key: cord-0313127-hf3249xu
authors: Ninnemann, Justus; Budzinski, Lisa; Bondareva, Marina; Witkowski, Mario; Angermair, Stefan; Kreye, Jakob; Durek, Pawel; Reincke, S. Momsen; Sánchez-Sendin, Elisa; Yilmaz, Selin; Sempert, Toni; Heinz, Gitta Anne; Tizian, Caroline; Raftery, Martin; Schönrich, Günther; Matyushkina, Daria; Smirnov, Ivan V.; Govorun, Vadim M.; Schrezenmeier, Eva; Dörner, Thomas; Zocche, Silvia; Viviano, Edoardo; Sehmsdorf, Katharina Johanna; Chang, Hyun-Dong; Enghard, Philipp; Treskatsch, Sascha; Radbruch, Andreas; Diefenbach, Andreas; Prüss, Harald; Mashreghi, Mir-Farzin; Kruglov, Andrey A.
title: Induction of cross-reactive antibody responses against the RBD domain of the spike protein of SARS-CoV-2 by commensal microbiota
date: 2021-08-08
journal: bioRxiv
DOI: 10.1101/2021.08.08.455272
sha: 8c6245ae5245d419e525338fcaf42ed5474c012c
doc_id: 313127
cord_uid: hf3249xu

The commensal microflora is a source for multiple antigens that may induce cross-reactive antibodies against host proteins and pathogens. However, whether commensal bacteria can induce cross-reactive antibodies against SARS-CoV-2 remains unknown. Here we report that several commensal bacteria contribute to the generation of cross-reactive IgA antibodies against the receptor-binding domain (RBD) of the SARS-CoV-2 Spike protein. We identified SARS-CoV-2 unexposed individuals with RBD-binding IgA antibodies at their mucosal surfaces. Conversely, neutralising monoclonal anti-RBD antibodies recognised distinct commensal bacterial species. Some of these bacteria, such as Streptococcus salivarius, induced a cross-reactive anti-RBD antibodies upon supplementation in mice. Conversely, severely ill COVID-19 patients showed reduction of Streptococcus and Veillonella in their oropharynx and feces and a reduction of anti-RBD IgA at mucosal surfaces. Altogether, distinct microbial species of the human microbiota can induce secretory IgA antibodies cross-reactive for the RBD of SARS-CoV-2.

SARS-CoV-2-RBD antibodies confers protection of the host against infection of target cells [5, 6] . Systemically distributed antibodies (mainly IgG, IgM, and IgA1) curtail virus propagation after productive infection of the host, while the presence of antigenspecific antibodies secreted at the mucosal surfaces (IgA2, IgA1, and IgM) may prevent initial infection of the host [7] . The absence of IgA2 antibodies specific for SARS-CoV-2 antigens in severely diseased COVID- 19 patients has also been demonstrated [8] , suggesting that mucosal anti-viral IgA antibodies may protect the host from a severe course of COVID-19. Several studies have reported the presence of RBD-binding antibodies in unexposed healthy individuals [9, 10, 11, 12, 13] .

Induction of such antibodies by previous infections with common cold coronaviruses has been postulated, but this link has not been formally proven. The original antigens inducing cross-reactive RBD-binding secretory IgA antibodies have remained obscure.

IgA antibodies at mucosal surfaces are mainly induced by commensal microbiota [14] .

It is estimated that the human microbiota contains several millions of genes [15] , thus potentially providing a plethora of epitopes for antibodies [16] . Some of such epitopes may resemble host proteins, potentially inducing autoimmunity [17, 18, 19, 20, 21] , while others may resemble proteins from other microorganisms and mediate crossreactive immunity [17, 22] . Microbiota-induced cross-reactive immunity also provides protection against microbial infections by Citrobacter rodentium, Clostridiodes difficile, Pseudomonas aeruginosa [23] and by viruses like influenza [24] . Protection is mediated by increasing fitness of the innate immune system, e.g. via tonic type I IFN production [25, 26] , and by cross-reactive adaptive antibody responses [23] .

Interestingly, cross-reactive antibodies targeting gp41 of HIV-1 are induced by commensal microbiota [27] . Here we describe the induction of cross-reactive antibody responses targeting SARS-CoV-2 by distinct members of the oral and gut microbiota.

We initially had analysed RBD-specific IgA in the fecal supernatants of age-matched healthy individuals and severely diseased COVID-19 patients (Table S1 ). Two out of 12 age-matched healthy donors, previously unexposed to SARS-CoV-2, as confirmed by lack of anti-NP SARS-CoV-2 IgG antibodies in their sera (Fig. S1A) , did have fecal IgA antibodies reactive to RBD (Fig. 1A) , 10 out of 21 severely diseased COVID-19 patients had fecal IgA specific for Spike protein RBD of SARS-CoV-2 ( Fig. 1B and Fig.   S1B ). Considering that age is an important risk factor for the development of severe COVID-19, we next determined the prevalence of RBD-binding IgA antibodies in young unexposed individuals (Fig. 1C , D and Table S1 ). We detected RBD-binding fecal IgA in approximately 50% of young healthy donors and the magnitude of the RBD-binding IgA responses in feces negatively correlated with the age of the donors (Fig. 1E) . Given the compositional complexity of fecal supernatant, we next purified IgA antibodies and tested whether the mucosal RBD-binding IgA can inhibit binding of RBD protein to the ACE2 receptor, thereby potentially blocking the entry of SARS-CoV-2 into the host cells. To this end, we expressed human ACE2 on 293T cells, then incubated the ACE2expressing cells with biotinylated RBD in the presence of purified mucosal IgA of various healthy donors (Fig. 1F , S1C). The fraction of bound RBD was analysed by flow cytometry using fluorescent streptavidin. Purified intestinal IgA from 5 out of 14 healthy donors inhibited RBD binding to ACE2 (Fig. 1F ). Of note, complete inhibition of ACE2-RBD interaction was not achieved even at 1:1 dilution, indicating a rather low concentration of neutralising anti-RBD IgA in the feces. Also, IgA from some donors with anti-RBD antibodies did not inhibit the RBD-ACE2 interaction, indicating that healthy individuals may harbor both inhibitory and non-inhibitory IgA antibodies directed against the RBD of SARS-CoV-2 (Fig. 1F) . Interestingly, healthy donors exhibited IgA2 antibodies specific for RBD in their feces, while severely diseased COVID-19 patients lacked fecal anti-RBD IgA2, consistent with a previous report [8] ( Fig. 1G ).

IgA is induced by microbiota and does bind to microbiota [28] . Thus we next analysed whether RBD-binding IgA also recognizes commensal microbiota. To this end, we first divided our healthy cohort (HC) in two groups based on the presence or absence of RBD-binding IgA in their fecal supernatants: HC RBD-IgA + and HC RBD-IgA -, respectively, and quantified the coating of bacteria by endogenous IgA. Both donor groups exhibited similar coating of their intestinal microbiota by mucosal IgA1 and IgA2 (Fig. 1H) . To identify the bacteria binding to mucosal IgA1 and IgA2, we isolated them by fluorescence-activated cell sorting and determined their taxonomic composition by 16S rRNA sequencing. Linear discriminant (LDA) combined with effect size (LefSE) analysis revealed distinct taxonomic differences of IgA coated bacteria of RBD-IgA + versus RBD-IgAhealthy donors. The IgA coated bacterial fraction of RBD IgA + donors were enriched for Parabacteroides, Sporobacter, Bilophila, and Vagococcus, while in RBD-IgAdonors the IgA coated fraction was enriched for Pseudomonas, Dorea, Soonwooa, Lachnospira, and Bacillus genera (Fig. 1I) . These data suggest that mucosal anti-RBD IgA is associated with recognition of distinct commensal microbiota by mucosal IgA.

To directly test whether anti-RBD antibodies bind to commensal bacteria, we stained the fecal microbiota of healthy individuals with neutralising anti-RBD antibodies that had either been generated in immunized rabbits or that had been cloned from hospitalised COVID-19 patients [29] . The neutralising rabbit antibody showed binding to a significant fraction of microbiota from HC ( Fig. 2A) . Furthermore, out of 15 monoclonal neutralising antibodies derived from hospitalized COVID-19 patients (for the details see [29] ) only two (HK CV07-287, HL CV07-250) showed no microbiota binding activity (Fig. 2B , C). The remaining antibodies recognised commensal bacteria, 9 of them also independently of pre-existing fecal anti-RBD IgA (Fig. 2B, C) . Of note, two clonally related antibodies, CV07-200 and CV07-283, showed distinct binding patterns ( Fig. 2C and Fig. S2 ). Co-staining of microbiota with rabbit and human monoclonal antibodies showed that both recognize similar as well as distinct fecal bacteria communities (Fig. S3) . Thus, most neutralising human anti-RBD SARS-CoV-2 antibodies tested in our study bind to distinct commensal bacteria.

To identify the bacteria recognized by neutraliing anti-RBD antibodies, we stained, sorted and sequenced antibody-bound fecal bacteria from 3 healthy donors using 4 different anti-RBD antibodies (Fig. 2D , E). Several genera with an abundance of more than 1% were bound by the respective antibodies, and the identified bacteria differed among various donors (Fig. 2E) , highlighting the inter-individual diversity in the bacterial composition. The binding of the anti-RBD IgG antibodies to microbiota was specific, since neither the secondary anti-IgG antibodies used to identify their binding ( Fig.2) , nor human IgG antibody with different specificity showed similar binding patterns towards microbiota (Fig. S3B) . The monoclonal human anti-RBD antibodies in particular showed reactivity towards Bacteroides. Some of them also recognised Clostridia species, Streptococci, Escherichia and Bifidobacteria (Fig. 2E ). Of the genera bound by IgA of HC RBD-IgA + donors, Parabacteroides and Bilophila also bound to the human anti-RBD IgG antibodies (Fig. 1I, 2E ).

By fluorescence-activated cell sorting we isolated bacteria recognised by the human anti-RBD IgG antibodies from 8 healthy donors, and cultured them using selective bacterial media and anaerobic culture conditions. Individual bacterial colonies were further expanded and their identity determined by 16S rRNA Sanger sequencing (Fig.   2F ). Two Bacilli species, three Streptococcus species, two Bifidobacterium species, two Enterococcus species, Veillonella parvula and Acidaminococcus intestinalis were identified as bacteria bound by anti-RBD antibodies (Fig. 2F) . Restaining of purified cultures confirmed their recognition by anti-RBD antibodies (Fig. S4A, B) . One of the isolated bacterial species was Streptococcus salivarius, bacteria living in the oropharynx, with probiotic activity. Indeed, S. salivarius K12, an established probiotic strain, is recognized by rabbit anti-RBD antibodies (Fig. S4A ). Of note, some bacterial cultures showed only partial staining with anti-RBD antibodies, probably reflecting the heterogeneity of bacteria during growth or community-dependent surface variability. Having shown that anti-RBD antibodies can cross-react with bacterial proteins, we tested whether the bacteria expressing these proteins can induce a cross-reactive anti-RBD antibody response. We immunised C57Bl/6 mice intraperitoneally once with heatkilled bacteria and analysed the antibody responses against RBD 14 days later. Mice immunized with heat-killed S. salivarius, but not those immunized with heat-killed B.

pseudocatenulatum, developed anti-RBD IgG antibodies in their sera (Fig. 3A) .

Veillonella parvulla also induced anti-RBD IgG upon immunization (Fig. 3B ). Sera from mice immunised with S. salivarius and V. parvulla could inhibit the binding of RBD to ACE2, as expressed in 293 T cells (Fig. 3C ). Closer to the physiological situation, the natural route of confrontation with bacteria of oropharyngeal microbiota, oral feeding with S. salivarius K12 and B. pseudocatenulatum, induced fecal IgA specific for RBD in C57Bl/6 mice (Fig. 3D) . Moreover, fecal supernatants from animals supplemented with bacteria inhibited binding of RBD to ACE2 (Fig. 3E ). To gain further insight on the specificity of antibodies induced by oral supplementation with bacteria, we next performed epitope mapping of the IgA induced in the gut against 564 peptides derived from the Spike protein of SARS-CoV-2. We observed that both B. pseudocatenulatum and S. salivarius induced antibodies bound to the peptide sequence GFNCYFPLQSYGFQPTNGV (Fig. 3F, Fig. S6 ), that corresponds to the receptor binding motif (RBM) of RBD, in line with ACE2 inhibition data. Also, the peptide recognition pattern of rabbit anti-RBD and HL CV07-200 antibodies overlapped: both antibodies had in their epitopes a similar sequences within the RBM motif (Fig. S6 ).

These data show that oral supplementation with S. salivarius K12 and B.

pseudocatenulatum can induce antibodies cross-reactive against the RBM motif of the spike protein of SARS-CoV-2.

In light of the ability of distinct oropharyngeal microbiota species to generate mucosal IgA cross-reactive to SARS-CoV-2, we compared the oral microbiota composition of healthy donors to that of COVID-19 patients, as well as of patients with flu-like symptoms, but negative for SARS-CoV-2 ( Fig. 4 and Table S2) (Fig. 4C) . Conversely, Veillonella and Streptococcus genera, but not Bifidobacteria genera, which we had identified as potential inducers of cross-reactive antibodies, were significantly reduced in patients with severe COVID-19 ( Fig. 4C, D) . Instead, these patients showed an increased abundance of the genera Enterococcus, Staphylococcus and Escherichia/Shigella in their oropharynx (Fig. 4C , D). This is not due to the treatment of severe COVID-19 patients with antibiotics (Abx), since our cohort includes both Abx naive and Abx-treated patients, and both groups showed the same prevalence of microbiota composition.

The differences in the oral microbiota composition also extend to the intestinal microbiota in severely affected COVID-19 patients ( Fig. S7A and Table S3 ). Also include presence of autoantibodies against type I IFN, genetic predisposition [30, 31] and preexisting disease conditions, such as diabetes, obesity, and ageing [32] .

Furthermore, while pre-existing memory T cells specific for SARS-CoV-2 may be protective, pre-existing low avidity memory T cells recognising SARS-CoV-2 antigens in the elderly may be a potential risk factor during COVID-19 [8, 33] . Here we report that healthy, unexposed individuals can have preexisting secretory IgA antibodies at mucosal surfaces, antibodies which also bind to the RBD of the S protein of SARS-CoV-2, and thus have the potential to neutralise the virus and prevent or ameliorate infection and COVID-19. This pre-existing mucosal immunity fades with age.

Microbiota may contribute to the protection of the host from infection via modulating the ACE2 receptor expression [34] , induction of tonic type I IFN responses [35] , and via tuning systemic and mucosal TGF-β1 levels, with TGF-β1 being the inductor of antibody class switch recombination to IgA [8, 36] . Here we have identified bacteria of the oropharyngeal microbiota that express protein antigens on their cell surface, which mimic epitopes of the RBD of the SARS-CoV-2 Spike protein, to an extent that they not only are recognised by anti-RBD antibodies of different origin but can themselves also trigger an antibody response capable of neutralising RBD in mice in vivo, both by intraperitoneal immunization and by oral feeding. Presence of these bacteria is associated with mucosal IgA antibodies recognizing RBD, and are capable of inhibiting its binding to ACE2, in healthy donors not previously exposed to SARS-CoV-2. It remains a challenge for future research, to determine how the bacteria induce such antibodies. Similar observations have been reported for the HIV-1 virus [27, 37, 38] .

In particular, a link between gp-41 and gp-120 reactive antibodies and their cross-reactivity against microbiota has been demonstrated [27, 37] . It is evident that bacteria of the microbiota provide a rich target proteome for the mucosal immune system, and that this can result in the generation of a cross-reactive, pre-existing mucosal immunity against distinct viruses and may explain heterogeneity of human subjects in susceptibility towards viral infection.

Apart from host-intrinsic factors, the initial virus load may affect disease outcome and severity [39, 40] , and there is an increasing evidence of microbiota changes during severe COVID-19 [41, 42] , suggesting that the microbiota composition may be a risk factor for the development of severe disease as well [41, 42, 43] . The data are conflicting in terms of the genera associated with disease severity, which is probably due to the heterogeneity of patient cohorts and differences in treatment. A common denominator is that acute COVID-19 is associated with the prevalence of opportunistic bacteria and depletion of immunomodulatory bacteria [42] . 

The recruitment of study subjects was conducted in accordance with the Ethics Committee of the Charité (EA 1/144/13 with EA 1/075/19, EA 2/066/20) and was in compliance with the Declaration of Helsinki. The linear discriminant analysis were performed using LEfSe, based on copy number adjusted counts normalized to 1M reads [46] . Raw sequence data were deposited at the NCBI Sequence Read Archive (SRA) under the accession number PRJNA738291.

The frozen microbiota stocks were topped up with 1 mL of autoclaved and sterile- The next day, DNA was isolated and the remaining bacteria were frozen in 40% glycerol LB medium in liquid nitrogen or -80 °C.

For the identification of the bacterial species bound to the neutralizing anti-RBD antibodies, the DNA from 200 µl of the grown bacteria was isolated with ethanol precipitation. The isolated DNA was subsequently amplified by the 16S rDNA specific primers LPW57 and LPW58 [47] . In brief, bacterial DNA was amplified with Taq 

Epitope mapping was performed using peptide microarray multiwell replitope SARS- 

The protein bands after 1D-PAGE were excised and washed twice with 100 mL of 0. 16 32 

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Cell entry mechanisms of SARS-CoV-2

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein

Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-Throughput Single-Cell Sequencing of Convalescent Patients' B Cells

Human neutralizing antibodies elicited by SARS-CoV-2 infection

Role of IgA versus IgG in the control of influenza viral infection in the murine respiratory tract

SARS-CoV-2 in severe COVID-19 induces a TGF-beta-dominated chronic immune response that does not target itself

Preexisting and de novo humoral immunity to SARS-CoV-2 in humans

HCoV-NL63 and SARS-CoV-2 Share Recognized Epitopes by the Humoral Response in Sera of People Collected Pre-and during CoV-2 Pandemic

A majority of uninfected adults show pre-existing antibody reactivity against SARS-CoV-2

Seasonal human coronavirus antibodies are boosted upon SARS-CoV-2 infection but not associated with protection

Maturation and persistence of the anti-SARS-CoV-2 memory B cell response

IgA Function in Relation to the Intestinal Microbiota

The Landscape of Genetic Content in the Gut and Oral Human Microbiome

More than 9,000,000 unique genes in human gut bacterial community: estimating gene numbers inside a human body

Gut microbiota as a source of a surrogate antigen that triggers autoimmunity in an immune privileged site

Microbiota-derived peptide mimics drive lethal inflammatory cardiomyopathy

Commensal orthologs of the human autoantigen Ro60 as triggers of autoimmunity in lupus

Cross-immunoreactivity between bacterial aquaporin-Z and human aquaporin-4: potential relevance to neuromyelitis optica

GDP-l-fucose synthase is a CD4(+) T cell-specific autoantigen in DRB3*02:02 patients with multiple sclerosis

Natural polyreactive IgA antibodies coat the intestinal microbiota

Antibiotic treatment-induced secondary IgA deficiency enhances susceptibility to Pseudomonas aeruginosa pneumonia

Microbiota regulates immune defense against respiratory tract influenza A virus infection

Microbiota-Induced Type I Interferons Instruct a Poised Basal State of Dendritic Cells

Commensal Microbiota Modulation of Natural Resistance to Virus Infection

HIV-1 envelope gp41 antibodies can originate from terminal ileum B cells that share cross-reactivity with commensal bacteria

IgA and the intestinal microbiota: the importance of being specific

A Therapeutic Non-self-reactive SARS-CoV-2 Antibody Protects from Lung Pathology in a COVID-19 Hamster Model

Autoantibodies against type I IFNs in patients with life-threatening COVID-19

The major genetic risk factor for severe COVID-19 is inherited from Neanderthals

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Low-Avidity CD4(+) T Cell Responses to SARS-CoV-2 in Unexposed Individuals and Humans with Severe COVID-19

Mining the Human Gut Microbiota for Immunomodulatory Organisms

Microbiota-Driven Tonic Interferon Signals in Lung Stromal Cells Protect from Influenza Virus Infection

Specific microbiota enhances intestinal IgA levels by inducing TGF-beta in T follicular helper cells of Peyer's patches in mice

The function and affinity maturation of HIV-1 gp120-specific monoclonal antibodies derived from colostral B cells

HIV-1 VACCINES. Diversion of HIV-1 vaccine-induced immunity by gp41-microbiota cross-reactive antibodies

SARS-CoV-2 viral load is associated with increased disease severity and mortality

Viral dynamics in mild and severe cases of COVID-19

Gut microbiota composition reflects disease severity and dysfunctional immune responses in patients with COVID-19

Alterations in Gut Microbiota of Patients With COVID-19 During Time of Hospitalization

Alterations of the Gut Microbiota in Patients with COVID-19 or H1N1 Influenza

Impaired humoral immunity to SARS-CoV-2 BNT162b2 vaccine in kidney transplant recipients and dialysis patients

Community Ecology Package

Metagenomic biomarker discovery and explanation

Identification by 16S ribosomal RNA gene sequencing of an Enterobacteriaceae species from a bone marrow transplant recipient

Competing interests: J.N., L. B., and A.K. have a pending patent application with regard to utilization of commensal bacteria for induction of antiviral immune responses.