developing LC, whereas otorhinolaryngological, pulmonary, musculoskeletal and other symptoms, in the acute phase, were associated with an increased risk. It is challenging to provide strong biological interpretations for these intriguing results, especially considering that the gastrointestinal and cardiovascular systems are potential reservoirs for SARS-CoV-267. Nonetheless, within the LC group, we detected several anti-SARS-CoV-2 antibody responses associated with protection against cardiovascular symptoms during LC. Moreover, cardiovascular and gastrointestinal symptoms were the least frequently observed LC symptoms (16% and 9.7%, respectively). In contrast, anti-SARS-CoV-2 immune responses triggered during acute infection appear insufficient to prevent neuropsychiatric LC symptoms, which were highly prevalent in our cohort (neuro: 95%, psychiatric: 55%).
Within the LC group, we detected differential anti-SARS-CoV-2 antibody responses associated with various LC symptoms. Overall, the detection frequency and/or levels of anti-SARS-CoV-2 antibodies and/or neutralizing capacity (VNT) were associated with both the risk of and protection against specific LC symptoms. All associations indicating potential deleterious effects of anti-SARS-CoV-2 antibodies were exclusively observed for LC symptoms involving non-vital organs (general and/or other symptoms). In contrast, for LC symptoms involving vital organs, we observed only associations suggesting protective effects of anti-SARS-CoV-2 antibodies, primarily against cardiovascular LC symptoms, at T1 and T2 and occasionally at T3. Specifically, cardiovascular, neurologic, psychiatric, pulmonary, gastrointestinal and musculoskeletal LC symptoms were associated with either lower levels or a lower detection frequency of certain anti-SARS-CoV-2 antibodies. Thus, their presence and/or higher levels could be beneficial in resolving or limiting the infection burden, protecting against damage to vital organs62.
It is interesting to connect these findings to our multivariate analysis results indicating that the occurrence of cardiovascular symptoms during COVID-19 is protective against development of LC. We suggest that anti-SARS-CoV-2 antibodies may contribute to the resolution of acute cardiovascular symptoms, likely providing later protection against cardiovascular LC symptoms. While our stratified analyses face sample size limitations, robustness testing revealed that unmeasured confounders would need to explain 7.6–18.1% of residual variance to nullify our observed associations—values substantially exceeding the original R2 contributions (0.12–17.8%). This suggests our models are particularly robust for otorhinolaryngological associations (robustness R2 = 18.1%) and gastrointestinal symptoms (11.4%). Nevertheless, we caution against overinterpreting subgroup findings with smaller effect sizes such as cardiovascular symptoms at 2.6%. Notably, a protective role against fatal outcomes was suggested for increased IgA and IgG anti-NP levels in patients with severe COVID-1968. These findings are significant, as cardiovascular involvement in severe COVID-19 can lead to fatal outcomes.
Another interesting finding was the protective association of antibody persistence at T3 with specific LC symptoms: IgM anti-NP against tachycardia and psychiatric symptoms and IgG anti-RBD against depression. The detection of these antibodies approximately 1 year after COVID-19 onset suggests that, even at a later timepoints, some antibodies may play a role in limiting or controlling central nervous system (CNS) distress. In contrast, the persistence of higher levels of neutralizing antibodies (> 1/160) was associated with the occurrence of general or other LC symptoms, suggesting that neutralization may also have excessive or deleterious effects in some tissues but not others, as previously reported68.
Although antibody-mediated protective mechanisms regarding LC phenotypes remain unclear, we highlight some potential connections involving specific SARS-CoV-2 antigens. NPs can form multiprotein complexes, aggregating with α-synuclein, possibly with amyloid fibrils55, and accumulate in the brain microvasculature69. If this accumulation is involved in psychiatric symptoms, anti-NP antibodies could potentially contribute to NP clearance, thereby limiting neuropsychiatric disturbances. Moreover, late detection of IgM anti-NP at T3 may indicate viral persistence, suggesting that these antibodies are likely beneficial in limiting virus-induced damage. This is plausible since the NP is an early-response antigen that induces the production of antibodies that decrease over time70. Notably, higher levels of IgM anti-NP at T1 or T2 were also associated with protection against cardiovascular LC symptoms.
Nevertheless, we still face many mechanistic unknowns71. It is challenging to interpret why the detection of IgA anti-NP at T1 is associated with protection against cardiovascular LC symptoms while simultaneously posing a risk for general and other symptoms. These distinct and seemingly opposing activities likely depend on various biological factors, including the phase of the disease process, antibody levels and potential cross-reactivity with self-antigens, which could favor pathological autoimmunity, or even the antibody ability to bind to the FcRIIB receptor (CD32)72, providing inhibitory signals, and downregulating inflammatory/effector functions. Additionally, the categorization of symptoms reflects the involvement of diverse tissues with varying sensitivities to viral tropism and its pathogenic effects, along with a complex array of self-antigens that may serve as potential targets for pathological autoimmunity.
Although our data reveal intriguing associations between anti-SARS-CoV-2 antibody profiles and LC phenotypes, we emphasize these findings are observational and mechanistic interpretations remain speculative. While antibodies may show dual roles, protective via viral clearance versus harmful via inflammation, the identified signatures—particularly IgA/IgG anti-NP links to cardiovascular protection—offer testable hypotheses. Future work on antibody functional activity such as ADCC and cytokine profiling and viral persistence measurements will contribute to understanding the diverse roles of anti-SARS-COV-2 antibodies in long covid.
Differential complex biological settings likely impact the predominance of protective or deleterious antibody effects and LC phenotypes. These factors include genetic and epigenetic backgrounds, past immunological experiences73, concomitant reactivation of other viral infections, as reported in LC74, and the beneficial effects of anti-SARS-CoV-2 vaccination5. It is important to note that the participants in our study were not previously vaccinated. Therefore, our findings may not be applicable to diverse populations worldwide, as exposure to various infections and multiple environmental challenges can influence immune system activity and, consequently, the capacity to manage SARS-CoV-2 infection and LC development.
Further underscoring the complexity of LC, several systemic immune profile perturbations may persist even six months following COVID-19, even in asymptomatic individuals recovered from severe COVID-19, indicating that homeostasis has not been reestablished75. This disturbed immunological profile supports the idea of some sort of silent long COVID, that may eventually manifest as critical clinical events, such as acute myocardial infarction or cerebral vascular accidents. These clinically silent biological perturbations may, at least in part, account for the excess number of deaths reported worldwide following the COVID-19 pandemic76.
Collectively, our data demonstrate that LC following predominantly mild COVID-19 presents diverse symptoms affecting major organ systems, including the central nervous, pulmonary and cardiovascular systems. We show that different system-category symptoms are differentially associated with anti-SARS-CoV-2 antibody profiles, both in early convalescence and later, during ongoing LC symptoms, indicating both potentially protective and deleterious effects of anti-SARS-CoV-2 antibodies.
We highlight that following predominantly mild COVID-19, diverse anti-SARS-CoV-2 antibody responses play important roles in protecting against vital organ-related LC symptoms, especially cardiovascular symptoms. Our data underscore the complexity of the potential involvement of anti-SARS-CoV-2 immune responses in either protecting against or contributing to the development of different LC phenotypes.
The complexity of long COVID emphasizes the need for investigations into potential mechanisms, integrating multiple biological features, as these factors are likely intertwined and may impact different clinical outcomes. A significant challenge ahead is to identify the major critical mechanisms involved in the LC and its phenotypes, as well as to develop strategies that promote or enhance the organism’s capacity to reorganize itself and reestablish homeostasis.
Data availability
All raw data on antibody levels, cellular responses and symptoms are available as a supplementary file.
References
Soriano, J. B., Murthy, S., Marshall, J. C., Relan, P. & Diaz, J. V. A clinical case definition of post-COVID-19 condition by a Delphi consensus. Lancet Infect. Dis. 22, e102–e107 (2022).
Fernández-de-las-Peñas, C. Long COVID: Current definition. Infection 50, 285–286 (2022).
Havervall, S. et al. Symptoms and functional impairment assessed 8 months after mild COVID-19 among health care workers. JAMA 325(19), 2015–2016. https://doi.org/10.1001/jama.2021.5612 (2021).
Huang, C. et al. 6-month consequences of COVID-19 in patients discharged from hospital: A cohort study. Lancet 397(10270), 220–232 (2021).
Trinh, N. T. et al. Effectiveness of COVID-19 vaccines to prevent long COVID: Data from Norway. Lancet Respir Med 12, e33–e34 (2024).
Zhang, H. et al. Data-driven identification of post-acute SARS-CoV-2 infection subphenotypes. Nat. Med. 29(1), 226–235 (2023).
Perumal, R. et al. Long COVID: A review and proposed visualization of the complexity of long COVID. Front. Immunol. https://doi.org/10.3389/fimmu.2023.1117464 (2023).
Cheung, C. C. L. et al. Residual SARS-CoV-2 viral antigens detected in GI and hepatic tissues from five recovered patients with COVID-19. Gut 71(1), 226 (2022).
Leppkes, M. & Neurath, M. F. Rear window—What can the gut tell us about long-COVID?. Gastroenterology 163(2), 376–378 (2022).
Swank, Z. et al. Persistent circulating severe acute respiratory syndrome coronavirus 2 spike is associated with post-acute coronavirus disease 2019 sequelae. Clin. Infect. Dis. 76(3), E487–E490 (2023).
Su, Y. et al. Multiple early factors anticipate post-acute COVID-19 sequelae. Cell 185(5), 881-895.e20 (2022).
Zubchenko, S., Kril, I., Nadizhko, O., Matsyura, O. & Chopyak, V. Herpesvirus infections and post-COVID-19 manifestations: A pilot observational study. Rheumatol Int. 42(9), 1523–1530 (2022).
Altmann, D. M., Whettlock, E. M., Liu, S., Arachchillage, D. J. & Boyton, R. J. The immunology of long COVID. Nat Rev Immunol. 23, 618–634 (2023).
Woodruff, M. C. et al. Chronic inflammation, neutrophil activity, and autoreactivity splits long COVID. Nat Commun. https://doi.org/10.1038/s41467-023-40012-7 (2023).
Haffke, M. et al. Endothelial dysfunction and altered endothelial biomarkers in patients with post-COVID-19 syndrome and chronic fatigue syndrome (ME/CFS). J Transl Med. 20(1), 138. https://doi.org/10.1186/s12967-022-03346-2 (2022).
Pretorius, E. et al. Persistent clotting protein pathology in Long COVID/Post-Acute Sequelae of COVID-19 (PASC) is accompanied by increased levels of antiplasmin. Cardiovasc. Diabetol. https://doi.org/10.1186/s12933-021-01359-7 (2021).
Ancona, G. et al. Gut and airway microbiota dysbiosis and their role in COVID-19 and long-COVID. Front. Immunol. https://doi.org/10.3389/fimmu.2023.1080043 (2023).
Zhu, Y., Sharma, L. & Chang, D. Pathophysiology and clinical management of coronavirus disease (COVID-19): a mini-review. Front. Immunol. 14, 1116131 (2023).
Chaves-Filho, A. M., Braniff, O., Angelova, A., Deng, Y. & Tremblay, M. È. Chronic inflammation, neuroglial dysfunction, and plasmalogen deficiency as a new pathobiological hypothesis addressing the overlap between post-COVID-19 symptoms and myalgic encephalomyelitis/chronic fatigue syndrome. Brain Res. Bull. 201, 110702 (2023).
Rasa, S. et al. Chronic viral infections in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). J. Transl. Med. https://doi.org/10.1186/s12967-018-1644-y (2018).
Augustin, M. et al. Post-COVID syndrome in non-hospitalised patients with COVID-19: A longitudinal prospective cohort study. Lancet Reg. Health - Eur. 1, 6 (2021).
Chen, N. et al. Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: A descriptive study. Lancet 395(10223), 507–513 (2020).
Vigón, L. et al. Impaired cytotoxic response in pbmcs from patients with COVID-19 admitted to the ICU: Biomarkers to predict disease severity. Front. Immunol. 26, 12 (2021).
Peghin, M. et al. Post–COVID-19 syndrome and humoral response association after 1 year in vaccinated and unvaccinated patients. Clin. Microbiol. Infect. 28(8), 1140–1148 (2022).
Mangge, H. et al. Immune responses against sars-cov-2—questions and experiences. Biomedicines 9, 1342 (2021).
Smatti, M. K. et al. Viruses and autoimmunity: A review on the potential interaction and molecular mechanisms. Viruses 11, 762 (2019).
Sundaresan, B., Shirafkan, F., Ripperger, K. & Rattay, K. The role of viral infections in the onset of autoimmune diseases. Viruses 15, 782 (2023).
Lehmann, P. V., Forsthuber, T., Miller, A. & Sercarz, E. E. Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358(6382), 155–157. https://doi.org/10.1038/358155a0 (1992).
Zhao, Z. S., Granucci, F., Yeh, L., Schaffer, P. A. & Cantor, H. Molecular mimicry by herpes simplex virus-type 1: Autoimmune disease after viral infection. Science 279(5355), 1344–1347. https://doi.org/10.1126/science.279.5355.1344 (1998).
Suliman, B. A. Potential clinical implications of molecular mimicry-induced autoimmunity. Immun. Inflam. Dis. https://doi.org/10.1002/iid3.1178 (2024).
Fujinami, R. S. & Oldstone, M. B. A. Amino acid homology between the encephalitogenic site of myelin basic protein and virus: Mechanism for autoimmunity. Science 230(4729), 1043–1045. https://doi.org/10.1126/science.2414848 (1985).
Arpaia, N. et al. A distinct function of regulatory T cells in tissue protection. Cell 162(5), 1078–1089 (2015).
Bilate, A. M. & Lafaille, J. J. Induced CD4 +Foxp3 + regulatory T cells in immune tolerance. Ann. Rev. Immunol. 30, 733–758 (2012).
Fernandes, E. R. et al. Time-dependent contraction of the SARS-CoV-2–specific T-cell responses in convalescent individuals. J. Allergy Clin. Immunol. Global. 1(3), 112–121 (2022).
Oliveira, J. R. et al. Immunodominant antibody responses directed to SARS-CoV-2 hotspot mutation sites and risk of immune escape. Front Immunol. 5, 13 (2023).
Wendel, S. et al. Screening for SARS-CoV-2 antibodies in convalescent plasma in Brazil: Preliminary lessons from a voluntary convalescent donor program. Transfusion 60(12), 2938–2951 (2020).
COVID-19 rapid guideline: managing the long-term effects of COVID-19. London: National Institute for Health and Care Excellence (NICE); (2020).
O’Mahoney, L. L. et al. The prevalence and long-term health effects of long covid among hospitalised and non-hospitalised populations: A systematic review and meta-analysis. EClinicalMedicine. 1, 55 (2023).
Magesh, S. et al. Disparities in COVID-19 outcomes by race, ethnicity, and socioeconomic status: A systematic review and meta-analysis. JAMA Netw. Open. 4(11), e2134147–e2134147. https://doi.org/10.1001/jamanetworkopen.2021.34147 (2021).
Subramanian, A. et al. Symptoms and risk factors for long COVID in non-hospitalized adults. Nat. Med. 28(8), 1706–1714. https://doi.org/10.1038/s41591-022-01909-w (2022).
Sudre, C. H. et al. Attributes and predictors of long COVID. Nat Med. 27(4), 626–631 (2021).
Lamers, M. M. & Haagmans, B. L. SARS-CoV-2 pathogenesis. Nat. Rev. Microbiol. 20, 270–284 (2022).
Maliha, S. T., Fatemi, R. & Araf, Y. COVID-19 and the brain: Understanding the pathogenesis and consequences of neurological damage. Mol. Biol. Rep. 51(1), 318 (2024).
Bussani, R. et al. Persistent SARS-CoV-2 infection in patients seemingly recovered from COVID-19. J. Pathol. 259(3), 254–263 (2023).
Stein, S. R. et al. SARS-CoV-2 infection and persistence in the human body and brain at autopsy. Nature 612(7941), 758–763 (2022).
Doerschug, K. C. & Schmidt, G. A. Pulmonary aspects of COVID-19. Ann. Rev. Med. https://doi.org/10.1146/annurev-med-042220 (2022).
Douaud. G., Lee, S., Alfaro-Almagro, F., Arthofer, C., Wang, C., McCarthy, P., et al. SARS-CoV-2 is associated with changes in brain structure in UK Biobank. medRxiv [Internet].2021.06.11.21258690 (2022). Available from: http://medrxiv.org/content/early/2022/03/02/2021.06.11.21258690.abstract.
Larsen, M. D. et al. Afucosylated IgG characterizes enveloped viral responses and correlates with COVID-19 severity. Science 371(6532), 8378. https://doi.org/10.1126/science.abc8378 (2021).
Peluso, M. J. et al. Long-term SARS-CoV-2-specific immune and inflammatory responses in individuals recovering from COVID-19 with and without post-acute symptoms. Cell Rep. 36(6), 109518 (2021).
Phetsouphanh, C. et al. Immunological dysfunction persists for 8 months following initial mild-to-moderate SARS-CoV-2 infection. Nat. Immunol. 23(2), 210–216 (2022).
Peluso, M. J. et al. Markers of immune activation and inflammation in individuals with postacute sequelae of severe acute respiratory syndrome coronavirus 2 infection. J. Infect. Dis. 224(11), 1839–1848 (2021).
Gaebler, C. et al. Evolution of antibody immunity to SARS-CoV-2. Nature 591(7851), 639–644 (2021).
Zhang, A. et al. Beyond neutralization: Fc-dependent antibody effector functions in SARS-CoV-2 infection. Nat. Rev. Immunol. 23, 381–396 (2023).
Gray-Rodriguez, S. et al. Multisystem screening reveals SARS-CoV-2 in neurons of the myenteric plexus and in megakaryocytes. J. Pathol. 257(2), 198–217. https://doi.org/10.1002/path.5878 (2022).
Semerdzhiev, S. A., Fakhree, M. A. A., Segers-Nolten, I., Blum, C. & Claessens, M. M. A. E. Interactions between SARS-CoV-2 N-Protein and α-Synuclein accelerate amyloid formation. ACS Chem. Neurosci. 13(1), 143–150 (2022).
Nakayama, E. E. et al. Anti-nucleocapsid antibodies enhance the production of IL-6 induced by SARS-CoV-2 N protein. Sci. Rep. https://doi.org/10.1038/s41598-022-12252-y (2022).
Usai, C. et al. The β-NGF/TrkA signalling pathway is associated with the production of anti-nucleoprotein IgG in convalescent COVID-19. Front. Immunol. 14, 12 (2022).
Minnone, G., De Benedetti, F. & Bracci-Laudiero, L. NGF and its receptors in the regulation of inflammatory response. Int. J. Mol. Sci. 18, 1028 (2017).
Theoharides, T. C. Could SARS-CoV-2 Spike protein be responsible for long-COVID syndrome?. Mol. Neurobiol. 59(3), 1850–1861. https://doi.org/10.1007/s12035-021-02696-0 (2022).
Wang, J. et al. IgG against human betacoronavirus spike proteins correlates with SARS-CoV-2 anti-spike IgG responses and COVID-19 disease severity. J. Infect. Dis. 226(3), 474–484 (2022).
Bye, A. P. et al. Aberrant glycosylation of anti-SARS-CoV-2 spike IgG is a prothrombotic stimulus for platelets. Blood 138(16), 1481–1489. https://doi.org/10.1182/blood.2021011871 (2021).
Klein, J. et al. Distinguishing features of long COVID identified through immune profiling. Nature 623(7985), 139–148 (2023).
Breedveld, A. & Van Egmond, M. IgA and FcαRI: Pathological roles and therapeutic opportunities. Front. Immunol. https://doi.org/10.3389/fimmu.2019.00553 (2019).
Lee, W. S., Wheatley, A. K., Kent, S. J. & DeKosky, B. J. Antibody-dependent enhancement and SARS-CoV-2 vaccines and therapies. Nat. Microbiol. 5(10), 1185–1191 (2020).
De Francia, S. et al. The influence of sex, gender, and age on COVID-19 data in the piedmont region (Northwest Italy): The virus prefers men. Life 12(5), 643 (2022).
Notarte, K. I. et al. Age, sex and previous comorbidities as risk factors not associated with SARS-CoV-2 infection for long COVID-19: A systematic review and meta-analysis. J. Clin. Med. 11, 7314 (2022).
Hany, M. et al. Incidence of persistent SARS-CoV-2 gut infection in patients with history of COVID-19: Insights from endoscopic examination. Endosc. Int. Open. 12(01), E11–E22 (2023).
Servian, C. D. et al. Distinct anti-NP, anti-RBD and anti-Spike antibody profiles discriminate death from survival in COVID-19. Front. Immunol. 14, 1206979 (2023).
Demarino, C. et al. Detection of SARS-CoV-2 nucleocapsid and microvascular disease in the brain: A case report. Neurology 100(13), 624–628 (2023).
Murrell, I. et al. Temporal development and neutralising potential of antibodies against SARS-cov-2 in hospitalised COVID-19 patients: An observational cohort study. PLoS ONE 16(1), e0245382 (2021).
Zhao, J. et al. Antibody responses to SARS-CoV-2 in patients with novel coronavirus disease 2019. Clin. Infect. Dis. 71(16), 2027–2034 (2020).
Morris, A. B. et al. Signaling through the Inhibitory Fc receptor FcγRIIB Induces CD8+ T cell apoptosis to limit T cell immunity. Immunity 52(1), 136-150.e6 (2020).
Domínguez-Andrées, J. et al. Trained immunity: Adaptation within innate immune mechanisms. Physiol. Rev. 103, 313–346 (2023).
Chen, B., Julg, B., Mohandas, S. & Bradfute, S. B. Viral persistence, reactivation, and mechanisms of long COVID. Elife https://doi.org/10.7554/eLife.86015 (2023).
Hocini, H. et al. Neutrophil activation and immune thrombosis profiles persist in convalescent COVID-19. J. Clin. Immunol. 43(5), 882–893 (2023).
Msemburi, W. et al. The WHO estimates of excess mortality associated with the COVID-19 pandemic. Nature 613(7942), 130–137 (2023).
Acknowledgements
We would like to acknowledge all the participants enrolled in this study. We would like to acknowledge the COVID-19 SP-Brazil Team, as listed in alphabetical order: André Kenji Honda, Cesar Sato, Cesar Remuzgo, Deibs Barbosa, Fernanda Romano Bruno, João Paulo Silva Nunes, Luis Antonio Rodrigues Carneiro, Marco Antonio Stephano, Marcos Camargo Knirsch, Marcio Massao Yamamoto, Maria Lucia Carnevale Marin, Mirian Pinheiro Bruni, Rafael Ribeiro Almeida, Raquel Elaine de Alencar, Ricardo José Giordano, Roberta Liberato Pagni, Samar Freschi de Barros, Sandra Maria Monteiro, Simone Regina dos Santos. We are deeply grateful to Jhosiene Yukari Magawa for her tireless dedication and invaluable contributions to this work. Her passion and kindness will forever remain in our hearts.
Funding
This project was supported by FAPESP projects N. 2020/05256-7, 2017/24769-2, 2016/20045-7 and 2020/06409-1 and FINEP grant 408518/2022-7. This study was partially supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES)—Finance Code 001 (JO, JM, and AU).
Author information
Authors and Affiliations
Contributions
K.S. and V.C.: Data curation, Investigation, Visualization, Formal Analysis, and Writing—original draft. J.M., J.K., K.S., E.C.-N., S.B., D.R., and E.D.: Data curation and Investigation. R.O., S.B., D.R. and J.M.: Writing—review & editing. K.S., V.C. and J.M.: Conceptualization, Formal Analysis, Investigation, Visualization, Methodology, Software, Writing—review & editing. M.F., C.R., and H.N.: Formal Analysis, Investigation, Methodology, Software, Writing—review & editing. R.M., J.M., I.D., A.U., G.S., and R.S.: Visualization, Methodology and Writing—review & editing. K.S. and V.C.: Conceptualization, Funding acquisition, Project administration, Writing—original draft, Writing—review & editing. L.J., M.F., J.O., J.M., A.K., A.L., G.S., F.A., G.M.: Investigation, Methodology, Statistical analysis and Writing—review & editing. D.O., E.D., P.B.: Data curation, Investigation, Writing—review & editing. D.S.R., E.C.-N. and J.K.: Conceptualization, Funding acquisition. All authors contributed to the article and approved the submitted version.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Verônica Coelho and Keity Souza Santos: Shared last authorship.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Yukari Magawa, J., Jacintho, L.C., Alves Ferreira, M. et al. Protective role of anti-SARS-CoV-2 antibody responses against vital organ related long COVID symptoms. Sci Rep 15, 23705 (2025). https://doi.org/10.1038/s41598-025-04152-8
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41598-025-04152-8
