Species-specificity of mycolic acids

How can we be so certain that these targets are species-specific? Reverse-phase HPLC separates mycolic acids from MTB principally according to their overall chain length. Contributions from the various functional groups (e.g. keto and methoxy groups) (Minnikin 1993) results in the characteristic "tight envelope" of peaks (see Figure 2) which enables MTB specific mycolates to be clearly distinguished from those of other mycobacteria (Butler et al. 1986; Butler et al. 1991; Glickman et al. 1994; Ramos 1994) that share this alpha-, methoxy- and ketomycolic acid pattern (Minnikin 1987). The further rpHPLC analysis (see Figure 4) of the separated individual alpha-, methoxy- and ketomycolate classes (see Figure 3) provides absolutely reliable evidence that these long-chain mycolates originate from M. tuberculosis.

Contamination from the burial environment

In this study, the bones were heavily contaminated with material from the burial environment, resulting in a much higher background signal, an extended "solvent front" that did not quite obscure the mycolic acids on first reverse phase (Figure 2). Further refinements in the extraction and chromatographic protocols should enable the removal of this signal overprint, eliminating this cosmetic deficiency.

Corroboration of osteological examination

Mycolic acids, characteristic of M. tuberculosis, have been confirmed in 5/21 ribs under investigation (23.8%) and this correlates with the recorded frequency of 27.1% for the Newcastle Infirmary. Serendipitously, one of the individuals (SK 31) giving positive results for mycolate analysis was subsequently found to have changes in the spine consistent with TB (Steinbock 1976). Lesions have been reported on the surface of archaeological ribs and although not yet proven, TB is suggested to be the most likely cause (Roberts et al. 1994). None of the rib samples in the Newcastle study had lesions.

Comparison with molecular methods

Why may there be a need for another biomarker approach to the diagnosis of ancient TB? Nucleic acids are already being used to diagnose ancient disease (e.g. Taylor et al. 1997). A brief comparison with the current molecular methods may be needed here. Molecular methods depend upon the demonstration of species-specific target sequences of nucleic acids, usually DNA. For most ancient studies, this requires that the target sequence be amplified using the polymerase chain reaction (PCR) method, so that detection is possible. In the case of ancient TB diagnosis, if this method is to be reliable, there must be no equivocal specificities in the target element, the target element must be carried by all strains and there must be assurance that disease, not infection, is being detected.

Currently the insertion element, IS6110, is taken as the "gold standard" for diagnosis of ancient M. tuberculosis (e.g. Salo et al. 1994; Baron et al. 1996; Taylor et al. 1996; Faerman et al. 1997; Donoghue et al. 1998). IS6110 is a repetitive insertion sequence that is usually present 6-20 times in the MTB genome. Certain regions of IS6110 are reported to be species-specific for MTB complex bacteria (Eisenach et al. 1993; Hellyer et al. 1996; Hellyer et al. 1997; Hellyer et al. 1998), but there is a suggestion that this may not be so (McHugh et al. 1995; Gillespie et al. 1997; Gillespie et al. 1998). Within IS6110, a 181base-pair area of homology exists which has been shown to be present in mycobacteria other than MTB (MOTT) (Gillespie et al. 1997). The majority of ancient work has used a set of primers which amplify a region (Eisenach et al. 1991) just outside this area of homology. Some workers, however, suggest that this latter target has a 3% false-positive rate in clinical samples for species of MOTT (Dalovisio et al. 1996). Additionally, not all modern strains of MTB complex carry this element (van Soolingen et al. 1993). Applying this to ancient tuberculosis detection, the percentage of ancient MTB complex strains lacking IS6110, and the number of ancient MOTT containing it (if any) must remain unknown.

If the host immune system fails (for whatever reason) to keep the infection in check, the numbers of mycobacteria increase (Ridley 1988, 211), and disease results. In disease, therefore, there are more organisms present than in infection. Amplification methods to detect DNA may be too sensitive, and the possibility that it is infection and not disease which is being amplified is a major cause for concern (Baron et al. 1996).

Mycolate detection does not involve amplification steps, so the number of skeletons containing mycolates may reflect the frequency of disease in the original population. It is interesting to note that the 23.8% prevalence of mycolates, recorded here, is higher than that expected in the general population (~20%) (Hardy 1994), but enhanced disease frequencies would be expected in a hospital site. This suggests that the detection of mycolates in archaeological skeletal material will give a true indication of the TB prevalence for a site, but it will be necessary to validate this method statistically by the examination of many samples, from sites with known TB prevalence. A biomarker that measures TB frequency in past populations will be a valuable adjunct for epidemiology. Thus far, mycolate analysis has shown the promise of being more certain for TB diagnosis than osteological examination and therefore it may be a reliable technique for the purpose.

Value as an indicator of "poverty"

Tuberculosis prevalence has a strong link with poverty in modern populations and the admissions and death registers show that the occupants of the Newcastle Infirmary Burial Ground - soldiers, sailors, pitmen and labourers - derive from the lowest economic strata of the developing industrial society of the 18th and early 19th centuries.

The chemical identification of species-specific MTB mycolic acids in the Newcastle Infirmary burials, where there is independent documentary evidence for the prevalence of the disease, suggests that this technique has the potential for the identification of broadly comparable levels of poverty in other historic cemetery populations. In mycolic acids we have, for the first time, a biomarker for a disease that may indicate how well-nourished our working-class ancestors were.


© Internet Archaeology URL:
Last updated: Tue Dec 1 1998