Genotyping of Mycobacterium tuberculosis and Mycobacterium leprae ancient DNA A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy in the Faculty of Science and Engineering 2020 Ammielle A. Kerudin School of Earth and Environmental Sciences
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Genotyping of Mycobacterium tuberculosis and
Mycobacterium leprae ancient DNA
A thesis submitted to The University of Manchester for the degree of
Doctor of Philosophy
in the Faculty of Science and Engineering
2020
Ammielle A. Kerudin
School of Earth and Environmental Sciences
2
Table of Contents
List of Tables ................................................................................................................... 5
List of Figures .................................................................................................................. 7
List of Abbreviations ..................................................................................................... 10
MIRU – microsatellites or microbial interspersed repetitive units
MOTT – Mycobacterium other than tuberculosis
MPC – Magnetic particle collector
MRCA – Most recent common ancestor
MTBC – Mycobacterium tuberculosis complex
N. dassonvillei – Nocardiopsis dassonvillei
N. dassonvillei – Nocardiopsis dassonvillei
N. dokdonensis – Nocardioides dokdonensis
N. moscoviensis – Nitrospora moscoviensis
NERC – National Environment Research Council
NGS – Next generation sequencing
PCR – Polymerase chain reaction
PGG – Principal genetic group
Pks 15/1 – Polyketide synthase 15/1
qPCR – Quantitative polymerase chain reaction
R. tataouinensis – Ramlibacter tataouinensis
RD2 – Region of difference 2
RD7 – Region of difference 7
RLEP – Mycobacterium leprae repetitive element
S. amylolyticus – Sandaracinus amylolyticus
S. avermitilis – Streptomyces avermitilis
S. bingchenggenesis – Streptomyces bingchenggenesis
S. cellulosum – Sorangium cellulosum
S. denitrificans – Steroidobacter denitrificans
S. espanaensis – Saccharothrix espanaensis
S. fulvissmus – Streptomyces fulvissmus
S. hindustanus – Streptoalloteichus hindustanus
S. laurentii – Streptomyces laurentii
S. roseum – Streptosporangium roseum
S. venezuelae – Streptomyces venezuelae
S. violaceusniger – Streptomyces violaceusniger
SAM – Sequence alignment mapping
SNPs – Single nucleotide polymorphisms
SPRI – Solid phase reverse immobilization
T. actinomycetes – Thermophilic actinomycetes
T. bispora – Thermobispora bispora
T. curvata – Thermomonospora curvata
12
TbD1 – M. tuberculosis specific deletion
Th1 – Helper T cell type 1
Th2 – Helper T cell type 2
TLR – Toll-like receptors
UV – Ultraviolet
V. paradoxus – Variovorax paradoxus
VNTRs – Variable number tandem repeats
WGS – Whole genome sequencing
WHO – World Health Organization
XDR-TB – Extensively drug-resistant tuberculosis
13
Abstract
The overall aim of this study was to employ a biomolecular technique – ancient DNA
(aDNA) – to study two ancient diseases that were endemic in Europe (and therefore
Britain) during the medieval period: tuberculosis and leprosy. In humans, the diseases
are caused by M. tuberculosis and M. leprae, respectively – both of which are members
of the M. tuberculosis complex (MTBC). Skeletal manifestations of both diseases may
develop in bone remains, which can be recognized using osteological analysis. In some
cases, however, the skeletal changes are ambiguous. Ancient DNA methods are used for
case confirmation and to answer historical questions such as the spread, origin and
evolution of disease. The first objective of this thesis was to determine whether the
MTBC aDNA detection frequency is high enough to plan a larger study to test
hypotheses such as possible strain differences in urban and rural areas, as it has been
suggested that urbanization assists the spread of tuberculosis, enhancing its virulence. To
meet this objective, 60 skeletal remains from 16 different locations in Yorkshire,
England were studied. All samples were screened for MTBC aDNA presence and 8
samples were selected for next-generation sequencing (NGS). In the PCR assay
screening, only 1 sample produced a positive MTBC amplification. However, when
subjected to NGS, this sample together with the other 7 samples did not produce enough
sequence reads to allow genome comparisons. An attempt to compare metagenomic
content between urban and rural sites was also performed. There was no specific
difference in metagenomic content between urban and rural samples. Based on the PCR
analysis, the sample St Andrew Fishergate 6, dated to the early 14th century AD, showed
evidence of possible tuberculosis infection. NGS analysis further revealed a possible M.
tuberculosis and M. leprae mixed infection, albeit with insufficient read coverage to
determine genome sequence polymorphism. The second objective was to use NGS to
determine the genotype of the M. leprae strains present in skeletons from two mediaeval
sites, at Chichester and Raund Furnells, both in England. This study served as a
continuation for the previous confirmation by PCR of leprosy in these skeletons. The
samples were further subjected to whole M. leprae genome target enrichment before
subsequent high-throughput sequencing. For all 3 historical M. leprae isolates, at least
70% genome sequence coverage was obtained, with a mean read depth of 4-10x. The
near-complete genome sequences that were obtained allowed subtype identification for
each of the ancient M. leprae isolates. Two mediaeval samples from Chichester
belonged to the 3I subtype, which is typical of ancient Northern European and
contemporary North American isolates. Meanwhile, an M. leprae isolate from Raunds
was identified as belonging to the 3K subtype – the first example of this subtype
identified in Britain. Transmission of the M. leprae 3K subtype to Britain is suggested to
have been associated with the travels of crusaders and pilgrims to the Holy Land during
the mediaeval period. The overall conclusion of the work is that although M. leprae
aDNA is well preserved in skeletal remains showing osteological signs of leprosy, the
same is not true for MTBC preservation in skeletons showing indications of
tuberculosis. To test hypotheses such as the effect of urbanisation on tuberculosis, a high
frequency of MTBC detection must be achieved, but this is complicated by the very
nature of ancient DNA itself – highly fragmented, low endogenous DNA copy, presence
of environmental contaminants – and by the possibility of low bacterial load in skeletons
at the time of death. In projects where the testing of a high number of samples is
required, more stringent selection criteria must be imposed to minimize the impact of
destructive analysis.
14
Declaration
No portion of the work referred to in the thesis has been submitted in support of an
application for another degree or qualification of this or any other university or other
institute of learning
Copyright
i. The author of this thesis (including any appendices and/or schedules to this
thesis) owns certain copyright or related rights in it (the “Copyright”) and
s/he has given The University of Manchester certain rights to use such
Copyright, including for administrative purposes.
ii. Copies of this thesis, either in full or in extracts and whether in hard or
electronic copy, may be made only in accordance with the Copyright,
Designs and Patents Act 1988 (as amended) and regulations issued under it
or, where appropriate, in accordance with licensing agreements which the
University has from time to time. This page must form part of any such
copies made.
iii. The ownership of certain Copyright, patents, designs, trademarks and other
intellectual property (the “Intellectual Property”) and any reproductions of
copyright works in the thesis, for example graphs and tables
(“Reproductions”), which may be described in this thesis, may not be owned
by the author and may be owned by third parties. Such Intellectual Property
and Reproductions cannot and must not be made available for use without the
prior written permission of the owner(s) of the relevant Intellectual Property
and/or Reproductions.
iv. Further information on the conditions under which disclosure, publication
and commercialisation of this thesis, the Copyright and any Intellectual
Property and/or Reproductions described in it may take place is available in
the University IP Policy (see
http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=24420), in any
relevant Thesis restriction declarations deposited in the University Library,
The University Library’s regulations (see
http://www.library.manchester.ac.uk/about/regulations/) and in the
University’s policy on Presentation of Theses.
15
Dedication and Acknowledgement
This work is dedicated to my parents, Kerudin Masintai and Rusinang Sibul who
inspired and taught me that nothing is more precious than education in life. You both
inspired me, prayed for me and supported me in a way that nobody could ever have done
and most importantly, it is your love that kept me going.
There are so many people whom I would like to thank – those who made my PhD
journey possible, bearable and enjoyable even – at times.
First, I submit my heartiest gratitude to my respected supervisor, Emeritus Prof Terry
Brown for his guidance, constant support and most of all his positivity and trust in me. I
could not have asked for a better supervisor. I would also like to thank my advisors, Dr
Richard Preziosi and Dr Russell Garwood for their guidance and support.
Secondly, I would like to show my appreciation to MARA who funded my PhD study –
which made it possible for me to pursue my PhD studies.
And to my family: brothers – Halley and Frankie, sister in laws, nephews and nieces; my
extended family and my family in law, thank you for your unwavering love, support and
encouragement which made it possible for me to go through some rough days along the
way.
To everyone in Brown’s group – thank you for your encouragement and support; for
treating me warmly for the last four years. Especially to Konstantina, Jannine, and Romy
who trained me in the lab; introduced and taught me about the “ancient DNA world”.
And to Kamalliawati for her lending ears, her house door that is always open for me and
for her constant encouragement. Most of all, thank you for making my thesis submission
possible – I wouldn’t have made it without you.
And to wrap this up, my biggest thanks to my husband, Benedict who have been
supportive from day one and for putting up with my rants and my emotional days. Most
of all, thank you for your patience throughout the years.
To my dear Estelle, you are my source of strength.
16
Chapter 1: Introduction
The history of mankind has been intertwined with infectious diseases. Tuberculosis and
leprosy, which are both contagious illnesses of ancient origins, have greatly affected the
course of world history. These diseases are still massive threats to human health to this
date, albeit their centuries of existence and the availability of treatments (World Health
Organization 2018a; World Health Organization 2018b). Their persistence in human
populations attracted studies from a wide range of fields. In sync to the notion: “the past
informs the present” (Brown & Barnes 2015, p.144), in the ever-growing discipline of
palaeopathology, ancient DNA is appropriately employed to answer historical questions
about palaeodiseases (Monot et al. 2005; Bos et al. 2014; Kay et al. 2015; Schuenemann
et al. 2018). The first isolation and detection of ancient DNA, from a dried muscle of
extinct Equus quagga in 1984, immediately ignited interest in the vast potential of this
tool in archaeology (Higuchi et al. 1984). Less than a decade later, Spigelman and
Lemma successfully isolated and detected the first microbial pathogen ancient DNA: M.
tuberculosis through the polymerase chain reaction (PCR) (Spigelman & Lemma 1993).
Since then, the study of ancient DNA in palaeopathology has not just been limited to
disease confirmation but also applied in answering broader and deeper questions, for
example, spread of disease through phylogeography and evolution of disease (Bos et al.
2014; Donoghue et al. 2015; Kay et al. 2015; Schuenemann et al. 2018).
1.1 Aims and Objectives
The overall aim of this study is to employ a biomolecular technique – ancient DNA – to
study two ancient diseases that were endemic in Europe (and therefore Britain) during
the medieval period: tuberculosis and leprosy. This chapter will focus on the background
of both ancient diseases studied: tuberculosis and leprosy. The biology of infection will
be described, followed by the manifestation of each disease in skeletons, which allows
the identification of the diseases in archaeological skeletons. The ability of tuberculosis
and leprosy to leave “fingerprint” changes in bones allows the identification of each
disease in skeletal remains using osteological methods. The limitation of this method
will be explained, which will then bring the ancient DNA method in to the picture. This
will be followed by the description of ancient DNA. The next section will describe the
ancient DNA study that has been done for both tuberculosis and leprosy to this date,
17
which includes case confirmation and evolutionary and disease origin studies. Chapter 2
describes the materials and methods performed as well as their relevance for this study.
The research described in the thesis had two objectives. The first objective was to
determine if the success rate of M. tuberculosis complex aDNA detection is high enough
to make it worthwhile to plan larger projects to test hypotheses such as possible strain
differences in urban and rural areas, as it has been suggested that urbanization assists the
spread of tuberculosis, enhancing its virulence (Comas & Gagneux 2011). Therefore, in
order to explore the association of urbanization and M. tuberculosis genotypes,
archaeological bones from urban and rural locations in Yorkshire, the majority with
tuberculosis indicative lesions which have been diagnosed by previous various
osteological studies, were examined. Yorkshire was selected because York was the
second largest city after London, during the 14th-15th centuries AD, and underwent rapid
urbanization during this time. The study of bones with tuberculosis lesions from
Yorkshire is split into 2 chapters: Chapter 3 and Chapter 4. The work described in
Chapter 3 aimed to screen all bone remains for the preservation of MTBC aDNA
through use of PCR assays of four targets selected based on their specificity for the
MTBC genome. The presence of MTBC aDNA is verified by cloning of PCR bands and
subsequent Sanger sequencing and matching the obtained sequences to the reference M.
tuberculosis H37Rv sequence. Based on the results obtained from Chapter 3, bone
remains were selected for next generation sequencing (NGS) using two sequencing
strategies, shotgun and target enrichment, which is described in Chapter 4. In Chapter 4,
the sequences obtained were then evaluated by various bioinformatic methods.
The work described in Chapters 3 and 4 showed that there is not extensive preservation
of MTBC aDNA in the bones from Yorkshire that were studied. The project was
therefore extended to include one additional objective. This additional objective aimed
to use NGS to determine the genotype of the M. leprae strains present in bones from
Chichester and Raund Furnells, both in England. This study serves as a continuation for
the previous confirmation by PCR of leprosy from Chichester and Raund Furnells, both
in England (Müller 2008). This work is described in Chapter 5. Leprosy was endemic in
Europe before it declined during the 16th century for unknown reasons, remaining only
in certain parts of Europe, human leprosy completely vanishing from Britain. Therefore,
archaeological remains are the only source of information on leprosy in Britain.
Identifying the genotypes of the M. leprae strains detected in the bones from Chichester
and Raund Furnells, as well as identifying new polymorphisms from each sequence in
18
comparison with the modern M. leprae TN strain (the reference genome that is used in
comparisons of M. leprae strain variability), will give an indication of the diversity of M.
leprae strains present in Britain during the medieval period, as well as giving clues about
the spread of leprosy to this part of the world. Finally, Chapter 6 describes the
conclusion of the works presented in this thesis.
1.2 The biology of tuberculosis and leprosy
Tuberculosis is one of the ten most common causes of global death (World Health
Organization 2018a). The World Health Organization (WHO) reported 1.7 million
fatalities caused by this disease alone, while there were 10.4 million people who
contracted the disease in 2016. Tuberculosis is also the leading cause of death among
HIV patients, accounting for 40% death in the same year (World Health Organization
2018a). This chronic granulomatous disease is caused by an obligate bacterium in
humans and animals (Brites & Gagneux 2015). There are seven closely related species
of Mycobacterium known with the ability of cause this disease: M. tuberculosis, M.
bovis, M. africanum, M. canettii, M. caprae, M. pinnipedii and M. microti (Wirth et al.
2008; Homolka et al. 2012). These pathogens are together known as the group members
of Mycobacterium tuberculosis complex (MTBC). The majority of infections in humans
and animals are caused by M. tuberculosis and M. bovis respectively (Bouwman et al.
2012). M. tuberculosis is transmitted through airborne droplets containing the pathogen
bacilli from an individual with active pulmonary tuberculosis infection, in which the
lung will be the primary site of infection (CDC 2013). On the other hand, the
transmission of M. bovis usually occurs through ingestion of contaminated milk and
meat, typically from cattle, where the gut will in turn become the main site of infection
(Atkins 2000; Waddington 2006). This is no longer a significant route of transmission
today, but bovine tuberculosis was a major problem in Britain from 1850 to 1950 where
it was responsible for approximately 800,000 deaths in the population (Atkins 2000;
Atkins 2008).
Tuberculosis bacilli are transmitted through droplet nuclei which will traverse the
mucociliary system and reach the alveoli of the lungs (CDC 2013). In this process, the
first line of defence would be the mucus-secreting goblet cells in the upper part of the
airways where the bacilli will be trapped on the mucus and subsequently “removed”
upward in accordance to the coordination by the cilia present on the cell surfaces
19
(Knechel 2009). The “unfiltered” bacteria will have the opportunity to invade the
alveolar spaces before they are subsequently phagocytized by the alveolar macrophages
(Pieters 2008). Following ingestion by macrophages, the mycobacteria will start to
multiply at a very slow rate (Knechel 2009). The macrophages will attempt to destroy
the bacteria by releasing cytokines and proteolytic enzymes (Flesch & Kaufmann 1993).
This cytokine production will trigger immune response from T cells through the
recognition of the mycobacteria antigen presented on the surface of macrophages. The
accumulation of immune cells including T cells, B cells and macrophages will aid the
formation of granulomas surrounding the M. tuberculosis, in an individual with good
cell-mediated immunity (Figure 1.1-a) (Pai et al. 2016). This will slow the replication of
the mycobacteria and prevents their spread. The micro-environment will contain the
bacteria but in the majority of cases, a complete eradication of the pathogen will not
occur but instead, they will adapt and survive (Forrellad et al. 2013). The low pH and
oxygen level conditions in the micro-environment will promote latency where the bacilli
survive within the granulomas but do not establish active infection to the host until the
immune system is compromised (Wayne & Hayes 1996; Cardona & Ruiz-Manzano
2004). The bacilli can stay in dormant stage, encapsulated in the calcified lesion, for
years, and even for a lifetime (Lin & Flynn 2010). In contrast, in hosts with a weaker
immune system, the infection will not be contained, therefore primary progressive
tuberculosis infection will be established (Figure 1.1-b) (Pai et al. 2016). The risks of
acquiring active tuberculosis infection is higher in HIV and diabetes patients,
malnourished people, as well as those who consume tobacco, with 8% of the cases
worldwide associated with smoking (World Health Organization 2018a). More than 95%
of tuberculosis cases and fatalities occur in the developing world; 45% and 25% of the
new cases are in Asia and Africa, respectively (World Health Organization 2018a).
Tuberculosis infection that occurs in the lungs is termed pulmonary tuberculosis
(Campbell & Bah-Sow 2006). The symptoms of pulmonary tuberculosis include cough
and sputum production, difficulties in breathing, haemoptysis, gradual wright loss, fever,
wasting, malaise, and anorexia in some cases (Campbell & Bah-Sow 2006). The typical
diagnostic methods for pulmonary tuberculosis include radiologic study, sputum direct
microscopy and culture and molecular detection (Ryu 2015). Although the lungs are the
primary site of tuberculosis infection, the bacilli can escape the lungs and spread to the
other parts of the body through the bloodstream and the lymphatic system (Ramirez-
Lapausa et al. 2015). Extrapulmonary tuberculosis can affect any part of the body but the
typical sites of infection are the lymph node, pleura, bones and joints or other skeletal
20
Figure 1.1: The mechanism of tuberculosis infection in cells and granulomas formation in active and latent tuberculosis infection. (a)
Inhaled tubercle bacilli enter the alveolar space of the lungs and bacteria elimination will be attempted by alveolar macrophages. In case of
unsuccessful bacteria elimination, the pathogens will invade the interstitial tissue of the lung. T cell priming will be initiated through the action of
either dendritic cells or monocytes – transporting the bacilli to pulmonary lymph nodes. This event will “attract” immune cells such as T cells and
B cells to the lung parenchyma leading to a subsequent granuloma formation. (b) M. tuberculosis replication occurs within the granuloma,
leading to high bacterial load. Failure of infection containment leads to bacteria dissemination to body organs. At this stage, the bacilli have the
ability to spread through the bloodstream or re-enter the respiratory tract to initiate similar immune reactions. In this infection phase, the host now
has active tuberculosis – will show symptoms and with the ability to infect other individuals (figure taken from Pai et al. 2016: 5).
21
parts, central nervous system, genitourinary system and abdomen (Golden & Vikram
2005). Extrapulmonary tuberculosis most frequently occurs in the lymph node and the
most frequent presentation is cervical lymphadenopathy but the involvement of the
inguinal, axillary, mesenteric, mediastinal and intramammary regions are also possible
(Golden & Vikram 2005). Approximately 35% of extrapulmonary tuberculosis cases
affect the skeletons; frequently spondylitis, followed by tuberculosis arthritis
concentrated on the joints (weight-bearing) and extraspinal manifestation (Golden &
Vikram 2005; Vanhoenacker et al. 2009). This type of tuberculosis involvement will be
discussed further in the next section. The regime treatment of a supposedly drug-
susceptible tuberculosis patient is a multi-drug combination consisting of rifampin,
isoniazid, pyrazinamide and ethambutol (Horsburgh et al. 2015). The inclusion of
ethambutol is not desired especially in young children due to its toxicity; it is usually
omitted in the treatment once drug-susceptibility is confirmed. The standard duration of
treatment for drug-susceptible patients is six months (Horsburgh et al. 2015). The multi-
drug combination regimen has been quite successful in treating tuberculosis patients; 53
million deaths were prevented between years 2000 and 2016 (World Health
Organization 2018a). However, the emergence of multi-drug resistance tuberculosis
(MDR-TB) and extensively drug-resistant tuberculosis (XDR-TB) halted the progression
towards zero tuberculosis cases (World Health Organization 2018a; World Health
Organization no date). According to the World Health Organization, there were
approximately 490,000 people with MDR-TB, with 600,000 newly reported cases of
rifampin drug resistance (World Health Organization 2018a). The treatment for
multidrug-resistant (MDR) tuberculosis is often complex and must be adjusted
specifically to the individuals by expert physicians according to drug-susceptibility
results performed using culture or DNA testing methods (Lange et al. 2014). Due to the
emergence of MDR-TB and XDR-TB, it is apparent that discoveries of new drugs and
vaccines are crucial. In addition, an effective public health system is also needed in order
to keep tuberculosis spread in check (Russell et al. 2010)
22
The second ancient disease of interest in this study is leprosy. This chronic infectious
disease causes manifestation on the skin, upper respiratory tract, peripheral nerves and
sometimes the eyes (Yassin et al. 1975; Lastória & de Abreu 2014). According to the
data gathered from 145 countries, there were 173,358 leprosy sufferers and 216,108 new
cases reported in 2016 alone (World Health Organization 2018b). Prior to 2008, M.
leprae was the sole leprosy pathogen known, before the identification of the second
leprosy causative pathogen: Mycobacterium lepromatosis (Han et al. 2008). Apart from
the majority infection observed in humans, leprosy infection has also been observed in
nine-banded armadillos in southern United States, chimpanzees, sootey mangabey
monkeys, and British Isles red squirrels (Meyers et al. 1991; Truman et al. 2011; Avanzi
et al. 2016). M. leprae is an extremely slow growing pathogen with a doubling time of
12.5 days (Jacobson & Krahenbuhl 1999). Similar to M. tuberculosis, this pathogen is
also an obligate intracellular bacterium (Groathouse et al. 2006). The study of M. leprae
is made difficult due to inability of its cultivation in normal cell culture media (Davis et
al. 2013). In vivo, leprosy has been studied using mouse foot-pad inoculation technique
and on immunocompromised mice, but neither systems are as successful as the
armadillo as an animal model for leprosy (Shepard 1960; Rees 1966; Kirchheimer &
Storrs 1971; Kirchheimer et al. 1972).
Although the precise path of M. leprae transmission is yet to be proven, an
overwhelming number of studies suggest that leprosy infection may spread through
nasal discharge containing M. leprae bacilli from untreated leprosy patients (Rees &
McDougall 1977; de Wit et al. 1993; Martinez et al. 2011). The bacilli will then make
their way into the healthy individual’s body through the respiratory route. Another
possible way for the bacilli to gain entry is through skin contact with a leprosy patient
(Satapathy et al. 2005). The bacilli enter the healthy individual’s body, possibly through
the skin and nose, and the host innate immune system will recognize the “foreign body”
and be triggered. This occurs through recognition of the mycobacterial lipoproteins by
the Toll-like receptors (TLRs) on macrophage and monocyte surfaces (Walker &
Lockwood 2006). In the case of M. leprae infection, recognition by the TLR2/1
heterodimer will allow the differentiation of monocytes into both dendritic cells and
macrophages (Modlin 2010). This will be followed by naïve T cell activation through
antigen presentation by dendritic cells to trigger more Th1 response (Walker &
Lockwood 2006).
23
Rather than M. leprae virulence, the clinical manifestation of this disease is determined
by the immune status of the individuals affected (Jin et al. 2018). One of the
classifications of leprosy discriminates between tuberculoid and lepromatous leprosy, as
well as the borderline intermediate infection, based on the resistance of the patient
towards the infection (Ridley & Jopling 1966).
An individual with stronger immune status will be able to “fight” the infection and will
only manifest the “mild” tuberculoid leprosy infection (Jacobson & Krahenbuhl 1999).
In this case, intense phagocytic activity will be triggered in lesions through the
secretions from Th1 T-cells (IL-2, lymphotoxin-α and interferon (IFN)-γ) (Walker &
Lockwood 2006; Bobosha et al. 2014). The cytokine-triggered macrophages will form
granulomas together with lymphocytes (Wang, Maeda, et al. 2013). Inside the
granulomas, T cells will produce granulysin, an antimicrobial protein (Walker &
Lockwood 2006). For these individuals, the CD4+ cells will outnumber the CD8+ cells
(Modlin 1994). Therefore, the infection will be halted by “trapping” the M. leprae inside
well-formed granulomas.
In contrast to tuberculoid leprosy, the granuloma is not so well-formed in lepromatous
leprosy, thus, making it difficult to “trap” the M. leprae bacilli (Wang, Maeda, et al.
2013). This is due to the different types of cytokines produced: IL-4, IL-5 and IL-10
(Yamamura et al. 1991). The cytokines IL-4 and IL-10 are capable of downregulating
TLR2 on monocytes while the latter will inhibit IL-12 production (secretion of which
will allow the activation of naïve T cells) (Zumla & James 1996; Walker & Lockwood
2006). The Th2 immune response will be activated while suppressing that of Th1, which
is responsible for initiation of the macrophage response (Modlin 1994). This type of
leprosy infection is characterized by the absence of granulomas and lack of successful
cell-mediated immunity.
The third type or degree of leprosy infection is intermediate or borderline infection
(Ridley & Jopling 1966). For this spectrum, the immunology response can change from
tuberculoid to lepromatous (White & Franco-Paredes 2015). Three types of borderline
leprosy infections are: borderline borderline (BB), borderline lepromatous (BL) and
borderline tuberculoid (BT) (Ridley & Jopling 1966). The borderline tuberculoid
spectrum of infections is the most common type of leprosy, although the exact
immunology is still not well-understood (Ankad 2018).
24
The clinical representation of leprosy depends on the strength of the host’s immune
system to fight M. leprae infection (Jin et al. 2018). In tuberculoid leprosy, there are
very few lesions found (Zumla & James 1996). The lesions are usually plaques and
macules with sharply demarcated edges (Lynnerup & Boldsen 2012). The appearance of
the lesions is usually hairless and hypoesthetic, which is due to the damage on the
dermal nerve fibres (Gunatilake & Settinayake 2004). The spread of infection to the
peripheral nerves such as the tibia, fibula and ulna is possible, however the damage is
limited. The effects on the nerves of a tuberculoid leprosy patient may include noticeable
enlargement of the area where the peripheral nerves are affected (Gunatilake &
Settinayake 2004). This is due to the granulomatous inflammation on the nerves which
in turn cause sensory and motor loss, resulting in the inability to detect pain (Walker &
Lockwood 2006). The swelling nerves could also experience further impairment as a
result of the entrapment inside the fibro-osseous tunnels.
In lepromatous leprosy, the early signs of disease include skin changes in the form of
macules distributed uniformly and widely on the face predominantly, and some are also
distributed on the upper limbs of the body (Lynnerup & Boldsen 2012). The edges of
these macules are almost undistinguishable and usually coupled with hypopigmentation
and redness on the skin (Gunatilake & Settinayake 2004; Reibel et al. 2015). In addition,
the face skin may thicken when the disease is neglected and untreated, and this results in
an appearance known as “leonine face” (Walker & Lockwood 2006). Dermal
involvement is pronounced in leprosy patients in this pole, which may give rise to
“glove and stocking” neuropathy (Sabin & Ebner 1969). The peripheral nerves may be
affected at a later stage (Lynnerup & Boldsen 2012). Skin lesions characteristic of
patients with borderline leprosy have characteristics in between those of tuberculoid and
lepromatous leprosy.
M. leprae shows a strong tropism in macrophages and Schwann’s cells of the peripheral
nervous systems (Reibel et al. 2015). Inside the cells, the pathogen bacilli will actively
proliferate, causing cell deterioration and failure to regenerate (Spierings et al. 2000).
Untreated leprosy nerve involvement will eventually result in deformity and disability,
sometimes irreversible (Lastória & de Abreu 2014). This occurs through sensory loss as
a result of nerve damage which in turn causes unnoticed injury, which leads to
secondary infections, causing damages in tissues (White & Franco-Paredes 2015). The
patient will eventually become disabled from the loss of motor function.
25
Multidrug combination regimen was introduced in 1982 which includes rifampicin,
clofazimine and dapsone (Jacobson & Krahenbuhl 1999; Reibel et al. 2015). The
duration of the course of treatment is still in debate; WHO has fixed the duration for 12
months, however many still argue that the duration is not sufficient for the infection to
resolve (Kumar et al. 2013; Malathi & Thappa 2013).
1.3 Skeletal changes in tuberculosis and leprosy
The field of palaeopathology has traditionally relied heavily on macroscopic inspection
to recognize ancient diseases in bone remains. Macroscopic or visual examination
provide direct evidence of such diagnosis in skeletal remains. It is crucial to understand
the effects of diseases in the human body, especially the skeleton, in a modern clinical
sense in order for the diagnosis to be made with skeletal remains (Roberts and
Manchester 2010). In bone remains, the osteological changes caused by infectious
diseases could be in the form of bone destructions, bone formation, or a mixture of both
(Roberts and Manchester 2010). Skeletal involvements are also common in modern
clinical samples, in the post-antibiotic era (Steyn et al. 2013; Steyn & Buskes 2016). The
lesion distribution on the skeleton is an important clue to identify the type of infection
the individual had suffered. Typically, palaeopathology researchers are performing the
osteological analysis using macroscopic and radiological methods but the histological
approach has become increasingly popular.
The skeletal changes in tuberculosis and leprosy are a manifestation of a chronic, long-
established infection in a relatively healthy individual (Zink et al. 2001; Tayles &
Buckley 2004; Adachi et al. 2006; Rubini, Zaio & Roberts 2014; Suzuki et al. 2014;
Steyn & Buskes 2016; Inskip et al. 2017). The tuberculosis bone changes occur in an
individual with a secondary tuberculosis infection, where the infection of a person with
latent tuberculosis is reactivated (Ortner 2003; Roberts & Manchester 2010). In
secondary tuberculosis infection, the tubercle bacilli spread haematogenously within the
bones to other parts of the body (Roberts & Manchester 2010). The bacilli will
predominantly reside in the skeletal areas where the circulatory and metabolic rates are
high inside the skeleton, e.g. in the haemopoietic or red marrow (Ortner 2003).
Regardless of the age of individuals, the red bone marrow can be found in all bones
including the vertebrae, sternum and ribs (Prabhakar et al. 2009).
26
Tuberculosis infection in skeletal remains can be recognized from bone destruction in
the spine, especially the lower thoracic and lumbar vertebrae (Roberts and Manchester
2010). In most cases, the bone changes are severe which involves abscess development
inside the vertebrae which in turn perforate into the abdomen or chest, and the affected
spine will finally collapse (Figure 1.2 – a) (Roberts & Manchester 2010; Holloway et al.
2013). In archaeological remains, result of the spinal collapse will cause an angular
deformity of the spine. This is the most common type of skeletal tuberculosis, which
accounts for 25-60% of infections (Roberts & Manchester 2010). Skeletal lesions on the
vertebrae are usually shown as osteolytic lesions on the frontal area of the lower thoracic
and the upper lumbar, in which the number of affected vertebrae is usually between one
to four (Holloway et al. 2011; Rasouli et al. 2012). Following the spine, the knee and hip
are the second part of the body that are frequently infected which make up to 10-20 to
15-30% of skeletal tuberculosis cases other than spine involvement, respectively (Ortner
2003; Roberts & Manchester 2010). The tuberculosis infections on these joints are
characterised by fibrous fixation also known as ankylosis (Roberts & Manchester 2010;
Saraf & Tuli 2015). However, great care has to be taken on confirming tuberculosis
diagnosis based on this evidence as it may also be caused by pyogenic osteomyelitis,
brucellosis and fungal infections (Holloway et al. 2011).
Other than that, there are other non-specific bone changes of tuberculosis. These include
new bone formation giving rise to rib lesions, which are increasingly common in post-
antibiotic era (Steyn & Buskes 2016). The tuberculosis lesions on the ribs could be the
result of chronic pulmonary tuberculosis (Figure 1.2 – b) (Kelley & Micozzi 1984;
Matos & Santos 2006). These lesions are typically subtle with the characteristics of
periosteal new bone formation as well as bone resorption (Roberts & Buikstra 2003).
The lesions usually result from bloodstream spread of bacilli, but can also be caused by
direct extension from neighbouring tuberculosis foci and extension from abscesses
within the paravertebral region (Ortner 2003; Roberts & Manchester 2010). New bone
formation is the result of the response of bone tissue to traumatic insults or pathological
infection. The bone infection will cause an inflammatory response that will trigger the
production of new bone on top of the periosteal bone surface, which sometimes can
result in a different colour to the bone ‘layer’ (Figure 1.2 -b). In 2003, Ortner
mentioned that rib tuberculous changes normally occur in 9% of tuberculosis sufferers
(Ortner 2003). In minor cases, dactylitis, which is toe or fingers inflammation, and skull
involvements are also characteristics used for tuberculosis diagnosis (Holloway et al.
2011; Steyn & Buskes 2016). Although these non-specific bone changes have been
27
Figure 1.2: The bone changes indicative of tuberculosis infection in skeletal remains.
(a) The destructive spinal lesions and collapse of vertebrae which is considered pathognomonic
for tuberculosis. The red arrow indicates where the vertebral bodies collapse occurred, which
results in spinal angular deformity in people suffering from spinal tuberculosis. (b) Rib bones
showing new bone formation, as shown by the yellow arrows. New bone formation is one of the
characteristics of probable tuberculosis infection in skeletal remains. (Images from Charlotte
Roberts’ personal collection, personal communication 2019).
(a)
(b)
28
considered more and more in future studies, especially the rib lesions, caution should be
taken as to never treat these as a definitive evidence of tuberculosis infection (Roberts
and Manchester 2010). The non-specific bone changes can also be caused by other
conditions. Endocranial new bone formation could also be caused by induced
meningitis. Similarly, rib lesions can also be attributed to other diseases such as
metastases, unspecific osteomyelitis, pneumonia, bronchiectasis and mycosis (Nicklish
et al. 2012).
Characterisation of bone changes in skeletal remains suspected to be infected by leprosy
was first undertaken by a Danish physician Vilhelm Møller-Christensen in 1953 (Møller-
Christensen 1953; Lynnerup & Boldsen 2012). Leprosy is a slow progressing, chronic
infection which explains its strong manifestation on the skeleton as written by Lynnerup
and Boldsen (2012), “leprosy is a disease one dies with rather than of”. Bone lesions of
leprosy infection in archaeological remains are recognizable on the skull, the extremities
of the hands and feet as well as the lower legs (Thappa et al. 1992; Roberts &
Manchester 2010; Mohammad et al. 2016). As previously mentioned, M. leprae bacilli
attack the peripheral nerves, causing functional loss (Gunatilake & Settinayake 2004).
This will subsequently cause deformities, with most affected regions being the upper and
lower limbs due to muscle paralysis on these regions. Consequently, anaesthesia will
occur on the hands and feet, thus allowing secondary infection to manifest, causing
tissue necrosis due to the unnoticed injuries. In turn, this will give rise to an appearance
of what is known as “clawed hands” (Ortner 2003; Gunawan et al. 2017). On the foot,
the loss of bones on the distal phalanges also may occur, giving rise to a noticeable
“pencil ends” configuration on the metatarsals (Figure 1.3) (Roberts & Manchester
2010). This appearance is associated with the bone atrophy that always takes place
starting from the distal margins of the hands or feet (Barnetson 1951). This bone atrophy
could then proceed to the proximal bones. For example, on the foot, bone absorption
may begin on the distal phalanx, and proceed to the proximal bones, resulting in the loss
of all or some of the phalanges. Further progress of the bone absorption could lead to the
pencilling of the metatarsals, as depicted in Figure 1.3. Although not as frequently as on
the foot, the same pattern of bone absorption can occur on the hand (Ankad et al. 2011).
The bone absorption that typically starts from the distal fingertip, will cause a defect that
looks like a V-shape or “sharpened” fingertip (Lu et al. 2015). Similarly, the worsening
and further progression of the bone absorption will eventually lead to the disappearance
29
of the distal phalanx or will leave behind irregular looking bone remains. The bone
absorption will progress to proximal bones.
Figure 1.3: The bone changes of leprosy on the foot.
The image below shows the ‘pencilling’ appearance of the foot metatarsals and loss of some of
the foot phalanges. As a comparison, the image on top is showing the normal foot appearance in
a modern skeleton. The blue brackets are showing the metatarsal bones, while the red brackets
are showing the phalanges in both normal and affected bones. The yellow arrows depict the
affected bones where the loss of phalanges and pencilling of metatarsal bones has occurred.
(Image from Charlotte Roberts’ personal collection, personal communication, 2019).
30
On the skull, the most recognizable leprosy change is what is termed as rhinomaxillary
syndrome (Ortner 2003). This term was introduced by Andersen and Manchester (1992)
in place of ‘facies leprosa’ proposed by Møller-Christensen in 1978. The former is a
preferred term in palaeopathology as the latter is also used to describe the soft tissue
changes and therefore will not best fit the palaeopathology contexts (Andersen &
Manchester 1992). Rhinomaxillary syndrome is exclusively the result of lepromatous
and near-lepromatous leprosy, therefore the presence of all elements of the bone changes
is pathognomonic of these conditions (Andersen & Manchester 1992; Nerlich & Zink
2008). The tendency of M. leprae to invade the cooler regions of the face explains the
lesion distribution on the mucosal membranes and the cooler exposed skin area
(Andersen & Manchester 1992). The bony lesions in rhinomaxillary syndrome can be
observed in the nasal cavity, the maxilla, on the oral surface of the palatine and alveolar
processes, without any effect shown on the mandible. The absence of lesions in the
mandibular region is explained by the difference in temperature in the regions affected.
Perforation is also possible on the palate, accompanying the inflammatory pitting on
both sides of it (Roberts & Manchester 2010). Apart from that, incisor teeth loss also
may occur as a result of the alveolar bone loss on the upper jaw in that region of the
teeth. The effects on the nasal spine and nasal aperture will include absorption and
remodelling, respectively.
There are always limitations in any discipline, and this is particularly the case with
palaeopathology. This has been highlighted by Wood and colleagues (1992), who
described the “osteological paradox”, which summarises the uncertainties that arise
when it is attempted to use skeletal evidence to study the health and disease status of a
past population. The author highlighted three major problems in palaeopathology. These
are the non-fixed state of the population (so a single or small group of skeletons from
one point in time is not representative of a large population that might be undergoing
change due to migration and mixing with other populations), selective mortality (not all
people with a disease died because of that disease), and the individual variations in the
risks of contracting disease and death (so skeletons with palaeopathological lesions
might represent only a biased proportion of population as a whole). These are the
limitations that are considered impossible to overcome. Other limitations that should be
taken into consideration is that the ‘dead populations’ being studied cannot be
considered to be representative of the living populations (Roberts and Manchester 2010).
Besides, the diagnosis confirmation is heavily dependent on the state of preservation of
31
the skeletal remains. Furthermore, bone changes in acute diseases might be difficult to
detect as the individuals might have died quickly from the disease without having
enough time to develop bone changes.
The most challenging aspect of ancient disease diagnosis based on osteological
observation is the lack of bone lesions specific for certain diseases. The abnormal bone
lesions observed could be the result of different diseases, therefore affecting the
accuracy of the differential diagnosis (Klaus 2017). This is particularly apparent in the
diagnosis of tuberculosis as the non-specific bone lesions could also be a result of other
diseases as previously described in this section. Furthermore, even if there are
pathognomonic bone changes that can diagnose a disease correctly, it would still be
dangerous to use this information to infer the disease status of the population from which
the studied skeleton originated. This information would still suffer from the osteological
paradox. Therefore, to overcome this, it is useful to incorporate other methods such as
molecular study of a skeleton to identify the disease or diseases suffered by that
individual in the past. The biomolecular study of skeletal remains, specifically the use of
ancient DNA, will be further described in the following section 1.6. This section
explains how such studies may complement and add precision to palaeopathological
studies of both tuberculosis and leprosy.
1.4 History of tuberculosis and leprosy based on historical
documents and skeletal changes evidence
1.4.1 History of tuberculosis
Historical documents – written or pictorial – are the key sources which can indicate the
antiquity of certain diseases, albeit the interpretations should be made with caution. The
first convincing historical evidence of tuberculosis appeared in the Chinese literature,
documented by Emperor Shennong of China, dating to approximately 2700 BCE
(Tripathy 2015). This medical text mentioned “xulao bing” which can be interpreted as
“weak consumption”, thought to describe tuberculosis. Medical papyrus from Egypt
such as the Ebers Papyrus (1500 BCE) is the other well-known ancient document which
outline diseases and treatments affecting populations in that area during that period
(Khalil & Richa 2014). A few authors are convinced that this document might be
32
describing tuberculosis. However, Chalke (1962), in much earlier writing is convinced
otherwise. The same differing opinions are held regarding the Bible references about
tuberculosis. Daniel and Daniel (1999) suggested that two verses from the Old
Testament are describing tuberculosis, but Chalke has stated that the descriptions from
the Bible are ambiguous and there is not sufficient information to support the description
of tuberculosis such as is present today (Chalke 1962; Daniel & Daniel 1999).
Throughout different periods, tuberculosis has been called by many names. Tuberculosis
was known during the Classical Greece period as ‘phthisis’, according to the records in
Greek ancient literature, during the period of Hippocrates (460-370 BC) (Moonan 2018).
He was one of the most well-known Greek physicians, who stated in his records that
‘phthisis’ was the most typical disease where death is almost inevitable during that time.
In addition, Hippocrates also described the symptoms of ‘phthisis’ in his Book 1 ‘of the
Epidemics’ including fever, coughing accompanied by concentrated sputa, colourless
urine, and loss in desire to consume food and drinks (Firth 2014). His book also noted
that people who severely suffering from ‘phthisis’ were in the age range between 18 and
35. Hippocrates was in agreement with others who believed at this time that the nature of
‘phthisis’ transmission was hereditary (Lakhtakia 2013). However, the famous Greek
philosopher Aristotle believed that this disease is contagious rather than hereditary, in
contrary to other contemporaries during that period. After Hippocrates, another
renowned physician, Galen, described phthisis, emphasising the presence of “lung
ulcers”, throat or thorax, and body consumption by pus in addition to the other
symptoms already described by Hippocrates (Rosenthal 2013). Some treatment methods
were also outlined such as the use of opium to induce sleep and to numb pains,
bloodletting and recommended diets (Tripathy 2015).
In Europe, tuberculosis was endemic in the 17th century AD and continued to be so for
two centuries, which is why it was known as the “Great White Plague” (Barberis et al.
2017). During this period, tuberculosis was the leading cause of death and whoever
contracted the disease was considered to have received a death sentence, as fatality
occured in almost all instances of infection. The spread of the disease agent was
aggravated by the living conditions in Europe during that time, which were overcrowded
and unsanitary (Roberts & Manchester 2010; Tripathy 2015). In the 17th century AD, the
Bills of Mortality were available in London as a record of mortalities together with their
diagnosis – with tuberculosis being the most common cause of death (Matossian 1985).
Later, from the 19th century onwards, the Registrar General’s statistics provide a quite
33
accurate representation of the disease progress in both Wales and England, which show a
stable drop in the prevalence of tuberculosis in these regions except during the World
Wars (Davies et al. 1999). The 19th century AD was the period of tuberculosis
romanticism (Daniel 2006). During this time, tuberculosis claimed many lives among
the glamorous, talented, smart and good-looking young individuals in society. The
characteristics of tuberculosis sufferers, pale and thin, added to the “romantic” aspect of
the disease, which was considered stylish during that time. In fact, tuberculosis is
depicted in many arts. Portrait of individuals with hunched-back and thin, pale looking
young women were often depicted during this era (Roberts & Buikstra, 2003).
There are many historical documents available throughout the centuries about
tuberculosis. However, precaution should be taken in using or verifying this information.
The accuracy of the tuberculosis diagnosis is questionable; there are instances where the
disease was wrongly assigned (Davies 1998). Diseases with similar symptoms should be
taken into consideration when confirming these diagnoses. For example, non-
tuberculous pneumonia and bronchitis may exhibit similar symptoms to tuberculosis
(Ukil 1940). In addition, the credibility of the individuals who assessed the disease must
also be taken into consideration (Roberts & Buikstra, 2003).
Other than historical documents, the antiquity of tuberculosis can also be inferred
through osteological analysis as described in section 1.3. The earliest skeletal evidence
in different locations based on the study of skeletal changes is described here. The
earliest skeletal evidence of tuberculosis is a female skeleton showing spondylitis found
in the Neolithic level of Arma dell’Aquila Cave in Liguria, Italy (Canci et al. 1996). The
age of the skeleton (5800±90 years BP) was determined based on the 14C radiocarbon
dating of a neighbouring grave located on the same horizon. There is also evidence of
skeletal tuberculosis in the Old World at the Late Neolithic tell settlement site in the
South of Hungary dated 4932-4602 cal BC (Masson et al. 2015). The earliest skeletal
evidence in Britain was found at Tarrant Hinton, Dorset, at an Iron Age site (400-230
BC) (Mays & Taylor 2003). Evidence of tuberculosis in Asia comes from relatively
recent sites, from China, dated to the second century BC, Japan (454 BC – 124 AD) and
Thailand (300 BC – 300 AD) (Suzuki & Inoue 2007; Stone et al. 2009). Plenty of
tuberculosis evidence is available from this period onwards including skeletal evidence
found in France, Lithuania and Austria dated to the 4th century (Stone et al. 2009). In the
Americas, skeletal evidence for tuberculosis mounted throughout the 20th century AD
(Stone et al. 2009). In South America, one of the earliest convincing pieces of evidence
34
of tuberculosis from human remains is from the Andean populations of Peru and
Northern Chile (Allison et al. 1981). Meanwhile in North America, all the skeletons with
evidence of tuberculosis are dated after 1000 AD (Roberts & Buikstra 2003).
Confirmation of ancient tuberculosis cases based on additional biomolecular evidence
will be discussed further in section 1.6.1.1.
1.4.2 History of leprosy
Similar to tuberculosis, leprosy also has existed since antiquity. The earliest mentions of
this disease can be traced back from ancient Egypt (1550 BC), India (600 BC); and
China (4th century BC – 217 BC) (Leung 2009; Robbins et al. 2009). Many authors
believe that the Ebers Papyrus from 16th century Egypt includes a description of leprosy
infection (Robbins et al. 2009). However, a few scholars are sceptical, based on the
accuracy of the translation of this manuscript (Mark 2002). In fact, the notion was
previously rejected and faced strong criticism due to a lack of consistency between the
disease characteristics as described and the symptoms of modern leprosy as we know
today. It was later agreed that the disease description is more fitting to gas gangrene
compared to leprosy (Mark 2002). Sushruta Samitha (600 BC), an ancient Indian
medical document, provided descriptions of disease characteristics that almost all agree
with advanced lepromatous leprosy symptoms (Dharmendra 1947; Jacob & Franco-
Paredes 2008). These include sensation loss, ulceration and disfigurement of limbs,
collapsing of the nose, and falling off of fingers (Roberts & Manchester 2010). Sushruta
Samitha is the only ancient medical writing that gives the most precise description of
leprosy. These texts were estimated to be compiled around 600 BC but the writings are
believed to contain information which was acquired from a much earlier period
(Dharmendra 1947). There are other Indian ancient literatures that might have mentioned
this disease; these are the Laws of Manu and the Atharava Veda (Jacob & Franco-
Paredes 2008). However clinical descriptions of the disease are absent, therefore there is
no significance confidence to confirm that the disease described was leprosy
(Dharmendra 1947). A description of leprosy from China has been acquired from a 3rd
century bamboo book which outlined similar disease characteristics as described in the
Sushruta Samitha (Roberts & Manchester 2010). The presence of these ancient medical
documents with such articulate descriptions of the disease raises the possibility that
leprosy might have existed in the Far East and India from a very early period, as the
35
authors seem to have a deep clinical understanding and judgement of the disease
(Roberts & Manchester 2010). Based on the presence of such detailed and accurate
medical writing and the presence of clinically skilled individuals, it is quite possible that
leprosy was prevalent in these regions during that period of time (Roberts & Manchester
2010). It has also been suggested that leprosy is referred to in the Book of Leviticus, in
the mention of the word tzaraat, but this proposal was later dismissed as a
mistranslation, and it is now accepted that the tzaarat mentioned in the bible was a range
of skin diseases and not the leprosy that we know of today (Trautman 1984; Mark 2002;
Grzybowski & Nita 2016).
In Europe, the prevalence of leprosy infection peaked in the 11th to 14th centuries AD
(Mendum et al. 2014). Leprosy was feared and stigmatized due to the disability and
disfigurement that it caused as well as the lack of understanding of this disease during
that time (Bennett et al. 2008; Sermrittirong & Van Brakel 2014). Leprosy was seen as a
punishment from God and the patients were isolated in a specific area or “facility” called
a leposarium, to avoid contagion from the infected individuals (Donoghue et al. 2015).
During the Middle Ages in Europe, people who were suffering from leprosy infection
were required to wear distinguishing attire, and had to ring bells to let people know that
they were close by (Bennett et al. 2008). In addition, they were also expected be on the
side of the road where the wind does not blow from, as people were so afraid of this
disease and they shunned whoever was suffering from it (Hussain 2007). Leprosy, due to
its unpleasant outcome, was viewed as a blasphemy and even thought to be inherited.
Due to its dramatic representation and high prevalence throughout history, leprosy has
inspired many art works (Gron 1973). Leprosy has been one of the most misunderstood
diseases and still is in certain regions of the world (Sermrittirong & Van Brakel 2014).
Luckily, the discovery of the causative pathogen M. leprae in 1874 by Dr. Gerhard
Armauer Hansen of Norway was the most important moment in the history of leprosy as
it opened a pathway to the understanding of this disease (Irgens 1984; Trautman 1984;
Ghosh & Chaudhuri 2015).
The earliest convincing evidence of leprosy in human remains is from Balathal India,
dated to 2,000 BC (Robbins et al. 2009). This might support the speculation that the
Mediterranean leprosy brought to Europe by the armies of Alexander the Great
originated in the Indo-Gangetic basin, which was not supported by any skeletal evidence
before in the review by Stone and colleagues (2009). One of the earliest evidence of
leprosy in skeletal remains has also been observed at the Dakleh Oasis in Egypt (250
36
BC) (Dzierzykray-Rogalski 1980) and in Israel and Nubia (Stone et al. 2009). More
leprosy bone studies will be elaborated further in section 1.6.2.
1.4.3 The concern in using historical documents as evidence of ancient
disease
Historical texts are crucial in the process of deciphering diseases in historic populations
(Metcalfe, 2007). The writers who lived in the past can provide a clue about the culture
of societies during that time, which then can assist the palaeopathological analysis of
human remains. In some historical texts, the signs and symptoms of diseases are
recorded clearly, and therefore can provide a direct source of information. In
palaeopathological studies, the exclusion of historical written documents will pose extra
challenges to the research itself and might cause inaccuracy in the interpretation
(Mitchell 2017). However, it is also worth noting that cautions must be taken in
interpreting historical documents as evidence of diseases in the past.
The differences in belief and perspective during the historic period might put the study at
risk of misinterpretation (Mitchell 2011). As the ancient medical texts are derived from
various civilizations, it is expected that they were written according to the perspective of
the belief in that geographical area during that time (Roberts 1971). Consequently, this
might give rise to a different description that can be easily misinterpreted from the
context of what we know in this modern medical period. Therefore, it is important to
consider the time when the historical documents were written and the interpretation must
be made by someone with an expertise in the beliefs and perspective of the ancient
civilization (Mitchell 2017).
Furthermore, the evidence of disease must be obtained by reading the original version of
the historical documents (Mitchell 2011; Mitchell 2017). Modern translations or any
quotes from secondary sources are best to be avoided where possible. The concern with
using documents from modern translation is that the translator might not have a medical
background, therefore affecting the accuracy of the disease’s descriptions. Therefore, it
is highly desirable that palaeopathologists work hand in hand with medical historians
who are sufficiently versed in the linguistics to make sure that the original texts are
properly translated and interpreted (Mitchell 2017). In addition, we also cannot be sure
that the medical writers have witnessed the disease during their lifetime or simply just
37
copied it from the previous records of their predecessors. If this is the case, it will give a
wrong indication of the timeline of the disease – or if it really has existed in the
population during the time the document was written. The skills and competency of the
medical personnel who produced the documents cannot be verified. It could be someone
who never had any medical training. By considering all these concerns, there is a lot of
room for disease misinterpretation.
The nature of the written sources and artwork also must be taken into account when
making an interpretation about past diseases. In the investigation of ancient diseases, it is
sensible to start by looking at historical medical texts. When using such document, the
historian must consider when they were written, the person who wrote them and the
purpose for producing or writing such a document (Mitchell et al. 2017). Typically, such
texts are written and being presented to someone high up. Therefore, the writer might
have the intention of impressing this person in power in order to gain something in
return, for example lucrative posts or career advancement (McVaugh, 2006). It is
possible that the text being presented might have been subject to alteration in order to
impress whoever the text was being presented to. Meanwhile, non-medical texts are
often written by non-medical practitioners with only basic medical knowledge
(Robinson 2003). Therefore, although they may not have as much bias as the medical
texts, the disease descriptions that were written could be quite vague and following
personal observations on only the obvious and clearest symptoms and signs. In the case
of histories, the text usually covers a long time range, therefore some of the chronicles
might have been copied from oral tales or much older written sources (Mitchell 2017).
Other documents that can be the source of disease information are biographies, personal
letters and diaries. These types of document can be highly illuminating as sometimes the
diseases are described in detail – how they, their friends or families feel or experience
when they suffered from the disease. Births and deaths registers could also provide good
information about the severity of a disease in the past – the statistic of disease in the past
in a particular region (Wrigley et al. 1997). However, these could also be the subject of
disease misinterpretation and alteration, depending on the reason for why such document
is being produced.
In short, there are many historical documents available which recorded diseases in the
past and they are no doubt great sources of information. But caution must be taken when
making interpretations by thinking carefully about what the document means to the
writer and to whom the document is being presented, and when was the document
38
written; artworks included. When using historical documents, it is also important that the
person is frank about how confident he or she is about the accuracy of the interpretation
(Mitchell 2017). Some documents might only mention a few symptoms and some
symptoms described could be ambiguous and could be present in more than one disease.
Therefore, the choice of words is important and remarks about other possible diseases
should be made.
1.5 Ancient DNA
1.5.1 Ancient DNA background
Genetic information from archaeological biological remains can be preserved and stored
in the form of ancient DNA (aDNA), which then can be assessed and studied (Higuchi et
al. 1984; Spigelman & Lemma 1993). The ancient DNA research field began in 1984
with the isolation of DNA from a preserved Equus quagga skin; this is a zebra-like
animal which became extinct 100 years previously (Higuchi et al. 1984). This
publication provoked interest to study aDNA in much wider types of archaeological
specimens but this was initially hindered by the difficulties in analysing the extremely
low amounts of fragmented and degraded nucleic acid materials from ancient specimens
(Brown & Brown 1992). The development of the polymerase chain reaction (PCR) by
Kary Mullis in 1985 opened a new pathway to study ancient DNA more efficiently as it
allows rapid amplification of a single fragment of DNA, which is perfect for aDNA
which typically exists at a very low copy number (Mullis et al. 1986). This led to the
first application of PCR to amplify aDNA from a 7000-year-old brain (Pääbo et al.
1988). Fast forward to this date, aDNA has been extracted from various archaeological
samples and studied using the current state-of-the-art sequencing technologies (Green &
Speller 2017).
1.5.2 Characteristic of ancient DNA
The DNA molecules of an organism are in constant threat of chemical “attack” at all
times. Fortunately, this is countered by enzymatic repair mechanisms when the organism
is still living to protect its genome integrity (Dabney, Meyer & Pääbo 2013). However,
upon the death of the organism, the genome becomes exposed to various “destroying
39
agents” that may affect its stability such as intracellular nucleases. This is termed
autolysis (Brown & Brown 2011). This, together with two other degradation factors,
environmental effects and microbial attack, influence the state of aDNA preservation
(Burger et al. 1999; Turner-Walker 2008; Kistler et al. 2017). Those biomolecules that
manage to escape the degrading autolysis activity will stand a possibility of being
archaeologically preserved (Brown & Brown 2011). However, they will be subjected to
other less rapid but persistent decaying processes due to the environment, both physical
and chemical factors (Turner-Walker 2008). The chemical agents include oxygen and
water, which are highly reactive and possess the ability to promote oxidative and
hydrolytic reactions, respectively. In addition, the by-products released from the
degradation of one type of biomolecule might themselves trigger decay of another
biomolecular type (Brown & Brown 2011). The physical factors involved in
biomolecular decay include geological, ultraviolet (UV) and cosmic radiations, which
possess the ability to significantly affect the integrity of biomolecules in both living cells
and post-mortem (Cadet & Wagner 2013). However, it is possible to disrupt the
complete destruction of the DNA molecules under favourable circumstances, for
example by freezing or desiccation of specimens (Nicholls 2005). The third degradation
factor, microbial, is also important as several types of microbe are attracted towards
organic material contained in biological remains as this may provide nutrients and an
energy source. In order to “feed” on the organic materials, microbes will secrete certain
enzyme(s) whose function is to break down the organic materials, including DNA,
leading to their degradation (Turner-Walker 2008). The extent of microbial-facilitated
biomolecular destruction depends on the microflora population existing in the burial
environment. Due to the presence of these destructive factors and the absence of repair
mechanisms after death, typical features of ancient DNA are (i) short fragment size due
to degradation; (ii) the presence of DNA replication blocking lesions; and (iii) the
presence of miscoding lesions (Dabney, Meyer & Pääbo 2013), as described below.
1.5.2.1 Fragmentation
Desiccation or anoxic conditions are thought to be the most suitable for DNA
preservation as these exclude the presence of DNA-damaging water and oxygen
(Nicholls 2005; Brown & Barnes 2015). Water targets the β-N-glycosidic bond resulting
in the release of purine or pyrimidine bases from the nucleotide sugar component,
40
consequently creating apurinic or apyrimidinic sites (Kim & Wilson 2012). These sites
are vulnerable to chemical attack, which will then lead to DNA fragmentation, creating
short fragments (Dabney, Meyer & Pääbo 2013). According to the results of in vitro
studies with contemporary DNA specimens, it has been proposed that DNA
fragmentation is caused by hydrolytic depurination. Subsequently, single-strand breaks
will occur due to hydrolysis of the phosphate backbone. These studies provide
supporting evidence to correlate temperature to fragmentation rate (Lindahl &
Andersson 1972; Lindahl & Nyberg 1972). Accordingly, thermal history was proved to
be a useful determinant in predicting the state of biomolecular survival in fossil bones;
the assumption is that DNA depurination is the main determinant of DNA degradation
(Smith et al. 2003). During the preservation period, DNA molecules that are contained
in archaeological remains undergo fragmentation. In result, the aDNA molecules
acquired from archaeological specimens are relatively short (Pääbo 1989). The average
fragment length of aDNA isolated from dried animal muscle tissue is between 100 to
200 base pairs; the maximum fragment length is 500 base pairs (Pääbo 1989). Studies
have shown that the size of aDNA fragments or the degree of fragmentation is
independent of the age of the archaeological specimen (Pääbo 1989; Hagelberg & Clegg
1991; Kistler et al. 2017). Instead of a time factor, it seems that the rapidity of tissue
desiccation after the time of death is more likely to be the determinant of the degree of
DNA fragmentation (Pääbo 1989). The aDNA decay model established for mammal
bones proposed by Kistler and colleagues (2017) suggests that DNA fragmentation
reaches a threshold rapidly after death and then slows down. In contrast, for DNA loss
throughout the preservation period, the process of bulk diffusion out of the sample might
be the major factor. By saying this, this group emphasized the importance of creating
optimum conditions in a closed system in order to improve the DNA preservation.
1.5.2.2 Miscoding lesions
Miscoding lesions are typical of aDNA as a result of hydrolytic deamination of
nucleotide bases (Dabney, Meyer & Pääbo 2013). These will cause errors of “reading”
by DNA polymerase during amplification resulting in incorporation of incorrect
nucleotides during PCR. Cytosine (C) has the highest susceptibility towards
deamination, which will produce uracil (U). During DNA duplication with a strand
containing a miscoding lesion used as template, uracil will direct adenine (A)
41
incorporation that subsequently causes C to T or G to A transition in the amplified PCR
product. Hofreiter and colleagues demonstrated the presence of modified deoxycytidine
in aDNA extracts from animal teeth and bones with ages ranging between 25,000 to
50,000 years (Hofreiter et al. 2001). The treatment of the aDNA template with uracil N-
glycosylase significantly lowered the numbers of substitutions. In their ancient DNA
damage model, using Neandertal mammoth DNA, Briggs and colleagues concluded that
nucleotide misincorporations in aDNA arise mainly because of cytosine deamination,
and are aggravated in single stranded molecules, which are typical for aDNA (Briggs et
al. 2007).
1.5.2.3 Blocking lesions
One of the effects of polynucleotide modification is inhibition of DNA polymerase
progress along the aDNA template strand (Brown & Brown 2011). This is due to
“blocking lesions” which inhibit DNA replication in living cells and may cause eventual
death to cells. However, in living cells, the repair mechanisms usually act accordingly
before such consequences occur (Karanam et al. 2013). However, these mechanisms will
cease to function post mortem, which results in the accumulation of blocking lesions
(Gilbert et al. 2003). If an aDNA sample with blocking lesions were to be used in PCR,
then shorter products are expected, as the Taq DNA polymerase will not be able to copy
the template to full length due to the inhibition from the blocking lesions (Feuillie et al.
2014). The majority of blocking lesions are caused by purine and pyrimidine oxidation
(Dabney, Meyer & Pääbo 2013). Some of the powerful oxidation agents known are
cosmic rays, and derivatives of geological radiation such as super-oxide, hydrogen
peroxide and hydroxyl radicals (Brown & Brown 2011). Oxygen is known as the less
reactive oxidant. Oxidation may disrupt the purine or pyrimidine ring structure and
break it open (Kim & Wilson 2012). This process also may cause purines on the
opposite strands to be dimerized. Maillard reactions may occur which may consequently
cause DNA-protein cross-links. These cross-links are effective blocking lesions which
can in turn prevent DNA amplification by Taq DNA polymerase (Brown & Brown
2011). Other than oxidation, modifications of guanines has also been suggested to be
one of the factors causing blocking lesions (Heyn et al. 2010). Other than cross-links,
the attachment of a peptide to a polynucleotide could also be sufficient to act as a
“blocker”.
42
1.6 Ancient DNA studies of tuberculosis and leprosy
Ancient DNA can be potentially recovered from any form of living remains: humans,
plants, animals. The sources of archaeological aDNA are reviewed by Green and Speller
(2017). Due to the variety of its sources, albeit hindered by preservation state and
molecule damage, aDNA has evolved as an important tool in archaeology. Ancient DNA
has been used to study myriad topics and answer historical questions including the
identification of sex and the relationship between archaeological skeletons; the study of
diets of the people in antiquity; the study of agricultural origin and spread; the
technology used in prehistoric periods; the history of early human origins and
migrations; and the study of ancient disease or palaeopathology (Brown & Brown 2011).
The aDNA application that will be discussed in detail here is the study of past diseases:
tuberculosis and leprosy. To this date, the study of both leprosy and tuberculosis has
been focused on case confirmation and the origin and evolution of both diseases. As
described in section 1.3, skeletal manifestations of both tuberculosis and leprosy are
providing a perfect opportunity to study both diseases using osteological methods.
However, this method does have limitations. For both diseases, not all infected
individuals show skeletal manifestations (Zink et al. 2001), and even for those skeletons
that do show bone changes, the lesions could be ambiguous and could be caused by
other types of disease that give rise to similar bone changes (Roberts & Manchester
2010). Therefore, aDNA studies may offer more precise disease identification by trying
to detect the presence of the pathogen that caused the infection. Evidently, the success of
aDNA disease detection methods depends on the preservation state of the skeletons
themselves.
1.6.1 Ancient DNA studies of tuberculosis
1.6.1.1 Case confirmation
The majority of aDNA studies of past diseases have been done on tuberculosis. This is
owing to the fact that tuberculosis is viewed as one of the most important palaeodiseases
in terms of its long co-existence with human populations as shown by the lesions in
archaeological remains, as well as the fatality burden that still exists today, despite the
previous success of the multi-drug treatments. Ancient tuberculosis has been the focus of
43
aDNA studies as it can cause identifiable bone changes and its DNA can persist in
archaeological remains despite long preservation periods (Spigelman & Lemma 1993).
The aDNA study of MTBC can provide an insight to the origin and evolution of this
ancient disease, which in turn might lead to a better understanding of this disease. As
described earlier, tuberculosis may cause skeletal changes in affected individuals. Out of
all individuals with active tuberculosis infections, only about 3-5% will develop skeletal
manifestation (Roberts & Buikstra 2003). From this 3-5%, approximately 50% will
develop spine tuberculosis. Therefore, the study of tuberculosis aDNA is typically
performed by extracting DNA from the vertebrae and sometimes the ribs – the ribs may
also show lesions in the case of tuberculosis infection (Mays & Taylor 2003; Taylor et
al. 2005).
The typical method of M. tuberculosis complex ancient DNA detection in archaeological
remains is by PCR assays utilising target regions that are thought to be specific to
MTBC members. The two earliest targets that have been used to screen for the presence
of MTBC are the multi-copy targets, insertion sequences IS6110 and IS1081 (Taylor et
al. 1996; Zink et al. 2001). These markers were thought to enhance the likelihood of
ancient TB detection as they are highly conserved within MTBC members as well as
their multi-copy nature. In each bacterium, IS6110 and IS1081 exist in zero to 26 copies
and five to six copies respectively (van Soolingen et al. 1992; Alonso et al. 2013).
However, there are some limitations in using IS6110 as the genetic marker. It is absent
in certain M. tuberculosis strains particularly in some Asian regions, where more than
10% of strains lack IS6110 (Brown & Brown 2011). Furthermore, there is a constant
suspicion about the lack of specificity of IS6110 as a marker (Müller et al. 2015). In
2015, Müller and colleagues published a study which confirmed the presence of IS6110
in ‘mycobacteria other than tuberculosis’ (MOTT) in archaeological remains from
Roman England and medieval Scotland. This study raised a doubt of using IS6110 as the
sole target in confirming archaeological detection of tuberculosis; usage of other more
specific tests is recommended when screening for the presence of ancient TB DNA.
Also, IS6110 and IS1081 do not allow discrimination between MTBC members, which
has prompted the study of single copy markers that might be to be used to identify
individual species in the MTBC group (Coros et al. 2008). One of these markers is the
oxyR pseudogene which harbours a single nucleotide polymorphism (SNP) that has the
discriminative power to distinguish M. bovis from the remaining members of the MTBC
(Sreevatsan et al. 1996). The presence of an A nucleotide at the 285th nucleotide position
of this gene confirms the M. bovis identity while the presence of G points out the
44
identity of the bacterium as being one of the other members of the complex (Sreevatsan
et al. 1996). More recently, two more genes harbouring distinguishing SNPs were
discovered and applied to ancient tuberculosis detection and discrimination: katG and
gyrA (Sreevatsan et al. 1997). The study of variable regions in 100 strains of M.
tuberculosis, M. bovis, M. africanum, M. microti and M. canettii led to the detection of
an M. tuberculosis specific deletion (TbD1) which allows discrimination between
“ancient” and “modern” strains of M. tuberculosis (Brosch et al. 2002). Another single
copy marker that is routinely used in M. tuberculosis aDNA detection is RD2 – currently
known as RD7 – although this is not as frequently used as IS6110 (Gordon et al. 1999).
This particular region can distinguish M. tuberculosis from M. bovis due to its intact
nature in M. tuberculosis but absence in the M. bovis genome.
PCR screening using IS6110, IS1081 and TbD1 confirmed the earliest case of
tuberculosis, dating from 9250 to 8160 years ago (Hershkovitz et al. 2008). The skeletal
remains belonged to a woman and an infant buried together, and were recovered from
Atlit-Yam in the East Mediterranean. Adult and infant aDNA were extracted from rib
and long bone, respectively. The study of aDNA in archaeological remains has brought
about exciting discoveries about this disease in the past. In 1994, Salo and colleagues
successfully amplified IS6110 from naturally mummified lung tissues from a body
which belonged to a 1000-year-old female recovered in Southern Peru (Salo et al. 1994).
This evidence pointed to the presence of human tuberculosis in the New World during
the pre-Columbian period.
1.6.1.2 Origin and evolution of tuberculosis
The study of MTBC variations has resulted in the attempts to reconstruct the origin and
evolution of tuberculosis. The genotyping studies performed in archaeological
specimens are benefiting from the genomic variation studies in contemporary MTBC
isolates.
One of such studies was performed by Sreevatsan and colleagues (1997) which
successfully identified two important SNP markers: katG and gyrA, which allow the
genotyping of MTBC strains into three main principal groups (PGG) : PGG1 to PGG3.
They looked at 26 structural genes in 842 modern MTBC global isolates and
successfully identified the two SNP markers. The discrimination lies in nucleotide
45
mutations in the codon positions 463 and 95 of the katG and gyrA genes, respectively.
The T to G nucleotide substitution on the 463rd position of katG will result in a leucine
to arginine amino acid change. Similarly, the C to G nucleotide substitution at the 95th
codon position of gyrA gene will result to a threonine to serine amino acid substitution.
The distribution of MTBC isolates is depicted in Figure 1.4 below. Interestingly, all M.
microti, M. africanum and M. bovis isolates were categorized specifically PGG1, while
M. tuberculosis isolates are spread across group PGG1 to group PGG3. It was then
highlighted that the species in PGG1 are ancestral to PGG2 and PGG3 as there is more
genetic variations observed – it was assumed that these pathogens were allowed to
accumulate more genetic variations by a longer time of evolution. However, the lack of
neutral mutations was highlighted in this study, as up to 95% of nonsynonymous
mutations are harboured in the antibiotic resistance genes which are under a strong
positive selection pressure (Sreevatsan et al. 1997).
Figure 1.4: The genotyping scheme of MTBC members using the katG463 and
gyrA95 markers. (taken from Sreevatsan et al. 1997: p. 9,871).
Later, a study performed by Bos and research members (2002) have discovered a
specific deletion in RD region and TbD1 gene that can be used to separate between
modern and ancient M. tuberculosis isolates. The classification is based on the absence
or presence of the deletion in the MTBC genome. Another important study of
contemporary MTBC isolates was published by Filliol and colleagues in 2006, where
they successfully classified MTBC into phylogenetically distinct groups made up of six
lineages described as SNP cluster groups (SCGs) as well as five subgroups (Filliol et al.
2006). The M. bovis isolates are contained in the seventh group.
46
Figure 1.5: MTBC phylogenetic lineages 1 to 7.
The ‘modern’ and ‘ancient’ MTBC can be distinguished based on the presence or absence of 7bp
deletions within the TbD1 region. Each branch is showing different MTBC lineages (L1-L7)
which is also associated to the geographical origins of the individuals where the MTBC isolates
were extracted from. (Taken from Brites et al. 2015).
Next, an additional phylogenetic lineage was recognized giving rise to 7 M. tuberculosis
lineages: L1-L7 (Figure 1.5), where M. bovis is now typically being called lineage 8
(Comas et al. 2013). It was revealed that there is a strong link between the place of birth
of the individual and the geographical origin of the M. tuberculosis sample. Lineage 1
typically occurs in the Indian Ocean and the Philippines, lineage 2 is linked to East Asia
tuberculosis, lineage 3 is assigned to both East Africa and India, lineage 4 is for
tuberculosis in Europe and North and South America, whereas lineage 5 occurs
predominantly in West Africa (Witas et al. 2015). Interestingly, lineage 4 is the most
typical group of M. tuberculosis and lineage 7 has only been found in the Ethiopian
region or among recent emigrants from Ethiopia. M. tuberculosis lineage genotyping has
been performed with several archaeological specimens. In 2012, Bouwman and
colleagues performed target enrichment for 247 SNPs of which 214 sites were
47
successfully covered in the subsequent sequencing. From this study, an individual from
the 19th century whose body was recovered from St George’s Crypt in Leeds, West
Yorkshire, England was identified to harbour a genotype belonging to lineage 2, SNP
type (ST) 14/40 and SCG 6 (Bouwman et al. 2012). Another British MTBC aDNA study
was also performed by Müller and colleagues (2014b) revealing that multiple genotypes
of M. tuberculosis were present in this region between the 18th – 19th centuries.
Many studies have been performed to decipher the origin and evolution of MTBC – both
using contemporary and ancient isolates. The most common misconceptions of the
evolution of tuberculosis is that human tubercle bacilli were derived from M. bovis
(Kapur et al. 1994; Stead et al. 1995). This assumption was initiated by the zoonotic
acquisition of tuberculosis infection by humans where the disease can be transmitted
through the ingestion of meat or milk products that have been contaminated by M. bovis.
This idea is propagated even further by the fact that M. tuberculosis has a highly
specialized niche in humans, while M. bovis have a wide host spectrum including cattle,
goats and sheep. The initial theory was that a strain of M. bovis successfully infected a
human, followed by successful spread within human populations. This led to co-
evolution of the tubercle bacillus with human hosts for millennia that resulted in a highly
specialized niche inside the human body, finally giving M. tuberculosis. Furthermore,
the earliest DNA evidence of human tuberculosis is only 9,250-8,160 years ago – which
is a skeleton from the now submerged Atlit-Yam site in the Eastern Mediterranean
(Hershkovitz et al. 2008). Meanwhile, the animal tuberculosis biomolecular evidence
points to a much older case – 17,000 before present (BP) isolated from the an extinct
long-horned bison recovered from the Natural Trap Cave in Wyoming, North America
(Rothschild et al. 2001). This theory however was proposed prior to the availability of
the whole genome sequence (WGS) of M. tuberculosis. The WGS of M. tuberculosis
allowed a comparative genomics study, which successfully uncovered variable genomic
regions among the MTBC members. Among the variable genomic regions uncovered is
the M. tuberculosis specific deletion 1 (TbD1) that is deleted in M. tuberculosis H37Rv
but is intact in the other members of the MTBC. In a set of experiments involving 100
strains of MTBC including M. tuberculosis, M. bovis, M. africanum, M. canettii and M.
microti, 20 variable regions were studied by Brosch and colleagues (2002), including the
TbD1. The MTBC member strains were collected from hosts from diverse global
locations. In this study, it was further found that some of the M. tuberculosis strains have
the deletion in the TbD1 region. In addition, this study also showed that M. bovis
harbours more deletions in its genome in comparison to M. tuberculosis; and based on
48
the M. bovis AF2122/97 (cattle strain) near-complete genome, it was shown that M.
bovis has a smaller genome compared to M. tuberculosis. It was speculated that the
tubercle bacilli that are most similar to the precursor of M. tuberculosis are human
pathogens, not an animal pathogen (M. bovis) – contrary to previous beliefs.
Meanwhile, MTBC origin is also being extensively studied using contemporary and
ancient isolates – with results for the date of origin varying in range from 70 kya to less
than 6,000 years ago. In an earlier publication, Kapur and colleagues (1994) studied
eight short target sequences in 31 modern M. tuberculosis isolates from all parts of the
world. The result revealed a very low nucleotide diversity among the M. tuberculosis
isolates. The calculation of evolutionary time scale based on the number of variations in
four genes suggested that the divergence among the modern isolates might have begun
15,300 to 20,400 years ago. This timeline coincides approximately to the
palaeomigration of humans into the New World. In addition, this time estimate also
coincides with the beginning of cattle domestication which convinced the authors that
rather than M. tuberculosis, M. bovis was the more ancestral species. However, this
study is not without its limitations. It did not take into account the diversity levels in the
other members of the MTBC, hence weakening the argument.
Later, a study by Sreevatsan and colleagues (1997) proposed a recent evolutionary
bottleneck event – with the speciation time predicted to have occurred around 15,000-
20,000 years ago. This assumption was based on the small allelic diversity observed in
the tubercle bacilli. An African origin of MTBC was proposed and several other studies
are also in line with this theory (Gutierrez et al. 2005; Hershberg et al. 2008; Wirth et al.
2008). The age of the M. tuberculosis last common ancestor was estimated by looking at
the accumulation rate of synonymous substitutions. By counting the synonymous
substitution rate, Gutierrez and colleagues (2005) estimated that M. tuberculosis has
existed since 3 million years ago. This coincides with the presence of hominids in East
Africa 3 million years ago – supporting the theory of the emergence of tubercle bacilli
from Africa. From there, it was speculated that the bacilli underwent early
diversification followed by the spread of successful clones to the other parts of the world
following human migrations out of Africa. However, the evidence from archaeological
specimens is not in line with this theory. Bos and colleagues (2014) proposed that the
age of the M. tuberculosis complex most recent common ancestor (MRCA) is less than
6,000 years ago. This conclusion was derived from an aDNA study of a Pre-Columbian
skeleton aged approximately 1,000-years-old. The Peruvian ancient genomes are more
49
closely related to animal lineages; closest to M. pinnipedii strains – restricted to seals
and sea lions. This provides a plausible theory of seal and sea lions as the possible route
of tuberculosis pathogen entry to the New World. The study confirmed pre-Columbian
MTBC infection in South America.
Another recent comprehensive study of ancient M. tuberculosis was published in 2015
by Kay and colleagues. This group has successfully obtained genome sequences of 14
ancient M. tuberculosis isolates through a shotgun metagenomics method, without prior
target enrichment (Kay et al. 2015). The 18th century Hungarian M. tuberculosis isolates
were extracted from human remains, most of which had undergone natural
mummification, recovered from a Dominican church in Vac. The genome sequences
depicted that all 14 ancient M. tuberculosis isolates belonged to phylogeny lineage 4.
This lineage characterization, which is also known as the Euro-American lineage, was
also confirmed by the 7 bp nucleotide deletion within the pks15/1 gene region. A
phylogenetic analysis performed with 1,582 unassembled genomes of other identified
lineage 4 members revealed that at least 12 M. tuberculosis strains were present in this
region during that time. Kay et al. (2015) also dated the divergence time for lineage 4 by
utilizing four high-coverage M. tuberculosis ancient genomes. The mutation rate of this
lineage is estimated to be 5.00 x 10-8 substitutions per nucleotide in one year. This is in
line with the hypothesis from the previous historical M. tuberculosis study (Bos et al.
2014), which suggested the age of the last common ancestor to be not more than 6,000
years ago. Mixed-infection from multiple M. tuberculosis genotypes was also shown in
this study.
In contrary to the hypothesis derived from contemporary MTBC isolates, in historical M.
tuberculosis isolates, Bos et al. (2014) and Kay et al. (2015) predicted the age of the last
common ancestor to be less than 6,000 years ago. However, this is not in line with
several historical tuberculosis cases that are older than 6,000 years old supported by
osteological and biomolecular evidence (Rothschild et al. 2001; Hershkovitz et al. 2008;
Nicklisch et al. 2012). This discrepancy could be influenced by the errors made in the
assumptions in the phylogenetic dating. Kay and colleagues (2015) also proposed that
contamination could also be the cause of this discrepancy. Looking at the current
theories of tuberculosis origin, it seems that more studies are needed to obtain a
conclusion that is consistent for both the contemporary and historical study of
tuberculosis.
50
1.6.2 Ancient DNA studies of leprosy
1.6.2.1 Case confirmation
Similar to tuberculosis, the earliest studies of leprosy ancient DNA were mainly
focussed on case confirmation as an extension of the osteological diagnosis approach.
Rafi and colleagues detected the first ancient M. leprae DNA in 1994 from a 600 AD
archaeological bone recovered from a grave at the Monastery of St John the Baptist,
located at the river Jordan which is where Jesus is believed to have been baptised by
John (Rafi et al. 1994). The detection was performed by PCR amplification of a part of
the 36 kDa antigen and 65 kDa protein encoding genes, the first of which is specific to
M. leprae and the second of which is found in all mycobacteria (Hartskeerl et al. 1989;
Telenti et al. 1993). PCR amplification of the 36 kDa gene gives a 530 bp product,
which is relatively large for ancient DNA amplification, considering the fragmented and
damaged nature of ancient DNA. In 2001, this problem was addressed by Donoghue and
colleagues by designing M. leprae specific primers for a nested PCR which targets the
repetitive element RLEP, and which gives an outer product of 129 bp and a 99 bp
product for the nested PCR (Donoghue et al. 2001). M. leprae detection by PCR of this
repetitive element was demonstrated by Woods and Cole (1990) with modern isolates,
and then subsequently used by Yoon and colleagues (Yoon et al. 1993) with clinical M.
leprae isolates. The repetitive element was first used for detection of M. leprae from
archaeological remains by Donoghue et al. (2001). Today, RLEP is the most frequently
used target to detect M. leprae in archaeological remains. The repetitive element RLEP
is dispersed in at least 28 copies throughout the genomes of modern M. leprae isolates
(Woods & Cole 1990), and is attractive for ancient M. leprae detection as its multi-copy
nature increases the chance of DNA fragment preservation.
The oldest case of leprosy confirmed with ancient DNA is from The Tomb of the Shroud
in Hinnom Valley, Mount Zion (Matheson et al. 2009). The location is associated with
the traditional Akeldama, mentioned in the Bible as the “Field of Blood”. The age of the
skeletal remains is estimated to be from 2025±28 years BP as estimated by the AMS
radiocarbon method. Interestingly, one of the skeletons is not only positive for M. leprae
DNA but also tested positive for M. tuberculosis DNA which indicates co-infection
(Matheson et al. 2009). The presence of these two pathogenic mycobacteria indicates
immune-suppression, which is one of the characteristics of a person with lepromatous or
51
multi-bacillary spectrum leprosy. This skeleton gave positive PCR results for both the
nested RLEP assay and the 18 kDA gene, although the product for the 18 kDa PCR was
in low amounts due to the single copy nature of the marker. The same marker was used
to detect M. leprae DNA in skeletal remains from a South German ossuary and a
Hungarian cemetery dated approximately to 1400-1800 AD and the 10th century AD,
respectively (Haas et al. 2000). The Germany samples were taken from skulls while the
Hungarian sample was from the hard palate and both gave positive RLEP1 and RLEP3
amplifications. In the same study, DNA from the hands and feet were also examined but
did not produce sufficient amplification products for M. leprae detection. The results
suggest that peripheral parts of a skeleton displaying signs of leprosy may contain
significantly fewer bacteria, or it could be that the peripheral bone changes are the result
of secondary infection. This also suggests that the rhinomaxillary alteration is a result of
direct M. leprae bacterial involvement. The first ancient leprosy case from Britain was
from Orkney, Scotland (Taylor et al. 2000). The archaeological remains recovered from
the 13-14th Norse Christian cemetery located at Newark Bay showed lepromatous
leprosy skeletal changes. Screening with the RLEP PCR produced positive amplification
for the extract from the skull bones but not the clavicle, left scapula, vertebra, or femur.
This again, seems to align to the result obtained by Haas and colleagues (Haas et al.
2000). On the other hand, a study of 12th century AD skeletal remains recovered from
the fortress site of Capilla y Castillo de San Jorge in Spain reported successful M. leprae
detection from metacarpal bones following RLEP PCR (Montiel et al. 2003). The
positive identification was also supported by restriction analysis with ClaI which detects
a site present in the RLEP region as well as Sanger sequencing.
Later studies have detected more leprosy cases from global locations. The earliest case
of leprosy confirmed by osteological evidence of high confidence is from the Roman
period (Rubini, Erdal, et al. 2014). A skeleton recovered from Martellona, Italy, was of a
child aged between 4 to 5 years with pronounced rhinomaxillary syndrome strongly
suggestive of a well-developed leprosy infection. This is the youngest individual with a
confirmed case of leprosy. It is quite surprising to observe convincing skeletal
manifestations in such young individuals, considering leprosy is a chronic and slow
progressive infection. However, one of the proposed explanations for this was that the
M. leprae strain that existed at this time was more aggressive compared to contemporary
M. leprae (Rubini, Erdal, et al. 2014). Another childhood leprosy case that was reported
in the same study, describing a 4-5 month old infant from Kovuklukaya, Turkey, with
some indication of chronic inflammation in the bone, is even more surprising as the
52
leprosy case was confirmed by molecular evidence (Rubini, Erdal, et al. 2014). These
skeletal remains are estimated to age from the 8th to 10th centuries AD, during the
Byzantine period. Both cases most probably result from in utero haematogenous
transmission of the infection from a pregnant mother who was suffering from leprosy
infection. In addition, past leprosy cases have also been confirmed through molecular
evidence in Japan, from archaeological remains recovered from the Hatanai site located
in the Honshu Island of Japan (Suzuki et al. 2010). The analysis of the excavation sites
estimated the age of the bone remains to be most likely from the mid-18th to the early
19th century AD. The bones show characteristics of the lepromatous type of leprosy,
with rhinomaxillary syndrome (Suzuki et al. 2010). Another study by Watson and
Lockwood (2009) confirmed leprosy cases in Croatia (8th-9th century AD), Denmark
(1275-1560 AD) and the United Kingdom (900-1000 AD). The positive detections were
from the rhino-maxillary area (Croatia), palatine bone (Denmark) and rhino-maxillary,
tibia, and metatarsal (United Kingdom). Positive M. leprae detection has also been
obtained from a Byzantine skeleton dated 300-600 AD from Bet Guvrin, Israel
(Spigelman & Donoghue 2001).
1.6.2.2 Origin and evolution of leprosy
Most of the past leprosy case confirmation studies described previously were performed
by amplification of the RLEP region. Whole genome sequencing of the modern M.
leprae TN strain from Tamil Nadu, India, provided an opportunity to explore additional
sites of the genome that could be studied and applied to ancient M. leprae isolates (Cole
et al. 2001). The 3,268,210 bp genome of M. leprae has a G+C content of only 57.8%,
which is much lower than that of M. tuberculosis, whose genome is 4,411,531 bp with
65.6% G+C content. In fact, the G+C content of M. leprae is the lowest of all known
mycobacteria (Singh & Cole 2011). From the genome sequence analysis done by Cole
and colleagues, only 49.5% of the M. leprae genome codes for proteins, while 27% and
23.5% of the genome are pseudogenes and non-coding DNA, respectively (Cole et al.
2001). From a comparison of the M. leprae and M. tuberculosis genomes, under the
assumption that these were at one point of time in a state of similar topology, it was
concluded that the M. leprae genome has undergone significant downsizing, and may
have lost around 2,000 genes since the last common ancestor of the two species
(Eiglmeier et al. 2001). The high number of pseudogenes together with the increasing
53
number of insertion sequences (IS) and reduced G+C content are attributes of reductive
evolution, which can be linked to the extreme obligate nature of M. leprae (Singh &
Cole 2011). The reason for the massive gene inactivation could be the highly specific
niches occupied by the bacteria, such as Schwann’s cells, where M. leprae will not be
subject to competitive pressure from other microbes. Gomez-Valero and colleagues
(2007) suggested that the shaping of M. leprae into a highly specialized pathogen was
due to massive gene reduction which most likely happened at the same time as the
pseudogenization events.
The sequencing of the whole M. leprae TN genome allowed further exploration of
regions and targets that can be used in genotyping. Selected genes, pseudogenes and
non-coding regions were compared between the genomes M. leprae TN and Brazilian
strains (Monot et al. 2005). This comparative genomic study resulted in the discovery of
three unique SNPs that can classify M. leprae isolates into four groups: SNPs type 1 to
4. These SNPs, at positions 14,676, 1,642,875 and 2,935,685, interestingly show strong
association with the geographic origins of patients. SNP type 1 is mostly shown by
patients from Asia, East Africa and the Pacific region, SNP type 4 is typically found in
the Caribbean and West Africa, while SNP type 3 is the typical of the M. leprae type
infecting the people in Europe, North Africa and the Americas, as demonstrated in
Figure 1.6 (Monot et al. 2005). Fascinatingly, SNP type 2 is the least typical type,
infection of M. leprae with this type being found only in Malawi, Ethiopia, Nepal or
North India, and New Caledonia (Monot et al. 2005). Based on this scheme, it is
suggested that leprosy originated in the Near East or in Eastern Africa. Subsequently,
successive human migrations have been responsible for the dissemination of this disease
worldwide. This scheme was soon used to classify ancient M. leprae isolates from
Europe, including the United Kingdom, Denmark and Croatia (Watson & Lockwood
2009). This study showed that all ancient M. leprae isolates genotyped from these three
European countries (476 AD to 1350 AD) belong to SNP type 3 (Watson & Lockwood
2009). A comprehensive study performed by Monot and colleagues (2009) involved the
comparative analysis of four complete genomes of modern M. leprae strains: M. leprae
TN from India, M. leprae Br4923 from Brazil, M. leprae Thai 53 from Thailand, and M.
leprae NHDP63 from the United States. The analysis revealed extremely low sequence
diversity among strains: they are
54
Figure 1.6: The distribution of four different leprosy SNP types around the world
and the prediction of human migrations.
Each SNP type is depicted in a different colour. SNP-type 1: yellow, SNP-type 2: orange, SNP-
type 3: purple, SNP-type 4: green. The arrows are showing the proposed routes of migrations,
inferred from SNP analysis – the colour is similar to the circles for each SNP-type. The
estimated time of migration (in years) is shown in grey arrows. The estimation was made based
on genetic, archaeological and anthropological studies (Figure taken from Monot et al. 2005).
99.995% identical. There were only 215 polymorphic sites and 5 pseudogenes identified.
The evaluation of polymorphic sites classified M. leprae isolates into 16 subtypes (1A-
1D, 2E-2H, 3I-3LM, 4N-4O) that are closely related to the geographical location of their
hosts. This subtyping scheme is able to shed light on the trade routes and migration
patterns of humans in the past, as depicted in Figure 1.6 (Monot et al. 2005). Based on
the construction of a phylogenetic tree of the present M. leprae isolate sequences, the
genotype of the ancestral M. leprae strain was estimated as between type 2 and 3. This
result supports the previous assumption that M. leprae in East Africa could belong to
type 2. This type 2 East African strain then produced the type 1 strain that moved east to
Asia following human migrations. From there, M. leprae spread to the west, towards the
Middle East then Europe, in type 3 form before finally giving rise to the type 4 that can
be found in West Africa, the latter movement believed to be through the slave trade. The
introduction of leprosy to the Americas was most probably through European
immigrants instead of crossings of humans via the Bering straits (Monot et al. 2009). A
genome-wide comparison of five medieval M. leprae strains, each with more than 80%
genome coverage, discriminated M. leprae isolates into five distinct branches, 0-4, each
55
associated with a particular subtype (Schuenemann et al. 2018). The most ancestral
branch 0 is associated with subtype 3K which can be found in contemporary M. leprae
strains isolated in China and New Caledonia.
56
Chapter 2: Materials and methods
2.1 Archaeological samples
The archaeological bones described here are the samples studied in Chapter 3 and 4. The
samples investigated in Chapter 5 is described separately in each respective chapter.
Sixty skeletal samples from sixty individuals dated from Roman to the post-Medieval
periods were studied (Table 2.1, Table S2.1). Rib fragments were mainly analysed from
the 60 skeletons from 16 different locations in Yorkshire. The samples originated from
urban and rural locations of Yorkshire. The exact location in each site is shown in
Figure 2.1. Seven of the sites are in York city, therefore the samples from the area are
classified as “urban”. The sample from East Heslington is still considered urban
although it is not located inside the city of York. This classification is based on its
proximity to York in comparison to the other rural sites. It could have been a separate
community in the past but the closeness to York made it possible for people to be
exposed to the same strain of MTBC that was circulating in the city during that period.
The archaeological bones were collected by Dr Darlene Weston as part of the National
Environment Research Council (NERC) project on “The Palaeopathology of
Tuberculosis in Britain and Europe” (NE/E015697/1, 2007–2010), subsequently
continued as “Palaeopopulation genomics of Mycobacterium tuberculosis
(NE/K012185/1, 2013–2016). Two from the 60 bone remains: 3 Driffield Terrace 13 and
East Heslington 229 have previously been studied and reported by Dr Romy Muller and
colleagues (2014a). The remaining 58 samples were studied for the first time in this
project. DNA extraction was performed from the bone powder in the aDNA lab facility
at the Manchester Institute of Biotechnology. In my work, PCR assays were used to test
for MTBC aDNA preservation, using markers which are supposedly specific to the
MTBC genome: the insertion sequence IS6110, and the gyrA, Rv0083 and Pks 15/1 loci
(Bouwman et al. 2012; Alonso et al. 2013; Müller et al. 2014a). Although the specificity
of the multi-copy region IS6110 is questionable, it is still the most robust target for
MTBC aDNA detection so far (Müller et al. 2015). Therefore, as per the
recommendation by Muller et al. (2015), the screening was accompanied by the other
three single copy regions, which are deemed specific to the MTBC genome. Nested PCR
was incorporated in the IS6110 assay to enhance the detection power and increase
57
detection specificity. The three single-copy targets have previously been used for SNP
(gyrA, Rv0083) and indel (Pks15/1) genotyping (Müller et al. 2014b).
58
(b)
(a)
Figure 2.1: The origin locations of the Yorkshire archaeological remains. (a) The rural sites
which are marked by red placemarks and one of the urban sites marked by the yellow
placemarks. The red circle marks the city of York location where the rest of the urban samples
were taken from, (b) The urban sites in the city of York marked by yellow placemarks. Some of
the placemarks overlap with each other. Images are extracted from Google Earth.
Melton Roman Rural 4297 Rib lesions Left rib Caffell & Holst 2008a
65
cemetery 2554 Rib lesions Right rib
5319 Control Left tibia
Hickleton Late
mediaeval –
early post
mediaeval
Rural 46 Spine Vertebrae Manchester & Roberts 1986
Table 2.1: A complete list of all bone remains studied for the preservation of MTBC aDNA. This table showing the skeleton dates, site origins, skeleton
parts showing lesion or non-specific lesions (and control) and skeleton parts sampled in the study. N/A: information not available
66
2.2 Authentication regimes
As discussed previously, the damaged, fragmented and low copy nature of ancient
DNA demands protocols to be performed in a very careful manner (Dabney, Meyer, et
al. 2013). DNA templates of these characteristics are very prone to exogenous DNA
contamination (Knapp et al. 2015). The typical sources of contamination are from the
environment and DNA of the personnel handling the archaeological remains which
effortlessly outcompete the very little and damaged endogenous aDNA recovered
(Noonan et al. 2005; Poinar et al. 2006). There is a risk to introduce contaminants at
any stage of the experiments (Gruber 2015). The study was carefully tailored to make
sure there is a balance between minimizing the risk of contamination in samples and
maximizing the yield of endogenous DNA.
All experimental protocols prior to PCR amplification in the PCR thermocycler were
performed in physically isolated clean rooms dedicated to ancient DNA at the
Manchester Institute of Biotechnology. The clean rooms are located on a different
floor (lower ground) to the laboratory where all the “modern” or “post PCR
amplification” experimental works is carried out on the second floor (previously first
floor) in the building. There are two clean rooms dedicated for experimental works
using archaeological bone remains as starting materials: a human bone aDNA
extraction room and a human bone aDNA PCR mixture preparation room. Both rooms
are physically isolated from each other. Access to these rooms is restricted to trained
lab personnel and the inevitable occasional equipment maintenance, and is locked at all
times except when in use. Each room is designed to have an antechamber at the anterior
of the clean room which acts as contamination buffer zone and gowning area before
entering the clean room. The room is supplied with ultra-filtered air which circulates in
the room in a positive displacement manner. This manner of air circulation will avoid
the uptake of possibly-contaminated unfiltered air from the outside of the clean room.
In addition, the clean rooms are also equipped with ultraviolet (UV) lights (254 nm);
the rooms are UV irradiated for at least four hours after each use.
67
Appropriate protective clothing is worn before entering the clean room; the gowning is
performed in the antechamber. The protective clothing includes disposable
IS6110 123 bp 134 49 Streptomyces lunaelactis 97.80% 2.72 e-35
Table 3.2: Sanger sequence result summary for 15 samples with a positive band in at least one of the PCR assays.
n.d. Not done. BLAST search was not performed as the product sequence completely matches the reference genome.
n.a. Not aligned. The sequence cannot be aligned to the reference genome in Geneious due a large number of mismatches.
95
3.2.1 St Andrew Fishergate 6
This sample produced a positive PCR amplification in the second nested step of the
IS6110 assay when the original undiluted DNA template was used (Figure 3.1-a).
Positive amplification was achieved in both the first and nested step of IS6110 where
the 10-fold diluted DNA was used as template (Figure 3.1-a). In this extraction, the
other three single-copy target PCR assays produced negative amplification (not shown).
A second DNA extraction was performed; positive amplification was observed in the
first and nested step of the IS6110 assay in both instances where undiluted and 10-fold
diluted DNA were used as templates (Figure 3.1-b). Positive amplification was
obtained from 10-fold diluted DNA in the gyrA PCR assay but not from the undiluted
DNA (Figure 3.1-b). A very faint band was shown in the agarose gel from the 10-fold
diluted DNA amplification for the Pks 15/1 PCR assay, and similar to the gyrA assay,
no band was produced when the undiluted DNA was used as template (Figure 3.1-b).
The Sanger sequences of the cloned IS6110 PCR product when aligned to the M.
tuberculosis H37Rv reference genome revealed a complete match in 5 clone sequences,
and one mismatch each at inconsistent positions in 4 clones for the first extraction. On
the other hand, all IS6110 nested PCR sequences show a complete match to the
reference genome (Figure 3.2-a). For the second St Andrew Fishergate 6 extraction, all
clone sequences for both the IS6110 123 bp and nested PCRs show a complete match
to the reference genome with no mismatch observed (Figure 3.2-b). The sequence
identities are also supported by BLAST analysis. The 3 clone sequences from the gyrA
PCR assay of the second DNA extraction revealed 6 mismatches from the 64 bp
sequence (after trimming of primers) in consistent positions for all the clones (Figure
3.3-a). BLAST search revealed the highest sequence similarity to Mycobacterium
peregrium partial gyrA sequence (Table 3.2). Meanwhile, the clone sequences from
Pks 15/1 PCR amplification of St Andrew Fishergate 6 second extraction contained
three consistent mismatches in all clones, and another one shared mismatch in 2 clones
(Figure 3.3-b). Based on the BLAST search result, the highest sequence match was
shown against M. tuberculosis and M. bovis (Table 3.2).
96
M A B C D
100 bp
200 bp
IS6110 - 123 bp IS6110 - 92 bp
M A B C D E F G H
200 bp
100 bp
IS6110 - 123 bp IS6110 - 92 bp gyrA Pks 15/1
Figure 3.1: Gel electrophoresis results for the sample St Andrew Fishergate 6 showing
positive bands for three markers. (a) DNA extraction 1. A and C: undiluted DNA as
template; B and D: 10-fold diluted DNA as template. (b) DNA extraction 2. A, C, E and G:
undiluted DNA as template, B, D, F and H: 10-fold diluted DNA as template. In both 1 and 2,
M: DNA ladder.
(a)
(b)
97
Figure 3.2: The alignment of St Andrew Fishergate 6 Sanger clone sequences of IS6110 123 bp and nested 92 bp PCR product against M. tuberculosis
H37Rv reference sequence. (a) St Andrew Fishergate 6 DNA extraction 1, (b) St Andrew Fishergate 6 DNA extraction 2. Primer sequences are omitted from
the alignment. Mismatches are highlighted in blue.
(a)
(b)
98
Figure 3.3: The alignment of St Andrew Fishergate House 6 Sanger clone sequences of gyrA and Pks 15/1 PCR product against M. tuberculosis H37Rv
reference sequence. (a) gyrA and (b) Pks 15/1 DNA extraction 2. Primer sequences are omitted from the alignment. Mismatches are highlighted in blue
(a)
(b)
99
3.2.2 St Andrew Fishergate 277
Skeleton St Andrew Fishergate House 277 produced multiple bands in the Rv0083
assay, one with a length close to the expected 110 bp PCR product from this assay
(Figure S3.1). The faint lower band was cut from the agarose gel, purified and cloned
into competent E. coli cells. The clone Sanger sequences show a PCR product of 109
bp, shorter than the expected fragment length (not shown). The alignment to the
reference M. tuberculosis H37Rv sequence shows only 74.2% pairwise identity.
BLAST search revealed that Nocardioides sp. JS614 as the closest match to the clone
sequences. However, the e-value for this match is only 1.31e-4 (Table 3.2).
3.2.3 St Andrew Fishergate 339
St Andrew Fishergate produced a positive band in the gyrA PCR assay, while showing
negative amplifications in the other PCR assays. The positive band was produced from
the amplification of the 10-fold diluted DNA template. It was faint (Figure S3.2) but
had an apparent length close to 104 bp. However, the sequence length spanned by both
forward and reverse primers had 6 mismatches with the reference sequence (Figure
3.4). The possible identity of the amplified sequence based on BLAST analysis is the
partial gyrA gene from an uncultured bacterium: ancient DNA from St Peter’s
Collegiate Church 28-a,b isolate (Müller et al. 2016).
Figure 3.4: The alignment of St Andrew Fishergate House 339 clone sequence of gyrA
PCR product against the reference M. tuberculosis H37Rv sequence. Primer sequences are
omitted from the alignment. Mismatches are highlighted in blue.
100
M A
200 bp
100 bp
3.2.4 St Helen on the walls 6003
Figure 3.5: gyrA PCR amplification for sample St Helen on the Walls 6003. A: undiluted
DNA template. M is the DNA ladder with the 100 bp and 200 bp marked in the figure. Faint
lower band is shown inside the red box.
St Helen on the Walls produced double bands in the gyrA assay when undiluted DNA
was tested, one of which with had a size close to the expected gyrA PCR product
(Figure 3.5). The 10-1 DNA dilution was not tested for this particular assay for this
sample. This sample tested negative for the other PCR assays. Clone sequences
revealed a 104 bp PCR product, which matched the expected product length of the
gyrA target. However, the sequence did not correspond to the reference M. tuberculosis
sequence. There are 8 consistent mismatches throughout all 9 clones, and one mismatch
that is present in 6 clones (Figure 3.6). Two clones show one mismatch each, in
different positions. BLAST search shows a closest match to Mycobacterium vaccae
(Table 3.2).
101
Figure 3.6: The alignment of St Helen on the Walls 6003 clone sequences from the gyrA PCR assay to the M. tuberculosis H37Rv reference sequence.
The reference sequence is shown at the top, while the mismatches are highlighted in blue. Primer sequences are omitted from the alignment and BLAST search.
Figure 3.7: The alignment of East Heslington 229 clone sequences from the IS6110 123 bp PCR assay to the M. tuberculosis H37Rv reference sequence.
The reference sequence is shown at the top, while the mismatches are highlighted in blue. Primer sequences are omitted from the alignment and BLAST search.
102
3.2.5 East Heslington 229
Skeleton East Heslington 229 produced multiple bands in the first step of the IS6110
PCR assay. The lowest band as marked in Figure S3.3 shows a faint possible positive
amplification. However, the subsequent IS6110 nested 92 bp PCR assay did not
produce positive amplification. The lowest band from the first step PCR assay was
excised from the gel, cloned and Sanger sequenced. Clone sequences revealed a
product with 123 bp length, however the sequence did not match the reference M.
tuberculosis H37Rv. There are 16 mismatches within the region spanned between the
forward and reverse primers, consistently found across all clones (Figure 3.7). The
closest match of this sequence is to the unidentified partial IS6110-like insertion
sequence; isolate IS6110-like sequence type B, which is reported in Müller et al.
(2015).
3.2.6 Wetwang Slack 2
The Wetwang Slack 2 skeleton produced multiple bands for the gyrA PCR
amplification, in which the length of the lower band can be approximated to 104 bp
(Figure S3.4). The other PCR assays tested for this sample produced negative
amplification – absence of band or bands with differing length to the expected product
size. The amplicon produced by the Wetwang Slack 2 sample is 104 bp as confirmed
by the clone Sanger sequences. However, the sequence does not correspond to the
reference. There are 11 mismatches to the reference within the 64 bp sequence – after
the primers are removed (Figure 3.8). Based on the BLAST search result, the closest
sequence match is to the uncultured bacterium partial gyrA gene of St Shchekavitsa 8-
a-e ancient isolates which was reported by Müller and colleagues (2015).
103
Figure 3.8: The alignment of Wetwang Slack 2 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence.
Disagreements to the reference sequence are highlighted in blue. Forward and reverse primer sequences are omitted from this figure as well as in the BLAST
search.
Figure 3.9: The alignment of Wetwang Slack 7 clone sequences from the IS6110 123 bp PCR assay against the M. tuberculosis H37Rv reference
sequence. Disagreements to the reference sequence are highlighted in blue while the gaps are shown in red. Forward and reverse primers are omitted from the
alignment and for the BLAST search.
104
3.2.7 Wetwang Slack 7
The undiluted DNA of the sample Wetwang Slack 7 produced a band with a size that
matches the expected 123 bp product with undiluted DNA as the template (Figure
S3.5). The subsequent nested PCR did not produce positive amplification. The clone
Sanger sequence however confirmed that the amplicon size is 118 bp, 5 bp shorter than
the expected product (Figure 3.9). The closest identity match of this sequence is to the
unidentified partial IS6110-like insertion sequence, isolate IS6110-like sequence type B
(Müller et al. 2015). Wetwang Slack 7 produced negative amplification for the IS6110
nested 92 bp PCR assay and the other 3 targets.
3.2.8 Wharram Percy 26 and Wharram Percy 1600
Both sample Wharram Percy 26 and Wharram Percy 1600 showed positive
amplification for the gyrA target (Figure S3.6) while showing negative amplification
for the other three targets. The positive amplification was obtained from the undiluted
DNA as template. The clone Sanger sequences revealed a 104 bp amplicon from both
samples, similar to the expected product length. However, the nucleotide sequences do
not correspond to the reference sequence. Wharram Percy 26 clone sequences show 7
mismatches to the reference sequence (Figure 3.10). The possible sequence identity
determined by BLAST is matched to Mycobacterium chelonae. The Wharram Percy
1600 clone sequence contain 9 mismatches against the reference sequence (Figure
3.11). Similar to Wharram Percy 26, the best BLAST match was assigned to M.
chelonae.
105
Figure 3.10: The alignment of Wharram Percy 26 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence. The
disagreements to the reference sequence are highlighted in blue. The forward and reverse primer sequences were trimmed from the alignment and were not
included in the BLAST search query.
Figure 3.11: The alignment of Wharram Percy 1600 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence.
The disagreements to the reference sequence are highlighted in blue. The forward and reverse primer sequences were trimmed from the alignment and were not
included in the BLAST search query.
106
3.2.9 Addingham 134 and Addingham 223
Both of samples Addingham 134 and Addingham 223 gave positive amplification in
the IS6110 first step, 123 bp PCR assay using the 10-fold diluted DNA as template
(Figure S3.7). The samples amplified multiple bands, with one of the bands showing a
length close to the expected amplicon size. These samples, however, produced negative
amplification in the subsequent nested IS6110 PCR assay and in PCR of the other 3
targets. Sanger sequences of product clones verified the amplicon size of 123 bp from
both samples, matching the expected fragment size. However, the nucleotide sequences
did not match the M. tuberculosis H37Rv reference. Both Addingham 134 and
Addingham 223 show 16 mismatches within the 83 bp amplicon (primer sequences
omitted) (Figure 3.12). BLAST search suggested the unidentified partial IS6110-like
insertion sequence, isolate IS6110-like sequence type B as the possible identity of the
amplicons from both samples.
3.2.10 Addingham 103
The sample Addingham 103 showed positive amplification in the gyrA PCR assay
using the 10-1 diluted DNA as template (Figure S3.8). This sample produced negative
amplification for the other targets. Sanger sequences from clones confirmed the
presence of a 104 bp amplicon which matches the expected product length. The
nucleotide sequences, however, do not correspond to the M. tuberculosis H37Rv
reference as there are 6 mismatches within the region spanned between the forward and
reverse primers (Figure 3.13). BLAST search revealed Mycobacterium peregrinum as
the possible sequence identity with 98.4% pairwise identity and e-value of 1.35e-21.
107
(a)
(b)
Figure 3.12: The alignment of Addingham 134 and Addingham 223 clone sequences from the IS6110 first step, 123 bp PCR assay against the M.
tuberculosis H37Rv reference sequence. (a) Addingham 134, (b) Addingham 223. The disagreements to the reference sequence are highlighted in blue.
Forward and reverse primer sequences are omitted from this figure as well as in the BLAST search.
108
Figure 3.13: The alignment of Addingham 103 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence. The
disagreements to the reference sequence are highlighted in blue. Forward and reverse sequence are omitted from this figure as well as in the BLAST search.
Figure 3.14: The alignment of Melton 5319 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence. The
disagreements to the reference sequence are highlighted in blue. Forward and reverse sequence are omitted from this figure as well as in the BLAST search.
109
3.2.11 Melton 5319
The amplification of the Melton 5319 undiluted DNA sample produced a faint band in
the agarose gel with length that matches the expected amplicon length (Figure S3.9).
The presence of a 104 bp product is confirmed by the clone Sanger sequences. Two
sequence variants are produced: one with 8 mismatches and one with 12 mismatches
against the reference sequence (Figure 3.14). Both do not correspond to M.
tuberculosis. The most possible identity for both sequence variants is Mycobacterium
chelonae as determined by BLAST search. The first sequence variant (with 8
mismatches) has 96.9% pairwise identity with 1.62e-20 e-value, while the second
sequence variant (with 12 mismatches) has 90.9% pairwise identity with 4.36e-15 e-
value.
3.2.12 Hickleton 46
Positive amplification of the gyrA target was produced from undiluted DNA of the
Hickleton 46 sample (Figure S3.10). A faint band at the expected position can be
observed after the amplification of the 10-1 diluted DNA of this sample. No positive
amplification was obtained from the other PCR assays. Sanger sequencing from clones
revealed the presence of a 104 bp amplicon, matching the expected PCR product
length. The nucleotide sequences, however, do not correspond to the M. tuberculosis
H37Rv reference as there are 8 mismatches within the region spanned between the
forward and reverse primers (Figure 3.15). The BLAST search suggests M. chelonae as
the most possible identity of the sequence rather than M. tuberculosis. This claim is
supported by 96.9% pairwise identity and 1.62e-20 e-value: the probability that the
matches occurred by chance.
110
Figure 3.15: The alignment of Hickleton 46 clone sequences from the gyrA PCR assay against the M. tuberculosis H37Rv reference sequence. The
disagreements to the reference sequence are highlighted in blue. Forward and reverse sequence are omitted from this figure as well as in the BLAST search.
111
IS6110
123 bp
IS6110
Nested 92 bp
M A B C D
200 bp
100 bp
3.2.13 St Giles by Brompton Bridge 1542
The sample St Giles by Brompton Bridge produced multiple bands in the IS6110 123 bp
PCR assay; one of the bands produced from the 10-1 DNA dilution template had a band
length close to the expected size (Figure S3.11). Sanger sequences revealed the amplicon
size to be 134 bp, which is 11 bp larger than the expected PCR product. In the alignment
against the M. tuberculosis H37Rv sequence, 49 disagreements were found, including
gaps (not shown). The closest sequence match in a BLAST search was Streptomyces
lunaelactis (Table 3.2). This bacterium has previously been isolated from moonmilk
speleothem from a cave (Maciejewska et al. 2015).
3.2.14 Sewerby 34
Figure 3.16: IS6110 123 bp and nested 92 bp PCR amplification of sample Sewerby 34. A
and C: undiluted DNA template, B and D: 10-1 diluted DNA template. M is showing the DNA
ladder with 100 bp and 200 bp bands indicated in the figure.
112
Positive amplification seemed to be produced from both the undiluted and 10-1 diluted
DNA template of Sewerby 34 for the IS6110 123 bp assay (Figure 3.16). However, the
sequence identity cannot be verified as the product could not be cloned. The cloning was
not pursued further as the nested PCR did not produce any positive amplification;
therefore, it is unlikely that the identity is MTBC.
3.3 Discussion
3.3.1 MTBC positive samples
In this study, a sample is considered MTBC-positive if positive amplification supported
by a Sanger sequence that matches the M. tuberculosis reference is obtained in at least
one of the assays. Exception is made for the small number of mismatches that can be
accounted for as miscoding lesions in the aDNA targets (Müller et al. 2014a). From the
60 archaeological remains studied, only seven of them produced positive amplification
in the first step of the IS6110 PCR assay: St Andrew Fishergate 6, East Heslington 229,
Wetwang Slack 7, Addingham 134, Addingham 223, St Giles by Brompton Bridge 1542
and Sewerby 34. In the subsequent nested PCR of this target, only St Andrew Fishergate
6 produced positive amplification, with a match to M. tuberculosis identity supported by
Sanger sequencing. The other six samples gave products that did not match the reference
sequence, albeit the product length was similar to the expected amplicon length with the
exception of Wetwang Slack 7 and St Giles by Brompton Bridge 1542. These other six
samples also produced negative amplification in the nested PCR, which can be attributed
to the dissimilarity of the sequence produced from the first IS6110 PCR step.
The results of the MTBC preservation screening therefore suggested that only one of the
60 samples contain MTBC aDNA: St Andrew Fishergate 6. The positive band from both
steps of the IS6110 PCR assay was confirmed as MTBC through Sanger sequencing.
There was one nucleotide discrepancy in each of the three clones for the 123 bp product
which is likely to be a result of miscoding lesions which are common in ancient DNA
templates (Brown & Brown 2011; Dabney, Meyer & Pääbo 2013). These mismatches
were only present in three of the clones and occurred at inconsistent positions. In
113
addition, these mismatches did not appear in the clone sequences derived from the
nested PCR amplicon. Therefore, the nucleotide discrepancies are very unlikely to be
true polymorphisms. This is further supported by the additional presence of clone Sanger
sequences for the amplicons from both steps of the IS6110 PCR from the second DNA
extraction which completely matched the reference sequence without any presence of
polymorphisms. The 7 bp deletion in the Pks15/1 region was also detected in this
sample. The absence of these 7 nucleotides is associated with the Euro-American lineage
of M. tuberculosis and has also been found in other ancient isolates (Bouwman et al.
2012). Three mismatches appeared in each of the clone sequences, which again can be
attributed to miscoding lesions (Müller et al. 2014a). Interestingly all the mismatches are
C → T substitutions which is a hallmark attribute of an ancient miscoding lesion.
Hydrolytic deamination will result in nucleotide modification, resulting in errors when
read by DNA polymerase during DNA replication (Dabney, Meyer, Pääbo 2013).
Cytosine is the primary target of such modification: deamination, which will form uracil
that will further incorporate adenine during the DNA duplication. This will cause either
a C → T or G → A substitution depending on which strand of the DNA is being
amplified (Dabney, Meyer, Pääbo 2013). In the gyrA PCR of St Andrew Fishergate 6,
however, despite the length of PCR product matching that of the expected amplicon, the
sequence did not agree with the M. tuberculosis reference. Instead, the most likely
identity as determined by BLAST analysis is to M. peregrinum. This sequence could
therefore have originated from a contaminant. M. peregrinum is a member of the
Mycobacterium fortuitum complex; one of the members of the rapidly growing non-
tuberculous mycobacteria. The M. fortuitum complex have previously been isolated from
the environment such as soil and water-related sources (Nagao et al. 2009).
3.3.2 Contamination
Three samples produced ambiguous results in the first round of amplification directed at
the insertion sequence IS6110. These samples produced near-exact length amplicons,
which is an indication of MTBC presence: the samples are East Heslington 229,
Addingham 134 and Addingham 223. However, all the samples failed to produce
positive amplification in the subsequent nested PCR round. Sanger sequences from all
three samples are similar to the MTBC reference sequence, but with 16 mismatches; the
positions of these mismatches consistent within the clones from each individual sample.
Table 4.12: Comparison of the 20 most abundant species in each sample as determined by BLAST and MEGAN analysis of the reads
obtained by shotgun sequencing. Yes: species is present abundantly in the sample, yellow: species is present abundantly in at least one skeleton in
all locations (urban and rural), blue: the abundant species is only present in the urban locations, pink: the abundant species is only present in the rural
locations, green: the presence of the abundant species is specific to a single location.
150
Pathogenic bacteria are present abundantly in two samples: St Helen-on-the-Walls 5494
and Wetwang Slack 185. The A. butzleri, a pathogenic bacterium species is abundantly
found in skeleton 5494 from St Helen-on-the-Walls. This microorganism has been
recognized as an emerging pathogen where it was associated with a persistent watery
diarrhoea case in a patient with chronic lymphocytic leukaemia (Arguello et al. 2015).
Two infectious bacteria species were identified in Wetwang Slack 185 sample. M. leprae
is the causative agent of leprosy, while K. rhizophila has been previously inflicted to a
persistent bloodstream infection and it was recognized as an emerging “micrococcus”
(Becker et al. 2008; Moissenet et al. 2012). The death of individuals St Helen on the
Walls 5494 and Wetwang Slack 185 could possibly be associated to these pathogens.
However, care should be taken in deriving such a conclusion. First, the manifestation
and nature of preservation of these pathogens in bone remains are not known. The
identities of these pathogens are not confirmed – the results may be a detection of
similar (but not yet identified) environmental bacteria. Secondly, the M. leprae and K.
rhizophila species account for only 6 and 4 reads in sample Wetwang Slack 185,
respectively. This is due to the low read number obtained from this sample as described
in section 4.4.1. This is not sufficient to draw a conclusion about the cause of death of
individual 185 from Wetwang Slack.
The most abundant read in sample Hickleton 46 is from the species S. album which was
shown to able to cause mycetoma infection in humans (Martin et al. 2004). Interestingly,
mycetoma or also previously known as Madura foot is an alternative diagnosis for
leprosy in skeletons (Hershkovitz et al. 1993). It is possible that the individuals St
Helen-on-the-Walls 5494, Wetwang Slack 185 and Hickleton 46 could have contracted
infections from the pathogens found in the samples but survived. However, the data
presented here are not sufficient to draw such a conclusion.
4.3.3 Target enrichment sequencing strategy
In sample St Andrew Fishergate House 6, the percentage of endogenous DNA recovered
after target enrichment did not differ that much in comparison to the shotgun sequencing
result. It was suggested that the success of target enrichment, especially in-solution
hybridization capture, is vastly dependent on the amount of endogenous DNA in the
original sample; an endogenous DNA amount of at least 1% in the overall DNA content
will be more likely to yield successful results after target enrichment (Cruz-Dávalos et
151
al. 2018). In addition, low clonality and low complexity of starting samples are
important factors in determining the success of target enrichment – especially for
archaeological specimens.
4.3.4 Possible mixed infection in sample St Andrew Fishergate House 6
The skeleton 6 from St Andrew Fishergate showed positive amplifications, verified by
Sanger sequencing as being M. tuberculosis, for at least two of the MTBC PCRs in each
of two DNA extraction batches. From the PCR result, it can be inferred that this sample
does contain MTBC aDNA. However, the result from the shotgun sequencing did not
show many reads mapped to the M. tuberculosis reference genome. The taxonomical
content determination of this sample showed that more reads were assigned to M. leprae
compared to M. tuberculosis. PCR was performed to screen for the presence on M.
leprae aDNA in the sample, but no amplification was shown. Separate target capture
was attempted for both M. tuberculosis and M. leprae, giving enriched libraries that
produced 1,108 and 4,104 reads that mapped to M. tuberculosis and M. leprae,
respectively, with a higher endogenous percentage for M. leprae. This suggests a
possible mixed infection in St Andrew Fishergate 6. Co-infection of M. tuberculosis and
M. leprosy has been demonstrated by Donoghue and colleagues (2005). The mixed-
infection was shown in bone remains from 1st century AD Israel, 4th century AD
(Roman) Egypt, 10th century AD Hungary and mediaeval (10-13th century AD) Sweden
(Donoghue et al. 2005). These cases are older than skeleton 6 from St Andrew
Fishergate (early 14th century AD). The latter is around the time where the prevalence of
leprosy peaked in Britain before it disappeared from this region. However, the number
of reads obtained was not enough to provide sufficient coverage to give any genotype
information for either Mycobacterium.
4.3.5 MTBC aDNA detection in bone remains from Yorkshire
The objective of the study performed in Chapter 3 and Chapter 4 was to determine if the
frequency of MTBC aDNA detection is high enough to plan for a much larger project to
test the hypothesis that there were strain differences between urban and rural areas.
Testing this hypothesis would require a much larger sample size and higher frequency of
positive results than was achieved in this thesis.
152
In the PCR screening, only one of the 60 samples produced a positive result of MTBC.
Other samples produced non-specific amplification, even though some of these skeletons
showed destructive lesions on the vertebrae – a specific bone change indicative of
tuberculosis. Eight samples were sequenced in order to compare the metagenomic
content and to see if there is any MTBC aDNA that can be detected by the NGS.
However, the NGS results showed an extremely low number of MTBC reads, even in
the PCR positive sample: St Andrew Fishergate 6. The frequency of the MTBC aDNA
detection is not high enough to enable endogenous sequence comparison between
different samples. The target capture is expected to increase the number of endogenous
read output from the samples. However, this is not the case in this study. Perhaps this is
due to extremely low copies of endogenous DNA in the sample to begin with. To study
broader hypotheses a much larger sample size would be required in order to obtain a
high enough aDNA detection frequency to allow different strains of MTBC to be
detected and compared. In addition, much better-preserved specimens might be
considered for such studies.
The study described in Chapters 3 and 4 highlight the difficulties in detecting and
isolating MTBC aDNA due to environmental contaminants. In addition, the study also
demonstrates the non-specific results of PCR assays that are supposedly specific to
MTBC alone. Caution should be taken in confirming the presence of tuberculosis
infection based on the skeletal lesions alone, especially the non-specific lesions.
Similarly, caution has to be taken when making disease identification based on positive
PCR results without sequencing authentication.
153
Chapter 5: Genotyping of Mycobacterium leprae ancient DNA
from mediaeval England
5.1 Background of study
This chapter focuses on a genotyping study of M. leprae isolates, which were
responsible for the leprosy endemic in mediaeval England. Leprosy was endemic in
Britain during the mediaeval period and the peak infection occurred in the 13th century
AD (Mendum et al. 2014). In the 15th century AD, human leprosy started to decline
before disappearing in the 16th century AD, although leprosy has recently been detected
in British red squirrels (Avanzi et al. 2016). The same timeline is reflected in Europe,
except for some regions where there is still some leprosy cases reported (Ramos et al.
2016). The absence of contemporary cases in Britain makes it difficult to know which
strains and genotypes were present in this region. Moreover, M. leprae is an extremely
host-dependent pathogen that cannot be cultured in normal culture medium, which
makes leprosy study even more difficult (Groathouse et al. 2006). Therefore, the ability
to study aDNA from archaeological remains provides an opportunity to access genomic
information of past M. leprae strains, especially where there is a zero modern human
leprosy cases as such in Britain.
Here, whole genome target enrichment was used as the NGS strategy, as it was shown to
substantially increase the proportion of M. leprae endogenous DNA in a sequencing
library in previous studies (Bos et al. 2014; Schuenemann et al. 2018). The three English
samples from Chichester and Raunds were tested positive for M. leprae aDNA in a
previous study, confirming the osteological observations (Müller 2008). This provided
the opportunity to retrieve genome sequences of the M. leprae strains present in this
region during that period of time. By comparing the Chichester and Raunds samples
with these other ancient samples, it will be possible to explore the diversity of M. leprae
infections in the past and the spread of different strains.
154
Here, this study is presented in form of a complete manuscript that has been submitted
and accepted, subject to revision for publication in Journal of Archaeological Science.
The roles of each author are outlined as follow.
Authors’ contributions: The osteological examination, DNA extraction and PCR
screening were performed by Dr Romy Muller as part of her Master’s project which was
supervised by Dr Christopher Knusel in 2008 at University of Bradford. She also
prepared the initial DNA library at Manchester Institute of Biotechnology during her
employment as Postdoctoral Research Associate in Brown’s lab. The following DNA
library target enrichment, final DNA library preparation and bioinformatics data analysis
and interpretation were performed by myself. The first draft of the article was written by
myself before being reviewed by my supervisor, Prof Terry Brown. The bone remains to
be studied were provided by Dr Jo Buckberry.
155
5.2 Publication draft
Ancient Mycobacterium leprae genomes from the mediaeval sites of Chichester and
Raunds in England
Ammielle Kerudina, Romy Müllera, Jo Buckberryb, Christopher J. Knüselc, Terence A.
Browna,*
a School of Earth and Environmental Sciences, Manchester Institute of Biotechnology,
University of Manchester, Manchester M1 7DN, UK
b Biological Anthropology Research Centre, School of Archaeological and Forensic
Sciences, University of Bradford, Bradford BD7 1DP, UK
c UMR5199 PACEA, Bâtiment B8, Allée Geoffroy Saint Hilaire, CS 50023, Pessac
We examined seven skeletons from mediaeval contexts from three sites in England for
the presence of Mycobacterium leprae DNA, each of the skeletons displaying
osteological indicators of leprosy. Polymerase chain reactions directed at the species-
specific RLEP multicopy sequence produced positive results with three skeletons, these
being among those with the clearest osteological signs of leprosy. Following in-solution
hybridization capture, sufficient sequence reads were obtained to cover >70% of the M.
leprae genomes from these three skeletons, with a mean read depth of 4–10. Two
skeletons from a mediaeval hospital in Chichester, UK, dating to the 14th–17th centuries
AD, contained M. leprae strains of subtype 3I, which has previously been reported in
mediaeval England. The third skeleton, from a churchyard cemetery at Raunds Furnells,
UK, dating to the 10th to mid-12th centuries AD, carried subtype 3K, which has been
recorded at 7th–13th century AD sites in Turkey, Hungary and Denmark, but not
previously in Britain. We suggest that crusaders or pilgrims to the Holy Land might have
been responsible for the transmission of subtype 3K from southeast Europe to Britain.
Keywords: Ancient DNA, Leprosy, Mediaeval England, Mycobacterium leprae,
Palaeopathology
5.2.1 Introduction
Leprosy is a slowly progressive, chronic granulomatous disease caused by
Mycobacterium leprae (Hansen, 1874) and potentially, in a minority of cases, by the
more recently characterised agent described as Mycobacterium lepromatosis (Han et al.,
2008). The primary symptoms are granulomas of the skin, peripheral nerves and
respiratory tract, but sometimes the eyes, bones and nasal cartilage are also affected
(Britton and Lockwood, 2004). The bacilli accumulate in the extremities of the body,
invading the Schwann cells causing nerve damage followed by a gradual sensory loss
and eventually leading to deformities and disabilities (Masaki et al., 2013). A multi-drug
regime comprising dapsone, rifampicin and clofazimine has been used successfully to
treat 16 million leprosy patients over the last twenty years, but new infections are
frequent with 210,671 leprosy cases reported in 2017 (World Health Organisation,
157
2018). With the highest incidence of new cases occurring in northeast South America,
central Africa and the Indian subcontinent, leprosy is classified as a ‘neglected tropical
disease’ (Lenk et al., 2018). Although the disease itself is curable, leprosy-related
deformities and disabilities are irreversible, especially when treatment has been delayed
(Britton and Lockwood, 2004). Some 2–3 million people worldwide display post-leprosy
disfigurements, and many are subject to the social discrimination referred to as leprosy
stigma, which in the past was driven by misunderstandings regarding transmission of the
disease, and which still persists today in some parts of the world (Grzybowski et al.,
2016).
Leprosy is one of the oldest diseases known to humankind. Although ambiguous,
textual references to skin diseases in the Indian Atharva Veda and Laws of Manu (2000–
1500 BC) (Bloomfield, 2004) and the Egyptian Ebers papyrus (1550 BC) (Hulse, 1972)
have been identified as leprosy, and there are more recent accounts of the disease dating
from the 6th century BC to 1st century AD from India (Bhishagratna, 1996), China
(McLeod and Yates, 1981; Leung, 2008), Greece (Pinhasi et al., 2005) and Rome
(Roberts and Manchester, 2010). Additional evidence is provided by palaeopathological
examination of archaeological skeletons for the osteological manifestations of the
disease that can be observed in the hands, feet, facial bones, tibiae and fibulae of
affected skeletons (Roberts and Manchester, 2010). The oldest skeleton displaying such
lesions dates to 2000 BC, from Rajasthan in northwest India (Robbins et al., 2009), in
accordance with the Indian textual references from the same period. It has been
suggested that the disease was brought to Europe and Northern Africa by the armies of
Alexander the Great, with their return from the Indian campaign in 327–326 C (Roberts
and Manchester, 2010). There is skeletal evidence of leprosy in Egypt at 200 BC
(Dzierzykray-Rogalski, 1980) and in western Europe from the 4th century AD (Reader,
1974). However, the disease appears to have been uncommon in Europe until the
mediaeval period, when skeletons displaying lesions become more abundant (Roberts
and Manchester, 2010). In Britain, the prevalence of leprosy peaks in the 13th century
AD and then declines during the 15th century AD before becoming uncommon again
from the 16th century AD onwards (for a review of the osteological evidence for Britain,
see Roberts, 2002), possibly because of improved social conditions combined with the
development of enhanced resistance to the disease among the human population
(Schuenemann et al., 2013). The decline is mirrored in continental Europe (Bennike,
2002), although the disease persisted in some parts of Norway and elsewhere until the
19th century AD (Boldsen, 2001).
158
About 5% of leprosy cases develop skeletal changes, and the lesions used in
osteological assessment of the disease can be ambiguous. An important adjunct to
palaeopathology has therefore been provided by the detection and sequencing of M.
leprae DNA, which is sometimes preserved in archaeological skeletons displaying
osteological lesions and has also occasionally been detected in skeletons free from such
lesions (Donoghue et al., 2017). Initially ancient DNA typing was used mainly to
support osteological identifications of leprosy (Rafi et al., 1994; Taylor et al., 2000,
2006; Donoghue et al., 2001, 2005, 2015; Inskip et al., 2015), but with increasing
knowledge of genomic diversity among extant M. leprae strains it has become possible
to contextualise ancient DNA data within an evolutionary scheme for the bacterium
(Schuenemann et al., 2018). The M. leprae genome is 3.27 Mb, substantially smaller
than the 4.42 Mb genome of Mycobacterium tuberculosis, and contains relatively high
number of pseudogenes, indicative of reductive evolution (Singh and Cole, 2011).
Different strains show high sequence similarity, with only a small number of variations
in the form of indels and single nucleotide polymorphisms (SNPs) (Monot et al., 2009).
The SNP variations were initially used to divide modern isolates into four main types
and 16 subtypes called 1A–1D, 2E–2H, 3I–3M and 4N–4P. With the addition of more
sequences, this classification has become elaborated into a phylogenetic scheme
comprising six main branches, with branches 1 and 2 corresponding to types 1 and 2,
respectively, branch 3 to subtype 3I, branch 4 to the type 4 strains and also subtypes 3L
and 3M, and branches 5 and 0 to different variants of subtype 3K (Schuenemann et al.,
2013, 2018). Among modern isolates, variants display geographical partitioning with
branch 1 associated with south and east Asia, branch 2 with south and southwest Asia,
branch 3 with Central and North America, branch 4 with west Africa and South
America, and branches 5 and 0 with east Asia (Monot et al., 2009; Schuenemann et al.,
2013). However, these present-day distributions do not reflect the full complexity of M.
leprae distribution in the past, especially in mediaeval Europe where subtypes within
branches 2, 3, 4 and 0 have been identified in skeletons dating from the 5th–14th centuries
AD (Singh and Cole, 2011).
Although M. leprae aDNA has been reported from a number of British sites
(reviewed by Donoghue et al., 2017), sufficient data for subtype identification has only
been obtained from six skeletons from the St Mary Magdalen leprosarium in Winchester
(Scheunemann et al., 2013; Taylor et al., 2013; Mendum et al., 2014; Roffey et al.,
2017) and one skeleton from a cemetery in Great Chesterford, Essex (Scheunemann et
al., 2018). Three of the Winchester skeletons yielded subtype 3I and the other three, as
159
well as the Great Chesterford sample, were subtype 2F. To extend the geographical
range of our knowledge of ancient M. leprae subtypes in Britain, we carried out a
biomolecular examination of seven skeletons from three sites from mediaeval England
(Fig. 5.1), each displaying pathological lesions indicative of leprosy though with varying
degrees of ambiguity. We report M. leprae genome sequences for three of these
skeletons. Two of the genomes correspond to subtype 3I, previously known in Britain,
but the third is novel to Britain and highlights the role that individual mobility might
have played in adding complexity to the phylogeography of M. leprae in mediaeval
Europe.
Figure 5.1: Locations of the sites from which skeletal samples were obtained.
5.2.2 Material and methods
5.2.2.1 Skeletons
Samples were selected, with permission, from the collection of the Biological
Anthropological Research Centre, University of Bradford, UK, based on various criteria.
160
First, skeletons that clearly show pathological rhinomaxillary changes indicative of
leprosy were identified. Given that these changes are pathogonomic for lepromatous
leprosy, bilateral and symmetrical non-specific changes in the lower limb and feet of
those skeletons are hypothesized to be associated with the disease as well. Additionally,
skeletons were sought that showed non-specific lesions in the lower limbs and feet as
commonly seen in leprosy but where no rhinomaxillary alterations could be recorded,
either because they were not present or they could not be observed due to the state of
preservation of the skeleton. The distribution of these lesions, as well as the fact that no
other alterations were found which would suggest a different aetiology, made the
differential diagnosis of leprosy for these skeletons likely. The decision about which
skeletal element and, in case of bilateral skeletal involvement, which side of the body
would be sampled, was based on whether or not destruction was justifiable given the
importance of the specimens for future studies.
Based on these criteria, samples were taken from seven skeletons from three sites
(Table S5.1, Supplementary Note). Skeletons C21, C35, C48 and C227 were excavated
in 1989 from a cemetery that had belonged to the Hospital of St James and St Mary
Magdalene, Chichester, UK. The original hospital was founded c.1118 AD and housed
leprosy sufferers until being dissolved in 1442 AD and converted to an almshouse when
the prevalence of leprosy declined in the UK (Magilton et al., 2008). Based on
osteological data, C21 was a young adult male, C35 an adult of indeterminate sex, C48 a
mature adult male, and C227 an adult, likely male. All four skeletons were dated to the
14th–17th century AD, based on historical documents and associated pottery (Magilton et
al., 2008). Skeletons R5046 and R5256 were males of 17–25 and 25–35 years,
respectively, from the late Anglo-Saxon churchyard cemetery at Raunds Furnells,
Northamptonshire, UK, excavated during 1977–1985. Stratigraphical analyses and
radiocarbon dating suggested that the churchyard cemetery was in use from the mid-10th
until mid-12th centuries (Boddington, 1996). Individual H3726 was a less well-preserved
and incomplete male skeleton of 25–35 years from a cemetery in the southwest area of
the cathedral at Hereford, UK, excavated in 1993 (Stone and Appleton-Fox, 1996). The
cemetery was in use from the 12th–16th centuries AD.
161
Table 5.1: Details of skeletons and samples that were taken.
Site Skeleton Sex, age at death Leprosy indicatorsa Elements sampleda
Rhinomaxillary
changes
Sub-periosteal new
bone formation
Other
changes
Chichester C21 Young adult male Yes Yes Yes Tibia, metatarsus
C35 Adult, indeterminate
sex
No Yes Yes Tibia
C48 Mature adult Yes Yes No Tibia, fibula
C227 Male(?) adult Yes Yes Yes Calcaneus, phalanx
Raunds R5046 Male, 17–25 years Yes Yes Yes Fibula
R5256 Male 25–35 years No Yes No Tibia, fibula, new bone
formation
Hereford H3726 Male 25–35 years Viscerocranium
absent
Yes Yes Tibia, fibula
a For details of pathology and elements sampled, see Supplementary Note (summary of archaeological sites, pathological lesions of skeletons, and elements that
were sampled) and Supplementary Table S5.1 (detailed osteological report).
162
Skeletons C21, C48 and C227 from Chichester and the Raunds skeleton R5046
displayed rhinomaxillary changes and other lesions indicative of leprosy (Table S5.1,
Supplementary Note, Supplementary Table S5.1). The fourth Chichester skeleton, C35,
and individual R5256 from Raunds did not show the typical rhinomaxillary changes but
both did have sub-periosteal new bone formation on other skeletal elements. Only parts
of the left lower limb were recovered from the Hereford skeleton H3726 and so
rhinomaxillary changes could not be assessed. However, the pattern of sub-periosteal
new bone formation on the lower limbs and feet suggested a possible leprous infection.
5.2.2.2 Ancient DNA regime
DNA extractions, PCRs and Illumina library preparations were performed in two
physically separated laboratories within the specialized ancient DNA research facility
at the University of Manchester. Each laboratory was supplied with ultra-filtered air
under positive displacement. After each use, benches and equipment were
decontaminated by UV irradiation and by cleaning with 5% hypochlorite acid, 70%
ethanol and DNA Away (Molecular Bioproducts). Small equipment, plasticware and
UV-stable reagents were decontaminated by UV irradiation (254 nm, 120,000 mJ cm–2
for 2 5 min, with 180° rotation between the two exposures) before use. Aqueous
solutions were similarly irradiated for 15 min. Personnel wore a disposable forensic
suit, face mask, hair net, goggles, two layers of gloves and disposable shoe covers at all
times. DNA extractions were accompanied by two blanks (normal extraction but
without skeletal material) per five samples and every set of 5–7 PCRs was
accompanied by at least two blanks (set up with water rather than DNA extract).
5.2.2.3 DNA extraction, PCR and sequencing
Bone samples were taken using a hacksaw or electronic drill by personnel wearing
protective clothing, including forensic suits, hair nets, face masks and two pairs of
sterile gloves. Samples were placed in sterile plastic bags and stored under dry and cool
conditions and transferred to the ancient DNA facility. The bone surfaces were
163
decontaminated by mechanical removal of the outer 1–2 mm of each sample, followed
by UV irradiation (254 nm, 120,000 mJ cm–2) for 2 5 min, with 180° rotation
between the two exposures (Bouwman et al., 2006). Bone samples were then placed in
a DNA-free plastic bag wrapped in a sterile piece of aluminium foil and crushed into
fine powder. DNA was extracted from 0.2 g of bone powder by standard methods
(method D of Bouwman and Brown, 2002; Dabney et al., 2013).
An initial screening for presence of M. leprae DNA was carried out by hemi-nested
PCR directed at the RLEP repetitive element, in 50 µl reactions comprising 2.5 µl of
DNA extract or 1.0 µl of first round PCR product, 1 AmpliTaq Gold PCR Master Mix
(ThermoFisher Scientific), 2 mM MgCl2, 200 µM dNTPs, 200 ng each primer, 1%
bovine serum albumin and 1.25 units AmpliTaq Gold DNA polymerase (ThermoFisher
Scientific). The primers for the first PCR (forward: 5´–