Health and fitness effects of Anaplasma species
infection in African buffalo (Syncerus caffer)
Danielle Rae Sisson
Thesis submitted for the degree of Master of Philosophy
Melbourne Veterinary School
Faculty of Veterinary and Agricultural Sciences
ORCID ID: https://orcid.org/0000-0002-1111-9635
September 2017 Abstract
Anaplasma marginale and A. centrale are intra-erythrocytic bacteria of domestic and wild ruminants and are mainly transmitted by ixodid ticks. Most of the work on anaplasmosis has been done on A. marginale infections in cattle, as it can cause disease with varying levels of severity, from icterus and anaemia, to abortions and death. However, wildlife, such as African buffalo (Syncerus caffer), appear to be only subclinically infected with A. marginale and A. centrale. This thesis aimed to characterise A. marginale and A. centrale in African buffalo from
Kruger National Park (KNP), South Africa, and investigate the effects of the burden of
Anaplasma species on the health and fitness of their host.
Firstly, the major surface protein 1α (msp1α) and heat-shock protein (groEL) genes were used to characterise A. marginale and A. centrale, respectively, from African buffalo. Sequence variation and phylogenetic analyses revealed that sequences of Anaplasma spp. from African buffalo were unique and that they grouped separately when compared with previously published sequences of both species. Sequencing the same species in cattle from the same area in the future will allow for more conclusive evidence as to whether African buffalo are a reservoir for anaplasmosis, thereby providing insights into the interface of domestic and wild ruminants.
Secondly, the burdens of A. marginale and A. centrale in blood samples from African buffalo were determined, using an established quantitative PCR, and then various statistical models were run to investigate associations between Anaplasma burden, co-infection dynamics and health outcomes for African buffalo. There appeared to be a time-lag between infection and host response, or co-infection response, for some of the parameters examined, showing the
i importance of considering such delays in studies of disease. Despite finding a positive association between the concurrent burdens of infection with the two Anaplasma species examined, once the time-lag was accounted for, there was a negative association between the species, possibly indicating resource competition or the development of cross-immunity.
African buffalo did not have an anaemic response to infection with either A. marginale or A. centrale; in contrast, for animals infected with A. marginale, there was an increase in haematocrit levels in response to infection. On the other hand, there were higher serum total protein levels associated with increased burdens of A. centrale, which may be due to the development of an immune response. Host responses to infection were also affected by external factors, including season and resource availability, and host factors such as gender. Younger buffalo appeared to be infected with higher burdens of A. marginale and A. centrale. In calves, infection with A. marginale appeared to occur before infection with A. centrale, and more frequently, which could be the result of different invasion and evasion techniques of the two species in this host.
This study provides an insight into the effects of a subclinical infection on a wildlife host, caused by a pathogen which may cause severe clinical disease in domesticated animals.
ii
Declaration
This is to certify that:
i) the thesis comprises only my original work towards the Master of Philosophy
(Veterinary Science), except where indicated in the preface;
ii) due acknowledgement has been made in the text to all other material used; and
iii) the thesis is fewer than the 40,000 words in length, exclusive of tables, maps,
bibliographies and appendices
…………………..
Danielle Sisson BSc(Hons)/BA, GradCertJour
iii
Preface
All work presented in this thesis was completed during the MPhil candidature and was completed solely by myself, Danielle Rae Sisson, excluding the following:
In chapters 2 and 3, Dr Anna Jolles and Vanessa Ezenwa are the principal investigators on the
NSF EEID (#DEB-1102493/EF-0723928, EF-0723918) grant which financially supported the field captures for longitudinal buffalo sampling from a free-ranging herd from 2008-2012. In chapters 2 and 3, Anna Jolles was the principal investigator on the USDA-NIFA-AFRI grant
(#2013-67015-21291), and Bryan Charleston the principal investigator of the UK
Biotechnology and Biological Sciences Research Council grant (#BB/L011085/1) that funded the managed herd project as part of the joint USDA-NSF-BBSRC Ecology and Evolution of
Infectious Diseases program. Anna Jolles’ USDA-NIFA-AFRI grant was also used to fund the laboratory work for this thesis.
In chapter 3, Brian Henrichs completed Reverse-line Blot Hybridisation for Anaplasma marginale and A. centrale from African buffalo serum from the free-ranging herd to determine presence of infection for his Masters dissertation. These results were used in analyses for this thesis, and to determine samples to be used in further laboratory work.
Anna Jolles, Brianna Beechler, Jasmin Hufschmid and Abdul Jabbar provided help with the conceptualisation of each chapter. All laboratory work was completed by Danielle Sisson.
Chapter 2 is a published paper (Sisson, D., Hufschmid, J., Jolles, A., Beechler, B. and Jabbar,
A. (2017). "Molecular characterisation of Anaplasma species from African buffalo (Syncerus
iv caffer) in Kruger National Park, South Africa." Ticks and Tick-borne Diseases 8(3): 400-406).
Phylogenetic analyses for this paper was completed by Danielle Sisson under the guidance of
Abdul Jabbar. The paper was written by Danielle Sisson with the assistance of Abdul Jabbar, and Jasmin Hufschmid, Anna Jolles and Brianna Beechler assisted in editing the manuscript.
Brianna Beechler and Mark Stevenson assisted with the statistical analyses, with all analyses being completed by Danielle Sisson for Chapter 3.
All thesis chapters were written by Danielle Sisson. Brianna Beechler, Jasmin Hufschmid and
Abdul Jabbar provided help in editing of this thesis.
Brian Henrichs, Caroline Glidden, Katherine Potgieter-Forssman, Henri Combrink, Hannah
Tavalire, Claire Couch, Brian Dugovich, Robert Spaan, Johannie Spaan, Julie Rushmore,
Courtney Coon, Erin Gorsich, Morgan Movius, Becca Sullivan, Emma Devereux, George
Meleleu, Abby Sage, Daniel Trovillion, Juliana Masseleux and Danielle Sisson were part of the buffalo research team, working on buffalo captures and sample processing which provided data on explanatory and predictive variables used in analyses. South African National Parks
(SANParks) granted permission to conduct this study in Kruger National Park, and Markus
Hofmeyr, Peter Buss and the entire SANParks Veterinary Wildlife Services Department assisted with animal captures and project logistics, as well as Kruger National Park DAFF veterinarians, including Lin Mari De-Klerk Lorist and Louis van Schalkwyk. Collaborators from the Pirbright Institute, including Eva Perez, Bryan Charleston and Fuquan Zhang were also involved in data and sample collection from buffalo captures.
v
Danielle Sisson was supported by an Australian Government Research Training Program
Scholarship.
vi
Acknowledgements
Firstly, thank you to my supervisors, Jasmin Hufschmid and Abdul Jabbar. I appreciate the time you made to go over and over and over the methods, and then over and over writing it up.
I have learnt so much working on this project in the past couple of years. Thanks to the Uni of
Melbourne for funding a trip back to South Africa and the U.S. as well, to collaborate with the other researchers working with some of this data. I do not think I could have pulled any of it together if I wasn’t able to sit in a room and pick the brains of people who designed the overall project and organised all the tests we ran on samples from the buffalo.
A huge thank you to the Jolles lab from Oregon State University, and in particular, Anna Jolles and Brianna Beechler. Thank you, Anna, for letting me a part of this project. Though we haven’t had too many cross-overs in person since I started invading your lab group, whenever
I have gotten to meet up with you, your advice and assistance has had an enormously greater impact than the time it took to impart it upon me. Bree, you’ve been an absolute champ throughout my degree, and you have more patience with annoying little wannabe researchers like me than I have or will have with anyone (which is why I’m still leaning towards the hermit life). Thank you for answering my constant flood of questions, helping me with setting up my database, cracking on with the stats, reading over drafts and giving me reliable and accurate feedback about my abilities and my work. I would have been much more lost than my usual base level of disorientation if it was not for your input.
Even though none of them will probably ever come across this paragraph, I’m extremely grateful to my hockey and Krav Maga clubs with which I was involved throughout my time in
Melbourne for keeping me fit, sane and socialised. Having to run out with you MUHC ladies
vii twice a week in freezing (sometimes wet) Melbourne winter nights, and then spend a fair chunk of each Saturday playing and watching hockey, helped me get through months and months of lab work that could have been finished in a couple of weeks if it was working, but, as goes with most people’s research, it mostly wasn’t…
Thank you to my friends (both in Aus and in the second home of my heart, South Africa) and family for keeping me happy throughout this Masters. Mum, Dad, Jay, Jem, Britt and my relatively fresh niece, Ella, thanks for having my back and checking in on me. Ella, you’ve been pretty useless at that actually, but I’ll forgive you for it because you’ve just turned one, and you’re an extremely entertaining time-absorber. Thanks especially to Dad, for letting me stay at home and be extremely selfish, focusing entirely on getting this done, as you cooked dinner, refilled my cup with spirits or tea as required, and put hot water bottles in my bed while
I carried on late at night staring deep into the abyss of my laptop. And thanks to all the dogs, especially ours, even though I’m pretty sure none of you know how to read. Your acceptance of my hugs and attention, despite recent research indicating that you do not particularly like this behaviour from humans, has somewhat softened my cynical heart. At least to your species.
You can live with me when I fulfil my hermit destiny.
viii
Table of Contents
Abstract I
Declaration iii
Preface iv
Acknowledgements vii
Table of Contents ix
List of tables xii
List of figures xvi
Chapter 1 – Literature Review 1
1. Introduction 1
1.1 Subclinical infections in wildlife 2
1.2 Health implications of subclinical infections in wildlife 4
1.3 Factors affecting the dynamics of subclinical disease 5
1.3.1 Environmental factors 5
1.3.2 Pathogen-related factors 6
1.3.3 Host-related factors 7
1.4 Anaplasma 10
1.5 Anaplasmosis in wildlife 13
1.6 African buffalo 14
1.7 Aims and structure of this thesis 17
Chapter 2 – Molecular characterisation of Anaplasma species from African buffalo
(Syncerus caffer) in Kruger National Park, South Africa 20
2.1 Introduction 20
ix
2.2 Materials and methods 23
2.2.1. Study site, animal characteristics and blood collection 23
2.2.2. Molecular characterisation 25
2.2.3. Phylogenetic analyses 26
2.3 Results 28
2.4 Discussion 34
Chapter 3 – Fitness and Health Consequences of Anaplasma marginale and A. centrale infections in African buffalo (Syncerus caffer) from Kruger National Park,
South Africa 38
3.1 Introduction 38
3.2 Materials and methods 41
3.2.1 Study site and animal characteristics 41
3.2.2 Sampling and measurements 43
3.2.3 DNA extraction and conventional PCR 44
3.2.4 Quantitative PCR 44
3.2.5 Statistical analyses 46
3.3 Results 50
3.3.1 Order of infection in calves from the managed herd 50
3.3.2 Effects of host and environmental factors on anaplasmosis 51
3.3.3 Fitness and health responses to infection 53
Body condition 53
Haematocrit 57
Total protein 60
Reproduction 64
x
3.3.4 Survival 65
3.4 Discussion 66
3.4.1 Factors affecting Anaplasma burden 66
3.4.2 Effects of Anaplasma infection on health parameters 71
Chapter 4 – General Discussion 77
4.1 Livestock: wildlife interface 77
4.2 Order of infection 78
4.3 Co-infection dynamics 79
4.4 Effects of age on infections 80
4.5 Seasonal variance 82
4.6 Fecundity and survival 83
4.7 Tolerance versus resistance host strategies 84
4.8 Future work 86
4.9 Conclusion 86
References 88
Supplementary Material 106
xi
List of Tables
Table 1.1 The cells infected, hosts and pathogenicity of different known
Anaplasma species. 11
Table 2.1 Pairwise differences (%) among the different partial major surface
protein 1β sequences of Anaplasma marginale from African
buffalo. 30
Table 2.2 Pairwise differences (%) among the different partial heat shock
protein (groEL) sequences of Anaplasma centrale from African
buffalo. 31
Table 3.1 Number of times individual African buffalo were sampled
throughout two longitudinal disease studies from Kruger National
Park. 43
Table 3.2 Mean infection intensities (copies/reaction) of Anaplasma
marginale and A. centrale in managed and free-ranging herds of
African buffalo from Kruger National Park. 50
Table 3.3 Generalised linear mixed (concurrent) model for the burden (log-
transformed) of Anaplasma marginale and/or A. centrale
(copies/reaction) in managed and free-ranging herds of African
buffalo from Kruger National Park. 51
Table 3.4 Generalised linear mixed (predictive) model for the burden of (log-
transformed) Anaplasma marginale and/or A. centrale
(copies/reaction) in managed and free-ranging herds of African
buffalo from Kruger National Park. 52
xii
Table 3.5 Generalised linear mixed (concurrent) model for body condition in
managed and free-ranging herds of African buffalo from Kruger
National Park. 53
Table 3.6 Generalised linear mixed (predictive) model for body condition in
managed and free-ranging herds of African buffalo from Kruger
National Park. 57
Table 3.7 Generalised linear mixed (predictive) model for haematocrit levels
in managed and free-ranging herds of African buffalo from Kruger
National Park. 58
Table 3.8 Generalised linear mixed (concurrent) model predicting serum total
protein concentrations in managed and free-ranging herds of
African buffalo from Kruger National Park. 60
Table 3.9 Generalised linear mixed (predictive) model for serum total protein
concentration in managed and free-ranging herds of African buffalo
from Kruger National Park. 61
Table 3.10 Mixed effects logistic regression model predicting pregnancy status
in managed and free-ranging herds of African buffalo from Kruger
National Park. 64
Table 3.11 Andersen-Gill survival model predicting survival outcome in
managed and free-ranging herds of African buffalo from Kruger
National Park. 65
xiii
Supplementary Tables
Supplementary Comparison of Anaplasma marginale and A. centrale infection
Table 3.1 detection via conventional PCR and reverses line blot hybridisation
(determining presence/absence of infection) with quantitative PCR
(qPCR) results (determining infection burden). 108
Supplementary Mixed effects logistic regression models predicting
Table 3.2 presence/absence of Anaplasma marginale and/or A. centrale in
managed and free-ranging herds of African buffalo from Kruger
National Park. 109
Supplementary Mixed effects logistic regression models predicting the previous
Table 3.3 capture’s presence/absence of Anaplasma marginale and/or A.
centrale in managed and free-ranging herds of African buffalo from
Kruger National Park. 110
Supplementary Generalised linear mixed model predicting body condition in
Table 3.4 managed and free-ranging herds of African buffalo from Kruger
National Park. 111
Supplementary Generalised linear mixed model predicting haematocrit volume in
Table 3.5 managed and free-ranging herds of African buffalo from Kruger
National Park. 112
Supplementary Generalised linear mixed model predicting total protein in managed
Table 3.6 and free-ranging herds of African buffalo from Kruger National
Park. 113
Supplementary Model selection for models used in this paper. 114
Table 3.7
xiv
Supplementary Generalised linear mixed model predicting haematocrit volume in
Table 3.8 managed and free-ranging herds of African buffalo from Kruger
National Park. 118
Supplementary Mixed effects and logistic regression model predicting pregnancy
Table 3.9 status in a free-ranging herd of African buffalo from Kruger
National Park. 119
xv
List of Figures
Figure 2.1 Map showing the study site in Kruger National Park in relation to
the rest of South Africa and the African continent. 24
Figure 2.2 Genetic relationships of Anaplasma marginale (A) and A. centrale
(B) detected from African buffalo (Syncerus caffer) from Kruger
National Park, South Africa (bold) with reference genotypes
selected from previous studies. 33
Figure 3.1 Mean body condition score with burden (log-transformed) of
Anaplasma centrale based on season (dry or wet), in a free-ranging
herd (n = 268) of African buffalo from Kruger National Park. 55
Figure 3.2 Mean body condition score with burden (log-transformed) of
Anaplasma marginale based on season (wet or dry) in free-ranging
African buffalo (n = 268) from Kruger National Park. 56
Figure 3.3 Haematocrit level with the previous capture’s infection intensity
(log-transformed) of Anaplasma marginale based on sex in a
managed herd of African buffalo (n = 226) from Kruger National
Park. 59
Figure 3.4 Total serum protein infection intensity (log-transformed) of
Anaplasma marginale (copies/reaction) and A. centrale in a
managed herd of African buffalo (n = 99) from Kruger National
Park. 62
xvi
Figure 3.5 Total serum protein and infection intensity (log-transformed) of
Anaplasma centrale (copies/reaction) based on sex in a managed
herd of African buffalo (n = 226) from Kruger National Park. 63
Supplementary Figures
Supplementary Alignment of partial major surface protein 1β sequences of
Figure 2.1 Anaplasma marginale determined herein. 106
Supplementary Alignment of partial heat shock protein (groEL) sequences of
Figure 2.2 Anaplasma centrale determined herein. 107
xvii
Chapter 1 - Literature Review
1. Introduction
Difficulties in studying wildlife diseases mean that they remain understudied in comparison to diseases in domesticated and production animals. While the focus of studies on diseases of domesticated animals is understandable, due to the relative ease in sampling and their economic importance, there are important zoonoses and livestock diseases that can be better managed by a more thorough understanding of diseases in wildlife (Ryser-Degiorgis, 2013). For instance, research conducted on gastrointestinal helminths and bovine tuberculosis (BTB) in African buffalo (Syncerus caffer) showed that anthelminthic treatment of co-infected buffalo resulted in longer host survival times and increased transmission opportunities of BTB to susceptible hosts, including livestock (Ezenwa et al., 2010). Infection dynamics elucidated in wildlife can thus be applied to the management of related infections in domesticated animals. An imbalanced depth of understanding, on the other hand, can cause oversights of interconnected disease persistence and transmission in people, their domesticated animals and wildlife.
In wildlife, subclinical infections receive even less attention than those resulting in overt clinical signs (Gunn & Irvine, 2003). Subclinical effects may be more obvious in domestic livestock due to selective pressures for high performance and productivity, thereby, offering more opportunities to observe these signs at a closer proximity (Gunn & Irvine, 2003). There are many infections and diseases that can occur in both domestic animals and wildlife; however, the impacts of these diseases are generally much better understood in the former than the latter (Ryser-Degiorgis, 2013).
1
The main aim of this review is to present an overview of the challenges around detecting and correctly interpreting subclinical infections in wildlife, the health implications of those subclinical infections and the factors associated with the dynamics of subclinical disease
(including environmental, pathogen- and host-related factors), before concentrating on
Anaplasma infections, focusing on wildlife, as well as describing the African buffalo host.
1.1 Subclinical infections in wildlife
Most disease outbreaks in wildlife, including those caused by overtly pathogenic infections, are detected because of a mass mortality event (Wobeser, 2006). A number of factors may be blamed for this, including a lack of targeted surveillance and lack of properly validated diagnostic techniques (Wobeser, 2006). For example, most reported cases of Babesia spp. in captive reindeer and caribou (Rangifer tarandus) are from fatal infections, while the prevalence of subclinical infection in live animals, although known to be present, is unclear (Bartlett et al.,
2009). It is therefore unsurprising that our understanding of the prevalence, dynamics and significance of subclinical disease in wildlife is even poorer than that of clinically manifesting disease.
One of the first obstacles to overcome may be to separate truly subclinical diseases from those that only appear to be subclinical, due to difficulties in detecting clinical signs in the species involved. For example, an infection may weaken the host’s immune system without producing obvious clinical signs, but nevertheless increase the likelihood of predation or the occurrence of secondary infections which might prove to be fatal (Davidson et al., 2015; Friend & Trainer,
1969; Guscetti et al., 2003; Ryser-Degiorgis, 2013). In such cases, the primary cause of morbidity and mortality may be misdiagnosed. In other cases, clinical signs may be present, but will only be observed on very close inspection of individual animals. A study on Sarcoptes
2 scabei infection in Norwegian red foxes (Vulpes vulpes) highlighted the difficulties in identifying subclinical infection, with confirmatory diagnosis requiring clearly visible lesions, resulting in an underestimation of diseased animals overall (Davidson et al., 2008).
The lack of clinical signs is a further, major, obstacle to the detection of subclinical infections in wildlife, and means that highly sensitive and specific diagnostic methods are needed to detect and correctly identify the pathogens involved. The diagnostic methods for diseases of domestic animals are usually well developed; however, very few diagnostic tests have been validated for wildlife diseases (Ryser-Degiorgis, 2013). This was also a conclusion from a study investigating chlamydiosis in koalas, stating that there was a lack of sensitive, simple and reasonably priced diagnostic tools (Griffith & Higgins, 2012). In addition, diagnosis of wildlife diseases is often complicated by a lack of suitable diagnostic samples; this is because the behaviour of wildlife species frequently complicates sample and field data collection (e.g. need for sedation), as can the transportation of either animals or samples to the designated laboratory, especially from remote study sites (Ryser-Degiorgis, 2013). The cyclical nature of some infectious agents can also lead to misdiagnosis, or lack of diagnosis, if the pathogen is in a trough of its cycle during the time of sampling (Kuiken & Danesik, 1999). This is the case in
Anaplasma spp. infections in cattle, which can have cyclic rickettsaemia leading to a persistent subclinical infection after the initial acute infection, but with varying infection intensities over time (Aubry & Geale, 2011).
Some wildlife species may indeed be infected with certain pathogens without experiencing any negative effects, e.g. Hendravirus in flying foxes (Pteropus spp.) (Williamson et al., 1998).
Such host-pathogen relationships may produce very competent reservoir hosts. Reservoir hosts of infectious agents typically show few obvious adverse effects to infection, but they can be
3 the most important part of the disease cycle to manage. For instance, larval amphibians, fish and aquatic turtles have lower susceptibility to a potentially lethal Ranavirus spp., but they allow the multiplication and persistence of viruses to more susceptible amphibians (e.g. mosquito fish (Gambusa affinis), red-eared slider turtle (Trachemys scripta elegans) and
Cope’s grey treefrog (Hyla chryscocelis) (Brenes et al., 2014). In another example, white-tailed deer (Odocoileus virginianus) are suggested to be reservoir hosts of Mycoplasma ovis which can cause severe disease in caprines (Boes et al., 2012).
If our ability to detect a disease thus relies on the observation of ‘typical’ clinical signs of a disease, this may well result in diagnosis occurring only when there is already a problem, with potentially no ability to prevent the spread of an infection (Griffith & Higgins, 2012). It will also restrict our ability to understand the true cost of subclinical infection to a wildlife host.
1.2 Health implications of subclinical infections in wildlife
Few studies have been conducted to elucidate the health effects of subclinical infection in wildlife, but based on studies of domestic livestock (e.g. Hage et al., 1998), it seems likely that some effects, potentially affecting less immediately obvious measures (such as reproductive success) do exist. Indeed, some studies have successfully demonstrated health impacts of subclinical infections in wildlife, resulting in decreased reproductive success and behavioural changes (Gunn & Irvine, 2003; McAllister, 2005). For instance, it has been shown that captive emperor tamarins (Saguinus imperator) have high rates of stillbirth and neonatal losses plausibly associated with a chronic subclinical infection with Encephalitozoon cuniculi
(Guscetti et al., 2003). A review on subclinical parasitism also indicated that some host behaviours evolve in response to negative effects on reproductive success, such as a change in forage strategy to minimise parasite exposure, as seen with post-calving migrations in reindeer
4
(Rangifer tarandus) to avoid infestations with the warble fly (Hypoderma tarandi) (Gunn &
Irvine, 2003). This parasite, in the absence of clinical signs, has been shown to be correlated with a decrease in body weight, as well as a reduced conception rate (Hughes et al., 2009). It has also been suggested that subclinical infections with Toxoplasma gondii in rodents may cause adverse behavioural changes, with infected rodents displaying lower levels of caution to new stimuli and the presence of cats, than uninfected rodents (McAllister, 2005). Nonetheless, our understanding of the impacts of subclinical disease in wildlife is still very limited.
Based on our current knowledge, it is thus reasonable to assume that subclinical infections may have significant impacts on their host, but much work is needed to elucidate these. To better understand the circumstances under which subclinical infection may affect host health and/or fitness, it is important to consider the factors which affect the effects and dynamics of all diseases: pathogen, host and environment (Wobeser, 2006).
1.3 Factors affecting the dynamics of subclinical disease
1.3.1 Environmental factors
Environmental factors, such as seasonality and geographic or topographic location, can impact on infectious agents and their vectors. For example, in Zambia, the prevalence of bovine tuberculosis in African buffalo changes throughout the year due to (i) yearly flooding that contributes to the propagation and dissemination of microorganisms, (ii) increased contact between individual animals during the mating season and (iii) at water points in the dry season when water sources are limited (De Garine-Wichatitsky et al., 2013). Further work on African buffalo in sub-Saharan Africa found that this clustering of buffalo at water points in the dry season can not only facilitate direct transmission of microorganisms, but also increases the transmission of a number of tick-transmitted infections (Eygelaar et al., 2015).
5
1.3.2 Pathogen-related factors
Pathogens can interact directly through the host immune system, or indirectly through resource competition (Jackson et al., 2006). The way different parasites interact changes depending on the composition of parasites and the order in which they infect the host (Jackson et al., 2006;
Kocan, 1995). Traditionally, studies have looked mainly at monospecific infections in wildlife
(Tompkins et al., 2011), but we now increasingly acknowledge that co-infection with two or more pathogens can further complicate already complex disease dynamics. Co-infections can have varying effects on host health, and so studies based on the interactions of multiple diseases are necessary to accurately understand the consequences of co-infection and develop effective control measures (Cattadori et al., 2008). There are finite economic resources available to livestock farmers or wildlife managers, increasing the importance of reliable data indicating which diseases or infections are the most detrimental to animal health and economic returns.
Parasites can interact directly through the host immune system, or indirectly through resource competition (Jackson et al., 2006; Lafferty, 2010). The way different parasites interact depends on the composition of parasites and the order in which they infect the host (Jackson et al., 2006;
Kocan, 1995). Interactions between different types of parasites (e.g. between endo- and ectoparasites, or between extra- and intracellular parasites) can activate mutually inhibitory immune responses (Th1 and Th2) (Craig et al., 2008). Alternatively, facilitation can occur between parasites, with unrelated parasites stretching the immune system responses, thereby making it easier for other parasites to invade the host, potentially resulting in more intense competition (Lafferty, 2010).
Parasite diversity increases with the diversity in available host species, but a diverse suite of parasites within an individual host does not necessarily increase the host’s disease risk as many
6 parasite species cause relatively little harm to their hosts (Johnson et al., 2013). The presence or absence of different parasites may affect the likelihood of co-occurrence with another parasite species, again emphasising the importance of considering multiple infections in a study population (Forbes et al., 1994). As well as parasites and pathogens from different guilds, closely related pathogens should be considered when investigating impacts on host health. For instance, infection with either Anaplasma marginale or A. centrale does not exclude infection with the other in cattle, even though both parasite species invade the same cells; however, they appear to have opposing impacts on the host (Shkap et al., 2008). How these closely related and co-occurring pathogens utilise host resources without negatively affecting each other is yet to be fully understood.
Another pathogen-related factor to consider is the ability for a pathogen to survive outside the host, as some pathogens can persist for extended periods of time, and/or be transmitted indirectly through fomites in the environment (Brenes et al., 2014). The hardiness of the pathogen itself may vary significantly but can make some diseases extremely difficult to eradicate. For example, anthrax (Bacillus anthracis) may persist almost indefinitely in the environment depending upon environmental conditions (Jensen et al., 2003). Similarly, the bacteria that can cause different types of plague (Yersinia spp.) can survive for prolonged periods in the soil, and then have an array of host species they can infect (Butler, 2013). The detection of these pathogens may thus require extensive environmental sampling in addition to animal surveillance.
1.3.3 Host-related factors
Host-related factors may determine an animal’s response to initial infection, including whether subclinical diseases will become clinical (Bartlett et al., 2009). In most systems, species, sex,
7 reproductive stage and age of the host have been found to be associated with variations in prevalence and intensity of infection with different parasites (Forbes et al., 1994). Other important host factors, such as stress, general health and previous infections, may further affect the response of an animal to infection (Bartlett et al., 2009).
Host species is an important determinant of susceptibility to disease, and some subclinical infections or mild diseases in a natural host species can become fatal when they are transmitted to a new, immunologically naïve, species (Bicknese et al., 2010). Herpesviruses, for example, usually cause subclinical infection or mild disease in their natural host species, but when transmitted to a new, immunologically naïve species, infection can be fatal (Bicknese et al.,
2010). This is also apparent with Mycoplasma ovis, a parasite of red blood cells in sheep and goats (Boes et al., 2012). Mycoplasma ovis does not seem to cause overt illness when transmitted among goats (Capra spp.), which are considered natural hosts, but when transmitted to sheep (Ovis aries) it causes acute haemotropic mycoplasmosis or chronic ill- thrift (Boes et al., 2012). In sub-Saharan Africa, canine babesiois (caused by Babesia rossi) can be fatal in domestic dogs, but black-backed jackals (Canis mesomelas) appear to only be subclinically infected and are considered the natural hosts of the piroplasm (Penzorn et al.,
2017).
Host sex and physiological status can also affect susceptibility to an infection. For example,
Davidson et al. (2015) reported that female moose (Alces alces) had higher mean abomasal nematode counts than males, and lactating females had higher counts than non-lactating females; in addition, higher parasite loads were found in pregnant than non-pregnant females.
These differences in parasite burden were thought to be the result of immunosuppression, caused by the high energetic costs of pregnancy and lactation, as well as costs associated with
8 hormonal production (e.g. oestrogen and progesterone) during different physiological stages
(Davidson et al., 2015). Conversely, higher parasite loads in males are often explained through the hypothesis of immunosuppression by testosterone (Foo et al., 2017).
Age is another important host factor that can affect the epidemiology of a disease. For example, in the same study group of moose (Davidson et al., 2015), younger animals were found to be more susceptible to parasitic infections than older animals. In contrast, calves infected with
Babesia spp. when they are less than nine months old, display signs of innate resistance to the babesiosis that affects older cattle (Zintl et al., 2005).
Illness can be more severe when exposure to a new pathogen occurs in an already stressed animal, as evident in Babesia spp. infection in captive reindeer (Bartlett et al., 2009). During translocations of these reindeer, the stressors associated with physical relocation were thought to have contributed to the appearance of clinical signs in previously subclinically-infected animals (Bartlett et al., 2009). Food availability is another potential stressor which can affect the shedding of pathogens, with malnourished animals likely to become immunosuppressed, resulting in higher mortalities as well as increased spread of the infectious agent (Bradley &
Altizer, 2007).
Finally, previous exposure to infectious agents may affect the susceptibility of hosts through the development of acquired immunity. For example, the susceptibility of two cetacean populations to morbillivirus was affected by previous exposure to this virus (Bossart et al.,
2010; Härkönen et al., 2006). Two disease outbreaks associated with morbilliviruses occurred in two populations of cetacean approximately five years apart. Following the second outbreak, it appeared that the immunologically naïve animals were most negatively affected by the
9 disease and that those that had survived the first outbreak were more likely to survive than naïve animals.
A large range of factors must thus be thoroughly explored to elucidate the true impacts and dynamics of subclinical infection in any population of wild animals. While prior infection with the same pathogen may result in priming of the adaptive immune system for subsequent re- infections, prior or current infection with other pathogens (co-infection), may result in a complex set of interactions with variable outcomes for host fitness and pathogen dynamics
(Ezenwa, 2016).
1.4 Anaplasma
Tick-borne diseases such as anaplasmosis, heartwater (caused by Ehrlichia sp.), theileriosis and babesiosis, are known to cause 18% of reported cattle mortalities in South Africa (De Waal,
2000). Cattle are also affected by Theileria parva, which causes East Coast fever, Corridor disease and January disease, all of which can be fatal (Debeila, 2012); Babesia bovis and B. bigemina, resulting in losses in meat and milk production, and in severe cases, death (Debeila,
2012); and heartwater, or cowdriosis, which has high mortality rates in a range of livestock species (Debeila, 2012). Anaplasmosis is estimated to be responsible for 3% of all cattle mortalities in South Africa (De Waal, 2000), and is considered one of the most important tick- borne diseases in sub-Saharan Africa, where it may affect both domesticated and wild ruminants (Eygelaar et al., 2015). In Saint Lucia, the prevalence of A. marginale infection in native cattle is 80%, which is thought to be representative of endemic regions where clinical disease is uncommon (Knowles et al., 1982). Of the total cattle population in South Africa,
99% are predicted to be at risk of developing anaplasmosis, as most cattle farming occurs in endemic and epidemic areas (De Waal, 2000).
10
While Anaplasma spp. may be transmitted mechanically through blood-contaminated fomites or biting flies, and transplacental transmission has been reported, the main mode of transmission appears to be through ticks (Potgieter & Stoltz 2004). Different species of
Anaplasma affect different cells in different animals, with the outcome ranging from mild illness to death (Table 1).
Table 1.1: The cells infected, hosts and pathogenicity of different known Anaplasma species
Anaplasma sp. Cells infected Host(s) Pathogenicity Reference(s)
A. ovis Red blood cells Small Mildly (Dumler, et al., ruminants pathogenic 2001; Visser, et al., 1991) A. bovis Monocytes Ruminants Pathogenic (Dumler, et al., 2001; Said, et al., 2015) A. phagocytophilum Neutrophil Ruminants, Pathogenic (Corona & granulocytes dogs, horses, Martínez, 2009; humans Eberts, et al., 2011; Rikihisa, 2010; Woldehiwet, 2010 A. platys Platelets Dogs Mildly Yabsley, et al., pathogenic 2008; Zobba, et al., 2014) A. buffeli Unknown Water buffalo Unknown (Kuttler, 1984)
A. mesaeterum Red blood cells Sheep Mildly (Kuttler, 1984; Lew, pathogenic et al., 2003)
A. marginale Red blood cells Wild ruminants, Pathogenic (Aubry & Geale, cattle 2011; Potgieter & Stoltz, 2004) A. centrale Red blood cells Wild ruminants, Mildly (Aubry & Geale, cattle pathogenic 2011; Potgieter & Stoltz, 2004) Anaplasma sp. Red blood cells Wild ruminants, Unknown (Debeila, 2012; Omatjenne cattle Henrichs, 2014)
11
Anaplasma spp. infection in livestock has been well documented, with A. marginale being the main causative agent of anaplasmosis, resulting in deaths of up to half of infected cattle over the age of two (Aubry & Geale, 2011; Kocan & de la Fuente, 2003). Anaplasma marginale was first identified in cattle from South Africa in the early 1900’s, and was named due to the marginal location of the organism in the host’s erythrocytes (Potgieter & Stoltz, 2004). Clinical signs in cattle include pale mucous membranes, icterus, anaemia and decreased body condition
(Jaswal et al., 2014; Potgieter & Stoltz, 2004). More severe cases can result in temporary infertility, anoestrus and abortion, and even death (Jaswal et al., 2014; Potgieter & Stoltz,
2004). A live vaccine developed from the less pathogenic A. centrale, named due to its central location within erythrocytes, has been used to protect against the severity of A. marginale infection in countries like Australia, Israel, and South Africa (Dalgliesh et al., 1990; Debeila,
2012; Kuttler, 1984). Besides vaccination, immunity to anaplasmosis can be reinforced naturally if hosts are repeatedly infected, starting when young (Dalgliesh et al., 1990).
In cattle, A. marginale invades erythrocytes, goes through numerous cycles of replication, is eliminated with non-infected erythrocytes by the reticuloendothelial system of the host, and invades erythrocytes again (Aubry & Geale, 2011). The average lifespan of a bovine red blood cell is 160 days, meaning continual re-invasion of cells is required to maintain a constant infection, which can occur within the host itself if some infected erythrocytes remain in the body (Aubry & Geale, 2011). The initial incubation period lasts from 7-60 days (de la Fuente et al., 2003; Kocan, 1995; Kocan, & de la Fuente, 2003). While there are many studies documenting the effects of Anaplasma spp. infection in domesticated animals, there is limited information on the effects of these infections in wildlife.
12
1.5 Anaplasmosis in wildlife
Infection with Anaplasma spp. in wildlife lacks the severity of the disease seen in livestock, with A. marginale infection rarely resulting in clinical signs (Berggoetz et al., 2014; Kuttler,
1984; Ngeranwa et al., 2008). Clinical anaplasmosis has only been described in giraffes
(Giraffa camelopardalis) (Kuttler, 1984; Ngeranwa et al., 2008). Three species of Anaplasma, all of which occur in cattle, have been identified in free-ranging African buffalo (Syncerus caffer) in South Africa: A. marginale, A. centrale and Anaplasma sp. Omatjenne (Debeila,
2012; Henrichs, 2014). While it is well known that Anaplasma spp. infection in cattle may result in severe clinical signs, the limited information available for African buffalo indicates a minimal response by this host to the same infections (Debeila, 2012; Henrichs, 2014). From the African continent, subclinical infections of Anaplasma (predominantly with A. marginale) have been described not only in the African buffalo, but also eland (Taurotragus oryx), black wildebeest (Connochaetes gnou), blue wildebeest (Connochaetes taurinus), grey duiker
(Sylvicapra gimmia), blesbok (Damaliscus dorcas phillipsi) and giant African rat (Cricetomys gambianus). However, subclinical infections of Anaplasma spp. in wildlife are found worldwide (Kuttler, 1984; Ngeranwa et al., 2008). Although wildlife species do not appear to be negatively affected by infection with Anaplasma spp., they are commonly implicated as potential reservoir hosts for the infections, potentially spreading the disease at the wildlife: livestock interface (Kuttler, 1984; Ngeranwa et al., 2008).
A study on Anaplasma spp. infection in wildlife would enhance our knowledge on how a subclinical infection affects a wildlife species, such as the African buffalo. In addition, it would improve our understanding of the potential of infection in free-ranging wildlife to influence epidemic and endemic infections in nearby populations of susceptible domesticated animals
(Kock, 2005). African buffalo have been implicated as reservoir hosts for a number of
13 important infectious diseases, leading to claims of buffalo threatening the health of livestock
(Eygelaar et al., 2015; Jolles et al., 2008; Kock, 2005). However, in reality, very little is known about the dynamics of infection with Anaplasma spp. in wild bovids such as African buffalo.
1.6 African buffalo
Wild ungulates have been shown to be useful species for the research in disease ecology, because their relatively long lifespan allows a time aspect to be incorporated into disease studies, and their interaction with domestic species can be used to estimate impacts on a broad range of host species (Jolles & Ezenwa, 2015). African buffalo are long-lived gregarious grazers, so their social system and lifespan allows for many opportunities for various diseases and infectious agents to be transmitted (Prins, 1996). Buffalo are thought to be reservoir hosts for some significant livestock diseases, including foot-and-mouth disease, Corridor disease (a form of theileriosis), babesiosis and heartwater (or cowdriosis) (Debeila, 2012; Henrichs et al.,
2016). Additionally, they are also hosts to many other infectious agents, including gastrointestinal helminths, protozoa, ticks, mites, and a range of viruses and bacteria (Debeila,
2012; Henrichs et al., 2016; Prins, 1996). There is evidence for increased mortality in African buffalo co-infected with both bovine tuberculosis (BTB) and gastrointestinal helminths (Jolles et al., 2008), but they appear to be asymptomatic carriers of many other infectious agents, including Anaplasma spp. (Aubry & Geale, 2011; Kocan et al., 2010; Potgieter & Stoltz, 2004).
In buffalo, ticks are the main vector for the transmission of Anaplasma spp. (Dumler et al.,
2001). The main tick species in sub-Saharan Africa is Rhipicephalus decoloratus (formally
Boophilus decoloratus) (Horack et al., 2006). Buffalo appear to have some acquired resistance
(rather than innate) to this tick species, probably from an ancient association similar to the resistance seen in Nguni cattle, a South African hybrid breed of Bos taurus and B. indicus
14
(Eygelaar et al., 2015; Horak et al., 2006). Similar results indicating resistance have been found with N’dama cattle in the Gambia, as there is an absence of A. marginale infection even though the vectors capable of transmitting it are present (Kuttler et al., 1988). This may mean that R. decoloratus, while being quite common in the region, may not be the major transmitter of
Anaplasma spp. in African buffalo, though more work is needed to identify the role of ticks in these infections. Other ticks capable of using buffalo as a host and transmitting Anaplasma spp. infections include Amblyomma hebraeum, R. appendiculatus, R. maculates, R. muehlensi, R. evertsi evertsi, R. simus, Haemaphysalis silaceae, H. marginatum and H. truncatum (Gallivan et al., 2011; Horack et al., 2011; Horack etal., 2007; Junker et al., 2015). Male ticks, in particular, are considered important for the spread of anaplasmosis, as their tendency to move between more animals results in greater opportunity to spread disease. They are also thought to be removed less easily than females during host grooming because female ticks engorge to a larger size (Dumler et al., 2001; Horak et al., 2006). There appears to be an upper limit to potential tick burden on most ungulates, as repeated feeding induces an immune response which reduces ticks’ feeding success and survival (Gallivan & Horak, 1997).
Buffalo tend to occur in herds which increase in size in the dry season when water sources are limited and shrink into smaller groups in the wet season when food and water are more abundant (Prins, 1996). These changes in herd density and size may result in seasonal variation of transmission frequency for Anaplasma spp. In addition, seasonal effects on resource availability may affect transmission dynamics through changes in foraging behaviour, which have been shown to affect the occurrence of tick-borne pathogens and parasites in African buffalo (Houston et al., 2007). More intense foraging means movement through more habitat, which increases risk of death by disease due to an increase in the opportunity for the attachment of disease-carrying ticks; this is more likely during periods of resource restriction, such as the
15 dry season (Spaan 2015). If the host is in better condition, and in an environment with higher food availability, less movement is required by the buffalo to feed and this can decrease exposure to ticks (Houston et al., 2007).
The movement of buffalo is controlled and severely restricted by humans in South Africa to avoid the spread of diseases (such as foot-and-mouth) to the cattle industry; however, commercial game farms and private game reserves are on the rise in South Africa, meaning wildlife are being moved into traditional livestock-farming areas (Debeila, 2012). Therefore, there is a need for more research into wildlife diseases that could cause disease outbreaks at these increasing wildlife: livestock interfaces (Debeila, 2012).
Previous work with Anaplasma spp. in African buffalo has focused on how the presence of a variety of parasite species affect each other and host health (Henrichs et al., 2016). Henrichs
(2016) found that the strongest effect upon a parasites’ ability to persist within a host is the proximity to other parasites within that same host. However, parasites in one type of body tissue (e.g. gastrointestinal tract) did not have a significant effect on the ability of parasites in another type of body tissue (e.g. haemoparasites in blood cells). Similar findings were observed in a study in humans, where associations were found between parasites if they clustered in one particular part of the body, suggesting hosts’ bodies may be divided into microhabitat modules
(Griffiths et al., 2014). The likelihood of a new parasite species successfully invading and persisting in a host may thus be determined by pathogens already present in the host (Henrichs,
2014). Surprisingly, previous work suggests that the closely related species A. marginale and
A. centrale do not appear to compete for resources or cause cross-immunity in buffalo (unlike what has been seen in cattle) even though they both replicate by binary fission, invade the same cells and are genetically very similar (Henrichs et al., 2016). These authors also found that
16
African buffalo infected with an Anaplasma spp. and with better body condition were more likely to be infected with A. centrale than A. marginale, indicating that there were some differences in host health depending on which species infected the host.
The co-infection dynamics of Anaplasma spp. in African buffalo are yet to be fully elucidated.
Previous work suggests the presence of interactions between Anaplasma spp. in this host, however, further work using more sensitive measures is needed to fully elucidate these. While infection with Anaplasma spp. does not appear to cause any clinical signs in African buffalo, the true health impacts of subclinical infections in wildlife are difficult to measure, but should not be overlooked.
1.7 Aims and structure of this thesis
This thesis aims to investigate the health effects of a subclinical infection on a wildlife host, focusing on the most common Anaplasma species found in African buffalo. More specifically, our aims were:
Aim 1: to conduct a molecular characterisation of A. marginale and A. centrale from African buffalo at Kruger National Park, including phylogenetic analysis.
Aim 2: To investigate the true health impacts of Anaplasma infection in African buffalo. We predicted that A. marginale would have greater negative health impacts on buffalo in terms of body condition and blood parameters, with a slightly anaemic response to A. marginale.
The first data chapter (Chapter 2) explores the phylogenetic relationships of A. marginale and
A. centrale from African buffalo in South Africa compared to sequences from other geographical areas (Aim 1).
17
The second chapter (Chapter 3) investigates the effects of the burden of infection with
Anaplasma spp. on health variables (body condition, haematocrit levels and total protein levels), as well as determining the effects of season, sex, location and age on infection levels
(Aim 2). Predictive models allow insight into the impact on the buffalo taking into consideration the time-lag between infection and host response, and survival models will indicate if A. marginale or A. centrale infections are correlated with higher rates of mortality than seen in non-infected hosts.
The last chapter (Chapter 4) discusses the results from the previous two chapters within the context of the overarching study question of what are important factors in subclinical infections in wildlife that need to be accounted for in wildlife disease ecology research, and provide suggestions for future work.
18
19
Chapter 2 – Molecular characterisation of Anaplasma species from African
buffalo (Syncerus caffer) in Kruger National Park, South Africa
Published chapter: Sisson, D., Hufschmid, J., Jolles, A., Beechler, B. and Jabbar, A. (2017).
"Molecular characterisation of Anaplasma species from African buffalo (Syncerus caffer) in
Kruger National Park, South Africa." Ticks and Tick-borne Diseases 8(3): 400-406
2.1 Introduction
Anaplasmosis is an important tick-borne disease (TBDs) of domestic and wild animals in tropical and subtropical regions of the world, and is caused by obligate intracellular bacteria of the genus Anaplasma (Rickettsiales: Anaplasmataceae). It is one of the four most detrimental tick-borne diseases (TBDs) of bovines in sub-Saharan Africa (others: babesiosis, cowdriosis and theileriosis) (Debeila, 2012), and is estimated to be responsible for 3% of all cattle mortalities in South Africa (Aubry & Geale, 2011; Brown, 2012; De Waal, 2000; Eygelaar et al., 2015). Amblyomma spp. and Rhipicephalus spp. are the main tick vectors of this disease in
South Africa (Gallivan, et al., 2011; Horak, et al., 2011; Horak, et al., 2007). Bovine anaplasmosis is mainly caused by Anaplasma marginale and, to a lesser extent, by A. centrale.
The disease is usually transmitted by ticks, but it can also be transmitted transplacentally, or mechanically by biting flies or blood-contaminated fomites (Aubry & Geale, 2011; Potgieter
& Stoltz, 2004). Major clinical signs in infected cattle include pyrexia, progressive anaemia, jaundice, anorexia, depression, reduced milk production, abortion in pregnant animals, and death, particularly in exotic breeds (Aubry & Geale, 2011; Debeila, 2012; Kocan, et al., 2010;
Potgieter & Stoltz, 2004).
20
While anaplasmosis is well documented in cattle, very little is known about the disease and its impact in wild bovids such as the African buffalo (Syncerus caffer). Tick-borne infections are common in buffalo, which serve as a reservoir for bovine theileriosis (Theileria parva)
(Debeila, 2012; Henrichs et al., 2016). Buffalo appear to be only mildly affected by Anaplasma infections (Berggoetz et al., 2014; Debeila, 2012; Kuttler, 1984), which raises the question whether they might also function as reservoirs for this group of parasites. To date, A. marginale,
A. centrale, A. sp. Omatjenne, A. bovis and A. phagocytophilum have been detected from
African buffalo using the 16S rRNA gene, though A. phagocytophilum has only been found once in one animal (Fyumagwa, et al., 2013; Henrichs et al., 2016). African buffalo are a useful wildlife species in which to study Anaplasma species dynamics as they are known to be wildlife reservoirs for numerous diseases such as foot-and-mouth disease, bovine tuberculosis and theileriosis, yet their role in maintaining Anaplasma species is unknown (Debeila, 2012).
Identification of wildlife reservoirs is essential to controlling infection in livestock which may have a direct or indirect contact with the reservoir population (Haydon, et al., 2002).
Various methods have been used to diagnose anaplasmosis, including microscopy, serology
(complement fixation, rapid card agglutination, indirect immunofluorescent antibody tests, capillary tube agglutination tests, enzyme-linked immunosorbent assays, latex agglutination and radioimmunassays) and molecular methods (Aubry & Geale, 2011; Potgieter & Stoltz,
2004). For molecular methods utilising polymerase chain reaction (PCR), a number of markers such as the 16S rRNA, major surface protein (msp1α, msp1β, msp2, msp3, msp4 and msp5), citrate synthase gltA and the heat shock protein groEL genes have been used for the detection of Anaplasma spp. (Carelli et al., 2007; Ceci et al., 2008; de la Fuente, et al., 2001; Lew, et al.,
2002; Lew, et al.,, 2003; Molad et al., 2006); whereas, msp1α, msp1β, msp4 and groEL have been used for finer scale differentiation of various Anaplasma species and strains.
21
Very little is known about the diversity of Anaplasma spp. from African buffalo and only one marker, the 16S rRNA gene, has been used so far to characterise them (Debeila, 2012; Henrichs et al., 2016). However, the most commonly used markers for characterisation of these species, such as genes for major surface proteins and heat shock protein groEL, allow for better resolution which might reveal significant genetic differences among Anaplasma spp. found in domestic and wild ruminants, and can be used in quantitative molecular methods. Therefore, the aim of this study was to characterise the two important species of Anaplasma (A. marginale and A. centrale) from African buffalo in South Africa using DNA sequences from the genes coding for a major surface protein and heat shock protein groEL, respectively. This study will allow future research into buffalo as a reservoir host of Anaplasma spp.
22
2.2 Materials and methods
2.2.1 Study site, animal characteristics and blood collection
Blood samples were collected as a part of a longitudinal disease study of African buffalo in
Kruger National Park (KNP), an area 360 km long (north-south) and 90 km wide (east-west) in the north-east corner of the Republic of South Africa (Figure 2.1). Kruger National Park is in a summer-rainfall area with an average temperature of 22°C (18-26°C) and average annual rainfall ranging from 458 to 746 mm (SA Weather Bureau; Zambatis, 2003). From February
2014 until August 2016, approximately 60 buffalo were sampled every two-to-three months
(12 captures in total) from a 900 ha predator-free enclosure near Satara rest camp (S 23°23’52”,
E31°46’40”) (Figure 2.1). At any given time depending on births and deaths, the herd consisted of 49 to 70 buffalo. As the buffalo are restricted in grazing grounds by the double-fenced enclosure, supplementary feed, including formulated feed and Lucerne hay, is occasionally supplied during the dry season, and a permanent man-made water source is available year- round as natural water sources tend to dry out seasonally.
23
Figure 2.1: Map showing the study site (right) in the Kruger National Park, in relation to the rest of South Africa, and the African continent. Rest camps within Kruger National Park are shown as black spots. Buffalo enclosure is located near the Satara rest camp (bold and underlined) (modified from SANParks).
24
For this study, blood samples (n = 747) collected over a two-year period (2014-2016) were used. Blood samples were taken from the jugular vein into EDTA-coated tubes by experienced, practicing veterinarians and stored at -80°C. The Animal Care and Use Proposal (ACUP) for this study was approved by the Institutional Animal Care and Use Committee (ACUP 4478) of the Oregon State University.
2.2.2 Molecular characterisation
DNA was extracted from EDTA bloods following an established Qiagen protocol (as per kit instructions) in African buffalo with a DNeasy Blood and Tissue Kit (Qiagen, USA) and stored at -80°C.
For the detection of A. marginale, the msp1β gene was amplified using a nested PCR as previously described by Molad et al. (2006). The external primers (AM456/AM1164) were used to amplify a PCR product of 700 bp while internal primers (AM100/AM101) yielded an amplicon of 246 bp (all primers ordered from GeneWorks). The PCR was performed in a final reaction volume of 25 μl, the PCR amplification mix contained 5–10 ng of purified genomic
DNA as template, 10 mM Tris-HCl (pH 8.4), 50 mM KCl (Promega), 3.5 mM MgCl2, 6.25 μM of each deoxynucleotide triphosphate (dNTP), 100 pmol of each primer, and 1 U of GoTaq polymerase (Promega, USA). PCR cycling conditions were an initial denaturation at 95°C for
5 min followed by 30 cycles at 95°C for 30 s, 60°C for 30 s (61°C for 10 s for nested PCR) and
72 °C for 30s, and 72 °C for 5 min.
Anaplasma centrale was detected using an amplification of a partial fragment of the groEL gene. PCR primers DS_Ac_F1 (forward 5’-GAGAAGATGCTGGTGGAGTT-3’) DS_Ac_R2
(reverse 5’-ACCACCGCATTCAAGGTCAT-3’) were designed to published groEL gene
25 sequences (GenBank accession numbers EF520691-EF52069 and AF414866-AF414867 (Ceci et al., 2008; A. E. Lew et al., 2003) and used to specifically PCR-amplify part (1,250 bp) of A. centrale. The reagents and PCR conditions were optimised in a series of experiments; the final
PCR was conducted in a 25 µl volume containing 10 mM Tris-HCl (pH 8.4), 50 mM KCl
(Promega, USA), 3.5 mM MgCl2, deoxynucleotide triphosphates (dNTPs; 200 µM each), primers (35 pmol each) and 1 U GoTaq polymerase (Promega) using the following protocol: 5 min at 94°C, followed by 30 cycles of 30 s at 94°C, 15 s at 55°C and 45 s at 72°C, followed by a final extension of 5 min at 72°C.
Negative and positive controls were included in each PCR. Following PCR, an aliquot (5 µl) of each amplicon was examined on 1.5% w/v agarose gels.
About 15% of PCR amplicons positive for each Anaplasma spp. were randomly selected for
DNA sequencing. Selected PCR amplicons of both A. marginale and A. centrale were treated with shrimp alkaline phosphatase and exonuclease I (Fermentas Inc., USA) (Werle, et al.,
1994) and subjected to direct bi-directional, automated sequencing (BigDye® Terminator v.3.1,
Applied Biosystems, USA) using the same primers used in PCR. The quality of sequences was assessed using the program Geneious Pro 2.0.10 (Larkin et al., 2007).
2.2.3 Phylogenetic analyses
Nucleotide sequences of the gene msp1β for A. marginale and groEL for A. centrale regions determined here were aligned using Mesquite v3.4 (Maddison & Maddison, 2015) with
MuscleAlign (Edgar, 2004), with those of homologous reference sequences for each respective
Anaplasma spp. in current databases (accessed on the 06 of June 2016; GenBank accession numbers AF111195-AF111197, AF348137, AF112479-AF112480, AF348138, AF110808-
26
AF110810, AF221692, EU281852, AY841153 for msp1β sequences of A. marginale;
EF520691, AF414867, and AF414866 groEL sequences of A. centrale and groEL sequences of A. marginale (JQ839014) and Anaplasma ovis (AF441131) were used as outgroup) and classified. Reference sequences for A. marginale were chosen based on their high sequence identity to sequences from this study when run through an nBLAST analysis
(http://blast.ncbi.nlm.nih.gov/Blast.cgi), and to compare a range of geographical areas. For A. centrale, as there were very few sequences available, all were included and those that had 100% sequence identity in the region used for sequence comparison were discarded. Sequence identities (in %) were calculated by pairwise comparison using the program BioEdit v7.2.5
(Hall, 1999). All sequences obtained for examining the msp1β and groEL regions were aligned over 345 and 881 nucleotide positions, respectively.
Phylogenetic analyses were performed using Bayesian Inference (BI) and Neighbor Joining
(NJ) methods. The BI was conducted, using Monte Carlo Markov Chain (MCMC) analysis in
MrBayes 3.1.2 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). The likelihood parameters for BI were based on the Akaike Information Criterion (AIC) test in jModeltest v2.1.5 (Darriba, et al., 2012; Guindon & Gascuel, 2003). The General Time
Reversible with a proportion of invariable sites (GTR+I), and transversion model with gamma distribution and (TVMef+G) were utilized for the analyses of the A. centrale and A. marginale sequence data, respectively. Four simultaneous tree-building chains were used to calculate posterior probabilities (pp) for 2,000,000 generations, saving every 100th tree produced. Based on the final 75% of trees generated, a consensus tree was constructed. The NJ analyses were performed using the program MEGA 6.0 (Tamura, et al., 2013), and the nodes were tested for robustness with 10,000 bootstrap replicates. The phylogenetic trees produced from the BI and
NJ analyses were compared for concordance in their topologies.
27
2.3 Results
PCR amplicons were of expected size (msp1β, 246; groEL, 1,250) when examined using agarose gel electrophoresis. PCR-sequencing analysis revealed the presence of single or mixed infections of Anaplasma spp. in African buffalo. Out of 747 samples tested, 129 (17.3%) and
98 (13.1%) were positive for single infection with A. marginale and A. centrale, respectively; whereas 113 (15.1%) were positive for both Anaplasma spp. Of the 103 individuals sampled at least once throughout the sampling period, only 17 were negative for both A. marginale and
A. centrale infection at all times, with the majority of those being calves (12/17). DNA sequencing of 41 PCR amplicons of A. marginale revealed four unique sequences of msp1β
(GenBank accession nos. KX714578- KX714581) from African buffalo while that of 36 PCR amplicons of A. centrale revealed seven unique sequences of groEL (GenBank accession nos.
KX714582- KX714588).
Subsequently, we studied sequence variation within the msp1β and groEL sequences. Within
A. marginale sequences of msp1β (aligned over 345 bp; GC content 52.6-55.6%), a nucleotide variability ranging from 1.6 to 8.5% (Table 2.1) was found, mainly attributed to transversions
(A ↔ G, T ↔ C, G ↔ C) at nucleotide positions 85, 147, 190, 213, 255, 271, 279, 306, and
321, and just one transition (T ↔ G) at nucleotide position 165 (see Supplementary Fig. 1).
Two msp1β sequences (KX714579 and KX714581) had three insertions at nucleotide positions
154 to 156, and one (KX714581) of these two sequences had an additional 15 nucleotide differences from positions 241 to 243 (see Supplementary Figure 2.1). Within groEL sequences of A. centrale (aligned over 881 bp; GC content 49.3-49.9%), a nucleotide difference of 0.3 to
2.4% was observed (Table 2.2) due to transitions of either A↔G at nucleotide positions 99,
114, 123, 126, 132, 138, 351, 402 and 429, or T↔C at positions 100, 136, 141, 145, 240, 324,
336, 420, 438, 549, 570 and 807 (see Supplementary Figure 2.2). We then compared sequences
28 of A. marginale and A. centrale, separately, with reference sequences available from GenBank
(Tables 2.1 and 2.2) to establish the nature and extent of nucleotide variation. All four msp1β sequences of A. marginale had the highest similarity (85.9 to 86.5%) with two reference sequences (GenBank accession no. AY841153, and AF111196 and M59845) from Israel and
USA, respectively, whereas they had the lowest similarity with another reference sequence
(CP000030) from the USA (Table 2.1). Similarly, the comparison of seven groEL sequences of A. centrale determined herein with available reference sequences revealed the highest similarity (93.9%) with a sequence from Italy (EF520691) while lowest with a sequence from
Australia (AF414866). All other sequences used for comparisons were from cattle.
29
Table 2.1: Pairwise differences (%) among the different partial major surface protein 1β sequences of Anaplasma marginale from African buffalo.
GenBank accession number KX714578 KX714579 KX714580 KX714581 AF110808 AF110809 AF110810 AF111195 AF111196 AF111197 AF112479 AF112480 AF221692 AF221693 AF348137 AF348138 AY841153 CP000030 EU281852 M59845 (location)
KX714578 (South
Africa)
KX714579 (South 3.5 Africa)
KX714580 (South 1.6 2.5 Africa)
KX714581 (South 8.5 5.0 7.5 Africa
AF110808 (U.S.A.) 23.4 22.3 22.5 26.1
AF110809 (U.S.A.) 23.4 22.3 22.5 26.1 0.0
AF110810 (U.S.A.) 23.4 22.3 22.5 26.1 0.0 0.0
AF111195 (U.S.A.) 20.2 17.6 19.3 14.1 17.0 17.0 17.0
AF111196 (U.S.A.) 19.3 16.7 18.5 13.2 14.4 14.4 14.4 4.7
AF111197 (U.S.A.) 18.8 16.1 17.9 12.6 16.1 16.1 16.1 1.5 3.3
AF112479 (U.S.A.) 24.8 23.7 23.9 24.6 4.6 4.6 4.6 14.4 14.1 13.5
AF112480 (U.S.A.) 19.3 16.7 18.5 13.2 15.3 15.3 15.3 3.3 1.5 2.4 14.4
AF221692 (U.S.A.) 23.4 22.3 22.5 26.1 0.0 0.0 0.0 17.0 14.4 16.1 4.6 15.3
AF221693 (U.S.A.) 22.8 20.1 21.9 23.9 24.0 24.0 24.0 21.9 18.4 21.0 27.7 19.3 24.0
AF348137 (U.S.A.) 18.5 15.8 17.6 12.3 15.5 15.5 15.5 2.4 3.0 0.9 13.5 2.1 15.5 20.2
AF348138 (U.S.A.) 23.9 22.9 23.1 23.7 3.7 3.7 3.7 13.5 13.8 12.6 1.6 14.1 3.7 26.3 12.0
AY841153 (Israel) 16.1 13.5 15.3 17.3 10.3 10.3 10.3 7.9 5.6 6.5 14.1 6.5 10.3 15.2 6.8 12.6
CP000030 (U.S.A.) 55.4 54.6 55.4 55.4 54.6 54.6 54.6 54.2 53.4 53.9 54.9 54.2 54.6 56.6 53.4 54.2 52.7
EU281852 (Brazil) 25.4 24.3 24.5 25.2 7.7 7.7 7.7 15.8 15.5 15.0 5.9 16.4 7.7 28.0 14.4 5.2 15.0 55.1
M59845 (U.S.A.) 16.1 13.5 15.3 17.3 10.0 10.0 10.0 7.9 5.0 7.1 13.8 5.3 10.0 14.0 6.2 12.3 1.2 52.7 14.7
30
Table 2.2: Pairwise differences (%) among the different partial heat shock protein (groEL) sequences of Anaplasma centrale from African buffalo.
GenBank accession number (location) KX714582 KX714583 KX714584 KX714585 KX714586 KX714587 KX714588 EF520691 AF414867 AF414866
KX714582 (South Africa)
KX714583 (South Africa) 0.8
KX714584 (South Africa) 2.2 2.4
KX714585 (South Africa) 1.8 2.0 0.6
KX714586 (South Africa) 0.4 0.6 1.9 1.4
KX714587 (South Africa) 0.6 1.1 2.0 1.5 0.5
KX714588 (South Africa) 0.4 0.8 1.9 1.4 0.3 0.3
EF520691 (Italy) 7.7 7.5 6.1 6.2 7.4 7.4 7.4
AF414867 (Australia) 7.9 7.8 6.3 6.4 7.7 7.7 7.7 1.2
AF414866 (Australia) 8.0 7.9 6.4 6.5 7.8 7.8 7.8 0.8 0.4
31
Phylogenetic relationships of A. marginale and A. centrale sequences were explored separately using selected reference sequences (Figure 2.2). The topology of the phylogenetic trees generated for both A. marginale and A. centrale sequences, employing BI and NJ methods were similar (data not shown); hence, the NJ trees for both Anaplasma spp. are presented here, with nodal support values given for both methods (Figure 2.2). For A. marginale, the tree revealed five major clades, where clades 1-3 and 5 contained msp1β sequences originated from Brazil,
Israel and the USA, whereas clade 4 consisted of four sequences (GenBank accession numbers
KX714578- KX714581) determined in this study. However, the nodal support for clade 4 was very weak (Figure 2.2A). For A. centrale, the tree comprised of two clades where seven groEL sequences of A. centrale determined here clustered together in clade 1, with strong statistical support (posterior probability value for BI, 1.0; bootstrap value for NJ, 100%), with the exclusion all reference sequences of A. centrale from Italy (EF520691) and Australia
(AF414867 and AF414867) grouping in clade 2 (Figure 2.2B). The groEL sequences of A. marginale and A. ovis formed separate clades (Figure 2.2B).
32
Figure 2.2: Genetic relationships of Anaplasma marginale (A) and Anaplasma centrale (B) detected from African buffalo (Syncerus caffer) from Kruger National Park, South Africa (bold) with reference genotypes selected from previous studies. The relationships were inferred based on the phylogenetic analysis of partial sequences of major surface protein 1β (A. marginale, A) and heat shock protein (groEL) (A. centrale, B) using Neighbour Joining (NJ) and Bayesian inference (BI) methods, and A. ovis was used as the outgroup for A. centrale. Nodal support values are indicated: posterior probability for BI (first), and bootstrap for NJ (second). The scale bar indicates distance.
33
2.4 Discussion
This is the first study to characterise A. marginale and A. centrale from African buffalo using species specific molecular markers, msp1β and groEL, respectively. Molecular-phylogenetic analyses of msp1β and groEL sequences of A. marginale and A. centrale, respectively, revealed that sequences of Anaplasma spp. from African buffalo were unique and they grouped separately when compared with previously published sequences of both species.
In the present study, we found that the overall sample prevalence of A. marginale (32.7%;
242/747) was similar to that of A. centrale (28.2%; 211/747) in African buffalo which does appear to differ from findings of previous studies reported from Botswana (Carmichael &
Hobday, 1975; Eygelaar et al., 2015) and South Africa (Debeila, 2012; Henrichs et al., 2016), but not statistically significant. For example, Eygelaar et al. (2015) tested blood samples from
African buffalo located in Chobe National Park and the Okavango Delta from Botswana by employing reverse line blot hybridization analysis (RLB) using 16s rRNA gene and found that the sample prevalence of A. centrale (32.7%; 36/110) was higher than A. marginale (21.8%;
24/110). Similarly, using the same RLB method, two recent studies by Debeila (2012) and
Henrichs et al. (2016) from South Africa (Kruger National and Hluhluwe-Imfolozi Parks) also found that the sample prevalence of A. centrale (49-75% and 14.3-61.2%, respectively) was higher than A. marginale (24-42% and 20.9-53.8%, respectively). These differences in the sample prevalence of Anaplasma spp. between previous studies and this one could be due to different markers used in PCR as we used Anaplasma species specific markers for A. centrale
(groEL) and A. marginale (msp1β) whereas other studies utilised genus specific marker (16S rRNA gene) followed by species-specific probes. In addition, as Anaplasma spp. infection occurs in a cyclical nature, the presence of the parasite has the potential to be missed if only single sampling events occur in an individual host (Aubry & Geale, 2011; Potgieter & Stoltz,
34
2004). The repeated sampling of buffalo may increase the probability of identifying Anaplasma spp. infection in the host, even if the rickettsia is below the detection level in some sampling periods. These differences in prevalence may also be due to a difference in sampling strategy, uneven representation of individuals in repeat sampling, the vector, individual characteristics and health status, the environment in which these buffalo are contained and co-infection interactions, with all of these factors warranting further investigation.
This study highlights the need to use multiple markers to characterise Anaplasma spp. from wild ruminants such as African buffalo as a ‘one gene fits all’ approach does not capture the extent of diversity in an organism. To date, only reverse line blot hybridization analysis employing 16S rRNA gene has been used to detect Anaplasma spp. (A. marginale, A. centrale,
A. sp. Omatjenne, A. bovis and A. phagocytophilum) from African buffalo (Fyumagwa et al.,
2013; Henrichs et al., 2016). The full extent of the diversity of Anaplasma spp. found in African buffalo is still not explored as we focussed only on two economically important species (A. marginale and A. centrale) of bovine anaplasmosis. Future studies should focus on the characterisation of A. sp. Omatjenne, A. bovis and A. phagocytophilum species from African buffalo using multiple species specific molecular markers.
Phylogenetic analyses revealed that four A. marginale (GenBank accession nos. KX714578-
KX714581) and seven A. centrale (KX714582- KX714588) sequences found herein were unique and that within each species analysis, there was clustering of these sequences from
South Africa (see Fig. 2A and B), with the exclusion of respective reference sequences.
Previously, Debeila (2012) used RLB analysis using 16S rRNA gene to characterise A. centrale and A. marginale from African buffalo and found eight novel sequences for A. centrale and four for A. marginale and this level of variation is very similar to what we observed in A.
35 centrale (no. of sequences = 7) and A. marginale (n = 4). However, Debeila (2012) reported that some (e.g., HIP/A8/b) of 16S rRNA sequences of A. centrale found in African buffalo were 100% identical to previously published sequences (e.g., AF309869) from cattle.
Contrarily, in this study, comparison of seven groEL sequences of A. centrale (KX714582-
KX714588) with those of reference sequences (EF520691, AF414867 and AF414867) revealed a pairwise difference of 6.1 to 8.0% and they grouped together with the exclusion of previously published sequences (Fig. 2B). Similarly, for four 16S rRNA sequences of A. marginale, Debeila (2012) observed a high similarity (just two nucleotide differences over 726 bp) with reference sequences, whereas we found a high nucleotide variation (13.2 to 55.4%) among four msp1β sequences of A. marginale and previously published sequences (n = 16).
The uniqueness of A. centrale and A. marginale sequences from African buffalo found herein might be due to different molecular markers used in this study (groEL and msp1β) as compared to that (16s rRNA) used by Debeila (2012). In addition, a higher pairwise difference in A. marginale sequences might be due to shorter sequences used for comparison (See supplementary Fig. 1) as compared to 726 used by Debeila (2012).
Given that only msp1β and groEL genes have been used to quantify the burden of A. marginale and A. centrale, respectively, (Carelli et al., 2007; Decaro et al., 2008) and both of these
Anaplasma species have been characterised using the two genes in this study, this detailed molecular analysis will allow for the development of a qPCR assay that can be used to quantify burdens and understand consequences of infection to the host.
In conclusion, this is the first study to characterise A. marginale and A. centrale from African buffalo using species specific molecular markers. Overall, 45.5% blood samples collected from buffalo kept in KNP, South Africa, were positive for single or multiple infections with A.
36 marginale and A. centrale. Sequence variation and phylogenetic analyses of msp1β and groEL sequences of A. marginale and A. centrale, respectively, revealed that sequences of Anaplasma spp. from African buffalo were unique and they grouped separately when compared with previously published sequences of both species, including those from cattle in sub-Saharan
Africa. To our knowledge, no sequences are available from cattle hosts for these Anaplasma species in the same location. Sequencing these species in cattle in the same area would allow for more conclusive evidence as to whether African buffalo are a reservoir for anaplasmosis or not. This study will pave the way for future studies to assess genetic variation among
Anaplasma spp. from wild ruminants using molecular markers that are able to distinguish more fine scale genetic variation, which is useful for molecular epidemiology studies that identify reservoirs and across species transmission, and help to undertake health and fitness studies and host-parasite dynamics using quantitative molecular tools.
37
Chapter 3: Fitness and health consequences of Anaplasma marginale and A.
centrale infections in African buffalo (Syncerus caffer) from Kruger
National Park, South Africa
3.1 Introduction
African buffalo (Syncerus caffer) are hosts to many infections and diseases that make them a useful animal to study disease ecology (Jolles & Ezenwa, 2015). A number of Anaplasma
(Rickettsiales: Anaplasmataceae) species can infect African buffalo, including Anaplasma marginale, A. centrale, A. sp. Omatjenne, A. phagocytophilum and A. bovis (Aubry & Geale,
2011; Debeila, 2012; Dumler et al., 2001; Henrichs et al., 2016; Potgieter & Stoltz, 2004).
Anaplasma spp. are gram-negative, obligate intracellular bacteria that invade different cells in different hosts, with varying levels of pathogenicity (Debeila, 2012; Henrichs et al., 2016;
Kocan et al 2010; Potgieter & Stoltz, 2004). They can be transmitted between domesticated and wild ruminants, typically using a tick vector (Debeila, 2012; Henrichs et al., 2016; Kocan, et al., 2010; Potgieter & Stoltz, 2004). Cattle can be severely affected by A. marginale, resulting in losses in productivity and even mortalities, but they appear less affected by A. centrale
(Aubry & Geale 2011; Potgieter & Stoltz 2004). African buffalo, on the other hand, do not appear to show any clinical signs (such as marked anaemia, icterus, abortions, anorexia) when infected with either A. marginale and/or A. centrale (Henrichs et al., 2016).
The apparent lack of morbidity and mortality in African buffalo does not mean that they are unaffected by Anaplasma. Previous exposure to A. marginale has been shown to negatively affect fecundity (pregnancy and lactation) in African buffalo (Henrichs et al., 2016).
Furthermore, buffalo infected with A. centrale were in better body condition than those infected with A. marginale (Henrichs et al. 2016). This suggests some differences in pathogenicity, the
38 extent of which requires further investigation (Henrichs et al., 2016). One of the key consequences of A. marginale infection in cattle is a reduction in red blood cells, causing anaemia and even death (Aubry & Geale, 2011; Brown, 2012). In African buffalo, on the other hand, the presence of A. centrale has previously been found to result in increased red blood cell counts and haematocrit (Henrichs, 2014). The mechanisms involved are unclear and the relationship thus needs to be determined at a finer scale, e.g. by estimating the burden level instead of just using presence/absence data, as well as assessing biochemical parameters. For instance, elevated total serum protein levels (consisting of alpha, beta and gamma globulins, and albumin) can indicate a host immune response in the absence of clear clinical signs (Couch et al., 2017; Kushner, 1982). While gamma globulins increase as part of the antibody response, beta and alpha globulins (comprised of proteins such as haptoglobin, fibrinogen and serum amyloid A) increase in response to inflammation and albumin decreases with inflammation, shock and malnutrition (Couch et al., 2017; Kushner, 1982). The combination of these biochemical changes, as measured by total protein, can be an important first step in understanding the host’s response to infection.
An estimate of intensity of infection can further provide an opportunity to elucidate co- infection dynamics between the infectious agents, which is important in the understanding of the mechanisms of disease relating to host response and direct damage caused by the infectious agents (Henrichs et al., 2016). Previous work further suggests that a range of host (e.g. age, sex, co-infection status) and environmental factors (e.g. location, season) may significantly affect the probability of infection with Anaplasma spp. (Ezenwa & Jolles, 2015; Henrichs et al., 2016). While presence/absence data of Anaplasma spp. infection is important and useful in interpreting the effects of infection, a quantitative measure of the infection intensity would allow better insights into co-infection dynamics within the host (Henrichs et al., 2016).
39
Using a quantitative approach, this study aims to better understand the health and fitness effects of subclinical infection with A. marginale and/or A. centrale in African buffalo, as well as to further investigate the epidemiology of these infections in buffalo. We hypothesise that infection with A. marginale in African buffalo will be more detrimental to the health of the host than that of A. centrale. In addition, the impact of a higher intensity of infection with A. marginale may result in a decreased body condition, decreased haematocrit, and increased total serum protein concentration. This study will provide unique detailed longitudinal data on a subclinical infection in a wildlife host, which is a greatly understudied area, and it will improve our understanding of the impact of anaplasmosis on African buffalo.
40
3.2 Materials and methods
3.2.1 Study site and animal characteristics
Blood samples used in this study were collected from herds of African buffalo used in two previous studies, designated as the managed and the free-ranging herds (Couch et al., 2017;
Ezenwa & Jolles, 2015), conducted at Kruger National Park (KNP), South Africa (see Chapter
2, Figure 2.1). The study for the free-ranging herd was conducted as a collaboration between
Oregon State University (OSU) and University of Georgia (UGA), while the managed herd was a collaboration between the same research team from OSU and the Pirbright Institute.
Samples from the free-ranging herd were previously investigated for the presence of a range of blood parasites, including Anaplasma spp. (Henrichs et al, 2016), however, no quantitative estimates of infection, were obtained previous to the present study.
Blood samples from buffalo in the managed herd were collected from February 2014 until July
2016, with approximately 60 buffalo being sampled every two-to-three months from a double- fenced predator-free enclosure near Satara rest camp (S 23°23’52”, E 31°46’40; see Chapter 2,
Figure 2.1), resulting in a total of 12 captures over the full study period. There were 49 to 70 buffalo at any given time in the study, with 103 animals (67 females and 36 males) being sampled at least once (Table 3.1). Based on age at first capture, 38% of the buffalo were calves
(under 1 year old), 28% sub-adult (1-5.5 years old), 26% adult (>5.5-15 years old) and 8% geriatric animals (over 15 years old). Supplementary feed, including glucose pellets and
Lucerne hay, was occasionally supplied to counteract the restriction in grazing grounds that resulted from being fenced in, and a permanent man-made water trough was available year- round.
41
Blood samples from the free-ranging herd were collected from mid-2009 to mid-2012. This herd was composed of sub-adult and adult females (3-5 years old at the time of initial capture), which were located near the Crocodile Bridge (S 25°21’30”, E 31°53’32”) and Lower Sabie
(S 25°7’16”, E 31°55’2”) rest camps in the lower half of KNP (see Chapter 2, Figure 2.1).
Buffalo from the free-ranging herd were sampled twice a year with 150-200 buffalo being sampled at every capture, resulting in 304 individuals being sampled at least once throughout the full study period (Table 3.1). Because of the difficulties associated with capturing individuals roaming freely over relatively large areas, each capture event stretched over three months for each of the two sub-herd, resulting in year-round captures.
We used samples from buffalo calves from the managed herd (no samples had been collected from calves in the free-ranging herd) to assess the order of first infection with A. marginale or
A. centrale. Only calves that had been sampled within the first six months of their life and were born within the enclosure were included in analyses. This cut-off was chosen because in cattle, calves are protected against Anaplasma spp. by maternal antibodies until weaning at approximately 4-6 months (Potgieter & Stoltz, 2004).
42
Table 3.1: Number of times individual African buffalo were sampled throughout two longitudinal disease studies from Kruger National Park.
Managed Herd Free Ranging Herd* Times sampled (n) (n) 12 30 NA 11 9 NA 10 8 NA 9 0 87 8 1 30 7 3 31 6 8 21 5 5 18 4 11 20 3 8 33 2 10 31 1 10 30 TOTAL 103 301 *The free-ranging herd was sampled nine times in total.
3.2.2 Sampling and measurements
Blood samples from 404 individual animals (103 from the managed and 301 from the free- ranging herd) were collected via jugular venepuncture in vacutainer tubes coated with lithium heparin (plasma), CA EDTA (whole blood) or no additives (serum). This resulted in 747 samples from the managed herd and 1740 from the free-ranging herd over the study periods mentioned above. Serum and plasma were collected following centrifugation at 5,000 x g, for
10 min, and stored at -80°C. Haematocrit (through an Automated Impedance Cell Counter,
Model ABC-Vet, as per Beechler et al. (2009) and total serum protein (Abaxis Vetscan VS2
Chemistry Analyser; Abaxis Inc, Union City, CA, USA) were determined immediately after the samples arrived at the laboratory as per Couch et al. (2017). An average body condition score was calculated from individual scores given for ribs, spine, hips and tail base (each on a scale of 1-5 in 0.5 increments) as per Ezenwa et al. (2009). A qualified veterinarian assessed the pregnancy status of each female buffalo over the age of two-to-three years at every capture
43 event through rectal palpation. Age was estimated using incisor teeth, based on number and wear of adult teeth (Jolles, 2007).
3.2.3 DNA extraction and conventional PCR
DNA was extracted from the blood samples following an established protocol using a DNeasy
Blood and Tissue Kit (Qiagen, USA) (as per the kit protocol) for further molecular diagnostics.
For the managed herd, a conventional PCR was run using major surface protein one beta
(msp1β) for A. marginale and the heat-shock protein groEL for A. centrale, to identify which samples were positive for single or mixed infections with these pathogens (see Chapter 2). An established protocol was used to amplify msp1β gene of A. marginale using a nested PCR
(Molad et al., 2006) while new primers were designed to amplify the groEL gene of A. centrale
(see Chapter 2).
For the free-ranging herd, Henrichs et al. (2016) had already determined the presence or absence of A. marginale and A. centrale infection using 16SrRNA gene employing the Reverse
Line Blot (RLB) Hybridisation method.
3.2.4 Quantitative PCR
An established multiplex qPCR protocol using TaqMan probes (Decaro et al., 2008) was employed to quantify A. marginale and A. centrale, targeting msp1β and groEL genes, respectively, in samples from both herds that were positive for conventional PCR or RLB hybridisation. Quantitative PCR had not previously been completed with samples from the free-ranging or managed herds.
44
For the construction of an A. marginale standard curve, a plasmid that contained a 729 base pair (bp) fragment of the msp1b gene was used (Carelli et al., 2007), whereas for A. centrale, a 488 bp fragment of the groEL gene was amplified using the primers groEL-ACF and groEL-
ACR (Decaro et al., 2008). These PCR fragments were 1.5% agarose gel-purified using the
QIAquick Gel Extraction Kit, (Qiagen, USA) and cloned into the pGEM®-T Easy Vector
System (Promega, USA) as per manufacturers’ instructions. Plasmid DNA was purified from transformed cells using Wizard Plus Midiprep (Promega) and quantified by spectrophotometrical analysis at 260 nm, and then subjected to bi-directional, automated sequencing using the same primers used in PCR. The quality of the sequences was assessed using the program Geneious Pro 2.0.10 (Kearse et al., 2012).
The multiplex qPCR reaction for the simultaneous detection and quantification of A. marginale and A. centrale DNA was performed on the Rotor-Gene Q real time machine (Qiagen, USA) using TaqMan probes. For A. marginale, the primers AM-For/AM-Rev targeting a 95 bp product within the msp1β gene and a probe (AM-Pb) with 6-Carboxyfluorescein (6-FAM) and
Black Hole Quencher 1 (BHQ1) dye were used. For A. centrale, the primers AC-For/AC-Rev to amplify a 77 bp region of the groEL gene and a probe (AC-Pb) with Hexachlorofluorescein
(HEX) and BHQ1 dye were used (Decaro et al., 2008). QIAgility (Qiagen) was used to load samples into a 100-well ring. Briefly, the reaction mixture included 12.5μl of Promega GoTaq probe qPCR Master Mix (Promega, USA), primers at a concentration of 600 nmol/l (A. marginale) or 900 nmol/l (A. centrale), and each probe at a concentration of 200 nmol/l, and
10 μl of template (1:10 dilution) or plasmid DNA (for the standard curve). The reaction consisted of an initial reaction period at 95°C for 10 min, followed by 40 cycles of 95°C denaturation for 1 min, and 1 min of 60°C of annealing/extension. Samples were run in
45 duplicates and each plasmid DNA was run in triplicate. A no-template control was included in each assay.
A 10-fold standard curve was created to assess sensitivities and dynamic ranges using standard
DNA ranging from 1010-100. A coefficient of variation (CV) was assessed between and within assays by multiple measurements of the absolute copy numbers obtained from standard DNA samples of high, intermediate and low concentrations within the same runs and between runs.
Cycle threshold (CT) values were determined automatically with the Rotor-Gene Q Series
Software Version 2.3.1 (Build 49) (Qiagen). The intensity of Anaplasma spp. from the qPCRs were only included as a positive reading if they exceeded a variable cut-off, specific to each run, which was set to be the higher reading out of either the non-template controls or the lowest standard in the standard curve (Supplementary Table 3.1). If a qPCR result did not meet the cut-off criteria, it was classified as negative. Burdens were also corrected in relation to conventional PCR readings. If there was an insufficient qPCR reading and a positive conventional PCR result, this entry was converted to ‘NA’ (i.e. data not used); however, if there was an insufficient qPCR reading and a negative conventional PCR result, this entry was converted to a ‘0’ for a negative infection burden (Supplementary Table 3.1).
3.2.5 Statistical analyses
There were important differences between the managed and free-ranging herds, including sampling intervals, population structure and management input. Therefore, it was decided to analyse the two herds separately. Similarly, two different types of models were run to investigate associations between Anaplasma burden, co-infection dynamics and various health outcomes: models using current Anaplasma burden as predictor variable (= “concurrent models”) and models using Anaplasma burden from the immediately previous capture (=
46
“predictive models”), which was two-to-three months previous for the managed herd, and six months for the free-ranging herd. The predictive models were included to detect associations that occur over extended time periods, however, variation in sampling intervals and time of year of sampling between individuals are likely to affect the results of this part of the analysis.
Models were run for both data on Anaplasma burden and presence/absence of infection; however, because the former was the focus of this study, except for the pregnancy and survival models, only the burden data are presented in the result section. Nevertheless, the models for the presence/absence data can be seen in the Supplementary section of this chapter
(Supplementary Tables 3.2-3.6). Burden models were run as single infections, with health response variables run separately for A. marginale burden and A. centrale burden. This was due to a loss of data in models that initially contained both infections as a predictor variable, or as a co-infection interaction, because of entries being omitted by the statistical program if either one of the species’ burdens had an entry of ‘NA’. These losses in data would have severely impacted the quality of the results.
Anaplasma burden was log-transformed. A chi-square test was used to assess the significance associated with the order of infection in calves. Generalised linear mixed models (GLMM) were used to investigate relationships between Anaplasma infection burden and measures of health (body condition, total protein and haematocrit), and a binomial error structure and a logit link were added where the outcome was presence/absence of infection. An individual identification number (Animal ID) was included as a random effect due to repeat captures of individuals in both projects. Animal ID and capture number were included as random effects for pregnancy analyses to minimise pseudo-replication of animals that had been repeatedly sampled throughout the one pregnancy. This was to account for lower reliability in accurate
47 pregnancy testing in earlier stages of pregnancy, the calving season running for three-to-four months, abortions and animals being missed in sampling events.
Independent variables of season, animal traits (age, body condition and reproductive status) and co-infection with A. marginale or A. centrale were added into the model as main effects if they were significant at the p < 0.05 level after univariable screening. Multivariable model selection was performed to minimise the Akaike Information Criterion (AIC) for fitness response models (Akaike et al., 1973). Individual terms were dropped if this improved the AIC by two or more points, and the final model was selected if no more terms were to be dropped that met this criterion (Supplementary Table 3.7). All statistically non-significant interaction terms were removed regardless of AIC. After model selection occurred, a forced co-infection interaction term (i.e. A. centrale*A. marginale) was added to the simplest model to evaluate whether co-infection was important to the parameter of interest; this was done to avoid the loss of data associated with “NA” values mentioned above. If the forced co-infection term was not significant, it was not included in the final models. Where Anaplasma spp. infection was the outcome, predictor variables were selected primarily based on likely biological significance.
Season was coded as a two-level binary variable (wet season = November to April, dry season
= May to October). Continuous variables (age, total protein and haematocrit) were centred by subtracting the mean (to avoid multicollinearity issues when these variables initially tested as interaction terms in the models) and rescaled by dividing by two standard deviations (to assist with coefficient interpretation by having variables on the same scale).
Burden data were unable to be modelled with the pregnancy data from the managed herd due to the omission of samples that were missing burden or pregnancy data; however, burden data could be analysed from the free-ranging herd, and the presence/absence data from both herds
48 could be used for this analysis. Similarly, burden data could not be used for the Andersen-Gill
(AG) survival model (a form of Cox Proportional Hazards model that accounts for multiple events), which was used to analyse mortality with both presence/absence from both A. marginale and A. centrale in both herds. This was due to a loss in sample size and an insufficient number of events in the burden data. Anaplasma marginale and A. centrale infections were a time-dependent variable.
All statistical analyses were performed using R Studio (R Core Team, 2016), with packages lme4 (Douglas et al., 2015), lmerTest (Kuznetsova et al., 2015) and nlme (Pinheiro et al., 2017).
The level of statistical significance was p < 0.05.
49
3.3 Results
For the managed herd, 220 samples (29.4%) fulfilled our criteria for a valid A. marginale burden reading and 181 for A. centrale (24.2%) (Table 3.2). For the free-ranging herd, the same was true for 476 samples (27.3%) for A. marginale burden reading and 426 (24.5%) for A. centrale (Table 3.2).
Table 3.2: Mean infection intensities (copies/reaction) of A. marginale and A. centrale in managed and free-ranging herds of African buffalo from Kruger National Park.
Managed herd Free-ranging herd
A. marginale A. centrale A. marginale A. centrale N 220 181 476 426
Mean 4,809.446a 2,783.622b 3,851.945a 724.673b
SD 16,660.440 17,210.860 20,565.020 2,613.317 Standard deviation (SD) and sample size (n) for each species in each study is included. The superscripts a,b within row indicate significantly different values.
Using a general linear mixed model, and including other variables such as sex, age and season, there was no significant difference in overall infection intensity between herds for either infection with A. marginale or A. centrale. The burden of infection with A. marginale was, however, higher than with A. centrale in both herds (p < 0.001) (Table 3.2)
3.3.1 Order of infection in calves from the managed herd
Fifteen calves under the age of six months at first capture were analysed for the order of infection, with 14 (out of 15) showing initial infection with A. marginale. The remaining calf was coinfected with A. marginale and A. centrale. Five calves were infected with A. marginale on their first capture after birth. A chi-square test showed that A. marginale occurred before A. centrale more frequently than was expected at random (p < 0.001, X2 = 13.133). The average
50 age of initial A. marginale infection was eight months (SD = ± 6 months), and the average age of initial A. centrale infection was 17 months (SD = ± 14 months).
3.3.2 Effects of host and environmental factors on anaplasmosis
In the managed herd, the burdens of both A. marginale and A. centrale decreased significantly with increasing age (p = 0.004 and p = 0.043, respectively). In addition, the intensity of infection with A. centrale was greater during the dry season (p = 0.013) and in males (p =
0.004) (Table 3.3). In the free-ranging herd, the sub-herd from the Lower Sabie area had significantly greater A. marginale burdens (p < 0.001) than that from Crocodile Bridge, and a higher burden of A. marginale was associated with a higher burden of A. centrale (Table 3.3).
Table 3.3: Generalised linear mixed (concurrent) model for the burden (log-transformed) of A. marginale and/or A. centrale (copies/reaction) in managed and free-ranging herds of African buffalo from Kruger National Park.
A. marginale A. centrale
Coefficient SE p Coefficient SE p MANAGED HERD n = 84 n = 55 Intercept 8.034 0.314 7.334 0.669 Age -0.356 0.116 0.004 -0.217 0.100 0.043 Season (Wet) -0.297 0.297 0.321 -1.132 0.437 0.013 Sex (Male) 0.143 0.380 0.710 1.780 0.492 0.004 Co-infection burden* 0.027 0.041 0.509 0.008 0.707 0.906
FREE-RANGING HERD n = 205 n = 138 Intercept 6.177 0.502 5.986 0.666 Age -0.104 0.072 0.153 -0.068 0.088 0.441 Season (Wet) 0.319 0.173 0.067 0.236 0.193 0.226 Sub-herd (Lower Sabie) 1.768 0.271 <0.001 -0.592 0.399 0.144 Co-infection burden* 0.058 0.043 0.185 0.181 0.060 0.003 *Co-infection burden refers to the other Anaplasma species. Significant p-values (<0.05) are bold; SE – standard error
The predictive model, which used variables from the previous capture period to predict current
Anaplasma burden, showed some of the same patterns in relation to age. In both herds, younger
51 animals had higher burdens of A. marginale in the subsequent sampling period than older ones
(p < 0.001 in managed herd, and p = 0.045 in free-ranging herd) (Table 3.4). However, there were some additional relationships, not seen in the concurrent model. In the managed herd, buffalo showed a significant negative association between current and previous Anaplasma spp. burdens, i.e. current A. centrale burden was higher if previous A. marginale burden was lower (p = 0.049) and current A. marginale burden was higher if previous A. centrale burden was lower (p < 0.001) (Table 3.4).
Table 3.4: Generalised linear mixed (predictive) model for the burden of (log-transformed) A. marginale and/or A. centrale (copies/reaction) in managed and free-ranging herds of African buffalo from Kruger National Park.
Previous A. marginale Previous A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 78 n = 78 Intercept 3.733 0.726 4.641 1.111 Age -5.531 0.906 <0.001 -2.454 1.412 0.088 Sex (Male) 0.310 0.682 0.655 1.309 1.099 0.243 Season (Wet) -0.207 0.621 0.740 0.668 0.658 0.307 Co-infection burden* -0.167 0.083 0.049 -0.553 0.121 <0.001
FREE-RANGING HERD n = 148 n = 154 Intercept 1.264 0.588 3.451 0.489 Age -2.200 1.088 0.045 -4.646 1.180 <0.001 Sub-herd (Lower Sabie) 2.306 0.668 <0.001 -0.188 0.624 0.763 Season (Wet) 1.168 0.519 0.026 -0.170 0.473 0.720 Co-infection burden* 0.143 0.089 0.110 -0.028 0.074 0.708 Co-infection burden refers to current other Anaplasma species; significant p-values (<0.05) are bold; SE – standard error.
Furthermore, in the free-ranging herd, animals from the Lower Sabie sub-herd had higher A. marginale burdens in the subsequent sampling period than those from Crocodile Bridge (p <
0.001), and animals sampled in the wet season had higher A. marginale burdens in the previous sampling period than those sampled in the dry season (p = 0.026).
52
3.3.3 Fitness and health responses to infection
Body Condition
A higher burden of A. centrale was associated with a higher body condition score in the managed herd (p = 0.039) (Table 3.5).
Table 3.5: Generalised linear mixed (concurrent) model for body condition in managed and free- ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 124 n = 124 Intercept 2.954 0.159 2.891 0.141 Season (Wet) 0.208 0.134 0.123 0.218 0.131 0.099 Anaplasma spp. burden 0.014 0.022 0.517 0.038 0.018 0.039 Sex (Male) 0.010 0.142 0.942 -0.068 0.147 0.645 Age 0.498 0.171 0.005 0.423 0.157 0.009
FREE-RANGING HERD n = 268 n = 268 Intercept 2.903 0.092 3.083 0.090 Season (Wet) 0.131 0.110 0.235 -0.356 0.110 0.002 Anaplasma spp. burden 0.035 0.015 0.020 -0.007 0.015 0.645 Sub-herd (Lower Sabie) -0.045 0.085 0.597 -0.003 0.080 0.966 Age -0.182 0.131 0.168 -0.175 0.129 0.174 Season (Wet): Anaplasma spp. -0.053 0.018 0.003 0.048 0.021 0.023 infection Significant p-values (<0.05) are bold; SE – standard error
In the free-ranging herd, however, there was an interaction between body condition, Anaplasma sp. burden and season (p = 0.003 and p = 0.003, for A. marginale and A. centrale, respectively)
(Figure 3.1). There appeared to be larger fluctuations in body condition in the wet season, with a tendency for increased body condition with increased A. centrale burdens. Animals in the dry season had better body condition overall and appeared to have little association with A. centrale. An increase in A. marginale burden was associated with better body condition in the dry season, but worse body condition in the wet season (Figure 3.2). Additional analysis showed no significant interaction between Anaplasma spp. burden and body condition when only dry or wet season samples were included, and it is therefore difficult to ascertain any real
53 association between the two. Further analysis also indicated that relationships between body condition and seasonality could be distorted, with the subsets used by the models for the Lower
Sabie sub-herd being skewed towards measurements from the early dry season when conditions were still good, and the early wet season when conditions were still poor. For the Crocodile
Bridge sub-herd, the subsets used by the models had more animals included from the late wet season when conditions were good and mid-late dry season when conditions were poor.
The only other significant predictor of body condition was age, with older animals being in better body condition in the managed herd (p = 0.005 and p = 0.009 for A. marginale and A. centrale, respectively) (Table 3.5).
54
Figure 3.1: Mean body condition score with burden (log-transformed) of Anaplasma centrale based on season (wet or dry), in a free-ranging herd (n = 268) of
African buffalo from Kruger National Park.
55
Figure 3.2: Mean body condition score with burden (log-transformed) of Anaplasma marginale based on season (wet or dry) in free-ranging African buffalo
(n = 268) from Kruger National Park.
56
In the predictive model (i.e. predictors included current body condition, season, age, sex/herd and previous Anaplasma burden), there was no significant association between A. marginale or A. centrale burden and body condition in either the managed or free-ranging herds (Table
3.6). There was also a negative association between concurrent age and body condition in the free-ranging herd (p = 0.019 for both Anaplasma spp) (Table 3.6).
Table 3.6: Generalised linear mixed (predictive) model for body condition in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 101 n = 101 Intercept 2.928 0.175 2.897 0.141 Season (Wet) 0.151 0.144 0.299 0.164 0.143 0.253 Previous Anaplasma spp. infection 0.009 0.024 0.720 0.021 0.020 0.274 Sex (Male) -0.021 0.148 0.890 -0.068 0.151 0.652 Age 0.256 0.177 0.152 0.208 0.161 0.200
FREE-RANGING HERD n = 226 n = 226 Intercept 3.066 0.0733 3.009 0.083 Season (Wet) -0.048 0.065 0.463 -0.046 0.065 0.487 Previous Anaplasma spp. infection -0.009 0.010 0.342 0.004 0.011 0.688 Sub-herd (Lower Sabie) 0.104 0.081 0.203 0.090 0.080 0.260 Age -0.323 -0.128 0.013 -0.301 0.128 0.019 Significant p-values (<0.05) are bold. SE – standard error.
Haematocrit
Neither herd showed a significant association between concurrent haematocrit and Anaplasma spp. infection burden (Supplementary Table 3.8), nor was Anaplasma spp. burden correlated with haematocrit in the predictive models in the free-ranging herd (Table 3.7).
57
Table 3.7: Generalised linear mixed (predictive) model for haematocrit levels in managed and free- ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 101 n = 101 Intercept -0.076 0.088 -0.185 0.065 Season (Wet) 0.028 0.059 0.629 0.016 0.059 0.791 Previous Anaplasma spp. burden -0.014 0.013 0.261 0.005 0.009 0.548 Sex (Male) -0.268 0.124 0.035 -0.060 0.075 0.430 Age 0.340 0.078 <0.001 0.327 0.075 <0.001 Sex (Male): Previous Anaplasma 0.040 0.019 0.040 NA NA NA spp. burden
FREE-RANGING HERD n = 207 n = 207 Intercept -0.037 0.071 0.010 0.082 Season (Wet) -0.011 0.076 0.885 -0.012 0.077 0.877 Previous Anaplasma spp. burden 0.006 0.010 0.530 -0.005 0.011 0.689 Sub-herd (Lower Sabie) 0.239 0.076 0.002 0.249 0.074 0.001 Age 0.145 0.131 0.271 0.131 0.130 0.316 Significant p-values (<0.05) are bold. SE – standard error. SE – standard error.
However, there were some significant associations between A. marginale and haematocrit in the managed herd in the predictive model. There was a significant interaction between haematocrit, sex and A. marginale burden (p = 0.040) (Table 3.7), with higher overall haematocrit levels and a larger increase in haematocrit with increasing A. marginale burden in the previous capture in females compared to males (Figure 3.3).
58
Figure 3.3: Haematocrit level with the previous capture’s infection intensity (log-transformed) of Anaplasma marginale based on sex in a managed herd of African buffalo (n = 226) from Kruger National Park. F = female, M = male.
59
Age was significantly associated with haematocrit in the managed herd in the predictive models, with older animals having a higher haematocrit (p < 0.001 for both Anaplasma spp. from the managed herd) (Table 3.7). A higher haematocrit was found in the Lower Sabie than the Crocodile Bridge sub-herd in both the A. marginale and A. centrale models (p = 0.002 for the A. marginale model and p = 0.001 for the A. centrale model) (Table 3.7).
Total protein
There was no direct association between serum total protein levels and A. marginale burden
(in either concurrent or predictive models) in either herd. In the free-ranging herd, higher serum total protein levels were associated with decreased concurrent A. centrale burden (p = 0.049)
(Table 3.8).
Table 3.8: Generalised linear mixed (concurrent) model predicting serum total protein concentrations in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 99 n = 99 Intercept 0.225 0.055 0.108 0.028 Season (Wet) 0.034 0.020 0.102 0.040 0.020 0.047 A. marginale burden -0.014 0.007 0.058 NA NA NA A. centrale burden -0.007 0.007 0.321 0.014 0.004 <0.001 Sex (Male) -0.038 0.033 0.254 0.014 0.040 0.737 Age 0.169 0.032 <0.001 0.175 0.031 <0.001 Co-infection Burden 0.002 0.001 0.025 NA NA NA Sex (Male): Anaplasma spp. burden NA NA NA -0.013 0.006 0.031
FREE-RANGING HERD n = 141 n = 141 Intercept -0.065 0.099 0.047 0.105 Season (Wet) 0.075 0.073 0.305 0.087 0.071 0.220 Anaplasma spp. burden -0.001 0.013 0.917 -0.026 0.013 0.049 Sub-herd (Lower Sabie) -0.322 0.108 0.004 -0.338 0.108 0.002 Age 0.277 0.153 0.073 0.267 0.151 0.079 Significant p-values (<0.05) are bold. SE – standard error. Co-infection burden stands for an interaction effect of A. marginale and A. centrale
In the managed herd, however, there was a significant interaction between concurrent A. marginale and A. centrale (co-infection burden) and total protein levels (p = 0.025) (Table 3.8).
60
Overall, there was an increase in A. centrale infection burden with both low and high A. marginale burdens (Figure 3.4). Anaplasma centrale appeared to have little effect on total protein when A. marginale was not present at all. There was also a significant interaction between sex and concurrent A. centrale burden and serum total protein concentration, with females appearing to have higher overall total protein levels and male total protein levels appearing to decrease at a greater rate with a higher A. centrale infection burden (p = 0.031)
(Figure 3.5; Table 3.8).
The predictive models for the managed herd showed that total protein levels increased with increasing A. centrale burden (p = 0.035) (Table 3.9).
Table 3.9: Generalised linear mixed (predictive) model for serum total protein concentration in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 76 n = 76
Intercept 0.144 0.034 0.100 0.026 Season (Wet) 0.038 0.026 0.145 0.055 0.025 0.036 Anaplasma spp. burden -0.003 0.004 0.484 0.007 0.003 0.035 Sex (Male) -0.034 0.035 0.336 -0.039 0.032 0.242 Age 0.109 0.033 0.002 0.112 0.302 <0.001
FREE-RANGING HERD n = 106 n = 106 Intercept -0.168 0.103 -0.158 0.117 Season (Wet) -0.001 0.099 0.995 -0.004 0.101 0.967 Anaplasma spp. burden 0.010 0.015 0.507 0.007 0.109 0.700 Sub-herd (Lower Sabie) -0.183 0.115 0.116 -0.169 0.109 0.128 Age 0.105 0.176 0.554 0.090 0.174 0.607 Significant p-values (<0.05) are bold. SE – standard error.
61
Figure 3.4: Total serum protein and infection intensity (log-transformed) of Anaplasma marginale (copies/reaction) and A. centrale in a managed herd of African buffalo (n = 99) from Kruger National Park. Anaplasma centrale burdens were converted into categories based on whether they were above or below the median of positive quantitative infection intensity, or had a negative reading. None = no infection, Low = infection intensity below the median, High = infection intensity above the median.
62
Figure 3.5: Total serum protein and infection intensity (log-transformed) of Anaplasma centrale (copies/reaction) based on sex in a managed herd of African buffalo (n = 226) from Kruger National Park. F = female, M = male.
63
Reproduction
There were no significant associations between pregnancy and burden of Anaplasma spp. in the free-ranging herd (Supplementary Table 3.9), and such data was not available for pregnancy in the managed herd. However, based on presence/absence of infection data in the managed herd, animals infected with A. centrale were less likely to be pregnant than those uninfected
(OR = 0.369; p = 0.012) (Table 3.10). The probability of being pregnant was not associated with the presence of A. marginale in either the managed or the free-ranging herd (Table 3.10).
Table 3.10: Mixed effects logistic regression model predicting pregnancy status in managed and free- ranging herds of African buffalo from Kruger National Park.
Coefficient SE p OR MANAGED HERD n = 294 Intercept 0.813 A. marginale presence 0.394 0.372 0.290 1.483 (-0.335 to 1.123) A. centrale presence -0.996 0.397 0.012 0.369 (-1.774 to -0.218) Age 0.806 0.336 0.017 2.239 (0.147 to 1.465) Sub-herd (Lower Sabie) NA NA NA NA
FREE-RANGING HERD n = 973 Intercept 0.128 A. marginale presence -0.134 0.172 0.435 0.874 (-0.471 to 0.203) A. centrale presence -0.152 0.170 0.368 0.858 (-0.185 to 0.181) Age 1.108 0.197 <0.001 1.700 (0.722 to 1.566) Sub-herd (Lower Sabie) 0.531 0.169 0.002 3.027 (0.200 to 0.862) Significant p-values (0.05) are bold. SE – standard error, OR – odds ratio.
There was a significant positive association between age and the odds of being pregnant in both herds (p = 0.012 and p < 0.001 for managed and free-ranging herd, respectively), and buffalo from the Lower Sabie sub-herd were more likely to be pregnant than those from the Crocodile
Bridge sub-herd (p = 0.002).
64
3.3.4 Survival
Neither the presence/absence of A. marginale nor A. centrale appeared to have any effect on survival in either the managed or the free-ranging herd (Table 3.11).
Table 3.11: Andersen-Gill survival model predicting survival outcome in managed and free-ranging herds of African buffalo from Kruger National Park.
Subjects Failed Coefficient (SE) p Hazard ratio (95%CI)
MANAGED HERD A. marginale presence 660 18 0.4555 (0.5246) 0.284 2.94 (0.69 to 3.62) A. centrale presence 660 18 0.5743 (0.5417) 0.201 3.55 (0.74 to 4.28) Age (years) 660 18 -0.5461 (0.4420) 0.309 1.46 (0.20 to 1.66) a Sex: Male 214 6 Reference 1.00 Female 446 12 -0.1595 (0.5163) 0.762 2.09 (0.30 to 2.39) Season: Wet 283 6 Reference 1.00 Dry 377 12 -0.6465 (0.5739) 0.315 1.70 (0.15 to 1.85)
R2 = 0.011 Likelihood ratio test 7.33; 5 df; p = 0.1972 a Interpretation: One unit increases in age at the time of sampling increased the 2-3 monthly hazard of failure by a factor of 1.46 (95% CI 0.20 to 1.66).
FREE-RANGING HERD A. marginale presence 1128 30 0.6988 (0.3736) 0.065 3.26 (0.96 to 4.22) A. centrale presence 1128 30 -0.3805 (0.3960) 0.341 1.18 (0.31 to 1.50) Age (years) 1128 30 -0.2499 (0.7789) 0.717 2.81 (0.20 to 3.01) b Sub-herd: Lower Sabie 541 10 Reference 1.00 Croc Bridge 587 20 -0.4332 (0.4170) 0.309 1.21 (0.28 to 1.49) Season: Wet 496 8 Reference 1.00 Dry 632 22 -0.7307 (0.4602) 0.108 0.98 (0.20 to 1.18)
R2 = 0.007 Likelihood ratio test = 8.11 on 5 df, p = 0.1503 b Interpretation: One unit increases in age at the time of sampling increased the six-monthly hazard of failure by a factor of 1.46 (95% CI 0.20 to 1.66). Significant p-values (<0.05) are bold. CI = Confidence Interval.
65
3.4 Discussion
Using both concurrent and predictive models, we could see how infections with A. marginale and/or A. centrale can affect African buffalo relatively quickly (concurrent), or after a period of months (predictive). The two herds also offered different insights into the interactions of A. marginale and A. centrale in the buffalo. The results from this study further suggest that, like cattle, African buffalo react differently depending on whether they are infected with A. marginale or A. centrale. Despite a lack of obvious clinical signs, there were significant associations between anaplasmosis and several indicators of general health, such as body condition, haematocrit and total proteins.
A loss of burden data was a frequent problem throughout many of the initial analyses which included both A. marginale and A. centrale in the models investigating co-infections and associations with pregnancy, as well as the survival analysis. The only burden model that had a significant co-infection term (after co-infection terms were forced into single infection models) was concurrent A. marginale infection with total protein. The lack of significant co- infection interactions in the fitness response models could thus be due to A. marginale and A. centrale presence or burden having no impact upon each other in the same host, or could be a result of that reduction in data.
3.4.1 Factors affecting Anaplasma burden
Younger buffalo were more likely to have higher A. centrale and A. marginale burdens, which may be related to a decrease in infection intensity over time because of acquired immunity
(Henrichs et al., 2016), which has previously been demonstrated in cattle as well (Aubry &
Geale, 2011; Kocan et al., 2010; Potgieter & Stoltz, 2004). The average age of initial A. marginale infection seen in the current study was approximately eight months, whereas it was
66
17 months for A. centrale infection. Based on this, it appears that most Anaplasma infections were not transmitted transplacentally, although five of the 15 calves tested positive for infection in their first capture after birth. Although transmission through ticks is thought to be the main mode of transmission, transplacental transmission cannot be excluded as a possible cause of infection in at least those five calves, and has been shown to occur in cattle, (Grau, et al., 2013;
Potgieter & Stoltz, 2004). However, the possibility of transplacental transmission was beyond the scope of this study, and would require data to be collected on calves at younger age then observed here. It is possible that calves are more likely to be infected with A. marginale first because they are more likely to be exposed to A. marginale. There was an overall infection presence with A. centrale of 40%, and 100% for A. marginale, by the time the calves turned six months of age. This could be the result of a higher presence or intensity of infection of A. marginale than A. centrale in the adult population, the latter but not the former of which is supported by our data. Ticks could also be more likely to transmit A. marginale than A. centrale; it is known that there are significant differences in the transmissability of rickettsial species and strains, including Anaplasma spp. (Ueti et al., 2007). However, the present study did not have the data available to test this. Alternatively, the order of infection could also indicate that A. marginale is facilitating A. centrale infections, however this would entail resource competition and potential displacement of the first species. The facilitation of a pathogen by another related pathogen to enter the host has been seen to occur with malarial pathogens and nematodes in birds (Clark, et al., 2016). However, to the best of our knowledge, this phenomenon has not previously been described for Anaplasma spp. in bovines.
Mean A. marginale infection burden also appeared overall to be higher than mean A. centrale burden. Little work has been done on quantifying A. marginale and A. centrale burdens to evaluate their interactions beyond vaccine studies against A. marginale, or for studies on
67 pathogenesis (Decaro et al., 2008). Anaplasma marginale infections may be higher due to their tendency to infect younger animals before their adaptive immune system has fully developed
(Aubry & Geale, 2011; Henrichs, et al., 2016; Kocan et al., 2010; Potgieter & Stoltz, 2004).
Anaplasma marginale may also be better at hiding from the immune system (Han, et al., 2010) and so has more time to replicate before detection, and could possibly replicate faster (Ueti, et al, 2007), however more work needs to be completed to determine the reasons for this result.
An increase in the infection intensity of A. centrale in buffalo from the managed herd during the dry season could be related to alterations to immune function due to altered nutritional intake, as seen in experiments with bobwhite chicks with lower protein diets resulting in poorer immune responses (Lochmiller et al. 1993). Previous studies of the free-ranging herd have shown that alterations in immune function can also occur with buffalo with bovine tuberculosis, in which decreased innate immunity and an increased pro-inflammatory immune response have been reported (Beechler et al., 2015; Jolles et al., 2015). Other studies have reported increased levels of gastrointestinal parasite infection in African buffalo, with higher faecal egg counts in the dry season or drier conditions (Ezenwa, 2004). In addition, being more restricted in their ability to select optimal grazing habitat during periods of limited resources than the free- ranging herd (Spaan, 2015), buffalo in the managed herd may be less able to avoid areas with high tick numbers, and thus Anaplasma spp. Another possible explanation is the tendency for buffalo (from both herds) to form larger herds during that time because of decreased grazing choice and water availability (Prins, 1996; Spaan, 2015), which results in greater contact, potentially leading to more frequent tick infestation of buffalo.
Males from the managed herd were infected with higher burdens of A. centrale than females.
While there may be a sex-bias in infection intensity, it is important to consider the skewed age
68 of males in the managed herd. Actual sex-bias is difficult to interpret in the managed herd, as males are not able to act naturally with their usual dispersal patterns creating sub-adult to adult bachelor herds, and sexually mature males are underrepresented due to the small herd size. To our knowledge, a male bias in Anaplasma spp. infection has not previously been reported. This sex-based susceptibility to Anaplasma infections could be due to the greater exposure faced by the males, as they tend to be on the edges of the herd where they are more likely to encounter ticks on vegetation (Prins, 1996), or immunosuppression. Various studies have demonstrated a significant suppressive effect of testosterone on immune function in a range of mammal species
(Foo et al., 2017). The relative value of either of these explanations for the observed relationship cannot be confirmed based on our data, however there is scope for this to be examined in these herds in the future.
Animals associated with the free-ranging Lower Sabie sub-herd were more likely to have a higher infection intensity of A. marginale than those from Crocodile Bridge. Previous work has shown that they also tended to be in overall better condition due to better resource availability in the Lower Sabie area (i.e. more consistent nutritional value of feed, feed availability, and access to water sources) (Henrichs, 2014; Spaan, 2015), although that was not apparent in our models. Better body condition may increase the ability of buffalo to tolerate an
Anaplasma spp. infection, as buffalo in better condition may have the resources to continue replacing red-blood cells rather than mounting a strong immune response to the infection, meaning the pathogen’s replication is not hindered by the host. It has previously been shown that, in certain circumstances, such tolerance responses may be more beneficial to the host than a strong immune response (Fenton & Perkins, 2010). For example, laboratory mice infected with rodent malaria have been found to develop variable and competing tolerance and
69 resistance responses measured by anaemia and weight-loss, and with a hereditary component
(Råberg et al., 2007).
The present study provides evidence of delayed, but not concurrent competition between the two Anaplasma species investigated. In the managed herd, a higher burden of A. centrale was associated with a higher concurrent burden of A. marginale. This is consistent with previous work on African buffalo which found no negative relationship between the presence of A. marginale and A. centrale (Henrichs et al., 2016). Similarly, Shkap et al. (2008) demonstrated experimental co-infection in cattle with A. centrale and A. marginale. Henrichs et al. (2016) suggested that a lack of resource competition between these species could be due to the resupply rate of erythrocytes exceeding the use of parasites, and the shared evolutionary history of the two species being shaped by invasion of the same host. Multiple studies on anaplasmosis in buffalo also indicated that A. centrale is more likely to be detected by the immune system than A. marginale, the latter of which is better at changing antigen presentation (Aubry &
Geale, 2011; Henrichs, 2014; Kocan et al., 2010). This appears to result in an increase in co- occurrence of the two species, possibly due to A. centrale infecting the same cells as A. marginale to avoid detection and persist within the host (Henrichs et al., 2016). There was a lack of reciprocity between concurrent co-infection results (with the A. marginale single infection model showing no significant association with A. centrale, but the A. centrale model showing a significant association with A. marginale), which may have been due to a difference in the subset of animals after model selection is imposed and co-infection terms were forced in. The results from the present study thus did not detect downward regulation of concurrent infection intensity of either species by the other. However, it is possible, that other co-infection effects are seen after these parasites have had time to become established in the host and cause a host response.
70
There was, in fact, evidence from the predictive model in the managed herd, that Anaplasma marginale and A. centrale eventually compete within the host. This relationship appears only to be apparent if examined within an appropriate time-frame, i.e. neither too far apart (as for the predictive model in the free-ranging herd, with 6-monthly sampling intervals) nor too close together (as for the concurrent models). This is the first report of a negative correlation between
A. marginale and A. centrale in African buffalo. Most work on A. marginale and A. centrale interactions have been in cattle, investigating cross-immunity through immunisations (whether natural or through a vaccination scheme). For example, Molad et al. (2004) suggested that the cross-protection against A. marginale due to A. centrale was a result of the conserved surface proteins that are shared between A. marginale and A. centrale; however, work on the interactions between the two species and their infection burdens is limited.
3.4.2 Effects of Anaplasma infection on health parameters
The present results suggest that there are significant relationships between concurrent
Anaplasma burden and body condition; however, these are complex and may not be readily explained and are further complicated when considering seasonal impacts. Better body condition in animals with higher Anaplasma burdens could be the result of the parasite being more likely to successfully invade the cells of otherwise healthy hosts, and is consistent with previous work on the free-ranging herd, which showed that body condition was positively associated with the presence of A. centrale (Henrichs, 2014). There could be several possible explanations for a positive association between Anaplasma infection and body condition; it could be the result of (i) A. centrale protecting buffalo hosts against the effects of infection with the more pathogenic A. marginale, as is the case in cattle (Aubry & Geale, 2011; Dalgliesh et al., 1990; Kocan et al., 2010; Potgieter & Stoltz, 2004) and/or (ii) a higher availability of red
71 blood cells in animals with better condition for Anaplasma to invade (Ezenwa & Jolles, 2008;
Prins, 1996).
The seasonal relationships with body condition may be a consequence of the subsets of the models which resulted in data being more concentrated in different times of the seasons, potentially skewing the results. The seasonal variation in the relationship between body condition and infection in the free-ranging herd, particularly the negative association found during the wet season, is in contrast to the findings by Henrichs (2014), which could be a result of this skew due to the model subsets. Biological reasons for the findings presented here could be a result of the combination of a delay in recognisable impacts on body condition after infection and the 6-monthly sampling pattern in that herd, with the effect of season on impact of infection only being recognised in the next sampling period. There could also be seasonal variation in infection exposure, potentially resulting in higher tick burdens in the wet season
(Anderson et al., 2013). Tick burden has previously been shown to have no association with haemoparasitic infection in African buffalo, however, a potential seasonality in infectivity of ticks in this host species has not been investigated (Henrichs, 2014). Reproductive cycles, immunity changes or differences in exposure can also affect seasonality, and so interpreting seasonal infection interactions can be difficult.
Surprisingly, there was no direct indication of anaemia (as measured by haematocrit) associated with the burden of infection of either A. marginale or A. centrale; however, there were positive associations between haematocrit and the previous capture’s burden of A. marginale in the managed herd. The actual intensity of infection with Anaplasma in African buffalo described here appears to be relatively similar to what has previously been found in cattle, indicating that buffalo have a significantly different host response compared to cattle, given the same level of infection (Decaro, et al., 2008; Kieser et al., 1990). A lack of anaemia could be one of the
72 reasons Anaplasma spp. infection is only subclinical in buffalo, but clinical in cattle; cattle acutely infected with A. marginale infection can become severely anaemic due to the destruction of red blood cells (Aubry & Geale, 2011; Potgieter & Stoltz, 2004). However, previous work with A. marginale infections and immunisations in cattle also indicated that a reduction in rickettsaemia did not result in less severe anaemia, signifying that the presence of
A. marginale infection, not burden, may be associated with red blood cell volume (Palmer et al., 1994). If this is the case, A. marginale is likely to have the same impact on the host’s red blood cell count regardless of the infection intensity. Increases in haematocrit are usually due to either a relative or absolute increase in red blood cell numbers (polycythaemia) or dehydration (e.g. Clark, 2004). It is possible that, in contrast to cattle, buffalo respond with polycythaemia to infection with A. marginale, and such a response could in fact by explained by two alternate hypotheses. The cyclic nature of Anaplasma spp. infections is known to result in fluctuating red blood cell levels in cattle in response to the destruction of infected cells by the invasion of the pathogen and the host’s own immune response (Aubry & Geale, 2011).
Increasing haematocrit levels associated with a greater previous infection intensity of A. marginale could thus be an indication of a delay in the red blood cell production after a severe
A. marginale infection in buffalo (Roland et al., 2014). Alternatively, it could suggest a tolerance response, where an animal makes more red blood cells rather than clearing the infection to allow oxygen to be distributed in the body even though some of those cells will be infected (Henrichs, 2014). Tolerance responses have been seen in other pathogen-host systems, such as with Human Immunodeficiency Virus (HIV) in humans (Regoes et al., 2014). That study showed that the alleles of younger people appeared to code for an increase in tolerance of an HIV infection, resulting in continual replacement of CD4+ T lymphocytes, as opposed to the more standard situation with HIV infection where CD4+ T lymphocyte cells decline leading to disease and even death (Regoes et al., 2014). However, the increase in haematocrit observed
73 in the present study could also have been associated with some level of dehydration.
Anaplasmosis is well known to affect the red blood cells of affected cattle (e.g. Aubrey &
Geale, 2011), however, based on the data available in this study, it is not possible to definitely differentiate between the potential causes of increased haematocrit as a result of A. marginale.
Future work in this area could investigate the relationship between (MCV) and A. marginale infection to clarify if there is an increase in red blood cell size in animals with affected haematocrit levels, and include reliable red blood cell counts.
The managed herd showed higher serum total protein concentrations with an increase in A. centrale, using both the concurrent and previous infection data. This indicates that total protein is both immediately higher with a higher A. centrale infection, as well as after a time delay between infection and total protein response, allowing the infection to become established, and then affect the host. There are numerous reasons for higher total protein in animals, including higher protein intake, chronic infection and autoimmune disease (Couch et al., 2017). A higher level of total protein could thus be an indication of the buffalo’s immune response to being infected with A. centrale, particularly if the increase is due to increased immunoglobulin and cytokines, as has been described for in Palacios et al. (2014). A study of calves experimentally infected with A. marginale demonstrated rising IgG levels approximately three weeks after inoculation, accompanied by increased serum total protein concentrations (Dimopoullos et al.,
1960). The lack of relationship between total protein and A. marginale in African buffalo, however, could indicate that A. marginale may be evading the immune system in this host, or that the preferable host strategy is a tolerance response of increasing red blood cell production rather than mounting an immune response (as discussed earlier). Unfortunately, it was not within the scope of this study to assess haptoglobin, fibrinogen, albumin and immunoglobulin levels separately, nor to calculate the albumin: immunoglobulin ratio, so it is not possible to be
74 certain that the increase in total serum protein concentrations was due to an immune response.
The present results do, however, indicate that such an investigation would be a valuable consideration for future studies. Considering the positive association between A. centrale and body condition discussed earlier, there could alternatively be an unexplained relationship between A. centrale burden, dietary protein intake and body condition, which would require further investigation. The interaction between sex and total protein may again be an indication of immunosuppression via sex hormones (Foo et al., 2017), with a limited total protein response from males when infected with higher burdens of A. centrale, however testing this was again outside the scope of this study.
The managed herd showed that the presence of A. centrale was associated with a lower probability of pregnancy. Previous work has shown no impact of the presence of infection with
A. centrale on pregnancy status; however, there was a decreased likelihood with the presence of A. marginale (Henrichs, 2014). The reasons for this difference in findings are not clear; however, the decreased likelihood of A. centrale infection in pregnant animals may be because buffalo are less likely to fall pregnant if they have an active infection, though in other species abortions and temporary infertility usually occurs only with an A. marginale infection (Aubry
& Geale, 2011; Potgieter & Stoltz, 2004). The present study did not find any evidence for an association between pregnancy and infection with A. marginale. No significant relationship was found between infection burden of either Anaplasma spp. and the probability of being pregnant. This suggests that any relationships between anaplasmosis and pregnancy in African buffalo are related to presence, rather than intensity of infection.
Previous work with Anaplasma spp. infections in buffalo has shown that in the dry season, host condition and infection with A. centrale was associated with a higher probability of survival
75
(Henrichs, 2014). However, in the present study, the presence of A. marginale and A. centrale did not appear to affect survival. These results are, however, not altogether surprising as there is no existing evidence of mortalities of African buffalo ascribed to A. marginale or A. centrale infections (Berggoetz et al 2014; Kuttler 1984).
Future work should investigate in more detail the age dynamics of Anaplasma spp. infection, to explain why younger buffalo are more likely to have higher infection intensity. This could be achieved through the analysis of immune responses using the presence of cytokines which could also help to confirm that elevated total protein levels were a result of inflammation. Tick burden (i.e. number of ticks on a single buffalo host), prevalence of Anaplasma spp. in the different tick species affecting African buffalo, and seasonal variations in the infectivity of ticks are also important factors to assess if exposure to infectious agents through tick transmission is what drives infection dynamics. The negative correlation between A. marginale and A. centrale in the predictive co-infection analyses indicates that there may be eventual competition between the two species, which is also an important factor to investigate further.
In cattle, cross-immunity from A. centrale against A. marginale has been utilised in the form of a live-vaccine to minimise the severity of infection with A. marginale, and so to minimise the incidence of anaplasmosis (Aubry & Geale, 2011; Dalgliesh et al., 1990; Kocan et al., 2010;
Potgieter & Stoltz, 2004). A similar mechanism may be also present in buffalo; however, based on the data from the present study, it is not clear whether infection with A. centrale is less costly to African buffalo than with A. marginale. The presence of co-infection relationships in combination with age-related effects does suggest a role for acquired immunity and the potential for cross-immunity.
76
Chapter 4 – General Discussion
This thesis investigated the molecular characterisation of Anaplasma spp. from African buffalo in Kruger National Park, South Africa. DNA sequences of A. marginale (msp1β gene) and A. centrale (groEL gene) obtained in this study were compared with previously published sequences of these species using different phylogenetic methods. In addition, this thesis examined infection and co-infection dynamics of A. centrale and A. marginale within African buffalo, and the relationships between the infection intensities of Anaplasma spp. and a range of health parameters.
4.1 Livestock: wildlife interface
The results from the present study confirm that African buffalo are an important reservoir of
Anaplasma, at least in the Kruger National Park region, with 84% of individuals sampled in the managed herd testing positive for Anaplasma infection at least once over the two-and-a- half-year study period, and an overall sample presence of approximately 17% for A. marginale and 13% for A. centrale (see Chapter 2). However, to determine whether transmission between
African buffalo and cattle does occur, it is also important to examine whether they are infected with the same strain. If movement restriction plans are to be incorporated into a livestock: wildlife management plan to minimise the spread of infections, as is currently the case for
African buffalo in certain areas of South Africa, it is essential to incorporate all known animals involved in the spread of this disease (De Garine-Wichatitsky et al., 2013; Gallivan & Horak,
1997; Gortazar et al., 2014). The phylogenetic analysis of A. marginale and A. centrale sequences presented in Chapter 2 showed that the sequences from African buffalo examined grouped separately from sequences from other geographic areas in the world. Molecular methods, such as single nucleotide polymorphisms (SNPs), or sequencing, provide a specific
77 measure of the connectedness of infections between different populations of wildlife and domestic animals, through genetic comparison of the infection itself. This method has been used to assess the relatedness of Anaplasma and Ehrlichia spp. infections in an array of wildlife species in the Brazilian Amazon (Soares et al., 2017). The results from the present study indicate that while there is enough similarity between the sequences from the African buffalo and those from cattle from different areas of the world (there were no other sequences available for buffalo) to be identified as A. marginale or A. centrale, there are also some differences indicating geographical differences in sequences of these infections. To fully understand the role of African buffalo in facilitating Anaplasma spp. infection of nearby livestock populations, future work should look at sequencing samples from cattle near the buffalo sample populations to assess their similarity. The results could then be used to assess whether the double-fence currently separating the buffalo population in KNP from primary cattle production areas is effective in preventing transmission events between the two species. The determination of how and where Anaplasma infections are being transmitted between different host species will allow for optimal management decisions at the wildlife: livestock interface.
4.2 Order of infection
This study has presented new information regarding the co-infection relationship between A. marginale and A. centrale. There appears to be an order of infection with Anaplasma in the buffalo, with A. marginale infection occurring first and with less variability around age of first infection than A. centrale (Chapter 3). The order in which infections successfully invade a host has previously been shown to influence the ability of new infectious agents to penetrate the same host’s defences (Thomas et al., 2003). For example, mucous membrane lesions caused by endoparasites, when entering a host, like the herpes simplex virus type-2 (HSV-2), have subsequently allowed other parasites (HIV infection in this case) to invade the host more easily
78
(Vaumourin, et al., 2015). The findings presented in the present study may be an indication of
A. marginale infections in buffalo facilitating A. centrale infections, however higher co- occurrence might be expected to be observed if this were the case. The Allee effect could explain this facilitation as seen in a study on the survival of ants infected with Metarhizium spp., where there is an invasion threshold for an infection to be successful (Hughes et al., 2004).
There also may be a genetic component to this facilitation. Some of the genetic differences between A. marginale and A. centrale occur in the major surface protein 1 alpha (msp1α) and beta (msp1β) genes, which are genes affiliated with bovine and tick erythrocyte adhesion, or just the bovine erythrocytes, respectively (Bowie, et al., 2002; Cabezas-Cruz & de la Fuente,
2015; Tamekuni et al., 2009; Zivkovic, et al., 2007). These differences in sequence could give
A. marginale an advantage in successfully invading a host, with A. centrale potentially favouring a certain level of the invasion threshold to be reached by the closely related A. marginale to facilitate its invasion (Brown, et al., 2001).
4.3 Co-infection dynamics
In addition to A. marginale appearing to facilitate infection with A. centrale in the buffalo host, current burden of A. centrale in the buffalo was also positively correlated with current burden of A. marginale (Chapter 3). This correlation further supports the hypothesis that infection with
A. marginale increases the ability of A. centrale to successfully persist within the buffalo host
(Chapter 3). A review on co-infection of the tick-borne pathogens Babesia microti and Borrelia burgdorferi similarly showed that the two pathogens were positively correlated, with co- infection potentially providing a survival advantage for both (Diuk-Wasser, et al., 2016). An infectious agent can aid in the penetration and persistence of another infectious agent within the host, for instance, by reducing resistance of the immune system to infections by stretching
79 the immune system resources by responding to a diverse array of pathogens and parasites
(Cattadori, et al., 2008; Cox, 2001).
While the data on concurrent burdens presented in Chapter 3 suggested that infections of A. marginale and A. centrale were positively correlated with each other, a different relationship was observed when allowing for a time delay, i.e. examining the inter-species associations between current and previous A. marginale and/or A. centrale infections. Infectious agents can be negatively correlated with each other, through cross-immunity, immune suppression or resource competition (Bordes & Morand, 2011; Cox, 2001). In the case of Anaplasma co- infection in African buffalo, there appears to be a time-lag in interactions between the two species, with higher previous infection burden of one species being correlated with lower subsequent infection burden of the other (Chapter 3). A negative correlation between the two species indicates that they eventually start to compete in the host. Future work can involve distinguishing if this negative interaction is a result of resource competition and/or cross- immunity through using immunological measures, such as presence of different cytokines, to confirm if the immune system has been activated.
Co-infections can also change host response in terms of health and fitness factors, for instance, in koalas infected with chlamydiosis, koala retrovirus and/or bone marrow disease had significantly lower body condition when coinfected with Trypanosoma gilletti compared to when there was no T. gilletti co-infection (McInnes, et al., 2011). Multiple infectious agents within the one host are an important consideration when working with wildlife, and further work is needed to further estimate the full effects of co-infection between Anaplasma spp., as well as Anaplasma spp. with other, non-related infections.
80
4.4 Effects of age on infections
Younger African buffalo had higher burdens of infection than older animals, however their age did not appear to affect their response to infection with Anaplasma spp. (Chapter 3). The age at first infection can have a serious impact upon the severity of the infection on the host (Aubry
& Geale, 2011; Kocan et al., 2010; Potgieter & Stoltz, 2004), but the results from the present study support previous findings that buffalo of any age appear to display few clinical signs.
This differs quite significantly from cattle, where animals younger than two tend to be less susceptible to the disease, infection in cattle under six months of age rarely results in illness, from six months to a year occasionally results in a mild form of the disease, and from one to two years results in an acute but rarely fatal disease (Aubry & Geale, 2011; Kocan & de la
Fuente, 2003). Anaplasma marginale infection in cattle over the age of two can result in death of the animal (Aubry & Geale, 2011; Potgieter & Stoltz, 2004).
The finding of greater Anaplasma burdens in younger African buffalo in the present study supports the idea of acquired immunity to Anaplasma spp. occurring in this host species. This is also seen in cattle infected with Anaplasma spp., where innate immunity and maternal antibodies in the colostrum further minimise the negative impact upon the calf’s health until its immune system develops an immune response to the infection, either through exposure to the infection naturally, or through vaccination (Dalgliesh, et al., 1990). It is not known whether such immune-transfer occurs in African buffalo. Although domesticated species may share infections with wildlife, these differences in host health exemplify the need to specify the effects of infections related to specific hosts and to not assume that extrapolating what is known about domesticated animals will fit with what happens with wildlife species.
81
4.5 Seasonal variance
The study presented here suggests that the dry season, which is a time period during which food and water resources become limited, overall may be associated with higher burdens of
Anaplasma spp. (Henrichs, et al. 2016; Spaan, 2015). Seasonal variation can affect host animals with impacts on available food quantity and quality, as well as non-nutritional causes such as seasonality in vectors and reproduction (Spaan, 2015). For example, a study on gastrointestinal infections in moose (Alces alces) showed a decline in moose carcass weight over 20 years, with reasons including a decline in forage quality and warmer summers, which resulted in a reduction in food availability and quality (Davidson et al., 2015). African buffalo tend to be found in small, closely-related groups in the wet season, when resources are more plentiful, and cluster together into larger herds in the dry season when water sources are more limited
(Prins, 1996). The results from the present study thus might be explained by buffalo converging into fewer, but bigger, herds in the dry season, where infections may be spread more readily between animals.
Unfortunately, the assessment of tick burdens and prevalence of Anaplasma infection in ticks was outside the scope of the present study. However, the seasonality of vectors is likely to be highly important when considering seasonal variation of vector-borne haemoparasites such as
Anaplasma spp. Changes in host connectivity due to season can influence the patterns and frequency of transmission of infections between individuals (Eygelaar et al., 2015) and seasonal variation has further been incriminated in influencing distribution and abundance of infective stages of a vector, and the vector itself (Forbes et al., 1994). West Nile Virus has seasonal transmission to greater sage-grouse (Centrocerus urophasianus) during the winter in the U.S., related to the environmental conditions affecting the mosquito vector (Dusek et al.,
2014). Risk factors for tick-borne diseases may include tick vector distribution, population
82 density of hosts, host resistance to both ticks and the tick-borne parasites, the age of the host, and host movement or migratory patterns (Eygelaar et al., 2015). Other studies have demonstrated that higher burdens of Rhipicephalus and Amblyomma ticks occur in the free- ranging herd in the wet season (Henrichs, 2014), however, based on the results in the present thesis, this is not reflected by higher Anaplasma spp. burden results at that time of year (Chapter
3). The infectivity of ticks during seasonal fluctuations may thus also offer more insight into the seasonality of anaplasmosis. Ticks are perhaps the most important factor for disease control as they are vectors of disease to non-human vertebrates, and so further investigation into their impact upon spreading diseases like anaplasmosis remain essential (Kocan, 1995).
Domesticated animals are more likely to be buffered from seasonal variation than free-ranging wildlife species, which must manage infections with the added stress of seasonal changes, such as decreases in food availability, increases in vector burdens and seasonal breeding patterns
(Kock et al., 2014). Research into the infections of wildlife can allow insight into how infections affect hosts throughout varying levels of stress in response to the variance in resource availability and other seasonal factors.
4.6 Fecundity and survival
The results on fecundity from the present study contrast with previous work on the free-ranging herd, which found A. marginale-positive animals to be less likely to conceive (Henrichs et al.,
2016) (see Chapter 3). On the other hand, the findings from the managed herd indicate that animals with A. centrale infections were less likely to become pregnant (see Chapter 3), which is similarly at odds with previous work in buffalo (Henrichs et al., 2016). The lack of effect on fecundity in these buffalo is in contrast with A. marginale infections being correlated with abortions, temporary infertility and anoestrous in cattle (Aubry & Geale, 2011; Kocan et al.,
83
2010; Potgieter & Stoltz, 2004). This is not altogether surprising, as even though haematological and serological measures (haematocrit and total protein) were shown to have a significant association with Anaplasma spp. infections in buffalo, the effects were relatively small and the infection remains subclinical.
In the present study, the survival of African buffalo did not appear to be impacted by
Anaplasma spp. infection (see Chapter 3), a result which differs from previous work on buffalo which suggested that host condition, A. centrale infection and the dry season were correlated with a higher probability of survival (Henrichs, 2014). The lack of mortalities in response to
Anaplasma spp. infection is, however, perhaps not entirely surprising, as no directly
Anaplasma-related fatalities have been reported in this host in the past (Berggoetz et al., 2014;
Kuttler, 1984).
4.7 Tolerance versus resistance host strategies
African buffalo appeared to be using a different strategy for responding to A. marginale than to A. centrale infections. Infection with A. marginale was associated with an eventual increase in haematocrit, while infection with A. centrale was correlated with higher total protein levels
(see Chapter 3). Host response to infection can vary depending on what resources are available at the time. The increase in haematocrit as a response to a parasite invading red blood cells could represent a tolerance strategy where the host continually replaces infected cells rather than mounting an immune response to the infection. On the other hand, based on the present data, is not possible to differentiate between such a response and dehydration or increased red blood cell counts as the cause of the increased haematocrit observed. It would be necessary to perform reliable red blood cell counts to be able to further explain this change, which, unfortunately, was not the case in the present study. Nevertheless, the increase in haematocrit
84 observed in response to A. marginale was greater in the Lower Sabie sub-herd, which had better access to resources and were better able to move around the park in response to seasonal food availability than either the other free-ranging sub-herd or the managed herd (Spaan 2015).
These buffalo also tended to have higher body condition, indicating that they had more resources able to manage extra stressors, such as a combination of poor feed and higher pathogen diversity and intensity. This would suggest that dehydration is a less likely cause of haematocrit increase, at least in this sub-herd.
The increase in total protein in response to infection with A. centrale could be the result of an immune response; rising serum total protein levels were also reported as a result of increasing globulin concentrations in calves infected with A. marginale (Dimopoullos et al., 1960). One possible explanation for the observed difference in host response to the different Anaplasma spp. is that infection with A. centrale in African buffalo might carry with it greater costs for the host than infection with A. marginale, in contrast to what happens in cattle (Aubry & Geale
2011; Potgieter & Stoltz, 2004). This would mean that it would could be more beneficial in the long term to mount an immune response to the former, while infection with the latter could be countered by up regulation of the production of affected blood cells. However, the current results do not support any significant costs to buffalo infected with A. centrale, and more work to characterise the immune, or tolerance, response of African buffalo to this pathogen is needed to investigate this hypothesis.
The differences in host response, i.e. mortalities in cattle and mild-to-no clinical signs in buffalo, indicate that wildlife and livestock could be managing infections in different ways.
This again emphasises that results from infections in livestock should not automatically be assumed to be relevant to wildlife.
85
4.8 Future work
Future research with A. marginale and A. centrale should involve an investigation of how infection intensity with each species interacts with other important infections, such as bovine tuberculosis (BTB), brucellosis and foot-and-mouth disease, as well as other parasites such as gastrointestinal helminths. Previous research on the free-ranging herd has investigated the interactions between gastrointestinal helminths and BTB, caused by the bacterium
Mycobacterium bovis (Ezenwa & Jolles, 2015). A comparison between BTB-positive buffalo that had been treated with anthelminthics and those that had no treatment revealed that the treated buffalo lived longer and became ‘super spreaders’, transmitting BTB to a higher number of buffalo than those from the non-treated group (Ezenwa & Jolles, 2015).
Additionally, more work needs to be undertaken on vector ecology and infectivity, to determine what conditions increase the likelihood of Anaplasma spp. infection being transmitted by ticks.
Host strategies in response to infection also require further research, as it currently appears that infections with A. marginale or A. centrale, although being very closely related pathogens, evoke different host responses. Unravelling these differences may resolve the different reactions to the infections seen in cattle.
4.9 Conclusion
African buffalo are affected by Anaplasma spp. infections, and the work presented here supports previous suggestions that infections remain subclinical. While Anaplasma spp. infections cause mortalities in cattle, the effect on buffalo is not as clear. There are clear impacts upon condition and blood measures, however, they are small and do not appear to influence fecundity or survival. The data from this study also re-emphasises the importance of examining
86 the impact of infection as the result of a combination of external and internal factors (e.g. co- infections, host age, location, season and host strategies i.e. tolerance versus resistance), all of which significantly appeared to affect the response of African buffalo to infection with
Anaplasma spp.
The findings presented in this thesis significantly contribute to our understanding of the health effects of a subclinical infection in a wildlife species, the African buffalo. The ability to more closely monitor and sample wildlife will provide us with more confidence in deciding whether infections are truly subclinical, or do in fact present difficulty in detecting negative health effects.
87
References
Akaike, H., Petrov, B., & Csaki, F. (1973). Information theory and an extension of the
maximum likelihood principle. Second International Symposium on Information
Theory, Akademiai Kiado, Budapest, 267-81.
Anderson, K., Ezenwa, V. O., & Jolles, A. E. (2013). Tick infestation patterns in free ranging
African buffalo (Syncerus caffer): Effects of host innate immunity and niche
segregation among tick species. International Journal for Parasitology: Parasites and
Wildlife, 2, 1-9.
Aubry, P., & Geale, D. W. (2011). A review of bovine anaplasmosis. Transboundary and
Emerging Diseases, 58, 1-30.
Bartlett, S. L., Abou-Madi, N., Messick, J. B., Birkenheuer, A., & Kollias, G. V. (2009).
Diagnosis and treatment of Babesia odocoilei in captive reindeer (Rangifer tarandus
tarandus) and recognition of three novel host species. Journal of Zoo and Wildlife
Medicine, 40, 152-159.
Beechler, B., Jolles, A., & Ezenwa, V. (2009). Evaluation of hematologic values in free-
ranging African buffalo (Syncerus caffer). Journal of Wildlife Diseases, 45, 57-66.
Beechler, B. R., Manore, C. A., Reininghaus, B., O'Neal, D., Gorsich, E. E., Ezenwa, V. O., &
Jolles, A. E. (2015). Enemies and turncoats: bovine tuberculosis exposes pathogenic
potential of Rift Valley fever virus in a common host, African buffalo (Syncerus
caffer). Proceedings of the Royal Society of London B: Biological Sciences, 282,
20142942.
Berggoetz, M., Schmid, M., Ston, D., Wyss, V., Chevillon, C., Pretorius, A. M., & Gern, L.
(2014). Tick-borne pathogens in the blood of wild and domestic ungulates in South
Africa: interplay of game and livestock. Ticks and Tick-borne Diseases, 5, 166-175.
88
Bicknese, E. J., Childress, A. L., & Wellehan Jr, J. F. (2010). A novel herpesvirus of the
proposed genus Chelonivirus from an asymptomatic bowsprit tortoise (Chersina
angulata). Journal of Zoo and Wildlife Medicine, 41, 353-358.
Boes, K. M., Goncarovs, K. O., Thompson, C. A., Halik, L. A., Santos, A. P., Guimaraes, A.,
Feutz, M. M., Holman, P. J., Vemulapalli, R., & Messick, J. B. (2012). Identification
of a Mycoplasma ovis-like organism in a herd of farmed white-tailed deer (Odocoileus
virginianus) in rural Indiana. Veterinary Clinical Pathology, 41, 77-83.
Bordes, F., & Morand, S. (2011). The impact of multiple infections on wild animal hosts: a
review. Infection Ecology and Epidemiology, 1, 7346-7356.
Bossart, G. D., Reif, J. S., Schaefer, A. M., Goldstein, J., Fair, P. A., & Saliki, J. T. (2010).
Morbillivirus infection in free-ranging Atlantic bottlenose dolphins (Tursiops
truncatus) from the Southeastern United States: seroepidemiologic and pathologic
evidence of subclinical infection. Veterinary Microbiology, 143, 160-166.
Bowie, M. V., de la Fuente, J., Kocan, K. M., Blouin, E. F., & Barbet, A. F. (2002).
Conservation of major surface protein 1 genes of Anaplasma marginale during cyclic
transmission between ticks and cattle. Gene, 282, 95-102.
Bradley, C. A., & Altizer, S. (2007). Urbanization and the ecology of wildlife diseases. Trends
in Ecology and Evolution, 22, 95-102.
Brenes, R., Gray, M. J., Waltzek, T. B., Wilkes, R. P., & Miller, D. L. (2014). Transmission of
ranavirus between ectothermic vertebrate hosts. PloS One, 9, e92476.
Brown, W. C. (2012). Adaptive immunity to Anaplasma pathogens and immune dysregulation:
Implications for bacterial persistence. Comparative Immunology, Microbiology and
Infectious Diseases, 35, 241-252.
Brown, W. C., Palmer, G. H., Lewin, H. A., & McGuire, T. C. (2001). CD4+ T lymphocytes
from calves immunized with Anaplasma marginale Major Surface Protein 1 (MSP1), a
89
heteromeric complex of MSP1a and MSP1b, preferentially recognize the MSP1a
carboxyl terminus that is conserved among strains. Infection and Immunity, 69, 6853-
6862.
Butler, T. (2013). Plague gives surprises in the first decade of the 21st century in the United
States and worldwide. The American Journal of Tropical Medicine and Hygiene, 89,
788-793.
Cabezas-Cruz, A., & de la Fuente, J. (2015). Anaplasma marginale major surface protein 1a:
a marker of strain diversity with implications for control of bovine anaplasmosis. Ticks
and Tick-borne Diseases, 6, 205-210.
Carelli, G., Decaro, N., Lorusso, A., Elia, G., Lorusso, E., Mari, V., Buonavoglia, C. (2007).
Detection and quantification of Anaplasma marginale DNA in blood samples of cattle
by real-time PCR. Veterinary Microbiology, 124, 107-114.
Carmichael, I., & Hobday, E. (1975). Blood parasites of some wild bovidae in Botswana.
Onderstepoort Journal of Veterinary Research, 42, 55-62.
Cattadori, I. M., Boag, B., & Hudson, P. J. (2008). Parasite co-infection and interaction as
drivers of host heterogeneity. International Journal for Parasitology, 38, 371-380.
Ceci, L., Decaro, N., Lorusso, E., Paradies, P., Elia, G., Martella, V., Carelli, G. (2008). First
report of bovine anaplasmosis by Anaplasma centrale in Europe, molecular
identification and phylogenetic analysis. Veterinary Research Communications, 32,
263-266.
Clark, P. (2004). Haematology of Australian Mammals. Melbourne, CSIRO Publishing.
Clark, N. J., Wells, K., Dimitrov, D., & Clegg, S. M. (2016). Co-infections and environmental
conditions drive the distributions of blood parasites in wild birds. Journal of Animal
Ecology, 85, 1461-1470.
90
Corona, B., & Martínez, S. (2009). Differences and similarities between Anaplasma marginale
and Anaplasma phagocytophylum. Revista de Salud Animal, 31, 1-7.
Couch, C. E., Movius, M. A., Jolles, A. E., Gorman, M. E., Rigas, J. D., & Beechler, B. R.
(2017). Serum biochemistry panels in African buffalo: defining reference intervals and
assessing variability across season, age and sex. PloS One, 12, e0176830.
Cox, F. (2001). Concomitant infections, parasites and immune responses. Parasitology, 122,
S23-S38.
Craig, B., Tempest, L., Pilkington, J., & Pemberton, J. (2008). Metazoan-protozoan parasite
co-infections and host body weight in St Kilda Soay sheep. Parasitology, 135, 433-441.
Dalgliesh, R., Jorgensen, W., & de Vos, A. J. (1990). Australian frozen vaccines for the control
of babesiosis and anaplasmosis in cattle – a review. Tropical Animal Health and
Production, 22, 44-52.
Darriba, D., Taboada, G., Doalla, R., & Posada, D. (2012). jModelTest 2: more models, new
heuristics and parallel computing. Nature Methods, 9, 772.
Davidson, R. K., Bornstein, S., & Handeland, K. (2008). Long-term study of Sarcoptes scabiei
infection in Norwegian red foxes (Vulpes vulpes) indicating host/parasite adaptation.
Veterinary Parasitology, 156, 277-283.
Davidson, R. K., Ličina, T., Gorini, L., & Milner, J. M. (2015). Endoparasites in a Norwegian
moose (Alces alces) population – Faunal diversity, abundance and body condition.
International Journal for Parasitology: Parasites and Wildlife, 4, 29-36.
De Garine-Wichatitsky, M., Caron, A., Kock, R., Tschopp, R., Munyeme, M., Hofmeyr, M.,
& Michel, A. (2013). A review of bovine tuberculosis at the wildlife–livestock–human
interface in sub-Saharan Africa. Epidemiology and Infection, 141, 1342-1356.
91
De la Fuente, J., Van Den Bussche, R. A., & Kocan, K. M. (2001). Molecular phylogeny and
biogeography of North American isolates of Anaplasma marginale (Rickettsiaceae:
Ehrlichieae). Veterinary Parasitology, 97, 65-76.
De Waal, D. T. (2000). Anaplasmosis control and diagnosis in South Africa. Annals of the New
York Academy of Sciences, 916, 474-483.
Debeila, E. M. (2012). Occurrence of Anaplasma and Ehrlichia species in African buffalo
(Syncerus caffer) in Kruger National Park and Hluhluwe-iMfolozi Park in South Africa.
Masters Thesis, University of Pretoria, 98pp.
Decaro, N., Carelli, G., Lorusso, E., Lucente, M. S., Greco, G., Lorusso, A., Radogna, A., Ceci,
L., Buonavoglia, C. (2008). Duplex real-time polymerase chain reaction for
simultaneous detection and quantification of Anaplasma marginale and Anaplasma
centrale. Journal of Veterinary Diagnostic Investigation, 20, 606-611.
Dimopoullos, G. T., Schrader, G. T., & Foote, L. E. (1960). Electrophoretic studies of bovine
serum. III. Serum protein changes in anaplasmosis. American Journal of Veterinary
Research, 21, 222.
Diuk-Wasser, M. A., Vannier, E., & Krause, P. J. (2016). Coinfection by Ixodes tick-borne
pathogens: ecological, epidemiological, and clinical consequences. Trends in
Parasitology, 32, 30-42.
Douglas, B., Maechler, M., Bolker, B., Walker, S., Christensen, R., Singmann, H., & Dai, B.
(2015). Fitting linear-mised effects models using lme4. Journal of Statistical Software,
67, 1-48.
Dumler, J. S., Barbet, A. F., Bekker, C., Dasch, G. A., Palmer, G. H., Ray, S. C., Rikihisa, Y.,
& Rurangirwa, F. R. (2001). Reorganization of genera in the families Rickettsiaceae
and Anaplasmataceae in the order Rickettsiales: unification of some species of
Ehrlichia with Anaplasma, Cowdria with Ehrlichia and Ehrlichia with Neorickettsia,
92
descriptions of six new species combinations and designation of Ehrlichia equi and
'HGE agent' as subjective synonyms of Ehrlichia phagocytophila. International
Journal of Systematic and Evolutionary Microbiology, 51, 2145-2165.
Dusek, R. J., Hagen, C. A., Franson, J. C., Budeau, D. A., & Hofmeister, E. K. (2014). Utilizing
hunter harvest effort to survey for wildlife disease: a case study of West Nile virus in
greater sage‐grouse. Wildlife Society Bulletin, 38, 721-727.
Eberts, M. D., de Paiva Diniz, P. P. V., Beall, M. J., Stillman, B. A., Chandrashekar, R., &
Breitschwerdt, E. B. (2011). Typical and atypical manifestations of Anaplasma
phagocytophilum infection in dogs. Journal of the American Animal Hospital
Association, 47, 86-94.
Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high
throughput. Nucleic Acids Research, 32, 1792-1797.
Eygelaar, D., Jori, F., Mokopasetso, M., Sibeko, K. P., Collins, N. E., Vorster, I., Troskie, M.,
& Oosthuizen, M. C. (2015). Tick-borne haemoparasites in African buffalo (Syncerus
caffer) from two wildlife areas in Northern Botswana. Parasites and Vectors, 8, 1-11.
Ezenwa, V.O. (2004). Interactions among host diet, nutritional status and gastrointestinal
parasite infection in wild bovids. International Journal for Parasitology, 34, 535-542.
Ezenwa, V. O. (2016). Helminth–microparasite co-infection in wildlife: lessons from
ruminants, rodents and rabbits. Parasite Immunology, 38, 527-534.
Ezenwa, V. O., & Jolles, A. E. (2008). Horns honestly advertise parasite infection in male and
female African buffalo. Animal Behaviour, 75, 2013-2021.
Ezenwa, V. O., & Jolles, A. E. (2015). Opposite effects of anthelmintic treatment on microbial
infection at individual versus population scales. Science, 347, 175-177.
93
Ezenwa, V. O., Jolles, A. E., & O’Brien, M. P. (2009). A reliable body condition scoring
technique for estimating condition in African buffalo. African Journal of Ecology, 47,
476-481.
Ezenwa, V. O., Etienne, R. S., Luikart, G., Beja-Pereira, A., & Jolles, A. E. (2010). Hidden
consequences of living in a wormy world: nematode-induced immune suppression
facilitates tuberculosis invasion in African buffalo. The American Naturalist, 176, 613-
624.
Fenton, A., & Perkins, S. E. (2010). Applying predator-prey theory to modelling immune-
mediated, within-host interspecific parasite interactions. Parasitology, 137, 1027-1038.
Foo, Y. Z., Nakagawa, S., Rhodes, G., & Simmons, L. W. (2017). The effects of sex hormones
on immune function: a meta‐analysis. Biological Reviews, 92, 551-571.
Forbes, M., Weatherhead, P., & Bennett, G. (1994). Blood parasites of blue grouse: variation
in prevalence and patterns of interspecific association. Oecologia, 97, 520-525.
Friend, M., & Trainer, D. O. (1969). Aspergillosis in captive herring gulls. Bulletin of the
Wildlife Disease Association, 5, 271-275.
Fyumagwa, R. D., Mdaki, M. L., Soombe, C., & Nyaki, A. (2013). African buffalo is an
important reservior of Anaplasma bovis in the Ngorongoro Crater, Tanzania. Research
Opinions in Animal and Veterinary Sciences, 3, 105-110.
Gallivan, G. J., Spickett, A., Heyne, H., Spickett, A. M., & Horak, I. G. (2011). The dynamics
of questing ticks collected for 164 consecutive months off the vegetation of two
landscape zones in the Kruger National Park (1988–2002). Part III. The less commonly
collected species. Onderstepoort Journal of Veterinary Research, 78, 1-9.
Gallivan, G., & Horak, I. (1997). Body size and habitat as determinants of tick infestations of
wild ungulates in South Africa. South African Journal of Wildlife Research, 27, 63-70.
94
Gortazar, C., Diez-Delgado, I., Barasona, J. A., Vicente, J., de La Fuente, J., & Boadella, M.
(2014). The wild side of disease control at the wildlife-livestock-human interface: a
review. Frontiers in Veterinary Science, 1, 1-12.
Grau, H. E. G., Cunha Filho, N. A. D., Pappen, F. G., & Farias, N. A. D. R. (2013).
Transplacental transmission of Anaplasma marginale in beef cattle chronically infected
in southern Brazil. Revista Brasileira de Parasitologia Veterinária, 22, 189-193.
Griffith, J., & Higgins, D. (2012). Diagnosis, treatment and outcomes for koala chlamydiosis
at a rehabilitation facility (1995–2005). Australian Veterinary Journal, 90, 457-463.
Griffiths, E. C., Pedersen, A. B., Fenton, A., & Petchey, O. L. (2014). Analysis of a summary
network of co-infection in humans reveals that parasites interact most via shared
resources. Proceedings of the Royal Society of London B: Biological Sciences, 281,
200-206.
Guindon, S., & Gascuel, O. (2003). A simple, fast and accurate method to estimate large
phylogenies by maximum likelihood. Systems Biology, 52, 696-704.
Gunn, A., & Irvine, R. J. (2003). Subclinical parasitism and ruminant foraging strategies: a
review. Wildlife Society Bulletin, 31, 117-126.
Guscetti, F., Mathis, A., Hatt, J. M., & Deplazes, P. (2003). Overt fatal and chronic subclinical
Encephalitozoon cuniculi microsporidiosis in a colony of captive emperor tamarins
(Saguinus imperator). Journal of Medical Primatology, 32, 111-119.
Hage, J., Schukken, Y., Dijkstra, T., Barkema, H., Van Valkengoed, P., & Wentink, G. (1998).
Milk production and reproduction during a subclinical bovine herpesvirus 1 infection
on a dairy farm. Preventive Veterinary Medicine, 34, 97-106.
Hall, T. A. (1999). BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. In Nucleic Acids Symposium Series, 41, 95-98.
95
Han, S., Norimine, J., Brayton, K. A., Palmer, G. H., Scoles, G. A., & Brown, W. C. (2010).
Anaplasma marginale infection with persistent high-load bacteremia induces a
dysfunctional memory CD4+ T lymphocyte response but sustained high IgG
titers. Clinical and Vaccine Immunology, 17, 1881-1890.
Härkönen, T., Dietz, R., Reijnders, P., Teilmann, J., Harding, K., Hall, A., Brasseur, S., Siebert,
U., Goodman, S. J., Jepson, P. D., Rasmussen, T. D., & Thompson, P. (2006). The 1988
and 2002 phocine distemper virus epidemics in European harbour seals. Diseases of
Aquatic Organisms, 68, 115-130.
Haydon, D. T., Cleaveland, S., Taylor, L. H., & Laurenson, M. K. (2002). Identifying reservoirs
of infection: a conceptual and practical challenge. Emerging Infectious Diseases, 8,
1468-1473.
Henrichs, B. (2014). Hidden Dynamics: Community Ecology and Costs of Parasite Infection
in African Buffalo (Syncerus Caffer). Masters thesis. Oregon State University, 88 pp.
Henrichs, B., Oosthuizen, M. C., Troskie, M., Gorsich, E., Gondhalekar, C., Beechler, B. R.,
Ezenwa, V. O., Jolles, A. E. (2016). Within guild co-infections influence parasite
community membership: a longitudinal study in African Buffalo. Journal of Animal
Ecology, 85, 1025-1034.
Horak, I., Golezardy, H., & Uys, A. (2006). The host status of African buffaloes, Syncerus
caffer, for Rhipicephalus (Boophilus) decoloratus. Onderstepoort Journal of
Veterinary Research, 73, 193-198.
Horak, I. G., Golezardy, H., & Uys, A. C. (2007). Ticks associated with the three largest wild
ruminant species in southern Africa. Onderstepoort Journal of Veterinary Research,
74, 231-242.
Horak, I. G., Gallivan, G. J., & Spickett, A. M. (2011). The dynamics of questing ticks collected
for 164 consecutive months off the vegetation of two landscape zones in the Kruger
96
National Park (1988-2002). Part I. Total ticks, Amblyomma hebraeum and
Rhipicephalus decoloratus. Onderstepoort Journal of Veterinary Research, 78, 8-17.
Houston, A. I., McNamara, J. M., Barta, Z., & Klasing, K. C. (2007). The effect of energy
reserves and food availability on optimal immune defence. Proceedings of the Royal
Society of London B: Biological Sciences, 274, 2835-2842.
Huelsenbeck, J., & Ronquist, F. (2001). MRBAYES: Bayesian inference of phylogenetic trees.
Bioinformatics, 17, 754-755.
Hughes, W. O., Petersen, K. S., Ugelvig, L. V., Pedersen, D., Thomsen, L., Poulsen, M., &
Boomsma, J. J. (2004). Density-dependence and within-host competition in a
semelparous parasite of leaf-cutting ants. BMC Evolutionary Biology, 4, 45.
Hughes, J., Albon, S., Irvine, R., & Woodin, S. (2009). Is there a cost of parasites to caribou?
Parasitology, 136, 253-265.
Jackson, J., Pleass, R., Cable, J., Bradley, J., & Tinsley, R. (2006). Heterogenous interspecific
interactions in a host–parasite system. International Journal for Parasitology, 36,
1341-1349.
Jaswal, H., Bal, M. S., Singla, L., Amrita, K. P., Mukhopadhyay, J. P., & Juyal, P. (2014).
Application of msp1β PCR and 16S rRNA semi nested PCR-RFLP for detection of
persistent anaplasmosis in tick infested cattle. International Journal of Advanced
Research, 2, 188-196.
Jensen, G., Hansen, B., Eilenberg, J., & Mahillon, J. (2003). The hidden lifestyles of Bacillus
cereus and relatives. Environmental Microbiology, 5, 631-640.
Johnson, P. T., Preston, D. L., Hoverman, J. T., & LaFonte, B. E. (2013). Host and parasite
diversity jointly control disease risk in complex communities. Proceedings of the
National Academy of Sciences, 110, 16916-16921.
97
Jolles, A. E. (2007). Population biology of African buffalo (Syncerus caffer) at Hluhluwe‐
iMfolozi Park, South Africa. African Journal of Ecology, 45, 398-406.
Jolles, A. E., & Ezenwa, V. O. (2015). Ungulates as model systems for the study of disease
processes in natural populations. Journal of Mammalogy, 96, 4-15.
Jolles, A. E., Ezenwa, V. O., Etienne, R. S., Turner, W. C., & Olff, H. (2008). Interactions
between macroparasites and microparasites drive infection patterns in free-ranging
African buffalo. Ecology, 89, 2239-2250.
Jolles, A. E., Beechler, B. R., & Dolan, B. P. (2015). Beyond mice and men: environmental
change, immunity and infections in wild ungulates. Parasite Immunology, 37, 255-266.
Junker, K., Horak, I. G., Penzhorn, B., (2015). History and development of research on wildlife
parasites in southern Africa, with emphasis on terrestrial mammals, especially
ungulates. International Journal for Parasitology: Parasites and Wildlife, 4, 50-70.
Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., Buxton, S.,
Cooper, A., Markowitz, S., & Duran, C. (2012). Geneious Basic: an integrated and
extendable desktop software platform for the organization and analysis of sequence
data. Bioinformatics, 28, 1647-1649.
Kieser, S. T., Eriks, I. S., & Palmer, G. H. (1990). Cyclic rickettsemia during persistent
Anaplasma marginale infection of cattle. Infection and Immunity, 58, 1117-1119.
Knowles, R., Montrose, M., Craig, T., Wagner, G., & Long, R. (1982). Clinical and serological
evidence of bovine babesiosis and anaplasmosis in St. Lucia. Veterinary Parasitology,
10, 307-311.
Kocan, K. M. (1995). Targeting ticks for control of selected hemoparasitic diseases of cattle.
Veterinary Parasitology, 57, 121-151.
Kocan, K. M., & de la Fuente, J. (2003). Co-feeding studies of ticks infected with Anaplasma
marginale. Veterinary Parasitology, 112, 295-305.
98
Kocan, K. M., de la Fuente, J., Guglielmone, A. A., & Meléndez, R. D. (2003). Antigens and
alternatives for control of Anaplasma marginale infection in cattle. Clinical
Microbiology Reviews, 16, 698-712.
Kocan, K. M., de la Fuente, J., Blouin, E. F., Coetzee, J. F., & Ewing, S. (2010). The natural
history of Anaplasma marginale. Veterinary Parasitology, 167, 95-107.
Kock, R. (2005). What is this infamous “wildlife/livestock disease interface?” A review of
current knowledge for the African continent. Conservation and Development
Interventions at the Wildlife/Livestock Interface: Implications for Wildlife, Livestock
and Human Health, 30, 1-13.
Kock, R., Kock, M., de Garine-Wichatitsky, M., Chardonnet, P., & Caron, A. (2014). Livestock
and buffalo (Syncerus caffer) interfaces in Africa: ecology of disease transmission and
implications for conservation and development. Ecology, Evolution and Behaviour of
Wild Cattle Implication for Conservation. Cambridge University Press, Cambridge,
UK, 608 pp.
Kuiken, T., & Danesik, K. L. (1999). Hematology and serum chemistry of captive juvenile
double-crested cormorants (Phalacrocorax auritus). The Canadian Veterinary Journal,
40, 493-496.
Kushner, I. (1982). The phenomenon of the acute phase response. Annals of the New York
Academy of Sciences, 389, 39-48.
Kuttler, K. (1984). Anaplasma infections in wild and domestic ruminants: a review. Journal of
Wildlife Diseases, 20, 12-20.
Kuttler, K., Clifford, D., & Touray, B. (1988). Prevalence of anaplasmosis and babesiosis in
N'Dama cattle of The Gambia. Tropical Animal Health and Production, 20, 37-41.
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2015). Package ‘lmerTest’. R
Package Version, 2(0).
99
Lafferty, K. D. (2010). Interacting parasites. Science, 330, 187-188.
Larkin, M. A., Blackshields, G., Brown, N. P., Chenna, R., McGettigan, P. A., McWilliam, H.,
Higgins, D. G. (2007). Clustal W and Clustal X version 2.0. Bioinformatics, 23, 2947-
2948.
Lew, A., Bock, R., Minchin, C., & Masaka, S. (2002). A msp1α polymerase chain reaction
assay for specific detection and differentiation of Anaplasma marginale isolates.
Veterinary Microbiology, 86, 325-335.
Lew, A. E., Gale, K. R., Minchin, C. M., Shkap, V., & de Waal, D. T. (2003). Phylogenetic
analysis of the erythrocytic Anaplasma species based on 16S rDNA and GroEL
(HSP60) sequences of A. marginale, A. centrale, and A. ovis and the specific detection
of A. centrale vaccine strain. Veterinary Microbiology, 92, 145-160.
Lochmiller, R. L., Vestey, M. R., & Boren, J. C. (1993). Relationship between protein
nutritional status and immunocompetence in northern bobwhite chicks. The Auk, 110
503-510.
Maddison, W. P., Maddison, D.R., (2015). Mesquite: a modular system for evolutionary
analysis. Version 3.04. http://mesquiteproject.org
McAllister, M. M. (2005). A decade of discoveries in veterinary protozoology changes our
concept of ‘subclinical’ toxoplasmosis. Veterinary Parasitology, 132, 241-247.
McInnes, L., Gillett, A., Hanger, J., Reid, S., & Ryan, U. (2011). The potential impact of native
Australian trypanosome infections on the health of koalas (Phascolarctos cinereus).
Parasitology, 138, 873-883.
Molad, T., Brayton, K. A., Palmer, G. H., Michaeli, S., Shkap, V. (2004). Molecular
conservation of MSP4 and MSP5 in Anaplasma marginale and A. centrale vaccine
strain. Veterinary Microbiology, 100, 55-64.
100
Molad, T., Mazuz, M., Fleiderovitz, L., Fish, L., Savitsky, I., Krigel, Y., Leibovitz, B., Molloy,
J., Jongejan, F., & Shkap, V. (2006). Molecular and serological detection of A. centrale
and A. marginale-infected cattle grazing within an endemic area. Veterinary
Microbiology, 113, 55-62.
Ngeranwa, J., Shompole, S., Venter, E., Wambugu, A., Crafford, J. E., & Penzhorn, B. L.
(2008). Detection of Anaplasma antibodies in wildlife and domestic species in wildlife-
livestock interface areas of Kenya by major surface protein 5 competitive inhibition
enzyme-linked immunosorbent assay. Onderstepoort Journal of Veterinary Research,
75, 199-205.
Palacios, C., de Echaide, S. T., & Mattion, N. (2014). Evaluation of the immune response to
Anaplasma marginale MSP5 protein using a HSV-1 amplicon vector system or
recombinant protein. Research in Veterinary Science, 97, 514-520
Palmer, G. H., Munodzana, D., Tebele, N., Ushe, T., & McElwain, T. F. (1994). Heterologous
strain challenge of cattle immunized with Anaplasma marginale outer membranes.
Veterinary Immunology and Immunopathology, 42, 265-273.
Penzhorn, B. L., Vorster, I., Harrison-White, R. F., & Oosthuizen, M. C. (2017). Black-backed
jackals (Canis mesomelas) are natural hosts of Babesia rossi, the virulent causative
agent of canine babesiosis in sub-Saharan Africa. Parasites and Vectors, 10, 124.
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., Heisterkamp, S., Van Willigen, B., &
Maintainer, R. (2017). Package ‘nlme’. Linear and Nonlinear Mixed Effects Models,
version 3-1.
Potgieter, F., & Stoltz, W. (2004). Bovine Anaplasmosis. In J. Coetzer & R. Tustin (Eds.),
Infectious Diseases of Livestock (2nd ed., pp. 594-616). Cape Town, South Africa:
Oxford University Press.
101
Prins, H. (1996). Ecology and Behaviour of the African Buffalo: Social Inequality and Decision
Making (Vol. 1, 293 pp.): London, United Kingdon: Chapman & Hall.
R Core Team. (2016). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing. Retrieved from http://www.R-
project.org/
Råberg, L., Sim, D., & Read, A. F. (2007). Disentangling genetic variation for resistance and
tolerance to infectious diseases in animals. Science, 318, 812-814.
Regoes, R. R., McLaren, P. J., Battegay, M., Bernasconi, E., Calmy, A., Günthard, H. F.,
Hoffman, M., Rauch, A., Telenti, A., & Fellay, J. (2014). Disentangling human
tolerance and resistance against HIV. PLoS Biology, 12, e1001951.
Rikihisa, Y. (2010). Molecular events involved in cellular invasion by Ehrlichia chaffeensis
and Anaplasma phagocytophilum. Veterinary Parasitology, 167, 155-166.
Roland, L., Drillich, M., & Iwersen, M. (2014). Hematology as a diagnostic tool in bovine
medicine. Journal of Veterinary Diagnostic Investigation, 26, 592-598.
Ronquist, F., & Huelsenbeck, J. (2003). MRBAYES 3: Bayesian phylogenetic inference under
mixed models. Bioinformatics, 19, 1572-1574.
Ryser-Degiorgis, M.-P. (2013). Wildlife health investigations: needs, challenges and
recommendations. BMC Veterinary Research, 9, 223.
SA Weather Bureau. WB42 Climate statistics of South Africa. 1961-1990 (unpublished).
Said, M. B., Belkahia, H., Karaoud, M., Bousrih, M., Yahiaoui, M., Daaloul-Jedidi, M., &
Messadi, L. (2015). First molecular survey of Anaplasma bovis in small ruminants from
Tunisia. Veterinary Microbiology, 179, 322-326.
SANParks. Kruger in Relation. Retrieved from www.sanparks.co.za/images/
parks/kruger/conservation/scientific/maps/map_images/Africa_SA_knp.jpg
102
Shkap, V., Leibovitz, B., Krigel, Y., Molad, T., Fish, L., Mazuz, M., Fleiderovitz, L., &
Savitsky, I. (2008). Concomitant infection of cattle with the vaccine strain Anaplasma
marginale ss centrale and field strains of A. marginale. Veterinary Microbiology, 130,
277-284.
Soares, H. S., Marcili, A., Barbieri, A. R., Minervino, A. H., Malheiros, A. F., Gennari, S. M.,
& Labruna, M. B. (2017). Novel Anaplasma and Ehrlichia organisms infecting the
wildlife of two regions of the Brazilian Amazon. Acta Tropica, 174, 82-87.
Spaan, R. S. (2015). Dispersal behavior in African buffalo (Syncerus caffer): trade-offs
between nutritional resources and disease exposure. Masters Thesis, Oregon State
University, Corvallis, U.S.A., 93 pp.
Tamekuni, K., Kano, F. S., Ataliba, A. C., Marana, E. R., Venâncio, E. J., Vidotto, M. C.,
Garcia, J.L., Headley, S.A., Vidotto, O. (2009). Cloning, expression, and
characterization of the MSP1a and MSP1b recombinant proteins from PR1 Anaplasma
marginale strain, Brazil. Research in Veterinary Science, 86, 98-107.
Tamura, K., Stecher, G., Peterson, D., Filipski, A., & Kumar, S. (2013). MEGA 6: molecular
evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30, 2725-
2729.
Thomas, M. B., Watson, E. L., & Valverde-Garcia, P. (2003). Mixed infections and insect–
pathogen interactions. Ecology Letters, 6, 183-188.
Tompkins, D. M., Dunn, A. M., Smith, M. J., & Telfer, S. (2011). Wildlife diseases: from
individuals to ecosystems. Journal of Animal Ecology, 80, 19-38.
Ueti, M. W., Reagan, J. O., Knowles, D. P., Scoles, G. A., Shkap, V., & Palmer, G. H. (2007).
Identification of midgut and salivary glands as specific and distinct barriers to efficient
tick-borne transmission of Anaplasma marginale. Infection and Immunity, 75, 2959-
2964.
103
Vaumourin, E., Vourc’h, G., Gasqui, P., & Vayssier-Taussat, M. (2015). The importance of
multiparasitism: examining the consequences of co-infections for human and animal
health. Parasites and Vectors, 8, 545.
Visser, E. S., Ambrosio, R. E., & De Waal, D. T. (1991). An Anaplasma centrale DNA probe
that differentiates between Anaplasma ovis and Anaplasma marginale DNA.
Veterinary Microbiology, 28, 313-325.
Werle, E., Schneider, C., Renner, M., Völker, M., & Fiehn, W. (1994). Convenient single-step,
one tube purification of PCR products for direct sequencing. Nucleic Acids Research,
22, 4354–4355.
Williamson, M., Hooper, P., Selleck, P., Gleeson, L., Daniels, P., Westbury, H., & Murray, P.
(1998). Transmission studies of Hendra virus (equine morbillivirus) in fruit bats, horses
and cats. Australian Veterinary Journal, 76, 813-818.
Wobeser, G.A. (2006). Chapter 2. What is disease? In Essentials of Disease in Wild
Animals, (pp. 20). Blackwell Publishing, Ames, USA.
Woldehiwet, Z. (2010). The natural history of Anaplasma phagocytophilum. Veterinary
Parasitology, 167, 108-122.
Yabsley, M. J., McKibben, J., Macpherson, C. N., Cattan, P. F., Cherry, N. A., Hegarty, B. C.,
Breitschwerdt, E. B., O’Connor, T., Chandrashekar, R., Paterson, T., Perea, M. L., Ball,
G., Friesen, S., Goedde, J., Henderson, B., & Sylvester, W. (2008). Prevalence of
Ehrlichia canis, Anaplasma platys, Babesia canis vogeli, Hepatozoon canis, Bartonella
vinsonii berkhoffii, and Rickettsia spp. in dogs from Grenada. Veterinary Parasitology,
151, 279-285.
Zambatis, N. (2003). Determinants of grass production and composition in the Kruger National
Park. Masters Thesis, University of Natal, Pietermaritzburg, 254 pp.
104
Zintl, A., Gray, J. S., Skerrett, H. E., & Mulcahy, G. (2005). Possible mechanisms underlying
age-related resistance to bovine babesiosis. Parasite Immunology, 27, 115-120.
Zobba, R., Anfossi, A. G., Parpaglia, M. L. P., Dore, G. M., Chessa, B., Spezzigu, A., Rocca,
S., Visco, S., Pittau, M., & Alberti, A. (2014). Molecular investigation and phylogeny
of Anaplasma spp. in Mediterranean ruminants reveal the presence of neutrophil-tropic
strains closely related to A. platys. Applied and Environmental Microbiology, 80, 271-
280.
105
Supplementary Material
10 20 30 40 50 60 70 80 90 100 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714578 TCCATCTCGGCCGTATTCCAGCGCACAGCAATCGACGCGAAGGCAGATGCCAAATACGAGAGTGTGGGGCTACGTGCTAAAGCAACTGCAGCATTGGGTA KX714579 ...... G...... KX714580 ...... G...... KX714581 ...... G......
110 120 130 140 150 160 170 180 190 200 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714578 ATCTCGGGCGGCTTGTCGCTCGTGGTAAACTCACAAGCTCAGATGCACCCAAG---CTTGACCATAGTATTGACAACCTGCCGCTCATGGATGAAGCGCC KX714579 ...... GAC...... G...... A...... KX714580 ...... ---...... G...... KX714581 ...... G...... GAC...... G...... A......
210 220 230 240 250 260 270 280 290 300 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714578 TGACACTGGTGAAAAGATTGAAGTACCAGCAGGTGAAGAGCAAGAATTTGGCAAAGCAGCAGCTTGGGGTTTAGCAGGTTTCAAGCGTAAAGTGGATGAA KX714579 ...... G...... G...... C...... C...... KX714580 ...... C...... KX714581 ...... GCG.G...TT..G.AGTAT..C..G...AG...... G...... C...... C......
310 320 330 340 ....|....|....|....|....|....|....|....|....| KX714578 AGCCTCGAGATGCTAGACCGAGGCATGAACATGCTCGCGGAAGGC KX714579 .....G...... KX714580 .....G...... G...... KX714581 .....G......
Supplementary Figure 2.1: Alignment of partial major surface protein 1β sequences of Anaplasma marginale determined herein. A dot indicates an identical nucleotide with respect to the top sequence. International Union of Pure and Applied Chemistry codes indicate polymorphic positions in the sequences.
106
10 20 30 40 50 60 70 80 90 100 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 GAGTTCGAGAATCCGTACATATTCTTGACCGAAAAGAAAATCAATCTAGTGCAAAGCATACTGCCAGTACTGGAAAACGTCGCTAGGTCAGGAAGGCCGC KX714583 ...... KX714584 ...... AT KX714585 ...... AT KX714586 ...... KX714587 ...... KX714588 ......
110 120 130 140 150 160 170 180 190 200 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 TGCTGATCATCGCAGAAGACGTGGAAGGTGAGGCACTGAGCACATTGGTGCTTAACAAGCTTCGTGGGGGCCTGCAAGTCGCTGCTGTTAAGGCACCGGG KX714583 ...... KX714584 ...... G...... A..G.....A...T.A..T...C...... KX714585 ...... G...... A..G.....A...T.A..Y...C...... KX714586 ...... KX714587 ...... R...... KX714588 ......
210 220 230 240 250 260 270 280 290 300 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 TTTCGGGGATAGAAGAAAAGACATGCTTGGCGATATTGCTGTGATAGCCGGAGCGAAGTACGTGGTAAATGACGAGTTAGCGGTAAAGGTTGAGGACATC KX714583 ...... KX714584 ...... C...... KX714585 ...... KX714586 ...... KX714587 ...... KX714588 ......
310 320 330 340 350 360 370 380 390 400 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 ACCCTGGACGACCTCGGCACTGCTAAGACCGTGCGCATCACTAAGGATACGACCACAATTATAGGAAGCGTCGACAGTAACGCTGACAGCATTACCAGCA KX714583 ...... C...... T...... A...... KX714584 ...... KX714585 ...... KX714586 ...... KX714587 ...... KX714588 ......
410 420 430 440 450 460 470 480 490 500 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 GGATTAGCCAGATTAAGTCTCAGATTGAAGTTTCTTCCTCTGATTACGACAAAGAAAAACTAAAGGAGCGGTTGGCGAAGCTCTCAGGTGGGGTTGCTGT KX714583 ...... KX714584 .A...... C...... G...... T...... KX714585 .A...... KX714586 ...... KX714587 ...... KX714588 ......
510 520 530 540 550 560 570 580 590 600 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 GCTTAAGGTTGGCGGCTCTAGYGAAGTTGAGGTCAAGGAGCGTAAGGACAGGGTTGAAGATGCTTTGCATGCAACTAGGGCTGCGGTTGAGGAAGGCGTG KX714583 ...... T...... KX714584 ...... T...... C...... KX714585 ...... T...... C...... KX714586 ...... T...... KX714587 ...... C...... KX714588 ...... C......
610 620 630 640 650 660 670 680 690 700 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 GTACCTGGTGGWGGAGCTGCCCTGCTATACACGCTTGCATCTCTGGACGAGGTGAAGGGGAAGAATGATGATGAGCAATTGGGCATCAACATCATAAAGC KX714583 ...... KX714584 ...... KX714585 ...... KX714586 ...... KX714587 ...... R...... KX714588 ......
710 720 730 740 750 760 770 780 790 800 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....| KX714582 GCGCTGTCCGCGCTCCAATAAAGAGGATCATAAAAAATTCCGGATCCGAAGAGGCTCCGTGTGTCATTCAGCATCTGTTGAAGCAGAATGATAAGGAGCT KX714583 ...... KX714584 ...... KX714585 ...... KX714586 ...... KX714587 ...... KX714588 ......
810 820 830 840 850 860 870 880 ....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|....|. KX714582 CATCTTTAATGTTGACACCATGAACTACGCTAACGCGTTTGCTTCCGGAGTTATGGACCCTCTAAAGGTAGTACGTATCGC KX714583 ...... C...... KX714584 ...... KX714585 ...... KX714586 ...... KX714587 ...... KX714588 ......
Supplementary Figure 2.2: Alignment of partial heat shock protein (groEL) sequences of Anaplasma centrale determined herein. A dot indicates an identical nucleotide with respect to the top sequence.
International Union of Pure and Applied Chemistry codes indicate polymorphic positions in the sequences.
107
Supplementary Table 3.1: Comparison of Anaplasma marginale and A. centrale infection detection via conventional PCR and reverse line blot hybridisation (determining presence/absence of infection) with quantitative PCR (qPCR) results (determining infection burden). Infection burdens from qPCR was designated as a positive reading if they exceeded either the cut-off for what was higher (in terms of copies/reaction) out of fluorescence reading from the non-template controls, or the lowest standard in the standard curve in each qPCR run. If a sample produced no fluorescence during the reaction, or if a sample did fluoresce but did not meet the cut-off, it was designated as an insufficient reading and both were assigned a negative qPCR result.
A. marginale A. centrale N % n %
Infection presence with positive qPCR reading 323 32 266 26
Infection presence with negative qPCR reading 3061 30 3943 39
Infection absence with positive qPCR reading 95 9 48 5
Infection absence with negative qPCR reading 2812 28 2974 30 1 291 insufficient readings, 15 negative readings 2 269 insufficient readings, 12 negative readings 3 328 insufficient readings, 66 negative readings 4 215 insufficient readings, 82 negative readings
108
Supplementary Table 3.2: Mixed effects logistic regression models predicting the presence/absence of Anaplasma marginale and/or A. centrale in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p OR Coefficient SE p OR MANAGED n = 735 n = 735 HERD Intercept -1.123 0.002 -1.989 0.281 Age -0.829 0.002 <0.001 0.436 -0.248 0.273 0.364 0.781 Sex (Male) 0.059 0.002 <0.001 1.061 0.486 0.396 0.219 1.626 Season (Wet) -0.254 0.002 <0.001 0.776 0.082 0.208 0.694 1.085 Co-infection* 1.111 0.002 <0.001 3.039 1.030 0.231 <0.001 4.398
FREE-RANGING n = 1318 n = 1318 HERD Intercept -1.060 0.123 -1.191 0.121 Age -0.758 0.210 <0.001 0.469 -0.622 0.223 0.005 0.537 Season (Wet) 0.323 0.130 0.003 1.481 0.967 0.135 <0.001 2.630 Sub-herd (Lower -0.038 0.143 0.790 0.963 0.089 0.164 0.589 1.093 Sabie) Co-infection* 0.383 0.134 0.004 1.467 0.378 0.141 0.007 1.460 *Co-infection presence refers to the Anaplasma species. Significant p-values (<0.05) are bold. SE – standard error, OR – odds ratio.*Co-infection presence refers to the Anaplasma species. Significant P- values (<0.05) are bold. SE – standard error, OR – odds ratio.
109
Supplementary Table 3.3: Mixed effects logistic regression models predicting the previous capture’s presence/absence of Anaplasma marginale and/or A. centrale infection in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p OR Coefficient SE p OR MANAGED n = 629 n = 630 HERD Intercept -1.081 0.196 -1.899 0.294 Age -1.081 0.211 <0.001 0.339 -0.331 0.290 0.253 0.718 Sex (Male) 0.197 0.297 0.506 1.218 0.479 0.433 0.269 1.614 Season (Wet) -0.085 0.199 0.668 0.918 0.223 0.219 0.309 1.250 Co-infection* 0.474 0.237 0.046 1.606 0.597 0.253 0.020 1.798
FREE-RANGING n = 1424 n = 1419 HERD Intercept -0.943 0.111 -0.750 0.119 Age -0.483 0.199 0.015 0.617 -0.346 0.206 0.093 0.708 Sub-herd (Lower -0.337 0.136 0.013 0.714 -0.061 0.148 0.683 0.941 Sabie) Season (Wet) 0.026 0.125 0.836 1.026 -0.479 0.128 <0.001 0.619 Co-infection 0.143 0.132 0.276 1.154 0.154 0.138 0.265 1.166 *Co-infection refers to the other Anaplasma species. Significant p-values (<0.05) are bold. SE – stand ard error, OR – odds ratio.
110
Supplementary Table 3.4: Generalised linear mixed model predicting body condition in managed and free-ranging herds of African buffalo from Kruger National Park. Concurrent capture’s infection Previous capture’s infection Coefficient SE p Coefficient SE p MANAGED HERD n = 737 n = 632 Intercept 2.618 0.052 2.540 0.059 A. marginale presence 0.064 0.063 0.306 0.146 0.071 0.041 A. centrale presence 0.207 0.066 0.002 0.180 0.074 0.016 Season (Wet) 0.204 0.054 <0.001 0.136 0.060 0.024 Sex (Male) 0.132 0.073 0.075 0.177 0.085 0.040 Age 0.025 0.048 0.610 0.096 0.056 0.091
FREE-RANGING HERD n = 1348 n = 1240 Intercept 2.878 0.036 2.890 0.037 Season (Wet) -0.084 0.031 0.007 -0.087 0.031 0.006 A. marginale presence 0.057 0.046 0.208 0.115 0.049 0.019 A. centrale presence 0.076 0.033 0.021 0.024 0.034 0.485 Age -0.369 0.054 <0.001 -0.371 0.056 <0.001 Sub-herd (Lower Sabie) 0.224 0.046 <0.001 0.237 0.047 <0.001 A. marginale: Sub-herd -0.166 0.066 0.011 -0.297 0.070 <0.001 (Lower Sabie) Significant p-values (<0.05) are bold. SE – standard error.
111
Supplementary Table 3.5: Generalised linear mixed model predicting haematocrit volume in managed and free-ranging herds of African buffalo from Kruger National Park. Concurrent capture’s infection Previous capture’s infection Coefficient SE p Coefficient SE p MANAGED HERD n = 724 n = 624 Intercept -0.289 0.033 -0.306 0.034 A. marginale presence 0.015 0.024 0.544 0.089 0.026 <0.001 A. centrale presence 0.045 0.026 0.086 0.013 0.028 0.630 Season (Wet) -0.045 0.019 0.019 -0.037 0.021 0.072 Sex (Male) -0.053 0.046 0.249 -0.016 0.047 0.729 Age 0.191 0.031 <0.001 0.166 0.033 <0.001
FREE-RANGING HERD n = 1125 n = 1039 Intercept 0.043 0.031 0.040 0.034 A. marginale presence -0.042 0.034 0.214 0.007 0.036 0.847 A. centrale presence 0.073 0.034 0.031 <0.000 0.034 0.984 Season (Wet) -0.012 0.032 0.710 0.010 0.033 0.772 Sub-herd (Lower Sabie) 0.148 0.035 <0.001 0.144 0.036 <0.001 Age 0.202 0.050 <0.001 0.203 0.051 <0.001 Significant p-values (<0.05) are bold. SE – standard error.
112
Supplementary Table 3.6: Generalised linear mixed model predicting total protein in managed and free-ranging herds of African buffalo from Kruger National Park. Concurrent capture’s infection Previous capture’s infection Coefficient SE p Coefficient SE p MANAGED HERD n = 605 n = 498 Intercept 0.082 0.012 0.071 0.014 A. marginale presence 0.009 0.011 0.388 0.018 0.012 0.131 A. centrale presence 0.039 0.012 0.001 0.017 0.013 0.185 Season (Wet) 0.069 0.009 <0.001 0.067 0.010 <0.001 Sex (Male) -0.032 0.020 0.122 -0.009 0.023 0.709 Age 0.095 0.012 <0.001 0.096 0.014 <0.001
FREE-RANGING HERD n = 699 n = 709 Intercept 0.005 0.048 -0.005 0.038 Season (Wet) 0.067 0.049 0.177 0.037 0.032 0.249 A. marginale presence -0.071 0.053 0.182 -0.036 0.038 0.349 A. centrale presence -0.014 0.051 0.790 0.016 0.036 0.658 Sub-herd (Lower Sabie) -0.250 0.055 <0.001 -0.251 0.046 <0.001 Age 0.161 0.077 0.037 0.105 0.060 0.079 Significant p-values (<0.05) are bold. SE – standard error
113
Supplementary Table 3.7: Model selection for models used in this paper. Model selection was conducted by forward selection starting with all possible interaction terms. Factors not producing a drop in >2 AIC were not included in the next step. If interaction terms were not significant, they were dropped regardless of AIC. Main explanatory variables that could be used included Anaplasma marginale burden (copies/reaction), A. marginale presence/absence, A. centrale burden (copies/reaction), A. centrale presence/absence, sub-herd, sex, season and age. The response variable is highlighted in bold. Anaplasma marginale and A. centrale burdens were log-transformed. Model Selection/Covariates AIC MH Average body condition score (C) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Sex*AM 1654.6 presence+Sex*AC presence+Age*AM presence+Age*AC presence+(1|Animal ID)) AM presence*AC presence+Season*AM presence+Sex*AM presence+Sex*AC 1652.6 presence+Age*AM presence+Age*AC presence+(1|Animal ID) AM presence*AC presence+Season*AM presence+Sex*AM presence+Sex*AC 1650.6 presence+Age*AM presence+(1|Animal ID) AM presence+AC presence+Season+Sex+Age+(1|Animal ID) 1641.9 MH Average body condition score (P) ~ previous AM presence*previous AC presence+Season*previous AM 1399.7 presence+Season*previous AC presence+Sex*previous AM presence+Sex*previous AC presence+Age*previous AM presence+ Age*previous AC presence + (1 | Animal ID) previous AM presence+previous AC presence+Season+Sex+Age*previous AM 1403.7 presence+(1 | Animal ID) Previous AM presence+previous AC presence+Season+Sex+Age+(1 | Animal ID) 1405.0 FH Average body condition score (C) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Age*AM 2277.0 presence+Age*AC presence+Sub-herd*AM presence+Sub-herd*AC presence+(1|Animal ID) AM presence*AC presence+Season*AM presence+Season*AC presence+Age*AM 2275.0 presence+Sub-herd*AM presence+Sub-herd*AC presence+(1 | Animal ID) Season*AM presence+Season*AC presence+Age*AM presence+Sub-herd*AM 2273.0 presence+Sub-herd*AC presence+(1|Animal ID) Season+AM presence+AC presence+Age+Sub-herd*AM presence+(1|Animal ID) 2270.1 FH Average body condition score (P) ~ previous AM presence*previous AC presence+Season*previous AC 2082.3 presence+Season*previous AC presence+Age*previous AM presence+Age*previous AC presence+Sub-herd*previous AM presence+Sub- herd*previous AC presence+(1|Animal ID) previous AM presence*previous AC presence+Season*previous AM 2080.3 presence+Season*previous AC presence+Age*previous AM presence+Sub- herd*previous AM presence+Sub-herd*previous AC presence+ (1|Animal ID) Season*previous AC presence+Age+Sub-herd*previous AM presence+(1|Animal 2079.3 ID) Season+previous AC presence+Age+Sub-herd*previous AM presence+(1|Animal 2080.2 ID) MH Average body condition score (C) ~ Season*AM burden+Sex*AM burden+Age*AM burden+(1|Animal ID) 283.32 Season+AM burden+Sex+Age+(1|Animal ID) 277.85 MH Average body condition score (P) ~ Season*previous AM burden+Sex*previous A. marginal burden+Age*previous AM 226.91 burden+(1|Animal ID)
114
Season+previous AM burden+Sex+Age+(1|Animal ID) 224.07 MH Average body condition score (C) ~ Season*AC burden+Sex*AC burden+Age*AC burden+(1|Animal ID)) 278.52 Season+AC burden+Sex+Age+(1|Animal ID) 274.23 MH Average body condition score (P) ~ Season*predictive AC burden+Sex*previous AC burden+Age*previous AC 223.39 burden+(1|Animal ID) Season+predictive AC burden+Sex+Age+(1|Animal ID) 222.94 FH Average body condition score (C) ~ Season*AM burden+Sub-herd*AM burden+Age*AM burden+ (1|Animal ID) 442.74 Season*AM burden+Sub-herd+Age+(1|Animal ID) 439.72 FH Average body condition score (P) ~ Season*previous AM burden+Sub-herd*previous AM burden+Age*previous AM 324.80 burden+(1|Animal ID) Season*previous AM burden+Sub-herd*previous AM burden+Age+(1|Animal ID) 322.81 Season*previous AM burden+Sub-herd+Age+(1|Animal ID) 324.55 Season+previous AM burden+Sub-herd+Age+(1|Animal ID) 327.11 FH Average body condition score (C) ~ Season*AC burden+Sub-herd*AC burden+Age*AC burden+(1|Animal ID) 443.71 Season*AC burden+Sub-herd+Age+(1|Animal ID) 442.80 FH Average body condition score (P) ~ Season*Previous AC burden+Sub-herd*Previous AC burden+Age*Previous AC 326.65 burden+(1|Animal ID) Season+Previous AC burden+Sub-herd+Age+(1|Animal ID) 324.21 MH Haematocrit (C) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Sex*AM 269.48 presence+Sex*AC presence+Age*AM presence+Age*AC presence+(1|Animal ID) AM presence+AC presence+Season+Sex+Age+(1|Animal ID) 261.37 MH Haematocrit (P) ~ previous AM presence*previous AC presence+ Season*previous AM 190.29 presence+Season*previous AC presence+Sex*Previous AM presence+Sex*previous AC presence+Age*previous AM presence+Age*previous AC burden+(1|Animal ID) previous AM presence+previous AC presence+Season+Sex+Age+(1|Animal ID) 179.82 FH Haematocrit (P) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Sub- 1763.2 herd*AM presence+Sub-Sub-herd*AC presence+Age*AM presence+Age*AC presence+(1|Animal ID) AM presence+AC presence+Season+Sub-Sub-herd+Age+(1|Animal ID) 1757.0 FH Haematocrit (P) ~ previous AM presence*previous AC presence+Season*previous AM 1613.1 presence+Season*previous AC presence+Sub-herd*previous AM presence+Sub- herd*previous AC presence+Age*previous AM infection+Age*previous AC presence+(1|Animal ID) previous AM presence*previous AC presence+Season*previous AM presence + 1611.1 Sub-herd * previous AM presence + Sub-herd * previous AC presence + Age * previous AM presence + Age * previous AC burden+ (1 | Animal ID) previous AM presence + previous AC presence + Season + Sub-herd + Age + (1 | 1603.0 Animal ID) MH Haematocrit (C) ~ Season*AM burden+ Sex*AM burden+Age*AM burden+(1|Animal ID) 73.663 Season+AM burden+Sex+Age+(1|Animal ID)) 73.952 MH Haematocrit (P) ~ Season*previous AM burden+Sex*previous AM burden+Age*previous AM 48.998 burden+(1|Animal ID)
115
Season+Sex*previous AM burden+Age+(1|Animal ID) 49.016 MH Haematocrit (C) ~ Season*AC burden+Sex*AC burden+Age*AC burden+(1|Animal ID) 75.723 Season+AC burden+Sex+Age+(1|Animal ID) 73.710 MH Haematocrit (P) ~ Season*previous AC burden+Sex*previous AC burden+Age*previous AC 52.196 burden+(1|Animal ID) Season+previous AC burden+Sex+Age+(1|Animal ID) 51.383 FH Haematocrit (C) ~ Season*AM burden+Sub-herd*AM burden+Age*AM burden+(1|Animal ID) 347.92 Season+AM burden+Sub-herd+Age+(1|Animal ID) 343.84 FH Haematocrit (P) ~ Season*previous AM burden+Sub-herd*previous AM burden+Age*previous AM 316.41 burden+(1|Animal ID) Sub-herd*previous AM burden+Age*previous AM burden+(1|Animal ID) 312.41 Sub-herd+previous AM burden+Age+(1|Animal ID) 310.53 FH Haematocrit (C) ~ Season*AC burden+Sub-herd*AC burden+Age*AC burden+(1|Animal ID) 352.57 Season+AC burden+Sub-herd+Age+(1|Animal ID) 349.31 FH Haematocrit (P) ~ Season*previous AC burden+Sub-herd*previous AC burden+Age*previous AC 323.69 burden+(1|Animal ID) Season+previous AC burden+Sub-herd+Age+(1|Animal ID) 317.85 MH Total protein (C) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Sex*AM -906.14 presence+Sex*AC presence+Age*AM presence+Age*AC presence+(1|Animal ID) AM presence+AC presence+Season+Sex+Age+ (1|Animal ID) -913.64 MH Total protein (P) ~ previous AM presence*previous AC presence+Season*previous AM -720.56 presence+Season*previous AC presence+Sex*previous AM presence+Sex*previous AC presence+Age*previous AM presence+Age*previous AC presence+(1|Animal ID) previous AM presence+previous AC presence+Season+Sex+Age+ (1|Animal ID) -727.91 FH Total protein (C) ~ AM presence*AC presence+Season*AM presence+Season*AC presence+Sub- 1388.7 herd*AM presence+Sub-herd*AC presence+Age*AM presence+Age*AC presence+(1|Animal ID) Season*AM presence+Season*AC presence+Sub-herd*AM presence+Sub-herd*AC 1386.7 presence+Age*AM presence+Age*AC presence+(1|Animal ID) Season*AM presence+Season*AC presence+Sub-herd*AC presence+Age*AM 1384.7 presence+Age *AC presence+(1|Animal ID) Season+AM presence+AC presence+Sub-herd+Age+ (1|Animal ID) 1375.4 FH Total protein (P) ~ previous AM presence*previous AC presence+Season*previous AM 860.54 presence+Season*previous AC presence+Sub-herd*previous AM presence+Sub- herd*previous AC presence+Age*previous AM presence+Age*previous AC presence+(1|Animal ID) previous AM presence*previous AC presence+Season previous AC presence+Sub- 858.54 herd*previous AM presence+Sub-herd*previous AC presence+Age*previous AM presence+Age*previous AC presence+(1|Animal ID) previous AM presence+previous AC presence+Season+Sub-herd+Age+(1|Animal 849.09 ID) MH Total protein (C) ~ Season*AM burden+Sex*AM burden+Age*AM burden+(1|Animal ID) -148.14
116
Season+AM burden+Sex+Age+(1|MBFTPChem$Anaplasma.AnimalID) -151.19 MH Total protein (P) ~ Season*previous AM burden+Sex*previous AM burden+Age*previous AM -110.10 burden+(1|Animal ID) Season+previous AM burden+Sex+Age+(1|Animal ID) -112.82 Season+AM burden*AC burden+Sex+Age+(1|Animal ID)) -158.59 MH Total protein (chemistry panel) (C) ~ Season*AC burden+Sex*AC burden+Age*AC burden+(1|Animal ID) -158.53 Season+Sex*AC burden+Age+(1|Animal ID) -160.39 MH Total protein (P) ~ Season*previous AC burden+Sex*previous AC burden+Age*previous AC -115.20 burden+(1|Animal ID) Season+previous AC burden+Sex+Age+(1|Animal ID) -116.97 FH Total protein (C) ~ Season*AM burden+Sub-herd*AM burden+Age*AM burden+(1|Animal ID) 204.90 Season+AM burden+Sub-herd+Age+(1|Animal ID) 199.73 FH Total protein (P) ~ previous AM burden + previous AC burden + Season + Sub-herd + Age + (1 | 165.53 Animal ID) previous AM burden+previous AC burden+Season+Sub-herd+Age+ (1|Animal ID) 165.53 FH Total protein (C) ~ Season*AC burden+Sub-herd*AC burden+Age*AC burden+(1|Animal ID) 201.81 Season+AC burden+Sub-herd+Age+(1|Animal ID) 196.03 FH Total protein (P) ~ Season*previous AC burden+Sub-herd*previous AC burden+Age*previous AC 167.58 burden+(1|Animal ID) Season+previous AC burden+Sub-herd+Age+(1|Animal ID) 164.10 MH Pregnancy status (C) ~ AM presence*AC presence+Age*AM presence+Age*AC presence+(1|Capture 391.27 Number)+(1|Animal ID) AM presence+AC presence+Age+(1|Capture Number)+(1|Animal ID) 388.49 FH Pregnancy status (C) ~ AM presence*AC presence+Herd *AM presence+Herd*AC presence+Age*AM 1207.0 presence+Age*AC presence+(1|Capture Number)+(1|Animal ID) AM presence+AC presence+Herd+Age+(1|Capture Number)+ (1|Animal ID) 1202.9 FH Pregnancy status (C) ~ AM burden*AC burden+Herd *AM burden+Herd*AC burden+Age*AM 191.07 burden+Age*AC burden+(1|Capture Number)+(1|Animal ID) AM burden+AC burden+Herd+Age+(1|Capture Number)+(1|Animal ID) 188.41 MH = Managed herd, FH = Free-ranging Herd, C = concurrent infection model, P = predictive infection model, AM = A. marginale infection, AC = A. centrale infection.
117
Supplementary Table 3.8: Generalised linear mixed model predicting haematocrit volume in managed and free-ranging herds of African buffalo from Kruger National Park. A. marginale A. centrale Coefficient SE p Coefficient SE p MANAGED HERD n = 124 n = 124 Intercept -0.103 0.080 -0.127 0.075 Season (Wet) 0.029 0.055 0.604 0.027 0.055 0.624 Anaplasma spp. burden -0.005 0.010 0.633 <0.000 0.008 0.944 Sex (Male) -0.092 0.087 0.296 -0.088 0.088 0.323 Age 0.419 0.089 <0.001 0.435 0.084 <0.001
FREE-RANGING HERD n = 223 n = 223 Intercept 0.120 0.079 0.033 0.084 Season (Wet) -0.093 0.070 0.186 -0.099 0.070 0.163 Anaplasma spp. burden -0.009 0.010 0.359 0.012 0.012 0.303 Sub-herd (Lower Sabie) 0.188 0.080 0.021 0.180 0.078 0.023 Age 0.357 0.136 0.009 0.386 0.136 0.005 Significant p-values (<0.05) are bold. SE – standard error.
118
Supplementary Table 3.9: Mixed effects logistic regression model predicting pregnancy status in a free-ranging herd of African buffalo from Kruger National Park. Free-ranging Herd Coefficient SE p OR MANAGED HERD n=129 Intercept 0.345 A. marginale burden -0.051 0.062 0.415 0.951 (-0.203 to -0.132) A. centrale burden -0.075 0.067 0.264 0.927 (-0.338 to -0.282) Age 0.742 0.607 0.222 2.100 (-0.448 to 1.932) Sub-herd (Lower Sabie) 0.274 0.417 0.511 1.316 (-0.543 to 1.091) Significant p-values (<0.05) are bold. SE – standard error, OR – odds ratio.
119
1
Minerva Access is the Institutional Repository of The University of Melbourne
Author/s: Sisson, Danielle
Title: Health and fitness effects of Anaplasma species infection in African buffalo (Syncerus caffer)
Date: 2017
Persistent Link: http://hdl.handle.net/11343/197971
File Description: Thesis
Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.