Role of human Pegivirus infections in whole P. falciparum sporozoite vaccination and 1 controlled human malaria infection in African volunteers

45 Background: Diverse vaccination outcomes and protection levels pose a serious challenge to 46 the development of an effective malaria vaccine. Co-infections are among many factors 47 associated with immune dysfunction and sub-optimal vaccination outcomes. Chronic, 48 asymptomatic viral infections can contribute to the modulation of vaccine efficacy through 49 various mechanisms. Human Pegivirus-1 (HPgV-1) persists in immune cells thereby potentially 50 modulating immune responses. We investigated whether Pegivirus infection influences 51 vaccine-induced responses and protection in African volunteers undergoing whole P. 52 falciparum sporozoites-based malaria vaccination and controlled human malaria infections 53 (CHMI). 54 Methods: HPgV-1 prevalence was quantified by RT-qPCR in plasma samples of 96 individuals 56 before, during and post vaccination with PfSPZ Vaccine in cohorts from Tanzania and 57 Equatorial Guinea. The impact of HPgV-1 infection was evaluated on (1) systemic cytokine 58 and chemokine levels measured by Luminex, (2) PfCSP-specific antibody titers quantified by 59 ELISA (3) asexual blood stage parasitemia and pre-patent periods with HPgV-1 infection status and (4) HPgV-1 RNA levels upon asexual blood stage parasitemia induced by CHMI. Results: The prevalence of HPgV-1 was 29.2% (28/96) and sequence analysis of the 5`UTR and E2 region revealed the predominance of genotypes 1, 2 and 5 in the positive volunteers. HPgV-1 elevated systemic levels of IL-2 and IL-17A. vaccine-induced anti-PfCSP antibody titers, blood stage and pre-patent periods HPgV-1 and negative level 71 Conclusions: Although HPgV-1 infection did not alter vaccine-elicited levels of PfCSP- 72 specific antibody responses and parasite multiplication rates, an ongoing infection appears to 73 improve some degree of protection against CHMI in PfSPZ-vaccinated individuals. This is 74 likely through modulation of immune system activation and systemic cytokines as higher levels 75 of IL-2 and IL17A were observed in HPgV-1 infected individuals. CHMI is safe and well 76 tolerated in HPgV-1 infected individuals. Identification of cell types and mechanisms of both 77 silent and productive infection in individuals will help to unravel the biology of this widely 78 present but largely under-researched virus. 79 80

Study population 145 Volunteers were enrolled in the BSPZV1 (NCT03420053), BSPZV2 (NCT02613520) BSPZV3 146 (NCT03420053) and EGSPZV2 studies (NCT02859350) that evaluated the safety, 147 immunogenicity and efficacy of live, cryopreserved, purified, irradiated-attenuated P. 148 falciparum sporozoites in malaria pre-exposed volunteers. Vaccine efficacy was evaluated by 149 homologous CHMI based on direct intravenous inoculation of 3200 fully infectious, aseptic 150 purified cryopreserved P. falciparum sporozoites. These whole sporozoites were suspended in 151 0.5 ml human serum and administered as a single dose 3 weeks after last vaccination.

153
Identification of Human Pegivirus RNA in RNA-Seq data from whole blood 154 Whole blood samples were used from a subset of participants (n=28) (Supplementary Figure   155 1A) that were enrolled into the BSPZV1 trial (NCT03420053). All volunteers were healthy 156 males, aged 18 to 35 years and confirmed as negative for HIV-1, Hepatitis B and C before 157 enrolment into the trial. Blood was collected and stored in Paxgene RNA tubes at different 158 timepoints including before vaccination (baseline), 2 days after first vaccination, 7 days after 159 the first and second vaccination as well as before CHMI, 2 and 9 days after CHMI. Each of the 160 placebo (n=6) and the vaccinees (n=22) had a total of 3 and 7 blood sampling timepoints 161 respectively resulting in 172 samples in total. All available samples (n=172) were subjected to 162 RNA-Seq analysis performed by Stuart Lab in Seattle. Briefly, RNAseq data was generated and 163 made from globin/rRNA cleared whole blood RNA that was fragmented and first strand cDNA approximately 3 million initial non-human unmapped paired end reads from each volunteer 173 were analysed. The unmapped reads were first searched for "suspected" viral hits by running 174 bowtie2 against the NCBI database containing more than 7424 viral genomes. Thereafter, low 175 quality and complexity reads as well as reads mapping to human genome, transcriptome and 176 repeat regions were removed from the resulting "suspected" viral reads using bowtie 2, knead 177 data and tandem repeat finder algorithms respectively. The "clean" viral reads were then  . In addition, human RNase P primers were also added as internal control.

219
(2 ul, 10 U/ul) and RNase free water (9 ul to 20 ul). Amplification conditions included 50 mins  Nucleotide sequence analysis and phylogenetic relatedness was performed in Geneious 235 software version 8.1.9. Chromatograms were examined for quality first, and only sequences 236 with quality above 86% were included in analysis. CLUSTALW algorithm was used to align 237 5` UTR nucleotide sequences from volunteers to selected reference sequences corresponding 238 to 5` UTR of HPgV-1 (genotype 1 to 7) available in the NCBI database. Thereafter phylogenetic 239 trees were constructed by neighbour joining method and the Kimura two parameter models.    were not analysed for multiple correction as we considered this question as exploratory. P-value 295 ≤ 0.05 was considered significant. Differences in viral diversity and abundances and prevalence 296 were assessed using LEfSe (Linear discriminant analysis effect size) [47] and GraphPad 297 Software (Prism V5) respectively.

301
Unbiased search for RNA molecules encoding human viruses in RNA-seq transcriptomics 302 data 303 We aimed to identify viruses present in our volunteers participating in PfSPZ Vaccine studies 304 by using a metagenomics approach. Analyses included samples from 28 participants collected 305 at multiple time points including before vaccination (baseline), 2 days after first vaccination, 7 306 days after the first and second vaccination as well as before CHMI, 2 and 9 days after CHMI.  (Fig. 1A). The number of reads for each of the identified viruses was quantified and is given in HPgV RNA with read counts ranging from low to high (Fig. 1C). Three out of 8 HPgV-positive 322 individuals were co-infected with CMV (Fig. 1C). This result indicated that HPgV is highly 323 prevalent (28.6%, 8/28) in Tanzanian adults. To reconfirm our findings, we extracted RNA 324 from plasma samples collected from these 8 volunteers and amplified HPgV-1 by RT-qPCR. 325 We were able to reconfirm in 2 out of 8 volunteers the in silico identified presence of HPgV 326 RNA. Interestingly, these 2 volunteers had the highest RNA read counts for HPgV in our 327 bioinformatics analysis.  341 Next, we quantified HPgV-1 viral load in plasma samples using RT-qPCR. HPgV-1 viral loads 342 were comparable between individuals from the two countries (Fig. 3A). However, based on 343 viral loads with a defined threshold of 10 6 viral RNA copies/ml of plasma, both cohorts were individuals, 2 were excluded due to poor quality of the sequences. Genotype 1 was found in 355 only 2 volunteers (7.7%). Surprisingly, genotype 2, described as dominating in Europe and 356 America, was found in 24 of 26 volunteers (92.3%) (Fig. 4). Most genotype 2 strains clustered 357 closely with the related genotype 2a sequences described from Venezuela (Fig. 4). To further 358 increase the resolution of the genetic relatedness of our isolates, we amplified in addition the 359 polymorphic E2 region of HPgV-1 virus. E2 RNA was successfully amplified and sequenced 360 in 9 out of 28 volunteers (32%). According to the E2-sequences of our HPgV-1 isolates, our 361 strains clustered within genotype 1, 2, and 5 (Fig. 5). In summary, these results show that a  (Supplementary Fig. 2). Although there was a trend of overall higher cytokine levels in HPgV-370 1 infected individuals, only IL-2 and IL-17A reached significance levels (Fig. 6). There was no 371 statistically significant difference in cytokine and chemokine levels when high and low viremic 372 individuals were compared. Also, we could not find differences in chemokine and cytokine  (Fig. 7 A-B). Similar results were observed when PfSPZ vaccine-induced antibody 383 responses were analysed as net (14 days post last vaccination-baseline) (Fig. 7 C) and as fold 384 change (14 past last vaccination/ baseline ratios) (Fig. 7 D). No significant correlation between 385 HPgV-1 infection status and measured anti-PfCSP antibody titers was found.  Supplementary Fig. 3). While none of the 393 placebo-receiving participants was protected (0/20), the overall protection in the vaccinated 394 group was 55% (26/47). The HPgV-1 prevalence was comparable in these 2 groups, 35% (7/20)  (Fig. 8A). Surprisingly, HPgV-1 positive vaccinees 399 16 showed higher protection after CHMI (62.5%; 10/16) than HPgV-1 negative individuals 400 (51.6%; 16/31). We also assessed anti-CSP antibodies titres at 14 days past last vaccination.

401
Slightly higher anti-CSP levels were seen in protected than non-protected individuals (Fig. 8B), 402 but without any statistical significance. However, these levels tended to be lower in the HPgV- is unknown. We evaluated parasite multiplication rates and pre-patent period in the control 409 volunteers (n=20) undergoing CHMI using PfSPZ challenge. Comparable asexual blood stage 410 multiplication rates and pre-patent periods were observed between HPgV-1 positive and 411 negative individuals (Fig. 9 A-B). positive before CHMI and 6 were positive at 28 days post CHMI (Fig. 9C).    The limitation of using only the amplification of the 5` UTR, a highly conserved region, to 474 discriminate closely related isolates is known [67,68]. We therefore amplified the E2 region in 475 addition to a successfully amplified 5` UTR. We were able to detect and sequence the E2 476 coding RNA only from subjects with high viremia (n=9). Based on the E2 sequences, these nine   [78,79]. The HPgV-1 envelope protein 2 506 (HPgV1-E2) has been implicated in these outcomes, due to its ability to inhibit T cell-receptor 507 mediated signalling and IL-2 signalling pathways [78,79].            were compared between groups and unpaired t-test was used to calculate significance.