A brief cover of the paleomicrobiology of Mycobacterium tuberculosis.

Tuberculosis is caused a by a group of bacteria that are closely phylogenetically related called the Mycobacterium tuberculosis complex. They are slow growing aerobic gram positive bacilli that due to its high lipid content in its one phospholipid bilayer, requires stain methods that facilitate the dye into their cell wall, such as the heat method of Ziehl Neelsen staining. It is estimated by the WHO that approximately one third of the world population have latent TB infections but only 5 to 10 percent of these individuals will develop the disease in their lifetime.

Skeletal Material

It was recently thought that TB was likely transferred to humans during the Neolithic period associated with increasing sedentary lifestyles and domestication of cattle, perhaps transferred from Mycobacterium bovis (Iseman 2478 : 1994), but with the exception of Mycobacterium canetti, members of the MTBC group do not show horizontal gene transfer (Gagneux 852 : 2012) most of the genetic differences in the MTBC are due to deletions. Thus comparison of human M tuberculosis and M bovis indicate that the M bovis genome is around 60 000 base pairs smaller than the human M tuberculosis genome, thus the human M tb lineage is older than M bovis and this also suggest TB in humans predates the Neolithic period (Gagneux 852 : 2012).

TB in the bone is the result of a post primary tuberculosis spread and is a chronic process, thus features remodelled marginal erosive lesions. In cases of active TB, 3 to 5 % of individuals develop bone lesions (Holloway et al 2011). Newly formed woven bone can be distinguished from mature lamellar sclerotic changes, that can suggest healing. The presence of woven bone indicates lesions active at time of death while sclerotic changes changes signal an inactive disease process (Roberts & Buikstra 88 : 2003). Mycobacterium bovis, is more likely to cause skeletal damage and is potentially up to 20 times more likely to lead to bone and joint tuberculosis in children.

Other diseases that show up in skeletal material include osteomyelitis, syphilis and leprosy. Indicated in Figure A.

Figure A

Because these diseases leave evidence in skeletal material much attention in paleopathology focused on these diseases in the historical record. Such classifications tend to be on the basis of most probable cause and when collections of skeletal material demonstrating pathology classifications has been re examined there have been many examples the initial classifications being challenged. (Picture from Infections in Palaeopathology : The Basis of classification according to most probably cause. By Juliet Rogers and Tony Waldron 1989).

Donoghue (2011) describes the gradual inclusion of bone specimens for TB analysis by DNA techniques that did not show bone pathology, the initial studies took this as indications of haematogenous spread of tuberculosis bacilli. There have been cases where not MTBC aDNA was found in bones demostrating pathology, most likely attributable to poor DNA preservation but at least one study has found brucellosis DNA sequences, IS6501 in vertebra negative for IS6110 and M bovis sequences oxyR pseudogene and mpt40 that showed pathology that could easily be classified as typical of M tuberculosis (Mutolo 2011).

Up until the 1960's it was still being argued that there was no Pre Columbian tuberculosis on the basis of skeletal material.

The cut off for pre-Columbian contact is 1492 AD, the exact date of the mummy is not given but is around this time. It is significant because an early theory for the absence of skeletal remains showing TB in the Americas was that TB was introduced by European contact, and was still being argued about in the 1960s by Dan Morse. An interesting part of his argument was that North Americans lived in small population densities that would not have supported the disease, making a comparison to then contemporary indian populations, he calculated that 2.24 percent of skeletal material should present evidence of bone tuberculosis and 0.67 percent with vertebral involvement (Roberts & Buikstra 2003 : 188).

The study conducted by Corthals et al (2012), titled “Detecting the immune system response of a 500 year old inca mummy” uses PCR amplification to detect the presence of Mycobacterium tuberculosis and “shotgun proteomics” to detect the protein expression in buccal swabs and cloth samples from 500 year old Andean mummies, two young children and an adolescent girl.

The mummies were buried 25 meters from the 6739 metre summit of Llullaillaco, a high elevation volcano in the province of Salta, Argentina. They were buried 50 cm underground and packed with volcanic ash. This ash inhibited the growth of decomposing bacteria & fungi, functioned as a barrier to moisture and airtight with snow and the freezing temperature, the corpses were preserved by the combination of mild humidity, anaerobic environment and natural disinfectants (see figure B).  

Figure B

The paper describes the three children as being sacrificed to Pachamama, the earth goddess, in the ritual of Capacocha. Subcutaneous fat in the bodies was converted to soap by a process called adipocere.  

The presence of pathogenic Mycobacterium was indicated as being probable by PCR using primers for the genus Mycobacterium and primers belonging to species in the Mycobacterium tuberculosis complex. For a lot of DNA techniques the significance of the presence of Mycobacterium DNA is questionable, part of the natural soil flora, a contaminant or an indication of the pathogen. Early work trying to confirm the presence or absence of a pathogen used a fragment of the M. tuberculosis complex MTBC specific repeat element, but this is for the whole MTBC complex, not just Mycobacterium tuberculosis. Because of the good preservation of the material the study has been able to use a wide range of primers, including species specific primers, as indicated in this figure (figure C).

Figure C

It has been argued that the mycobacterium cell wall properties can potentially enhance the preservation of nucleic acids (Donoghue & Spigelman 2006) but Wilbur et al 2009 argues that the same cell wall properties will enhance its degradation, so this is an issue to be resolved. More recent papers describe the use of a MALDITOF, Matrix Assisted Laser Desorption Ionization Time of Flight, mass spectrometry technique that detects the presence of the tuberculosis using the lipid components of the cell wall, mycolic acids. This is not without its issues, but it is potentially able to be used to detect Mycobacterium tuberculosis pathogen in less well preserved skeletal material.

The Corthals et al (2012) paper wanted to establish whether or not the mummy of the adolescent female had an active infection at the time of her sacrifice. It would be interesting to speculate whether this influenced the process of her being used in the sacrifice. The proteonomics technique used by the study was able to match up the mass spectrum of proteins from the lip swab sample of the 15 year old adolescent female to a library of immune system proteins.

They found normal serum proteins and some proteins not normally present in blood or saliva but are consistent with immune response to infectious disease. These included,

• Cathepsin G, a specialized neutrophilic polymorphonuclear leukocyte serine protease found in the azurophil granules, with its function liked to pathogenesis of diseases associated with inflammation and neutrophil infiltration of the airways.
• Alpha antitrypsin, a marker of chronic lung infection, a strong indicator of mycobacterial infection.
• Neutrophil defensin 1 and 3 are part of the defensin family of cysteine rich cationic proteins found in leukocytes and are specifically associated with macrophages involved in lung tissue inflammation response.  
• Proteins including apolipoproteins A1 & A2 and transthyretin, which are expressed in chronic and acute lung inflammation, and can be used as monitoring biomarkers for pulmonary related diseases.

The full complement of these proteins in the mouth swab of the 15 year old adolescent provides strong evidence of a response to a severe respiratory bacterial infection, the physical evidence and other immune response proteins supports the hypothesis of an infection caused by Mycobacterium.

After the proteomics technique the paper presents a PCR and a maximum likelihood phylogeny to indicate a higher probability that presence of Mycobacterium belonged to the pathogenic MTBC complex.

The ancient lineage of Mycobacterium tuberculosis is suggested from a study on a pre-Pottery Neolithic site in Atlit Yam in the Eastern Mediterranean dating from around 9250 to 8150 Before present era, bones belonging to a woman and infant demonstrated morphological changes, the detection of ancient DNA and MTBC specific cell wall lipid biomarkers (Donoghue 2011).

Helen Donoghue (Donoghue 82 : 2008) describes the use of cell wall lipid biomarkers for the purpose of identifying members of the MTBC, with the advantage that the lipid biomarkers are more stable, potentially lasting longer than DNA and the methods used, HPLC and Mass spectroscopy are sensitive enough for direct detection.

There are issues with the use of lipid biomarkers for identification of members of the MTBC, although the recent mycoloyl arabinoglactan- peptidoglycan complex is known, ancient MTBC mycolic acids as synthesized are potentially different due to adaptive evolution (Mark et al 112 :2011) and will undergo chemical reactions post mortem thus over time will change the mass spectrum peaks and fragment identification. The images on the screen are from Mark Lazlo's et al (2011) paper “Analysis of ancient mycolic acids by using MALDI TOF MS: response to “Essentials in the use of mycolic acid biomarkers for tuberculosis detection” by Minnikin et al 2010. Lazlo is referenced by Helen Dononhue (2011) in the detection of lipid biomarkers, the team is responding to a critique that the method is not sensitive enough for its application for archaeological material.

When it was being argued that there was no MTBC on the American continent, part of this argument was there was a founder effect bottleneck and a low population density did support endemic Mycobacterium tuberculosis. The low population density argument is shown to be flawed with evidence of high density sedentary Pre-Columbian populations in North America, without even discussing Meso and South America but also a different picture of the incidence of Mycobacterium tuberculosis in ancient American populations is emerging. Because of M tuberculosis biomarkers are being found in archaeological material without demonstrating pathological change the incidence of M tuberculosis in historical/ paleolithic populations is seen to be greater than previously estimated (Donoghue 821 : 2011). This is accompanied by more recent insights in the origin and evolution of M tuberculosis.

A paper by Holloway et al (2011) titled Evolution of human tuberculosis: A systematic review and meta-analysis of paleopathological evidence attempts to compare all known cases of TB showing bone involvement from the range of 7250 BCE to 1899 and does a Chi squared test against the null hypothesis that the frequency and location of bone lesions does not change over time. The papers data was reported cases in literature, which yielded 530 TB cases from 221 gravesites reported in 151 references (Holloway et al 407 : 2011). One of the interesting aspects of the paper was its classification of sites in terms of pre urbanised, urbanised and early modern time periods. In the Mediterranean early urbanization is associated with Phoenician and Greek expansion, the beginning date used was from 600 BCE (before the common era), while in Northern Europe early urbanization is from 800 CE (common era). The paper classifies early urbanization in the Americas as from 1500 CE, the point of European contact. The paper produced a bar chart, comparing time periods that shows a decrease in the frequency of skeletal lesions over time, and a decrease in the frequency of involvement of the spine with an increased involvement in other skeletal locations (Holloway et al 411 : 2011). The significance of this can be discussed with respect to an “osteological paradox” (Holloway et al 414 : 2011), where to develop bone lesions the individual has to be generally immunologically resistant enough and healthy enough to survive a chronic condition. Thus this decrease in the frequency of bone lesions could mean that the frequency of individuals with sufficient immunological resistance has decreased. They argue that because the average age of death was the same in all three time periods that this is unlikely and describe a general evolutionary process of decreased virulence of bacteria over time, although the cause of death may not have been due to TB.

Figure D

This decreased virulence over time is argued to be evidence of a form of sympatric speciation, a process of coevolution that is often demonstrated by reciprocal transplant experiments, where the performance of locally adapted pathogen variants (sympatric) are compared to non locally adapted pathogen variants (allopatric) ( Fenner et al 2 : 2013). An indirect way to observe this is to look at the performance of pathogen variants when the immune system of individuals of the sympatric host population is disrupted. A paper by Fenner, Egger, Bodmer, Furrer, Ballif et al 2013 called HIV infection disrupts sympatric host pathogen relationships in human tuberculosis has done this presenting a statistical argument that HIV infection disrupts sympatric host pathogen relationships in human tuberculosis. Currently Mycobacterium exists in six main phylogenetic lineages associated with geographical regions with sympatric human populations. Transmission tends to occur in higher frequency in sympatric host pathogen combinations and these associations remain in international cosmopolitan situations. The Fenner et al study focused on 233 European born patients from a population of 518 patients (21.6% had HIV) and defined a sympatric host-pathogen relationship as infection with lineage 4 and allopatric as infection with any other lineage. From this group of 233 European born patients, 36 had HIV and of these individuals 9 had an infection with a non lineage 4 Mycobacterium tuberculosis strain (Fenner et al 4 : 2013).

Figure E

Using a multivariate analysis there was a statistically significant associated between HIV infection and infection by allopatric M tuberculosis (unadjusted OR 7.0, 95% Confidence interval 2.5 -19.1, p value less than 0.0001). It also found that the strength of the association between HIV infection and allopatric lineage increased as CD4 T cell count went down (Fenner et al 6 : 2013). The study has attempted to adjust for age, sex, frequency of travelling, contact with foreign populations and other forms of immunosuppresion and found that the only strong adjusted variable was repeated travelling to low income countries with a decreased OR to 4.5 (95% CI 1.5 -13.6, p = 0.008). This data supports the idea of co-evolution of the Mycobacterium tuberculosis and human host populations.  

The time scale for the co evolutionary relationship of the Mycobacterium tuberculosis Complex with humans is suggested by genome sequencing of M tuberculosis to be at least 2.6 to 2.8 million years ago, earlier than the range of Homo habilis (2.3 to 1.4 million years ago), suggested by synonymous sequence diversity (Donoghue 825 : 2011), there is a possible material evidence of case of TB in Homo erectus dating to 490 -510 000 years ago, from a partial skeleton found in western Turkey (Donoghue 826 : 2011) .

The common image of Mycobacterium tuberculosis, with very good reason, is as a scourge of high density population centers characterised by the conditions of the early industrial revolution (20^th century) and urban populations characterised by poverty. It is the medical reasoning for antispitting laws that were enforced by New York City around 1910, recently Waltham Forest Council in Britain (Paul Cocozza, from The Guardian, September 2013) fined spitters and there is currently an anti spitting Bill is being introduced in the Phillipines. But given its long association with humans, predating the Neolithic, the predominant conditions providing the evolutionary selection pressure on it would be different from those now.

These pre Neolithic conditions would be low density populations, that are composed of small, groups organised around kinship structures that are predominantly mobile (Donoghue 826 : 2011). Because of this the organism would be under selection pressure to survive through the lifetime of the individual host, with transmission occurring from adults with reactivation of the disease to other individuals, including infants with immature immune systems. Reactivation would occur in situations of nutritional deprivement, age related failing immune systems, and physical and mental stress (Donoghue 826 : 2011). As such the organism would be selected towards a quasi-commensal host pathogen relationship, with increased latency in the majority of cases, over time there would be co evolutionary processes leading to sympatric host pathogen relationships.

Conclusions

One of the insights that this model provides is that given the increased population density, with a greater opportunity for transmission, the selection pressures favoring latent infection would be reduced and so selection pressures should favor an increased bacterial virulence (Donoghue 827 : 2011). There is evidence of evolutionary modern lineages that induce lower levels of early inflammatory response and faster progression to active disease (Donoghue 827 : 2011).

The use of antibiotics is a selection factor, and most papers describe the mutation rates of Mycobacterium tuberculosis in resistance to line 1 antibiotics, isoniazid, rifampicin and ethambutol as within the range of 1 X 10^-7 (ethambutol) to 2.3 X 10^-10 (rifampicin) and seem to describe Mycobacterium tuberculosis as predominantly transferring antibiotic resistance vertically, through generations (Gumbo 722 : 2013). In lineage 2, which includes the Beijin strains, there is a prevalence of MDR tuberculosis. It has been found that lineage 2 appears to have a tenfold increased acquisition of rifampicin resistance compared to lineage 4. This suggests another implication, that there are different rates of multidrug resistance (antibiotic resistance) in the different lineages.

Figure F

Comments.

Rabbit test models to determine when treponemal dissimination found that in later stages of the disease Treponomal DNA could not be isolated from the bone. Syphilis produces marks of severe inflammation on long bones and cranium, deep erosions and nodes that can be described as a moth eaten appearance and a thickening of the long bones. Sectioned bone reveals a spongy appearance with obliteration of the medullary cavities.

Holloway et al (2011) uses Pompeii as a way of checking whether the frequency of lesions in the age and sex distributions found in the cemeteries corresponded with the frequency of lesions in living population age and sex distributions. It was found that it did not differ in adult populations but did differ significantly in sub adult (child) populations (Holloway et al 414 : 2011), this is why the Holloway et al (2011) paper only discusses adult populations.

Mechanisms of action of antibiotics, for figure F.

Ethambutol inhibits arabinosyl transferase involved in cell wall biosynthesis, the enzyme is coded for by three homologous genes embCAB and resistance is associated with mutation of these loci

Isoniazid is a peroxide producing drug that relies on catalase activity to interfere with mycolic acid biosynthesis.

Rifampic prevents transcription of DNA dependant RNA polymerase, binds to beta subunit, resistance is conferred by mutations of the rhoB gene that encodes

Pyrazinamide interferes with fatty acid synthesis, lowering intracellular pH which inactivates fatty acid synthase

Bibliography

Donoghue, Helen D. (2008). Chapter 6 : Palaeomicrobiology of Tuberculosis. In D Raoult & M Drancourt (eds). Paleomicrobiology : Past Human Infections. Published by Springer-Verlag Berline Heidelberg. Pages 75 -97

Donoghue, H.D. (2011). Insights gained from palaeomicrobiology into ancient and modern tuberculosis. In the Journal of Clinical Microbiology Infectious Diseases. Volume 17. Pages 821 – 829.

Fenner, L; Egger, M; Bodmer, T; Furrer, H; Ballif, M et al. (2013). HIV Infection Disrupts the Sympatric Host-Pathogen Relationship in Human Tuberculosis. In PloS Genet Volume 9, Index 3. e1003318. doi:10.1371/journal.pgen.1003318

Gagneux, Sebastien. (2012). Host-pathogen coevolution in human tuberculosis. In Philosphoical transactions of The Royal Society. Volume 367. Pages 850 -859.

Gumbo, Tawanda. (2013). Biological variability and the emergence of multidrug-resistant tuberculosis. In Nature Genetics. Volume 45. Pages 720 -721.

Holloway, K.L; Henneberg, R. J; Lopes, M de Barros & Henneberg, M. (2011). Evolution of human tuberculosis: A systematic review and meta-analysis of paleopathological evidence. In Homo- Journal of Comparative Human Biology. Volume 62. Pages 402 -458.

Iseman, Michael. (1994). Evolution of drug resistant tuberculosis: A tale of two species. In Proceedigns of the National Academy of Sciences of the United States of America . Volume 91. Pages 2428 -2429.

Mutolo, Michael ; J. Jenny, Lindsey; Buszek, Amanda; Fenton, Todd W & Foran, David. (2012). Osteological and Molecular Identification of Brucellosis in Ancient Butrint, Albania. In the American Journal of Physical Anthropology. Volme 147. Pages 254 -263.