Mycobacterium tuberculosisComplex DNA in Ancient Human Bones

Journal of Archaeological Science (1996) 23, 667–671

Mycobacterium tuberculosis Complex DNA in Ancient Human Bones Heike Baron, Susanne Hummel and Bernd Herrmann Institut für Anthropologie der Universität Göttingen, Bürgerstrasse 50, 37073 Göttingen, Germany (Received 5 December 1994, revised manuscript accepted 15 July 1995) Mycobacterium tuberculosis complex DNA was detected in ancient human bone specimens from three individuals with bone tuberculosis. The reliability of PCR results and the applicability of the method to skeletal remains from soil were ensured by a system of control reactions. The pathogen’s DNA could not only be detected in affected bone tissue, but also in bone tissue without any tuberculous lesions. Therefore we expect that any infectious agent which is carried into bone with the bloodstream should be detectable by PCR. ? 1996 Academic Press Limited Keywords: DNA, ANCIENT, PCR, BONE, INFECTIOUS DISEASES, MYCOBACTERIUM TUBERCULOSIS.

Mycobacterium tuberculosis aDNA could already be amplified (Salo et al., 1994). There are also indications that it can be amplified from bone (Spigelman & Lemma, 1993; Dixon et al., 1994).

Introduction nfectious diseases have always been a major life threat in historical and prehistorical human populations. Paleoepidemiology of infectious diseases is important for reconstruction of living conditions since the spreading of infectious agents is strongly related to socioeconomical factors like nutritional status, population density and hygienic conditions. Paleoepidemiological data are also helpful to study human adaptation to infectious diseases when linking them to information about genetic variations in human populations. Epidemiological data from different ancient burial sites lead to models about the evolution of infectious agents, as shown by the model of interaction of tuberculosis and leprosy bacilli (Manchester, 1991). The collection of paleoepidemiological data depends on the diagnosis of infectious diseases on human remains which are mainly skeletal ones. Since most infectious diseases do not leave any traces on the skeleton, diagnosis is only possible by detection of the infectious agents themselves. Most infectious agents can be isolated from blood (Dalton et al., 1986), thus, in principle they should be retrievable in bone, too. However, their amount in bone can be assumed to be very small. Therefore the diagnosis of infectious diseases in human skeletal remains requires a highly sensitive method like the Polymerase Chain Reaction (Mullis & Faloona, 1987) which has been applied successfully to a wide variety of ancient biological sample materials containing ancient DNA (aDNA) (Herrmann & Hummel, 1993). Although untreated tuberculosis infectious may leave bone lesions, we selected tuberculosis as an example to investigate if it is generally possible to amplify aDNA of infectious agents from archaeological human bone. From mummified soft tissues,

I

Experimental Design Selection of specimens Since PCR based detection of pathogens from old bone samples is not a routine procedure yet, samples used in this study had to come from individuals with a confirmed diagnosis suffering from an infectious disease. This procedure ensures a discrimination of the negative PCR results, which are due to insufficient methodology only, from those, which are due to the absence of the disease instead. Therefore, tuberculosis was chosen as an example for this novel technical approach identifying diseases since it leaves morphological lesions on the bone in advanced stages. However, differential diagnosis of bone tuberculosis in ancient skeletal remains is not always possible because of morphologically overlapping signs with some other infections (Ortner & Putschar, 1985). Therefore, a confirmed diagnosis could be provided by autopsy specimens from patients where tuberculosis was diagnosed before death. As the specimens should contain aDNA to ensure the applicability of the method to ancient skeletal remains, we chose autopsy specimens from the 100 year old pathological collection of our Department. High background of human DNA DNA extracts from autopsy bone specimens with tuberculous lesions can be expected to contain a small amount of the infectious agents’ DNA in a high 667

0305-4403/96/050667+05 $18.00/0

? 1996 Academic Press Limited

668 H. Baron, S. Hummel and B. Herrmann

Figure 1. Tuberculosis of the distal femur. Object No. 13 K 37:2 from the historical pathological collection at the Institute of Anthropology, University of Göttingen. Whole scale bar=5 cm.

surplus of human DNA. A high background of nontemplate DNA may inhibit specific amplification (Hummel, 1992). In order to simulate this situation, we carried out control amplification reactions with mixtures of diluted DNA from M. tuberculosis or Mycobacterium bovis, respectively, and a high surplus of DNA from modern human bone. Control PCR reactions were done as well with human DNA only in order to exclude false-positive PCR results due to unspecific amplification of human DNA. Nontuberculous mycobacteria There are many soil-inhabiting mycobacterial species (for a review see Tsukamura, 1984) not belonging to the M. tuberculosis complex which consists of M. tuberculosis, M. bovis, M. bovis BCG, M. africanum and M. microti (Tsukamura, 1983). It is known that bacteria from soil can populate skeletons (Herrmann et al., 1990). Thus, when working with skeletal remains excavated from soil, DNA of soil-inhabiting mycobacteria might cause false-positive PCR results. Therefore we selected primers which are specific to the M. tuberculosis complex. In addition, negative control reactions were carried out with DNA extracted from soil and from cultures of soil-inhabiting mycobacteria.

Material and Methods Materials The investigation was carried out with specimens from a historical pathological collection from the files of the

Institute of Anthropology, University of Göttingen. The autopsy specimens had been collected from the end of the 19th century up to the beginning of World War II. We selected two femora and one skull from individuals who had suffered from bone tuberculosis (one is shown in Figure 1) and took samples from affected bone tissue. From two of the specimens we additionally took bone tissue that did not show any tuberculous lesions. Modern human DNA was extracted from autopsy bone specimens dating from 1990 which had been stored frozen until DNA extraction. DNA from M. tuberculosis, M. bovis and from the soilinhabiting species M. gordonae and M. aurum were extracted from bacteria cultures. Soil samples came from a burial site in Lünen/Wethmar, North Rhine Westphalia. Preparation of bone and soil samples Before aDNA extraction, the bone samples were irradiated with UV light (254 nm) to inhibit possible contaminating DNA (Ou, Moore & Schochetman, 1991). After this decontamination procedure they were powdered in a mill. Samples of 0·3 g bone powder each were mixed with 1·5 ml 0·5  EDTA, pH 8·0 and constantly shaken at room temperature for 12–24 h. The bone powder was removed by centrifugation (10 min, 5000 rpm) and the supernatant was submitted to an automated DNA extraction. Except for UV irradiation, soil samples were prepared in the same way as bone samples.

Mycobacterium tuberculosis Complex DNA in Ancient Human Bones 669

Automated DNA extraction 1200 ìl supernatant of the EDTA bone incubation or 250 ìl of bacteria culture, respectively, were submitted to DNA extraction. The following steps were carried out by an automatic DNA extractor (Applied Biosystems, Gene Pure Type 341 A): (1) 60 min shaking at 60)C with 1·5 ml sterile water (Ampuwa) and 500 ìl proteinase K (20 mg/ml); (2) extraction with 3 ml 70% phenol:chloroform:water; (3) extraction with 4 ml chloroform; (4) precipitation on small teflon filters supported by 100 ìl 2  sodium acetate, 3 ml isopropanol and 5 ìl of a silicate suspension (glassmilk); (5) washing with 4 ml ethanol. After the automated extraction the DNA was rinsed manually from the teflon filters and resolved in Ampuwa. Polymerase chain reaction The primers we used for amplification of M. tuberculosis complex DNA were described by Eisenach et al. (1990). They frame a 123 bp fragment unique to the members of the M. tuberculosis complex which exists as 10–16 copies per cell. This DNA fragment is part of a repetitive insertion element-like sequence, IS6110, described by Thierry et al. (1990a). PCR was carried out following the principles of biphasic amplification (‘‘Booster-PCR’’) described by Ruano & Kidd (1989), Ruano, Fenton & Kidd (1989) which has already been adapted to aDNA extracts (Hummel, Nordsiek & Herrmann, 1992; Hummel & Herrmann, 1993a,b). The first reaction stage was done in a 50 ìl reaction mixture containing 10 m Tris-HCl (pH 8·3), 50 m KCl, 2·5 m MgCl2, 175 ì each dNTP, and 0·4 fmol of each primer (final concentration: 8·0 p). After 7 min UVirradiation of the reaction tubes containing the reaction mixture and an overlying drop of mineral oil, DNA was added. Reactions with DNA from historical bone samples were done with 15–150 ng total DNA. DNA (150 ng) from modern bone were submitted to PCR. The reactions with bacterial DNA contained 300 fg (M. tuberculosis, M. bovis) or 30 pg (M. gordonae, M. aurum); simulations were carried out with a mixture of 150 ng human DNA and 300 fg DNA of M. tuberculosis or M. bovis. Reactions with DNA from soil contained 100 ng DNA. Taq-Polymerase (1·5 U) (Perkin Elmer) was added at 94)C, following the hot start technique. After 20 cycles of amplification with 94)C 1 min, 68)C 4 min, 72)C 1 min in a Thermal Cycler (Perkin Elmer Cetus), primer concentration was increased to 0·1 ì by adding 5 ìl of a mixture containing 10 m Tris-HCl (pH 8·3), 50 m KCl, 2·5 m MgCl2, and 5·5 pmol of each primer. Cycling parameters for the second stage were 94)C 1 min, 68)C 2 min, 72)C 1 min, 40 cycles.

Digestion of PCR products with the restriction endonuclease Sal I To ensure the identity of the PCR products, they were analysed with the restriction endonuclease Sal I for which the specific PCR product possesses an internal recognition site (Eisenach et al., 1990). Before analysis, PCR products from the historical bone samples were accumulated by an additional amplification reaction. Reaction mix and cycling parameters were nearly the same as in the second stage of biphasic amplification, with 4 pmol of each primer, 1·0 U Taq-Polymerase (Perkin Elmer), 2 ìl PCR product as template, and only 25 cycles of amplification. Incubation with Sal I was carried out for 1 h at 37)C with 17 ìl accumulated PCR product, 2 ìl incubation buffer H (Boehringer), and 10 U Sal I (Boehringer). After digestion with Sal I PCR products were run for 24 h at 130 V on a 15% (w/v) polyacrylamide gel, stained with ethidium bromide and photographed under UV light.

Results The 123 bp M. tuberculosis complex-specific DNA sequence could be amplified from all three specimens (Figure 2). Success rates were approximately equal for all three individuals; on the whole they were 33·7% for DNA from affected bone tissue (N=104 amplification reactions) and 25·0% for DNA from normal bone tissue which was not itself affected (N=48 amplification reactions). The success rates include reactions which were done with DNA amounts differing from those given above. When working only with DNA amounts as given above success rates are expected to be higher. Incubation of PCR products with the restriction endonuclease Sal I produced the expected fragments of 53 bp and 66 bp with cohesive ends of 4 bp (Figure 3). All control samples revealed the expected results, indicating correct PCR-conditions (Figure 2). DNA (300 fg) from M. tuberculosis or M. bovis which equals approximately 100 bacteria (Thierry et al., 1990b) revealed positive results, exhibiting the specific 123 bpbands, even after addition of a 500,000-fold surplus of human DNA. Amplification reactions with human DNA alone gave negative results as it was expected. Product sizes clearly ranging above 123 bp which is unspecific to the M. tuberculosis complex were amplified from DNA extracts of the soil-inhabiting species M. gordonae and M. aurum. Reactions with DNA extracted from soil gave no amplification products at all.

Discussion We have been able to detect a M. tuberculosis complexspecific aDNA sequence in three historical bone samples by using the Polymerase chain reaction. The reliability of these PCR results was ensured by

Mol. wt. LS V

DNA from soil

M. aurum

M. gordonae

DNA from 13 K 25:1 pathological samples 13 K 37:1

13 K 37:2

Human DNA

Human DNA + M. bovis

M. bovis

Human DNA + M. tuberculosis

M. tuberculosis

No template

Mol. wt. LS V

670 H. Baron, S. Hummel and B. Herrmann

267 bp 123 bp

Mol. wt.

DNA from pathological sample, Sal I

DNA from pathological sample

DNA from pathological sample, Sal I

DNA from pathological sample

M. bovis, Sal I

M. bovis

Mol. wt.

Figure 2. Results of amplification reactions for detection of M. tuberculosis complex DNA. PCR products (20 ìl each) were run for 45 min at 100 V on a 4% agarose gel, stained with ethidium bromide, and photographed under UV light.

123/124 bp 80 bp 64 bp 57 bp 51 bp

Figure 3. Results of the digestion of PCR products with the restriction endonuclease Sal I. Alternating each 20 ìl of the undigested and the digested PCR products were run for 24 h at 130 V on a 15% polyacrylamide gel, stained with ethidium bromide, and photographed under UV light.

restriction enzyme analysis and by a system of control reactions. The success rate of 33·7% respectively 25·0% should mainly be traced back to stochastic matters: not every reaction might have contained target DNA since its amount in the DNA extract is assumed to be very small. In addition, in Booster-PCRs with very low concentrations of primers and target DNA the amplification process become stochastic rather than deterministic (Ruano & Kidd, 1989). Both explanations are especially likely for amplification reactions carried out

with ancient DNA where not all targets are left intact (Golenberg, 1991; Eglington & Logan, 1991). A higher success rate might be achieved, for example, by lowering stringency parameters in the first few cycles to promote template-jumping (Pääbo, Irwin & Wilson, 1990). Since lowering the stringency parameters also promotes formation of unspecific products, it should be applied only when combined with the use of nested primers. An important part of our study was to test if our PCR assay was specific for the M. tuberculosis complex and could thus be applied to skeletal remains excavated from soil. Eisenach et al. (1990) have reported amplification of a 123 bp PCR product from one M. simiae strain with the IS6110 primer pair. Apart from this we did not find any experimental proof from literature that specific PCR products can be obtained with DNA of other mycobacterial species and the IS6110 primers. In our study, DNA from M. gordonae, M. aurum and from soil did not give any specific amplification product. Nevertheless the high molecular PCR products obtained with DNA from M. gordonae and M. aurum show that the genomes of mycobacteria not belonging to the M. tuberculosis complex may contain sequences with high homology to M. tuberculosis complex-specific primers. In conclusion, we assume that it is generally possible to apply our assay to skeletal remains excavated from soil; however, DNA of the surrounding soil should be extracted and submitted to PCR as a control. An important result was the detection of M. tuberculosis complex-specific DNA in bone tissue without any tuberculous lesions. The amount of infectious agents carried into bone with the bloodstream seemed to be sufficient for PCR. Therefore we expect that it is possible to detect any infectious agent in bone which is retrievable in blood even if it is not able to cause

Mycobacterium tuberculosis Complex DNA in Ancient Human Bones 671

lesions in bone tissue. Hence, Polymerase chain reaction might raise the possibility to detect infectious diseases like plague in ancient skeletal remains. The reliability and reproducibility of PCR-based detection of M. tuberculosis have been questioned because of considerable variations in the sensitivity and specificity among different laboratories (Noordhoek, van Embden & Kolk, 1993). In the pilot study presented here we developed a system of control reactions to check the sensitivity and specificity of our assay and to identify false-positive results when applying the assay to large numbers of unknown ancient specimens. However, complete exclusion of false-negative results is only possible when working with known specimens, as was done in our study. The use of our method for individual diagnosis is limited by the possibility of false-negative results and the impossibility to distinguish between silent M. tuberculosis infection and acute disease which does not necessarily follow infection. The potential of our method lies in the estimation of infection rates in different ancient populations and their linkage to information about the living conditions. In this way, PCR could open up new perspectives for paleopathology and paleoepidemiology of infectious diseases.

Acknowledgement This work was supported by the Bundesminister für Forschung und Technologie.

References Dalton, H. P., Archer, G. L., Slifkin, M., Harris, R. C., Welshimer, H. J., Sosnowski, K. M., Nottebart, H. C., Jr., Duma, R. J., Clark, R. B., Warren, N. G., Kerkering, T. M., Swenson, P. D., Ager, A. L., Jr. & May, R. G., Jr. (1986). Blood specimens. In (H. P. Dalton & H. C. Nottebart, Jr., Eds) Interpretive Medical Microbiology. London: Churchill Livingstone, pp. 4–187. Dixon, R., Ensor, S., Lewis, E. & Roberts, C. (1994). The detection of Mycobacterium tuberculosis by PCR from ancient human bone. Poster Presentation on the Xth European Meeting of the Paleopathology Association, Göttingen 1994. Eglington, G. & Logan, G. A. (1991). Molecular preservation. Philosophical Transactions of the Royal Society of London B 333, 315–328. Eisenach, K. D., Cave, M. D., Bates, J. H. & Crawford, J. T. (1990). Polymerase chain reaction amplification of a repetitive DNA sequence specific for Mycobacterium tuberculosis. The Journal of Infectious Diseases 161, 977–981. Golenberg, E. M. (1991). Amplification and analysis of Miocene plant fossil DNA. Philosophical Transactions of the Royal Society of London B 333, 419–427. Herrmann, B. & Hummel, S. (Eds) (1993). Ancient DNA. Recovery and Analysis of Genetic Material from Paleontological, Archeological, Museum, Medical, and Forensic Specimens. New York: Springer Verlag. Herrmann, B., Grupe, G., Hummel, S., Piepenbrink, H. & Schutkowski, H. (1990). Prähistorische Anthropologie. Heidelberg: Springer Verlag, pp. 10–14.

Hummel, S. (1992). Nachweis spezifisch Y-chromosomaler DNASequenzen aus menschlichem bodengelagerten Skelettmaterial unter Anwendung der Polymerase Chain Reaction. Ph.D. Thesis, University of Göttingen, Germany. Hummel, S. & Herrmann, B. (1993a). General aspects of sample preparation. In (B. Herrmann & S. Hummel, Eds) Ancient DNA Recovery and Analysis of Genetic Material from Paleontological, Archeological, Museum, Medical, and Forensiz Specimens. New York: Springer Verlag, pp. 59–68. Hummel, S. & Herrmann, B. (1993b). Y-chromosomal DNA from ancient bones. In (B. Herrmann & S. Hummel, Eds) Ancient DNA. Recovery and Analysis of Genetic Material from Paleontological, Archeological, Museum, Medical, and Forensic Specimens. New York: Springer Verlag, pp. 205–210. Hummel, S., Nordsiek, G. & Herrmann, B. (1992). Improved efficiency in amplification of ancient DNA and its sequence analysis. Naturwissenschaften 79, 359–360. Manchester, K. (1991). Tuberculosis and leprosy: Evidence for interaction of disease. In (D. J. Ortner & A. C. Aufderheide, Eds) Human Paleopathology. Washington, DC: Smithsonian Institution Press, pp. 23–35. Mullis, K. B. & Faloona, F. A. (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed chain reaction. Methods in Enzymology 155, 335–350. Noordhoek, G. T., van Embden, J. D. A. & Kolk, A. H. J. (1993). Questionable reliability of the polymerase chain reaction in the detection of Mycobacterium tuberculosis. The New England Journal of Medicine 329, 2036. Ortner, D. J. & Putschar, W. G. J. (1985). Identification of Pathological Conditions in Human Skeletal Remains. Washington, DC: Smithsonian Institution Press, pp. 144. Ou, C.-Y., Moore, J. L. & Schochetman, G. (1991). Use of UV irradiation to reduce false positivity in polymerase chain reaction. BioTechniques 10, 442–444. Pääbo, S., Irwin, D. M. & Wilson, A. C. (1990). DNA damage promotes jumping between templates during enzymatic amplification. The Journal of Biological Chemistry 265, 4718–4721. Ruano, G. & Kidd, K. K. (1989). Booster PCR: a biphasic paradigm for amplification of a few molecules of target. Amplifications 3, 12–13. Ruano, G., Fenton, W. & Kidd, K. K. (1989). Biphasic amplification of very dilute samples via ‘‘booster PCR’’. Nucleic Acids Research 17, 5407. Salo, W. L., Aufderheide, A., Buikstra, J. & Holcomb, T. A. (1994). Identification of Mycobacterium tuberculosis DNA in a preColumbian Peruvian Mummy. Proceedings of the National Academy of Sciences USA 91, 2091–2094. Spigelman, M. & Lemma, E. (1993). The use of the polymerase chain reaction to detect Mycobacterium tuberculosis in ancient skeletons. International Journal of Osteoarcheology 3, 137–143. Thierry, D., Cave, M. D., Eisenach, K. D., Crawford, J. T., Bates, J. H., Gicquel, B. & Guesdon, J. L. (1990a). IS6110, an IS-like element of Mycobacterium tuberculosis complex. Nucleic Acids Research 18, 188. Thierry, D., Brisson-Noël, A., Vincent-Lévy-Frébault, V., Nguyen, S., Guesdon, J. L. & Gicquel, B. (1990b). Characterization of a Mycobacterium tuberculosis insertion sequence, IS6110, and its application in diagnosis. Journal of Clinical Microbiology 28, 2668–2673. Tsukamura, M. (1983). Numerical classification of 280 strains of slowly growing mycobacteria. Microbiology and Immunology 27, 315–334. Tsukamura, M. (1984). The ‘‘non-pathogenic’’ species of mycobacteria: their distribution and ecology in non-living reservoirs. In (G. P. Kubica and L. G. Wayne, Eds) The Mycobacteria: A Sourcebook. Part B. New York: Marcel Dekker, pp. 1339–1359.