Inconsistent MHC Class II association in Beagles experimentally infected with 3 Leishmania infantum 4 5 6

20 The clinical outcome of Leishmania infantum infection in dogs varies from subclinical 21 infection to severe disease. Researchers attribute this variability in clinical manifestations to 22 the ability of the immune response to limit pathogen multiplication and dissemination, which 23 is, in part, likely determined by the immune response genes. The aim of this study was to test 24 the hypothesis that MHC class II genes are associated with disease outcome of experimental 25 L. infantum infection in Beagles. Dog leukocyte antigen (DLA) class II haplotypes were 26 characterised by sequence-based typing of Beagle dogs experimentally infected with L. 27 infantum during vaccine challenge studies. Variability of response to infection was 28 determined by clinical score, serology and quantification of L. infantum DNA in the bone 29 marrow over the study period. 30 31 Dogs showed limited DLA diversity and the DLA profiles of dogs recruited for the 32 different vaccine challenge studies differed. There were variable responses to infection, 33 despite the apparent restriction in genetic diversity. One haplotype DLA-DRB1*001:02-34 DQA1*001:01--DQB1*002:01 was associated with increased anti-Leishmania antibodies in 35 one infection model, but no DLA associations were found in other groups or with parasite 36 load or clinical score. Examination of this particular DLA haplotype in a larger number of 37 dogs is required to confirm whether an association exists with the immune or clinical 38 responses to L. infantum infection. 39 40


42
The protozoan parasite Leishmania infantum is most commonly transmitted between 43 mammalian hosts via biting female sandflies, belonging to the genera Phlebotomus or   (Baneth et al., 2008). In an experimental infection model utilised in vaccine studies, high 55 doses of amastigotes or promastigotes are given IV, but even under these circumstances, 56 some dogs do not develop clinical signs over the study period, despite a relatively large 57 challenge dose (Campino et al., 2000;Costa et al., 2013).

59
The non-specific clinical signs and the absence of a reference standard test complicate 60 the diagnosis of canine leishmaniosis (Rodriguez- Cortes et al., 2010). Serological testing is 61 often used for diagnostic purposes, to monitor the infection course and/or the response to 62 treatment. However, while Leishmania-specific antibody levels do not correlate with disease 63 protection, high antibody reactivity is associated with clinical disease (Reis et al., 2006). 64 Detection of Leishmania DNA in the tissues with PCR is a sensitive alternative technique for identifying infection (Cortes et al., 2004) and high parasite load in the tissues is associated 66 with clinical disease (Dos-Santos et al., 2008).

68
Previous canine studies appeared to confirm a role for T-cell mediated immunity 69 (CMI) in resistance to canine leishmaniosis, with IFN-γ, produced by stimulated lymphocytes 70 from subclinically infected dogs, able to lyse Leishmania infected macrophages, in contrast 71 with lymphocytes from clinically infected dogs (Pinelli et al., 1995). Despite several studies 72 examining cell mediated immunity in dogs, a clear picture of the T-helper phenotypes 73 associated with disease outcome has not emerged and results are often contradictory (Hosein 74 et al., 2017). Therefore, CMI assays are infrequently performed to diagnose clinical 75 leishmaniosis and their utility for predicting outcome of infection is not always reliable.

77
The genetic background of the host might play a role in determining the outcome of 78 infection with L. infantum and differences in susceptibility between different dog breeds has 79 been suggested. The Ibizan hound in particular has been identified as a potentially resistant 80 breed (Solano-Gallego et al., 2000). Other studies have suggested that the Cocker spaniel 81 and Boxer breeds might be more at risk of developing clinical disease (Franca-Silva et al., 82 2003). As the outcome to L. infantum infection is largely dependent on the host immune 83 response, much of the genetic research has focussed on immune response genes that might 84 determine the outcome of infection.

86
Dog leukocyte antigen (DLA) class II genes determine antigen presentation by MHC 87 class II molecules and influence the subsequent immune response; therefore, they might also 88 determine the ability to control L. infantum parasite numbers in tissues and clinical outcome.

89
A previous study has examined DLA genes in a naturally infected group of cross breed dogs 90 in Brazil and DLA-DRB1 015:02 was associated with increased risk of L. infantum infection 91 (Quinnell et al., 2003). with 1 x10 6 promastigotes, dogs in study C (n=23) were challenged with 5 x10 7 promastigotes 113 and dogs in study D (n=20) were challenged with 2 x10 8 amastigotes. Dogs were monitored 114 for a period of up to 2 years and were regularly examined and subjected to diagnostic testing 115 during this period. Dogs were allocated a clinical score (0-2/2), for clinical parameters which 116 included body condition, demeanour, skin lesions, mucous membrane colour, ocular lesions, 117 lymph node size (Supplementary Table 2). A combined clinical score was then allocated for 118 each time point, based on sum of the individual scores. Dogs were monitored for between 5 119 and 19 months and clinical scores were assigned every 2-8 weeks, depending on the study 120 design.  Quantitative PCR was performed on bone marrow samples as previously described 142 (Francino et al., 2006). Quantitative analysis was performed through absolute quantification 143 from a 6-point standard curve, with serial dilutions of a parasite culture, top standard 144 equivalent to 500 promastigotes, and values were expressed as genome copy/mL bone  Each study group was analysed separately, since challenge and dose was likely to 156 influence outcome. At each monthly time point studied and for each phenotype parameter 157 (clinical score, serology and parasite load determined by qPCR), dogs were ranked based on 158 whether they were positioned above or below the median score. Dogs consistently (>70% of 159 time points in study) above the median score for each phenotype were categorised as high.

160
Dogs consistently (>70% of time points in the study) below the median score for the 161 phenotype were categorised as low. All other dogs were categorised as medium for each 162 phenotype. Dogs that were euthanased as a result of L. infantum infection during the study 163 period were assigned to the high category regardless of phenotyping method.

164
Haplotype frequencies were calculated and were compared between groups using 166 Fisher's exact test, with Bonferroni correction when all haplotypes were examined 167 concurrently, in SPSS Statistics v22 (IBM, Hampshire, UK). Haplotype frequencies between 168 different groups were compared for each phenotyping method. Each group was analysed 169 separately and then all groups were analysed together.  PCR was performed using a G-Storm GS1 Thermal Cycler (Gene Technologies).

187
Reactions were heated to 95 °C for 10 min, followed by 35 cycles consisting of 94 °C for 40 188 s, 55 °C for 30 s for DQA1 or 60 °C for DRB1 and DQB1, and 72 °C for 1 min, with a final 189 extension step at 72 °C for 10 min. PCR products were processed using the GenElute PCR 190 Clean-up Kit (Sigma-Aldrich) and submitted for sequencing (Source Bioscience) using M13F 191 primer.

193
Sequencing results were analysed using CLC Workbench v 6.9.1 (CLC bio). DLA 194 alleles were assigned using SBT Engine 3.6.1 software (GenDx). Three locus haplotypes 195 were assembled from the assigned alleles, based on previous data regarding common 196 haplotypes in Beagle dogs (Soutter et al., 2015).  was not considered appropriate to combine the dogs into a single cohort. Therefore, it was 245 decided to undertake further analysis separately for each study group and to phenotype the 246 dogs independently, according to clinical score, serology or parasite load (Table 1). 247 248 DLA typing 249 We identified 10 different DLA haplotypes (found in two or more dogs) in this study, 250 plus three other haplotypes found in single dogs only, with substantial variation in the DLA 251 profile between study groups (Table 2). Two haplotypes were common between all four study 252 groups, while four haplotypes were only present in one of the four groups. The most   (Table 3). Specifically, the DLA-DRB1*001:02--DQA1*001:01--DQB1*002:01 haplotype 268 was more common in the high seropositivity group than the low seropositivity group       DNA by qPCR compared with peripheral blood samples (Francino et al., 2006;Hernandez et 357 al., 2015). Most dogs in our study were qPCR positive in the bone marrow by the end of the   The individual study designs were somewhat variable, which meant that phenotyping 382 dogs for a genetic association study was challenging and groups could not be easily  (Kennedy et al., 1998).