Oligosymptomatic long-term carriers of SARS-CoV-2 display impaired innate resistance but increased high-affinity anti-spike antibodies

Summary The vast spectrum of clinical features of COVID-19 keeps challenging scientists and clinicians. Low resistance to infection might result in long-term viral persistence, but the underlying mechanisms remain unclear. Here, we studied the immune response of immunocompetent COVID-19 patients with prolonged SARS-CoV-2 infection by immunophenotyping, cytokine and serological analysis. Despite viral loads and symptoms comparable to regular mildly symptomatic patients, long-term carriers displayed weaker systemic IFN-I responses and fewer circulating pDCs and NK cells at disease onset. Type 1 cytokines remained low, while type-3 cytokines were in turn enhanced. Of interest, we observed no defects in antigen-specific cytotoxic T cell responses, and circulating antibodies displayed higher affinity against different variants of SARS-CoV-2 Spike protein in these patients. The identification of distinct immune responses in long-term carriers adds up to our understanding of essential host protective mechanisms to ensure tissue damage control despite prolonged viral infection.


INTRODUCTION
Patients infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have clinical presentations ranging from asymptomatic-mildly symptomatic (70-90%) to severe and critical (10-30%). [1][2][3][4] These different clinical outcomes, including the risk of COVID-19-related death, have been associated with age, gender, and underlying comorbidities, such as obesity and diabetes. 2,5,6 Regardless of pathogen loads, critically ill COVID-19 patients present local and systemic inflammation leading to severe tissue dysfunction, characterized by an increase in inflammatory cytokines, monocytes, and neutrophils, and a marked decrease in lymphocytes compared to patients with mild disease. [7][8][9][10][11][12][13][14] Moreover, a significant fraction of patients with life-threatening COVID-19 present defects in type I IFNs (IFN-I) because of inborn mutations and auto-antibodies, pointing to a critical role of IFN-I in the immune response against SARS-CoV-2. 10-12 These distinct immune and inflammatory signatures are observed early after COVID-19 diagnosis, correlate with divergent disease trajectories and might have prognostic value. 9, 13,14 Alternatively, immunosuppressed individuals, who exemplify the paradigm of low host resistance, display a variety of clinical presentations, from asymptomatic to severe. [15][16][17][18] Low resistance might impact SARS-CoV-2 clearance in multiple ways, leading to high viral titers in the upper-respiratory tract (URT), dissemination to other tissues, especially the lungs, or long-term virus persistence. Although viral persistence has been more frequently described in immunosuppressed patients, persistent URT infection and long-term virus shedding have been documented in immunocompetent patients with asymptomatic or mild COVID-19 as well. [19][20][21][22][23] Most long-term carriers remained SARS-CoV-2 positive by qRT-PCR despite seroconversion, reinforcing the risk of continuous SARS-CoV-2 transmission. 20,24,25 Defects in antigen-specific cytotoxic T cell responses were found in such patients, 26 but the immune dynamics along the course of infection remain unclear. Thus, the aim of this study is to gain insights into the immune mechanisms associated with prolonged SARS-CoV-2 infection in oligosymptomatic, immunocompetent subjects. Overall,

SARS-CoV-2 persistency does not depend on viral load or mucosal cytokines
As SARS-CoV-2 enters the organism via the URT, early local immunity at the nasopharyngeal mucosa may be important for fast and efficient viral clearance. Therefore, we first analyzed early viral loads and immune parameters in the nasopharyngeal mucosa of patients that eventually developed prolonged SARS-CoV-2 infection. Viral titers, determined by RT-qPCR, in nasopharyngeal swabs of long-term carriers were similar to those of non-persistent COVID-19 patients during the first days of infection (%7 DSSO), but started differing after the first week of infection. Most long-term carriers presented detectable viral RNA in the URT for up to 4 weeks, albeit with slightly higher Ct values ( Figure 1D). Strikingly, a group of patients still tested positive by RT-qPCR for longer than 8 weeks, with the latest positivity result being at 134 DSSO in one patient ( Figure 1E).
Next, we used a multiplex assay to compare a set of 47 immune mediators, including IFNa, IL29, IFNg, IL-1b, IL-4, IL-13, IL-17A and IL-10, in nasopharyngeal swab samples collected at % 10 DSSO from a group of non-persistent patients (NP, n = 9) and long-term carriers (P, n = 13). The results did not show any differences in early cytokines between both groups ( Figures 1F and S1). These results suggest that early alterations of these mucosal cytokines at the entry site of SARS-CoV-2 do not drive viral persistence.
Type 1 responses shift to type 3 immunity in long-term carriers The immune system orchestrates distinct resistance mechanisms depending on the nature of the infectious agent, the site of infection and the time after infection onset. Type 1 immunity, mediated by IFNg, NK cells, T helper 1 (Th1) lymphocytes and cytotoxic T cells, is primarily induced in response to intracellular pathogens, such as viruses. Dysregulated systemic inflammatory and antiviral responses have been pointed out as potential drivers of distinct clinical progression of COVID-19. 9,10,13 Therefore, to gain insights into specific immune mechanisms leading to prolonged SARS-CoV-2 infection, we compared innate immune cells by flow cytometry longitudinally in non-persistent COVID-19 patients and long-term carriers, starting early after disease onset (% 10 DSSO). Frequencies of circulating monocyte populations, including classical (lin À CD14 + CD16 À ), intermediate (lin À CD14 + CD16 + ) and non-classical (lin À CD14 À CD16 + ) monocytes, were similar in both groups across the infection, and so were the frequencies of dendritic cells (DCs) (Figures 2A, S2,  iScience Article a significant reduction in the percentage of blood-circulating pDCs (lin À CD14 À CD304 + ) in long-term carriers. The percentage of circulating NK cells (lin À CD56 + ) was also reduced in patients with long-term infection compared to non-persistent ones and remained low over the infection period and on the convalescent phase. Altogether, these results suggest a deficient activation of type 1 immunity in long-term SARS-CoV-2 infection. iScience Article Next, we titrated plasma levels of tissue growth factors like PDGF-BB, basic FGF and VEGF, which are involved in tissue damage and repair and have been previously connected to disease tolerance. 27 We did not observe any increase in any of these factors in long-term carriers, indicating that systemic tissue repair mechanisms were not preferentially induced in these patients ( Figures 2B and S4A). While VEGF concentrations acutely dropped at the second week after symptom onset, some patients showed a transient increase in PDGF-BB and basic FGF around this time point and stabilized later on. The analysis of inflammatory mediators previously found to correlate with COVID-19 symptoms, such as TNFa, IL-1b or IL-1 RA, [7][8][9] showed that long-term carriers display inflammatory responses comparable to non-persistent patients at the onset of disease, which mildly increased at the third week of infection in some of the patients ( Figure 2C). These results suggest that the absence of sustained symptoms or tissue damage in prolonged SARS-CoV-2 infection is probably because of the lack of strong inflammation, and not related to an enhancement of disease tolerance through increased tissue repair.
Furthermore, we aimed to analyze a set of cytokines and chemokines related to type 1, 2 and 3 responses early after disease onset and longitudinally. Since pDCs are the first and major IFN-I producers during viral infections 28,29 and were less represented in long-term carriers, it was not unexpected that plasma concentrations of IFNa and the IFN-induced chemokine IP-10 (CXCL10) did not increase in this group of patients ( Figure 2C). Strikingly, other type 1 cytokines, such as IL-12, IFNg and IL-7, were significantly lower in longterm carriers compared with regular COVID-19 patients at disease onset (<7 DSSO). Type 2 cytokines were similarly reduced in long-term carriers, whereas type 3 cytokines related to neutrophil production and recruitment, such as IL-17A and G-CSF, displayed a significant enhancement. Altogether, these results indicate that patients with persistent SARS-CoV-2 infection present an altered immune profile skewed toward type 3 responses.
To visualize the interplay between the analyzed cytokines and viral loads, we depicted cytokine concentrations along with Ct values (represented as Ct-1 ) for the eight individual patients with more than three longitudinal samples ( Figure S5A). Instead of a unique immune profile, we found distinct cytokine and Ct dynamics, which suggests that each long-term carrier copes with viral persistence in a different manner.

Persistent SARS-CoV-2 infection does not affect effector or memory T cell responses
Next, we studied the functionality of the T cell compartment in long-term carriers. We analyzed the cytokine production capacity of circulating CD4 + and CD8 + T cells from the PBMCs of non-infected individuals, nonpersistent patients and persistently infected COVID-19 patients at early time points of infection (%10 DSSO). To that aim, we performed in vitro polyclonal stimulation of PBMCs with anti-CD3/anti-CD28 magnetic beads, and analyzed intracellular cytokine production by flow cytometry. Although the production of TNFa by CD4 + T cells and Granzyme B by CD8 + T cells was enhanced in long-term SARS-CoV-2 carriers compared with the healthy control group, no significant differences in cytokine production were found between non-persistent patients and long-term carriers ( Figure 3A). In addition, we analyzed whether the frequencies of circulating memory T cell subpopulations were altered during the course of disease, but we could not find differences between both groups of patients, neither during infection nor at the convalescent phase ( Figures 3B and S3). Only naive CD4 + T cells appeared to be reduced during the first week of infection in long-term carriers. Furthermore, to elucidate whether long-term infection had or not an iScience Article effect on the development of antigen-specific T cell responses, we took PBMCs at a DSSO corresponding to the first negative swab PCR result, corresponding to the convalescent phase, and stimulated them with a pool of peptides spanning the sequence of the S protein of the circulating variant of SARS-CoV-2 in Rio de Janeiro at the time of collection. CD4 + and CD8 + T cells from non-persistent and persistently-infected patients proliferated similarly in response to antigen-specific stimulation, according to the amount of Ki67, detected by flow cytometry ( Figure 3C). Production of effector cytokines, such as TNFa, IFNg, and granzyme B, was also comparable between both groups. However, the frequency of CD25 + IL-10 producing cells among the total pool of CD4 + cells, representative of Treg populations, was significantly lower in the PBMCs from long-term carriers, suggesting weaker T cell-mediated regulatory responses in this group of patients. Altogether, these results indicate that persistent infection does not interfere with the development of antigen-specific T cell-mediated immune memory.

Patients developing persistent infection display an early distinct immune signature
Unsupervised cluster analysis of plasma cytokines at % 10 DSSO revealed at least two groups of immune mediators that were differentially regulated in non-persistent patients compared to those with long-term infection ( Figure 4A). Cluster 2, composed of IL-6, IL-8, IL-17A, G-CSF, IL-15 and MIP-1a, was overrepresented in most long-term carriers, whereas the cytokines from cluster 4, containing IL-9, IL-10, IFNg and IL-12, were reduced. In addition, we combined plasma cytokine data with immunophenotyping data from the same time points (Figures 4B, S6A, and S6B). The correlation matrix of soluble proteins and immune cell subtypes revealed a positive correlation of effector CD8 + T cells with Th1 and Th2 cytokines which, in turn, inversely correlated with the activation of alternative monocytes and pDCs ( Figure 4B). Furthermore, unsupervised cluster analysis on cytokine data by t-distributed Stochastic Neighbor Embedding (t-SNE) identified four different clusters of patients ( Figure S6B). All non-infected controls gathered in one cluster, whereas long-term carriers were distributed in three different groups along with non-persistent patients, irrespective of age, gender or comorbidities. Finally, fold change importance analysis performed on NP versus P patients highlighted IL-17A, MIP1a, IL-15, IL-8 and IFNg as the top five cytokines characterizing SARS-CoV-2 persistency ( Figure 4C). Altogether, bioinformatic analysis strongly suggests that a combination of early blood markers correlates with prolonged SARS-CoV-2 infection.

SARS-CoV-2 persistency favors the generation of high affinity spike-specific antibodies
Adequate innate immune activation is essential for the development of adaptive immune responses, but our results showed that long-term carriers presented an altered innate immune profile. As B cell memory and SARS-CoV-2-specific antibody responses might also correlate with distinct disease trajectories, we additionally studied the development of antigen-specific humoral responses in long-term carriers. We started analyzing follicular helper T cells (T FH ), since they contribute to humoral immunity by delivering to B cells the necessary signals to enter the germinal center, going through class-switch recombination and affinity maturation. Although we did not find changes in the frequencies of circulating CD4 + or CD8 + T FH cells ( Figure 5A), polyclonal stimulation of PBMCs collected at % 14 DSSO revealed that this T cell population produced higher amounts of IL-21 in long-term carriers, being this cytokine essential for B cell help ( Figure 5B). Furthermore, we compared the composition of the B cell compartment and antibody development in patients with regular or delayed resolution of infection. We did not observe significantly different frequencies of circulating mature, class-switched B cells, although the proportion of circulating plasmablasts/plasma cells dropped in long-term carriers after 14 DSSO ( Figure 5C). Serological analysis did not identify significant alterations in the production of nucleocapsid (N) protein-specific IgG ( Figure S7A) or spike (S) protein-specific IgM, IgG and IgA antibodies over time (Figures 5D and  S7A). However, plasma of long-term carriers displayed higher titers of circulating high-affinity Spike-specific antibodies against all analyzed SARS-CoV-2 variants already three weeks after symptoms onset, and iScience Article at least until the patients reached the convalescent phase ( Figures 5E and S7B). These data suggest that generation of systemic N protein-specific or S protein-specific humoral responses do not seem to be sufficient for viral clearance during primary SARS-CoV-2 infection, although it might confer better protection against subsequent re-infections.

DISCUSSION
Our cohort comprises a group of oligosymptomatic patients with persistent infection. Here we show that low resistance in these patients is likely because of impaired innate antiviral immunity, as no major  iScience Article defects in adaptive immunity were found. Furthermore, we observed divergent cytokine profiles in the nasopharyngeal mucosa and in the plasma, in line with other studies recently published. 30 In contrast with systemic cytokines, we did not find changes in nasopharyngeal immune mediators. However, although relevant mucosal cytokines, such as IL-17A and IL-10, were not altered in the nasopharynx, we cannot exclude the contribution of cytokines with only transient induction or other mucosal mediators not studied here.
Long-term carriers displayed decreased frequencies of circulating pDCs and NK cells shortly after symptom onset and in later time points. The pDCs are considered essential for the control of viral replication through the rapid release of IFN-I. 28,29,31 Early IFN-I triggers different mechanisms targeting virally infected cells, namely the expression of antiviral proteins, the activation of NK cells and the initiation of Th1 responses. In our study, some of these elements were underrepresented in those COVID-19 patients who later on presented persistent infection, indicating that low antiviral innate immune responses may be a possible reason for viral persistency. iScience Article Major differences in systemic cytokines, chemokines and growth factors were found in patients with prolonged course of infection from early on. Long-term carriers displayed an immunological shift characterized by a low systemic Th1/2 signature, increasing in turn IL-17A, IL-8 and other neutrophil-recruiting chemokines. In line with our data, mouse models of chronic infection by TMEV also link Th17 cells to viral persistence. 32 The origin of type 3 cytokines in persistent patients is unclear at this point. Mucosal sites, including the gut, are a major source of IL-17A and highly dependent on the microbiota. 33 Considering that gut mucosal tissues are sites of active SARS-CoV-2 replication, 34 it is possible that the observed increase in type 3 cytokines might be influenced by the microbiome, and by viral persistence at these sites.
Impaired IFN-I responses and an enhanced type 3 signature are also traits that had been previously associated with severe COVID-19. 9-11 Conversely, inflammatory and Th1/2 responses are elevated in patients with severe symptoms but low in oligosymptomatic long-term carriers, which suggests their involvement in the pathogenesis of COVID-19. [7][8][9]13 All in all, our data suggest that early systemic immunological patterns may indicate future persistency of SARS-CoV-2 infection in immunocompetent patients. Further studies to better characterize the correlation of these immunological patterns with infectious disease outcome are warranted.
The engagement of disease tolerance mechanisms, like enhanced tissue repair, metabolic adaptations or immune regulation, may also lead to distinct disease trajectories. 35,36,37 Our data could not identify enhanced systemic tissue repair or regulatory responses during the course of prolonged SARS-CoV-2 infection. Actually, the immunoregulatory cytokine IL-10, which is commonly linked to immunological tolerance, but also a well-accepted disease marker in COVID-19, 38,39 was particularly low in long-term carriers.
In order to understand the physiological meaning of this effect and the involvement of additional disease tolerance mechanisms, further analysis should be conducted.
In our study, long-term SARS-CoV-2 carriers had higher titers of spike-specific antibodies already three weeks after symptom onset. First, this suggests that systemic high affinity antibodies against Spike/RBD may not be sufficient for viral clearance from the URT during primary infection. Type I IFNs might contribute to this effect, since they have been shown to regulate B cell responses and antibody production in the context of viral infection. 40,41 Furthermore, patients with severe COVID-19, who may remain infected for extended time periods and present dysregulated IFN-I and type 3 responses, have been shown to produce higher titers of neutralizing antibodies. 42 Together with our data, this suggests that dampened IFN-I responses and/or strong type 3 immunity in persistent SARS-CoV-2 infection might confer advantages against re-infection. On the other hand, these antibodies could also contribute to viral persistence via antibody-dependent enhancement or modulate COVID-19 immunopathology via Fc-receptors. Hence, analysis of the functionality of the antibodies elicited during long-term infection, and longer follow-up of the adaptive cellular and humoral immunity in these patients could be valuable for COVID-19 treatment and vaccine design and development.
Collectively, our study provides a thorough analysis of the immune dynamics during viral persistency in COVID-19 patients. Over the past three years, important progress has been made in controlling the pandemic, but the vaccines available, though successful in limiting disease, are not sterilizing and do not completely prevent transmission of new SARS-CoV-2 variants. 43 In COVID-19 as well as in other viral infections, oligosymptomatic and asymptomatic carriers represent the main vector of viral transmission. Therefore, long-term immunocompetent infected individuals are potential long-term spreaders and may, in addition, facilitate intra-host evolution of SARS-CoV-2 and other viruses. 15,19,20 This reinforces the need for a better understanding of the immune mechanisms of viral control in these patients. Shedding light on this issue, our study identifies a set of early plasma markers associated with prolonged infection and reveals alternative immunological strategies to deal with viral infection without overt tissue damage and pathology.

Limitations of the study
Given the out-care characteristic of our cohort and the techniques available, our investigation of viral RNA was restricted to nasopharyngeal samples. Assessment of viral titers in various tissues would allow for a better understanding of whether long-term carriers display systemic low resistance, or if this phenomenon is restricted to the nasopharyngeal mucosal site. Although we did not find increased damage and tissue repair markers in the blood or in the nasal mucosa, we cannot exclude that other SARS-CoV-2 target ll OPEN ACCESS iScience 26, 107219, July 21, 2023 iScience Article tissues, such as the lungs or the gut, have altered disease tolerance or support viral replication. To study disease tolerance and viral presence in such locations, animal models for SARS-CoV-2 would be helpful. Furthermore, our samples are limited and were collected in a specific timing and geographical context of the pandemic. Nonetheless, to our knowledge, the present study constitutes the characterization of immune parameters in the larger number of immunocompetent oligosymptomatic patients with persistent infection of SARS-CoV-2 so far. Still, sample size is critical when studying such a heterogeneous disease as COVID-19, and, therefore, validation of our data in larger, independent cohorts would be informative.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors declare no competing interests.

INCLUSION AND DIVERSITY
We support inclusive, diverse, and equitable conduct of research. We worked to ensure sex balance in the selection of non-human subjects. One or more of the authors of this paper self-identifies as a member of the LGBTQIA+ community. One or more of the authors of this paper self-identifies as a gender minority in their field of research.

EXPERIMENTAL MODEL AND STUDY PARTICIPANTS DETAILS
All patients included in the present study sought testing at the Diagnostic Screening Center for COVID-19 at the Federal University of Rio de Janeiro (CTD-UFRJ) and declared written informed consent. From April to December 2020, we enrolled two thousand seven hundred and fifty-nine patients who were tested for SARS-CoV-2 infection at the Diagnostic Screening Center for COVID-19 of the Federal University of Rio de Janeiro (CTD-UFRJ). Among them, 1,133 individuals (41.07%) tested positive for the presence of SARS-CoV-2 RNA by quantitative PCR with reverse transcription (RT-qPCR) on nasopharyngeal swab samples. Those individuals were offered weekly follow-up testing until SARS-CoV-2 RNA was no longer detected. Blood from those patients was collected in heparinized tubes for plasma and PBMC storage and further analysis. Symptoms, use of medication, comorbidities and demographic information were assessed by oral questionnaire performed by trained personnel. Based on blood sample availability, 33 patients were selected from those with persistent SARS-CoV-2 infection, defined as positive SARS-CoV-2 RT-qPCR in upper respiratory tract (URT) samples after 21 days after symptom onset. As such, 32 patients were selected from those who did not display persistent viral RNA, defined as a negative SARS-CoV-2 RT-qPCR in URT samples up to 21 days after symptom onset. The criterion was guided by the median of positivity duration in the overall cohort, which was around three weeks (Voloch et al, 2021). Twenty-five non-infected volunteers were included as controls, defined as a negative SARS-CoV-2 qRT-PCR, no history of a positive SARS-CoV-2 qRT-PCR and no seroconversion for SARS-CoV-2 epitopes. Since age and gender of the study subjects may influence the obtained results, all groups were matched by median age and male and female individuals were equally distributed (