Recurrence of COVID-19 associated with reduced T-cell responses in a monozygotic twin pair

Recurrence of COVID-19 in recovered patients has been increasingly reported. However, the immune mechanisms behind the recurrence have not been thoroughly investigated. The presence of neutralizing antibodies (nAbs) in recurrence/reinfection cases suggests that other types of immune response are involved in protection against recurrence. Here, we investigated the innate type I/III interferon (IFN) response, binding and nAb assays and T-cell responses to severe acute respiratory distress syndrome coronavirus 2 (SARS-CoV-2) with IFN gamma (IFNγ) enzyme-linked spot assay (ELISPOT) in three pairs of young adult monozygotic (MZ) twins with previous confirmed COVID-19, one of them presenting a severe recurrence four months after the initial infection. Twin studies have been of paramount importance to comprehend the immunogenetics of infectious diseases. Each MZ twin pair was previously exposed to SARS-CoV-2, as seen by clinical reports. The six individuals presented similar overall recovered immune responses except for the recurrence case, who presented a drastically reduced number of recognized SARS-CoV-2 T-cell epitopes on ELISPOT as compared to her twin sister and the other twin pairs. Our results suggest that the lack of a broad T-cell response to initial infection may have led to recurrence, emphasizing that an effective SARS-CoV-2-specific T-cell immune response is key for complete viral control and avoidance of clinical recurrence of COVID-19.

Recurrence of COVID-19 in recovered patients has been increasingly reported. However, the immune mechanisms behind the recurrence have not been thoroughly investigated. The presence of neutralizing antibodies (nAbs) in recurrence/reinfection cases suggests that other types of immune response are involved in protection against recurrence. Here, we investigated the innate type I/III interferon (IFN) response, binding and nAb assays and T-cell responses to severe acute respiratory distress syndrome coronavirus 2 (SARS-CoV-2) with IFN gamma (IFNγ) enzyme-linked spot assay (ELISPOT) in three pairs of young adult monozygotic (MZ) twins with previous confirmed COVID- 19, one of them presenting a severe recurrence four months after the initial infection. Twin studies have been of paramount importance to comprehend the immunogenetics of infectious diseases. Each MZ twin pair was previously exposed to SARS-CoV-2, as seen by clinical reports. The six individuals presented similar overall recovered immune responses except for the recurrence case, who presented a drastically reduced number of recognized SARS-CoV-2 T-cell epitopes on ELISPOT as compared to her twin sister and the other twin pairs. Our results suggest that the lack of a broad T-cell response to initial infection may have led to recurrence, emphasizing that an effective SARS-CoV-2-specific T-cell immune response is key for complete viral control and avoidance of clinical recurrence of COVID-19. Table 1. Demographic, clinical data and human leukocyte antigen (HLA) information of the participants. Each monozygotic twin pair shared the same bedroom and was previously exposed to SARS-CoV-2 at home, according to clinical reports.

Background
Clinical recurrence of PCR-confirmed COVID-19 in adults with previous infection has been increasingly reported [1]. This has been attributed to a viral relapse in a host that failed to completely eradicate the virus or to reinfection with a different viral genome [2,3]. The involvement of innate/Type I/III IFN response as well as the specific antibody and T-cell responses is well established in the protection against severe acute respiratory distress syndrome coronavirus 2 (SARS-CoV-2) [4]. However, the immune mechanisms underlying reinfection/virus relapse are mostly unexplored. In this context, investigation of anti-SARS-CoV-2 immunity in clinical recurrence/reinfection cases has only been directed to the humoral response [5].
Since COVID-19 reinfection can occur both even in the presence of significant neutralizing antibody (nAb) titres [6], additional immune responses apart from humoral response may be involved in controlling reinfection and recurrence. The investigation of COVID-19 reinfection/recurrence can thus provide key information regarding immune protection mechanisms and guide vaccine development [6]. Studies with monozygotic (MZ) twins regarding viral infections have been a valuable source since they allow deep analysis of the environmental and host influences from different infectious agents [7]. In this article, we present a comprehensive assessment of innate and adaptive immunity in three pairs of recovered young adult MZ twins with confirmed COVID-19 and one recurrence case requiring admission to an intensive care unit.

Participants recruitment
Three pairs of COVID-19 recovered young adult MZ twins from São Paulo (the most populous city in Brazil) who were living together were recruited at the Human Genome and Stem Cell Research Center (HUG-CELL): a pair of MZ twin sisters who acquired mild COVID-19 in early 2020 where one, a healthcare worker, displayed severe clinical recurrence of COVID-19 four months after initial infection (ID 01 and ID 02); a second pair of MZ twin brothers with concordant asymptomatic infection (ID 03 and ID 04); and a third pair of MZ twin brothers with discordant symptomatic infection (ID 05 and ID 06). Baseline characteristics of the six twins are shown in table 1.

Sample collection
Blood samples were collected for global immune profiling at two instances post infection. On the first blood draw, we assessed SARS-CoV-2 IgG, IgA and IgM against SARS-CoV-2 spike (S), receptor-binding domain (RBD) and nucleocapsid protein (NP), nAbs and antibodies against the RBD region of human endemic coronaviruses, at least four weeks after initial COVID-19 diagnosis and then in January 2021. Twins were followed up for 10 months after the first blood draw. Samples were taken in vacutainer tubes with sodium heparin (BD Biosciences, USA, catalogue no. 367874) to obtain peripheral blood mononuclear cells (PBMCs); tubes without additives to obtain serum (BD Biosciences, USA, catalogue no. 366703) and tubes with ethylene diamine tripotassium (BD Biosciences, USA, catalogue no. 360057) to obtain plasma and for DNA extraction. Plasma and serum were obtained by centrifugation for 10 min at 2000g at room temperature within 30 min after blood draw. After this, the supernatant was transferred in aliquots of 1.5 ml into cryo vials (Corning, USA, catalogue no. 430487), and samples were transferred to a -80°C freezer until the moment of use. PBMCs were obtained by centrifugation of PBS-

Humoral immune response
SARS-CoV-2 IgG and IgM were initially detected by the clinical laboratory using the chemiluminescence immunoassay MAGLUMI 2019-nCoV IgM/IgG assay (Shenzhen New Industries Biomedical Engineering Co., Ltd, China, catalogue Table 2. Primers used for reverse transcription polymerase chain reaction (RT-PCR) to assess the innate immune response.

IFNγ ELISPOT assay
SARS-CoV2 specific T-cell responses were assessed using human ex vivo IFNγ enzyme-linked spot assay (ELISPOT) against a set of 20 CD4+ and 26 CD8+ T-cell epitopes from 13 distinct SARS-CoV-2 proteins with high HLA allelic population coverage in isolated PBMCs of each volunteer, three to four months after the COVID-19 episode or hospital discharge in the recurrence case. We identified and synthesized the CD4+ T-cell epitopes by scanning the whole proteome in SARS-CoV-2 reference genome (RefSeq: NC_045512.2) using the promiscuous HLA-DR binding peptide approach [10]. The chosen CD8+ T-cell epitopes were known to bind stably (www.immunitrack.com) or to be directly recognized [11] in the context of the 10 most frequent HLA class I alleles. The world population coverage of HLAs predicted to bind to the 20 CD4+ T-cell epitopes and 26 CD8+ T-cell epitopes was 99.6% and 94%, respectively, according to the IEDB epitope database [12]. Peptide sequences are listed in

Genomic assays
Whole-exome sequencing was performed on peripheral blood DNA in Illumina NovaSeq platform at HUG-CELL facilities. Sequencing data were analysed following bwa-mem, and GATK Best Practices workflow, quality control and annotation were performed as previously described [13]. HLA genes were realigned and called using HLA-mapper, which reduces mapping and calling errors [14].

Results and discussion
Here, we investigated the immunological profiles of three pairs of recovered young adult MZ twins, living together  TLACFVLAAV  LALLLLDRL  LLLDRLNQL  SIIAYTMSL  ATSRTLSYY  CTDDNALAY  LTNDNTSRY  YLQPRTFLL  FLNRFTTTL  TISLAGSYK  VVNARLRAK  GVYFASTEK  ASMPTTIAK  KSAGFPFNK  TYACWHHSI  VYMPASWVM  YYKKDNSYF  GEAANFCAL  AELAKNVSL  AESHVDTDL  GFDYVYNPF  KRVDWTIEY  HANEYRLYL  MIAQYTSAL  FAPSASAFF   SFU per 10  The finding that all six tested volunteers including the COVID-19 recurrence case presented an early and strong type I/III IFN response indicates that recurrence was not associated with failure in the innate IFN response.
To investigate the humoral immune response, we performed serological assays for SARS-CoV-2 IgA, IgG and IgM through ELISA for S, RBD and NP proteins ( figure 2a-c). We observed that all MZ twins displayed IgG against the spike protein, confirming that all of them had been infected by SARS-CoV-2. However, detection of IgG-RBD and IgG-NP was variable among the MZ twins. As expected, the antibody titres against SARS-CoV-2 declined with time when comparing the two blood collections. On the other hand, IgG antibodies against the RBD of the four main circulating endemic coronaviruses, NL63, 229E, HKU1 and OC43, were similar for all six cases, displaying antibodies for three of four tested coronaviruses. The antibody profiles to human endemic coronaviruses were virtually identical in the MZ twin siblings (figure 2d).
In addition, nAbs titres were around 1 : 500 for the recurrent sister (ID 01), while the sister without recurrence (ID 02) presented titres similar to the negative control (figure 2e). For pair 2, nAb levels were similar to those of ID 01. On the other hand, the symptomatic brother from pair 3 (ID 05) displayed nAb titres around 1 : 500, while nAb titres similar to the negative control nAb were observed in the asymptomatic brother (ID 06). nAbs were detected in several cases of reinfection/recurrence [6], suggesting that responses beyond nAbs are important to control reinfection.
Finally, we assessed the T-cell responses to SARS-CoV-2 T-cell epitopes, by performing IFNγ ELISPOT assays (figure 3) on PBMC samples. We found that all recovered young adult MZ twins recognized greater than 70% of CD4+ and CD8+ epitopes except the patient with recurring infection (ID 01) who recognized only 7 of 46 CD4+ and CD8+ T-cell epitopes (15%), while her sibling recognized 36 CD4+ and CD8+ T-cell epitopes (78%; p < 0.0001, Fisher Exact Test). Thus, patient ID 01, who showed a COVID-19 recurrence, displayed a drastically reduced breadth (number of recognized epitopes) of both CD4+ and CD8+ SARS-CoV-2 T-cell epitopes as compared with her non-recurrent sibling and the others MZ twin pairs. Interestingly, we found that the T-cell response is the only immune parameter that was substantially lower in the COVID-19 recurrence case ID 01. To our knowledge, this is the first report of T-cell immune responses in the context of COVID-19 recurrence and reinfection. Of note, there were also fluctuations in some other immune parameters within each MZ twin pair.
Our finding that the CD4+ and CD8+ T-cell response of the COVID-19 recurrence case had a drastically reduced breadth four months after hospital discharge is indicative of a low SARS-CoV-2-specific T-cell response. Given the importance of T-cell responses associated with COVID-19 infection [15], a dampened CD4+ T-cell response can have important consequences for many aspects of anti-SARS-COV-2 immunity. Asymptomatic and mild cases of COVID-19 are correlated with specific CD4+ and CD8+ T-cell responses, but not with IgG or nAb, suggesting that T cells are the primary effectors controlling a primary SARS-CoV-2 infection [4,16,17]. The dominant cytokine produced by virus specific CD4+ T cells is IFNγ with a Th1 profile, associated with antiviral activity. CD4+ T cells protect mice from lethal SARS-CoV infection [18], and Th1 CD4+ T cells are important to provide help for the cytotoxic CD8+ T responses crucial for clearance of viral infections. CD4+ T follicular helper cells contribute for B cell responses, and IL-22-producing T cells observed in COVID-19 are keys for maintenance of mucosal repair, particularly gut and lung epithelial cells [4].
It is unlikely that the reduced T-cell responses observed in ID 01 are due to the deficient HLA presentation to T cells, since her MZ twin ID 02 carrying the same HLA alleles displayed a broad recognition profile. Also, it is not likely that the contrasting T-cell responses observed between the two siblings is a result of previous exposure to cross-reacting coronaviruses [19] since their IgG profile against human endemic coronaviruses' RBD was nearly identical. Our findings are also illustrative that adaptive immune responses and clinical presentations of COVID-19 can be drastically different within a MZ twin pair. The observed diversity is consistent with the fact that T-and B-cell repertoire developments are somatic DNA rearrangement events likely to differ even among MZ twins.
In short, our results suggest that the failure in inducing a broad T-cell response might have enhanced susceptibility to COVID-19 recurrence in patient ID 01. Our data may support a prime role for T cells in protection against reinfection. Given the increased concern that SARS-CoV-2 variants escaping antibody neutralization could give rise to a massive raise in reinfection [20,21], our case stresses the importance of T-cell immune responses in protection against reinfection. This is in line with the reported lack of deleterious effect of virus variants in the cellular immune response [22]. Further investigation in a larger cohort can shed light on whether T-cell dysfunction is a common mechanism for recurrence of COVID-19.

Ethics. The study was approved by the Committee for Ethics in
Research of the Institute of Biosciences at the University of São Paulo (CAAE 34786620.2.0000.5464), and all participants (or their parents) provided a written consent.