Immunoglobin and T cell receptor repertoire changes induced by a prototype vaccine against Chagas disease in naïve rhesus macaques

A vaccine against Trypanosoma cruzi, the agent of Chagas disease, would be an excellent additional tool for disease control. A recombinant vaccine based on Tc24 and TSA1 parasite antigens was found to be safe and immunogenic in naïve macaques. Here we performed a transcriptomic analysis of PBMC responses to vaccination, to shed light on the immunogenicity of this vaccine and guide the optimization of doses and formulation. RNA-sequencing analysis indicated a clear transcriptomic response of PBMCs from macaques after three vaccine doses, with the up-regulation of several immune cell activation pathways and a broad non-polarized immune profile. Analysis of the IgG repertoire showed that it had a rapid turnover with novel IgGs produced following each vaccine dose, while the TCR repertoire presented several persisting clones that were expanded after each vaccine dose. These data suggest that three vaccine doses may be needed for optimum immunogenecity and support the further evaluation of the protective efficacy of this vaccine.


INTRODUCTION
Chagas disease is a zoonotic parasitic disease of the Americas, caused by the protozoan parasite Trypanosoma cruzi, and transmitted primarily by hematophagous triatomine bugs.Disease burden reaches at least 6 millions cases in Latin America alone 1 and a global burden of 806,170 Diseaseadjusted life years (DALYs) and an associated $627.46 million in health-care costs 2 .
The disease starts with a short acute phase lasting a few weeks with u-like signs and symptoms, followed by a chronic phase initially asymptomatic, but 30-40% of infections will progress to symptomatic cardiac or digestive disease, sometimes decades after infection 3 .Only two drugs are available for treating patients, but their e cacy decreases dramatically as disease progresses, and their side effects lead to frequent treatment interruptions and non-compliance [4][5][6] .Structural barriers for accessing diagnostic testing and treatment also considerably restrict access to appropriate care for patients [7][8][9] .
Because of the limitations of current drug treatments, a vaccine would be an excellent additional tool for disease control.Many studies have shown the feasibility of multiple vaccine antigens and platforms in mice 10,11 , and a few vaccine candidates have been tested in other animal models including dogs and non-human primates.Among these, a therapeutic DNA vaccine based on Tc24 and TSA1 parasite antigens has been shown to be effective to prevent cardiac alterations caused by T. cruzi infection in experimentally-infected macaques 12 .A recombinant protein version of this vaccine is undergoing further development [13][14][15] and was found to be safe and immunogenic in naïve macaques 16 .Indeed, high titers of antibodies and antigen-speci c cytokine production by T cells were detected following subcutaneous immunization with three doses of recombinant Tc24-C4 and TSA1-C4 antigens formulated with a TLR4 adjuvant, warranting a more detailed evaluation of the immune response induced by vaccination.
Here we performed a transcriptomic analysis of PBMC responses to vaccination in naïve rhesus macaques, to shed light on the immunogenicity of this Chagas disease vaccine and guide the optimization of vaccine doses and formulation.In particular, we focus on identifying IgG and T cell receptor repertoires, which can inform in detail on the breadth of the immune response induced by vaccination [17][18][19][20] .

PBMC transcriptomic pro le following vaccination
We rst assessed the transcriptomic pro le of PBMCs from vaccinated macaques one month after each of three vaccine doses, to further assess the immunogenicity of this candidate vaccine against T. cruzi.RNA-sequencing yielded 12-35 million quality reads/sample and over 80% were successfully mapped to the macaque genome.Differential expression analysis indicated that the gene expression pro le of PBMCs was not signi cantly altered after the rst vaccine dose, and only 17 genes were differentially expressed compared to baseline levels after the second vaccine dose.On the other hand, after the third vaccine dose, a total of 639 genes were differentially expressed, with 283 that were up-regulated and 356 down-regulated (Figure 1).
Pathway analysis was performed based on differentially expressed genes after the third vaccine dose.
Biological pathways associated down-regulated genes following vaccination included multiple RNA processing pathways as well as several metabolic/catabolic pathways (Figure 2A).On the other hand, upregulated genes were associated with leukocytes and lymphocytes activation pathways and other immune system activation, as well as cell migration, adhesion or activation (Figure 2B).These data suggested a strong activation of PBMCs and of the immune response after the three vaccine doses.
For a better understanding of the orientation of the immune response and its potential polarization, we assessed the changes in cytokine expression pro le from the PBMCs.As shown in Figure 3A, limited changes were observed in the cytokine pro les after each vaccine dose.In particular, no marked polarization was detected as Th1, Th2 and Th17 cytokines genes remained evenly expressed following vaccination.Analysis of Ig subclass expression pro les also indicated limited subclass switch following the rst vaccine dose, and no major changes after subsequent doses (Figure 3B).Indeed, the rst vaccine dose induced an increase in IgM, mostly at the expense of IgA expression, which was reduced, and this pattern was maintained after subsequent vaccine doses.Importantly, high levels of expression of IgG1 were sustained following vaccination, while no IgE was induced.Together, these results are in agreement with the induction of a broad and non-polarized immune response following vaccination, and three vaccine doses appear to be needed for a strong transcriptomic response of PBMCs.
We then typed the major histocompatibility complex (MHC) of the vaccinated macaques to assess potential MHC restriction of vaccine response.MHC class I alleles were the most diverse among the three macaques, with a total 6 and 5 alleles identi ed for Mamu A and Mamu B genes, respectively, while MHC class II genes were less diverse, with 1-4 alleles among these macaques (Table 1).As all three macaques developed an immune response to the vaccine antigens, these were adequately processed through these diverse MHC.Changes in immunoglobulin G repertoire following vaccination Immunoglobulin G (IgG) diversity result from random combinations of variable (V), diversity (D), and joining (J) gene segments 21,22 , to generate unique complementarity-determining region 3 (CDR3) within the hypervariable region of the protein which is involved in antigen-binding speci city 23 .Thus, we assessed the changes in IgG repertoire and Ig heavy chain VDJ gene usage following vaccination.As shown in Figure 4A, changes in V and D gene usage could be detected after each vaccine dose, while J gene usage was mostly unaffected by vaccination.For example, IGHV1 and IGHD3 were the most frequently expressed genes prior to vaccination in macaque KL72, while after three vaccine doses IGHV4 genes were the most frequently used, as well as IGHD3 and IGHD5 genes (Figure 4A).Similar changes were detected in the other macaques.These data suggested that different IgGs are produced after each vaccine doses.Therefore, we analyzed Ig CDR3 domain sequence diversity and elaborated networks illustrating antibody repertoire of all three vaccinated macaques.At baseline, an extensive diversity of CDR3 sequences were observed, but most were uniques and only a few were present in more than one copy (Figure 4B).Following the rst vaccine dose, a dramatic reduction in CDR3 sequence diversity was observed, while several sequences were much more abundant, likely indicative of the clonal expansion of vaccine antigen-speci c antibodies (Figure 4B).For example, CDR3 sequences ARDSSGSWNWFDV, ARREGSSSSGYYFDY, ARGSRGSLLGDYLEF, AREGCSGGVCSLRFDV expanded after the rst vaccine dose.Interestingly, no CDR3 sequence detected at baseline persisted after the rst vaccine dose, and all CDR3 sequences were novel, in agreement with the changes in VD gene usage detected above in all macaques.
After the second vaccine dose, all CDR3 sequences induced by the rst vaccine dose were replaced by novel CDR3 sequences, indicating a major turnover of antibody-producing cell clones.CDR3 diversity remained limited compared to the baseline level, and new CDR3 sequences were further expanded.These included ARGGFCSDSGCSSFDY, ARDLSAAADLYNWFDV, ARQPTRRYSRYFEF, ARDQPWWPRGSFDV.This turn-over process appeared to be repeated after the third vaccine dose as well, and a new repertoire of CDR3 sequences was detected, although with more limited clonal expansion of CDR3 sequences (Figure 4B).Analysis of Richness (Figure 4C) and Shannon (Figure 4D) diversity indices over time con rmed that antibody diversity was reduced after the rst and second vaccine dose, likely associated with the clonal expansion of a few vaccine antigen-speci c IgGs.Although antibodies with novel CDR3 sequences were generated after the third vaccine dose, there were no further changes in the level of antibody diversity, suggesting that two vaccine doses may be su cient for the induction of an optimum humoral response.
Changes in T cell receptor repertoire following vaccination T cell receptors (TCRs) are responsible for detecting epitopes presented by MHC molecules and their diversity is also driven by rearrangements of TCR beta V, D and J gene fragments and TCR alpha V and J genes, to generate unique CDR3 regions that determine TCR epitope binding a nity and speci city 24,25 .Thus, TCR beta CDR3 diversity and VDJ gene usage were similarly analyzed.TRB VDJ gene usage resulted altered starting with the rst vaccine dose, particularly for V and J genes (Figure 5A).For example, TRBV11 and TRBV12 genes frequently used at baseline were replaced by TRBV4 and TRBV6 which became the most used genes after the rst vaccine dose in macaque LD53.These data are consistent with the induction of antigen-speci c TCRs following vaccination.Further changes in gene usage were observed after the second vaccine dose, but limited changes seemed to occur after the third vaccine dose and TRBV12 and TRBJ1 genes were the most frequently used after vaccination.
Network analysis of TCR beta CDR3 domain diversity revealed further changes in the repertoire following vaccination (Figure 5B).Indeed, vaccination had limited effects on overall TCR beta CDR3 diversity, although it tended to decrease after the rst and second dose of vaccine as assessed by Richness and Shannon indices (Figure 5E and F).Nonetheless, novel CDR3 sequences were induced after each vaccine dose and several.were present at high frequencies, likely representing antigen-speci c TCR sequences.Moore strikingly, a few sequences present a low frequency at baseline persisted and were present at increasing frequencies following each vaccine dose, including the third dose (Figure 5B, C and D).These corresponded to CDR3 sequences CASSPGTVMEKLFF, CASRPGHPYEQYF, CASSLADPGGVQNTQYF, or CASSLETGSTDPQYF, for example.One of these CDR3 sequence was also identi ed in two macaques (CASSLADPGGVQNTQYF, detected in macaques KL72 and LD53), indicating convergence of their immune response.These data evidenced a strong T cell response to vaccination, with the likely proliferation of cells with antigen-speci c TCR beta.

DISCUSSION
The development of a Chagas disease vaccine would be a key step towards a better control of this neglected disease, and a vaccine prototype based on Tc24 and TSA1 antigens is emerging as an attractive candidate for further development 10,13 .An initial evaluation of this vaccine in naïve macaque suggested that it could stimulate both B and T cell immunity 16 .We aimed here to expand these observations by assessing the transcriptomic response of PBMCs from vaccinated macaques, and assess changes in their IgG and TCR repertoires.
Transcriptomic analysis revealed that no changes in gene expression pro les were detected one month after the rst vaccine dose, and only marginal changes one month after the second dose, so that three vaccine doses were required to detect signi cant alterations in the PBMC gene expression pro le.This was somewhat unexpected as a large increase in antigen-speci c IgG could already be detected one month after the rst vaccine dose, indicating immunogenicity at this early time point 16 .However, the response to vaccines is largely asynchronous in humans, with early and delayed responses depending on the vaccine and its formulation 26 .Similarly, the transcriptomic response of PBMCs from macaques vaccinated with BCG is larger on day two after vaccination and decreases in the following weeks 27 .On the other hand, PBMCs from humans vaccinated with Plasmodium falciparum sporozoites present a transcriptomic response at day 27 post-immunization but not before 28 .Thus, the inclusion of additional time points would be needed for a ne scale analysis of the kinetics of transcriptomic changes to this Chagas disease vaccine, as it is unclear what changes may occur in the few days after immunization.Also, increasing sample size would also allow to reach a greater power to identify differentially expressed genes as we observed many genes presenting potential changes in expression levels after the rst vaccine dose, but none reached statistical signi cance.
On the other hand, there was a clear transcriptomic response of PBMCs after the third vaccine dose, with over 600 differentially expressed genes, con rming its immunogenicity.These genes were involved in several metabolic/catabolic functions which appeared down regulated, suggesting changes in cell activity/differentiation.Also, as expected, several pathways associated with the activation of immune cells from leukocytes to T cells were upregulated, indicative of an active immune response.The immunoglobin expression pro le indicated limited subclass switch, as the main change was an increase in the proportion of IgM while IgA was reduced, and the other subclasses remained unaltered.In particular, the production of IgG1 remained predominant among IgGs, and this isotype is associated with the highest effector function activity in macaques 29 .Notably, the expression of IgE, which is associated with allergic reactions 30 was negligible, suggesting a lack of allergic reaction to the vaccine, although the analysis of nasal secretion would be required for con rmation 31 .Analysis of the cytokine expression pro le also suggested the induction of a balanced immune response, including Th1, Th2 and Th17 cytokines.Such a balanced immune response would be indicated for an effective vaccine against T. cruzi, as a cellular response is critical for parasite control but hyperpolarization may lead to tissue damage 10,32 .These data support the further evaluation of vaccine e cacy against T. cruzi infection.
Analysis of the IgG heavy chain repertoire indicated changes in their VDJ gene usage one month after the rst vaccine dose, and after each following vaccine dose.These changes were associated with a renewal of the CDR3 repertoire after each vaccine dose, and no clone persisted over time.Rapid changes in Ig CDR3 repertoire have similarly been observed at 7 and 28 days post-vaccination with a pneumococcal vaccine in humans 33 .Also, the longitudinal follow-up of the human IgG repertoire showed that >85% of CDR3s can be detected only once out of 24 time points over a 11 months period 34 , indicating a very rapid turnover.Although vaccine antigens Tc24 and TSA1 are both highly conserved among T. cruzi strains 35,36 , the generation of diverse polyclonal antibodies against them may help broaden their binding a nity to accommodate for more sequence variants and/or increase their e cacy by targeting different regions of the antigens.
Vaccination also induced important changes in T cell receptor repertoire, with the notable persistence and expansion of speci c CDR3 sequences over at least 4 months, as well as some convergence in two of the vaccinated macaques.These expanded CDR3 clones are likely antigen speci c and further studies should help assess this speci city.These data evidence a clear T cell response to vaccination and are in agreement with the activation of T cells targeting multiple epitopes from the vaccine antigens, and their expansion over at least 4 months suggest the presence of memory T cells.Vaccination in humans has often been associated with increases in TCR beta CDR3 diversity, for example in response to a Rabies virus vaccine 37 or a Hepatitis B vaccine 38 , but our data suggest a more focused response to our vaccine candidate with no expansion of the TCR repertoire.This may be due to differences among vaccine antigens and their immune processing, or differences in the kinetics of these responses as more time points would be needed for a more detailed analysis of their time course, as mentioned above.Also, while two vaccine doses were su cient to induce changes in TCR repertoire and expansion of some CDR3 clones, the third vaccine dose brought further expansion of these clones and may thus be needed for a stronger immunogenicity of this vaccine.Future studies should help assess potential differences in protective e cacy after two or three vaccine doses.
Remarkably, all three vaccinated macaques presented transcriptomic responses and changes in their IgG and TCR repertoires following vaccination, in the context of a varied MHC background, particularly for class I MHC.This is also encouraging as it suggests that so far there are no MHC restrictions of vaccine immunogenicity 39,40 and future studies should include animals with additional MHC background to broaden this evaluation.
In conclusion, we identi ed here a clear transcriptomic response of PBMCs from macaques vaccinated a Chagas disease vaccine based on Tc24 and TSA1 parasite antigens, with the up-regulation of several immune cell activation pathways.These data con rming the immunogenicity of this vaccine, with a broad non-polarized immune pro le.While changes in IgG and TCR repertoires could be detected one month after the rst vaccine dose, further changes were observed after the second and third vaccine dose.The IgG repertoire showed a rapid turnover with new IgGs following each vaccine dose, while the TCR repertoire presented persisting clones that were expanded after each vaccine dose, suggesting that three vaccine doses may be needed for optimum e cacy.This work warrants the further characterization of the IgG and T cell receptor repertoires of Chagas disease patients 41 , particularly with different stages of disease progression/parasite control, to identify immune repertoire pro les associated with better disease outcomes and correlates for protection.Also, the evaluation of the protective e cacy of this vaccine candidate against T. cruzi infection will be key for vaccine development and the availability of naturally-infected macaques provide an ideal model for testing a therapeutic vaccine 42 .

Animals and vaccination
Animals were housed at the Tulane National Primate Research Center (TNPRC) under the care of TNPRC veterinarians, in accordance with the standards incorporated in the Guide for the Care and Use of Laboratory Animals and with the approval of Tulane Institutional Animal Care and Use Committee (IACUC).Three naïve male rhesus macaques (Macaca mulatta) 4-5 years old were vaccinated with three doses of the vaccine based on Tc24-C4 and TSA1-C4 antigens formulated with a TLR4 adjuvant, one month apart, as described before 16 .A blood sample was collected in EDTA at baseline and one month after each vaccine dose.PBMCs were isolated and cryopreserved until used for RNA puri cation and RNA-sequencing.
RNA puri cation and RNA sequencing Cryopreserved PBMCs 16 were checked for viability with trypan blue, and about 10 6 cells were used for RNA extraction using PerfectPure RNA Cultured Cell Kitä (5 Prime, Inc.) following manufacturer instructions.RNA integrity was assessed on an Agilent BioAnalyzer and all samples had a RIN>8.About 200 ng of RNA was used for library preparation and sequencing on an Illumina MiSeq platform, and about 12-35 million reads/sample were obtained after quality ltering.Raw reads have been deposited in NCBI SRA database under Bioproject #PRJNA1010169, Biosamples SAMN37182435-SAMN37182446.

Transcription pro le
Reads were mapped to the Rhesus macaque reference genome (Mmul10 accession: GCF_003339765.1) in Geneious 11.Read counts were then normalized and differentially expressed genes called using DESeq2 as implemented in iDEP1.1 43 .One sample was removed from further analysis (macaque KL72, second vaccine dose) due to a very low number of reads.Differences in gene expression levels >1.5 fold change were called at a signi cance alpha of 0.05 adjusted for multiple testing using the false discovery rate method.Volcano plots were used to visualize differentially expressed genes.Enriched functional pathway associated with up-regulated and down-regulated genes, respectively, were identi ed using ShinyGO 0.77 44 , based on Gene Ontology Biological Process database.

MHC typing
RNA-seq read counts mapping to MHC reference genes (Mamu A, Mamu B, Mamu DPA, Mamu DPB, Mamu DQA, Mamu DQB and Mamu DRA) for individual macaques were used for de-novo assembly in Geneious.Assembled MHC sequences were compared with the IDP-MHC allele database 45 using BLAST and the top two matches with >98% sequence identity for each MHC gene were retained.
Antibody isotype RNA-seq read counts mapping to unique 150 bp regions of IgA, IgE, IgG1, IgG2, IgG3, IgG4 and IgM subclasses were used to assess the relative expression level of the respective antibody isotype owing vaccination.These were normalized according to the total number of reads mapping to the macaque genome from each sample, to account for differences in sequencing depth and coverage and provide relative abundance level of Ig isotypes.

IgG antibody repertoire
Reads mapping to the variable CDR3 region of IgG heavy chain gene, which mediates antibody speci city, as well as anking regions were extracted for analysis of IgG variable domain repertoire using IgBLAST 46 .The combination of the variable (V) gene, the diversity (D) gene and the joining (J) gene from each sequence was determined to assess gene usage frequency for IgG production in individual macaques following vaccination.The translated CDR3 sequences were analyzed for frequency and similarity using the EFI Enzyme Similarity Tool 47 and similarity networks were elaborated in Cytoscape to visualize changes in antibody repertoire over time.CDR3 sequence diversity was further assessed using Shannon H and Richness S summary indices, which were calculated in Past4.0 48.

T cell receptor repertoire
Reads mapping to the CDR3 region of the TCR beta subunit gene, which mediates T cell epitope binding speci city, as well as anking regions were extracted for analysis of TCR variable domain repertoire using IMGT/HighV-QUEST 49,50 .The combination of the TCR beta V/D/J genes from each sequence was determined to assess TCR gene usage frequency and its changes following vaccination.The translated CDR3 sequences were analyzed for frequency, similarity and diversity as described above.Figure 5

Table 1 .
MHC typing of vaccinated macaques