Tick-borne Pathogens as Etiologies of Undifferentiated Acute Febrile Illness in Kenya: Geographical Distribution, Risk Factors and the Tick Bacteriome as a Bio-indicator of Tick - Borne Zoonosis

Abstract

. Background: Vector-borne pathogens, which include viruses, bacteria and protozoa, are among the most common etiological agents of acute febrile illness (AFI) worldwide and have reportedly caused epidemics of human disease in recent decades. Ticks are important vectors and they are ranked only second to mosquitoes, in vector-borne disease transmission. Studies carried out at Kenya Medical Research Institute / United States Army Medical Research Directorate – Africa (KEMRI/USAMRD-A) have shown a wide distribution of ticks in Kenya indicating potential for transmission of tick-borne pathogens including Rickettsia spp., Borrelia spp., Coxiella burnetii, Ehrlichia spp., Babesia spp., Anaplasma spp. and Crimean-Congo virus. Although these pathogens have been reported in Kenya, there has been no systematic study of the disease burden in the country. Within Kenya, AFI is a reason common for patients to go to health care facilities and patients are often empirically treated without laboratory evaluation for the causal agent. It is therefore, imperative to determine etiologies of acute febrile illness other than malaria and associated risk factors. Objective: The present study aimed at optimizing two multiplexed real time polymerase chain reaction assays for detection of eight tick-borne pathogens and using them to evaluate the distribution, risk factors and associated signs and symptoms of tick-borne pathogen infections as etiological agents of acute febrile illness in different geographical regions in Kenya. The study also aimed at determining whether ticks infesting domestic animals in Kenya harbor bacteria of zoonosis relevance that can be transmitted to human. Methodology: The present study was conducted on 6,207 archived samples that were collected from patients seeking healthcare aged one year and above with fever of equal or greater than 38⁰C without a diagnosis of the causative agent of fever after routine clinical diagnosis. Samples used in this study were collected between 2008 and 2017 in county and sub-county hospitals located in diverse geographical regions in Kenya including Lake Victoria region (Kisumu County hospital, Kombewa sub-county hospital, Alupe sub-county hospital, Busia), Kisii Highlands (Kisii county hospital), Rift valley region (Marigat sub-county hospital Baringo, Lodwar county hospital and Gilgil sub-county hospital, Nakuru) North Eastern Region (Iftin sub-county hospital, Garissa, Garissa police line dispensary, Garissa and Isiolo county hospital), Coastal region (Malindi sub-county hospital, Kilifi) and Cosmopolitan urban region (Eastleigh health center, Nairobi). To determine bacteria community in ticks that can be potentially transmitted to humans, archived genomic DNA samples from 460 adult tick samples were used. These ticks were collected between 2007 and 2008 from domestic animals including cattle sheep and goats from diverse geographical regions in Kenya. Two multiplex Real Time PCR assays and 16s rRNA metagenomics were used for detection of tick-borne pathogens in the human and tick specimens, respectively. Demographic and clinical data including age, gender, place of residence, signs and symptoms, contact with animals and type of animals, tick bite, collected at the time of sample collection was used to determine risk factors for human tick-borne infections. Results: Two real time PCR assays in a multiplex format were optimized and evaluated for eight tickborne zoonosis pathogens. The assays were organized in duo assays of 4-plex each. Assay 1 was optimized for Anaplasma phagocytophilum, Coxiella burnetii, Borrelia burgdoferi and Ehrlichia chaffeensis. Assay 2 was optimized for Rickettsia spp., Bartonella spp., other Borrelia spp. which are not B. burgdoferi and Babesia spp. Using synthetic plasmids it was

shown these assays can specifically detect all the target sequences in the same reaction tube. Assay 1 had a limit of detection of 2 copies for all target genes. Assay 2 was less sensitive and on average had a limit of detection of 18 copies of target genes. In replicate tests, both assays had intra-assay variation of less than two cycles. Using Bland-Altman analysis, the performance of the two 4-plex assay was similar to that of the singleplex assays qualifying them for simultaneous detection of four tick-borne zoonosis pathogens per assay. Whole blood samples collected from undifferentiated acute febrile illness patients were tested for presence of eight tick-borne zoonosis using the two 4-plex real time PCR. Of the patients accessed 3,082 (50.9%) were male and 2,979 949.2%) were female. The age of the patients ranged 1 to 80 years with a median of 5 years. Of these, all 6,207 were tested for Rickettsia, 4,017 for Coxiella, 3,615 for A. phagocytophilum, 3,659 for E. chaffeensis, and 3,657 for B. burgdoferi, 3,480 for Babesia spp., 3,463 for Bartonella spp and 3,463 for Borrelia spp. which are not B. burgdoferi. Overall, 1,017 (16%) were positive for at least 1 tick - borne zoonosis pathogen (confidence interval (CI :) 15.3 – 17.1). Of these 475 (8%) (CI: 7.9-8.3) were positive for Rickettsia, 271 (6%) were positive for Coxiella (CI: 5.6-7.0), 55 (1.5%) were positive for A. phagocytophilum (CI: 1.1 – 1.9), 111 (3%) were positive for E. chaffeensis (CI: 2.5 – 3.6), 56 (1.5%) were positive for B. burgdoferi, (CI: 1.1 – 1.9), 36 (1%) were positive for Babesia spp. (CI:: 0.2 – 1.4), 8 (0.2%) were positive for Bartonella spp. (CI: 0.07 – 0.4) and 135 (4%) were positive for other Borrelia spp. which are not B. burgdoferi (CI: 3.3 – 4.5). Out of the 1017 tick-borne zoonosis positive specimens 906 (89.1%) tested positive for a single tick-borne zoonosis pathogen and 99 (9.73 %) samples were positive for two tick-borne zoonosis pathogens, 11 (1.1%) samples were positive for three tick-borne zoonosis pathogens and 1 sample was positive for four tick-borne zoonosis pathogens. Majority of the co-infections

were combination of Borrelia spp. with Rickettsia spp. and Coxiella burnetii with Rickettsia spp. There was a higher prevalence of tick-borne zoonosis pathogen infection in the age categories 6 to 15 years (17.5%) and ≥ 16 years (19.4%) compared to 14.3% in the ≤ 5 years category (p-values = 0.02 and < 0.0001, respectively). Patients from a cosmopolitan urban region of Eastleigh, Nairobi County had higher prevalence of tick-borne zoonosis pathogens infections of 39% compared to Coastal region with 17%, Kisii highlands with12%, Lake Victoria region with 16%, and arid and semi – arid region with 17%. Patients who reported having contact with sheep and had a tick bite had higher odds of tick-borne zoonosis pathogen infection. In ticks collected from domestic animals, a total of 645 unique operational taxonomic units (OTUs) (bacteria genera) were detected and grouped into 27 bacteria phyla. Sequence reads in the phyla were skewed with four phyla contributing 96.2% of the sequences. Proteobacteria contributed the majority 61.2 % of the sequences that tarried to 33.8% OTUs, 15.9% for Firmicutes (23.4% OTUs), and 15.6% for Actinobacteria (20% OTUs), 4.7% for Bacteroidetes (11.6% OTUs). The remaining 22 phyla included 0.7% for Fusobacteria, 0.5% for TM7 (Saccharibacteria) 0.3% for Verrucomicrobia 0.2% for Acidobacteria, 0.2% for Deinococcus-Thermus, 0.2% for Planctomycetes, 0.1% for Chloroflexi, OD1 (Parcubacteria), Tenericutes and Gemmatimonadetes, 0.04% for Armatimonadetes, 0.03% for Spirochaetes, 0.02% for Aquificae, 0.01% for SR1 (Absconditabacteria), Lentisphaerae and BRC1, 0.004% for Chlamydiae and Nitrospira, 0.002% for Chlorobi, Synergistetes, Fibrobacteres, WS3 and Elusimicrobia contributed only 2.5 % of the sequences (11.2% OTUs). Potentially pathogenic bacterial genera identified include Coxiella 41.8%, Corynebacterium 13.6%, Acinetobacter 4.3%, Staphylococcus 3.9%, Bacillus 2.7%, Porphyromonas 1.6%, Ralstonia 1.5%, Streptococcus 1.3%, Moraxella 1.3%, Cloacibacterium 1.3%, Neisseria 1.2%, Escherichia_Shigella 1.2% and Proteus, Aerococcus, Alloiococcus, Stenotrophomonas 1% each. Coxiella genus was the most abundant constituting 41.8% (15,445,204 out of 36,973,934 total sequences). Other less abundant (<0.1%) but potentially pathogenic genera included Burkhoderia (0.4%), Klebsiella (0.3%), Escherichia- Shegella (0.3%), Achromobacter (0.2%), Rickettsia (0.1%), Haemophilus (0.1%), Legionella (0.1%), Campylobacter (0.04%), Treponema (0.03%), Francisella (0.02), Anaplasma (0.01), Elizabethkingia (0.006%), Mycoplasma (0.006%), Ehrlichia (0.005%), Bordetella (0.004%), Vibrio (0.002%), Borrelia (0.0008%) and Brucella (0.0002%). By Shannon diversity index, A. variegatum carried less diverse bacteria (mean Shannon diversity index of 2.69 ± 0.92) compared to 3.79 ± 1.10 for Amblyomma gemma, 3.71 ± 1.32 for A. hebraeum, 4.15 ± 1.08 other Amblyomma spp, 3.79 ± 1.37 for Hyalomma truncatum, 3.67 ± 1.38 for other Hyalomma spp, 3.86 ± 1.27 for Rhipicephalus annulatus, 3.56 ± 1.21 for Rh. appendiculatus, 3.65 ± 1.30 for Rh. Pulchellus, but the difference was not statistically significant (p=0.443). Conclusion: Eight tick-borne pathogens were detected in undifferentiated acute febrile illness patients from different geographical regions in Kenya using the two 4-plex real time PCR assays optimized in this study. This study has shown up to 16.4 % prevalence of tick-borne zoonosis pathogens in acute febrile illness patients in Kenya. Prevalence of tick - borne zoonosis infection was significantly higher in patients of 6 years and above compare to patients of less than 5 years. Patients living in an overcrowded cosmopolitan urban region and reporting a tick bite were at higher risk of tick-borne zoonotic infections. Patients with tick-borne zoonosis infections reported non-descriptive signs and symptoms which can lead to difficulties in definitive diagnosis of tick-borne zoonosis. Adult ticks collected from livestock animals including cattle,

sheep and goats in the Kenya harbored an array of bacterial community including pathogenic bacteria genera which suggests potential risk of transmission of tick - borne zoonosis infections in the country. Pathogens harbored and transmitted by ticks (Tick-borne pathogens) should be added to the list of etiologies considered for differential diagnosis of acute febrile illness patients and laboratory capability for their diagnosis should be enhanced for early detection and appropriate treatment in Kenya. Tick control effort in Kenya should embrace one health approach to include human, animals and environment factors.