The detection of ancient microbial DNA offers a new approach for the study of infectious diseases, their occurrence, frequency and host–pathogen interaction in historic times and populations. Moreover, data obtained from skeletal and mummified tissue may represent an important completion of contemporary phylogenetic analyses of pathogens. In the last few years, a variety of bacterial, protozoal and viral infections have been detected in ancient tissue samples by amplification and characterization of specific DNA fragments. This holds particularly true for the identification of the Mycobacterium tuberculosis complex, which seems to be more robust than other microbes due to its waxy, hydrophobic and lipid-rich cell wall. These observations provided useful information about the occurrence, but also the frequency of tuberculosis in former populations. Moreover, these studies suggest new evolutionary models and indicate the route of transmission between human and animals. Until now, other pathogens, such as Mycobacterium leprae, Yersinia pestis, Plasmodium falciparum and others, have occasionally been identified – mostly in single case studies or small sample sizes – as well, although much less information is available on these pathogens in ancient settings. The main reason therefore seems to be the degradation and modification of ancient DNA by progressive oxidative damage. Furthermore, the constant risk of contamination by recent DNA forces to take time and cost effective measures and renders the analysis of ancient microbes difficult. Nevertheless, the study of microbial ancient DNA significantly contributes to the understanding of transmission and spread of infectious diseases, and potentially to the evolution and phylogenetic pathways of pathogens.
The molecular detection of microbial infections in ancient human and animal remains represents an emerging field of research. This research developed rapidly within the last few years, from isolated reports on diagnosis of infectious diseases to much more profound studies on disease frequencies in ancient populations and evolutionary aspects of host–pathogen interaction. In the beginning of this new scientific branch modern molecular methods, like PCR-based amplification of specific DNA fragments, were used to amplify ancient DNA (aDNA) residues in human remains to provide evidence for a presumed infectious disease. The first successful studies were performed in the 1990s on a 1000 year old South American mummy and an Egyptian mummy from the New Kingdom (1550–1080), in which Mycobacterium tuberculosis DNA was detected, leading to the confirmation of a tuberculosis infection in these mummies. In the following years, the diagnostic ability of this approach was widened by the detection of several other microbes including Mycobacterium leprae, Yersinia pestis, Plasmodium falciparum, Trypanosoma cruzi, Treponema pallidum, Escherichia coli and Corynebacterium spp.. These studies were mostly based on single findings or on small series of selected samples. In more recent studies on the molecular detection of M. tuberculosis complex DNA, larger series were analyzed including specimens with non-specific or even without morphological alterations, suggesting infection with tuberculosis. Beside the benefit of a more precise diagnosis in ancient material, this approach paved the way to investigate both the frequency and the distribution of infectious diseases in various ancient populations. This new approach, recently termed ‘molecular paleoepidemiology’, rises above the satisfaction of a purely historic interest, and offers the possibility to compare modern data on tuberculosis epidemiology with the situation prevalent in previous times. In particular, direct insights into the evolution of pathogens and the interaction with their hosts may be obtained which are crucial to understand the transmission and spread of particular infectious diseases. Nevertheless, the extensive degradation and chemical modification of double stranded DNA in ancient samples clearly reduces the analytical potential of this approach and engenders the risk of producing non-specific or even false positive results. The usually low amount of authentic DNA in samples of up to several thousands years of age significantly increases the susceptibility of the technique to contamination with modern DNA. Therefore strict precautions to avoid contamination have to be imposed and there exists a general consensus on the clear necessity of strictly controlled conditions by all researchers working with aDNA.
2Molecular identification of pathogens
2.1The analysis of tuberculosis
Most of the molecular research on ancient microbial infections has been performed on the Mycobacterium tuberculosis complex (M. tuberculosis, Mycobacterium bovis, Mycobacterium africanum, Mycobacterium microti, Mycobacterium canettii), which causes human and animal tuberculosis. It can be assumed that mycobacteria are better preserved than other bacteria due to their special cell wall structure. They differ from most other bacteria due to the high content of lipids in the cell wall, which renders them highly resistant against environmental destruction. Since tuberculosis involves the skeletal system to a considerable extent and leads to characteristic macromorphological alterations (such as the severe angular kyphosis called gibbus), this disease can be identified with some certainty even in skeletal material (which represents by far the most abundant human biomaterial from historic populations). In addition, in mummified human remains, adhesions of the lung to the chest wall can provide evidence for pulmonary tuberculosis (Fig. 1). Such alterations are easily detected in mummies and represented the starting point for further molecular investigations. In most studies, a 123-bp fragment of the insertion sequence IS6110 was amplified by PCR and the specificity of the reaction was subsequently confirmed by direct sequencing or restriction enzyme digestion (Fig. 2). Occasionally, the specificity of IS6110 has been questioned, but in several studies it has been shown that the insertion sequence is highly specific for the detection of M. tuberculosis complex DNA, combined with an adequate sensitivity. Thereby, this molecular approach meets the requirements to study successfully ancient DNA, since the aDNA is degraded and therefore only small fragments of DNA can be amplified.
The molecular confirmation of supposed tuberculosis infections in human remains was complemented by more recent studies on material with non-specific or without morphological alterations (Fig. 3). This offers potential insight into the frequency of tuberculosis in ancient populations and generally allows evaluating the impact of infectious diseases on historic populations. In particular, the high mortality rate of young adults observed in a wide range of studies on e.g. ancient Egyptian, medieval and later populations may be explained by high incidences of infectious diseases, such as tuberculosis.
A further important point in analyzing ancient mycobacterial DNA is the differentiation of sub-types of the M. tuberculosis complex. First, studies using the spoligotyping technique have been performed successfully on ancient tissue samples. This technique is based on the variation of the DR (direct repeat) region in M. tuberculosis complex members and allows differentiation of strains up to a sub-species level. Recently, this approach has widely been used in diagnostic medical microbiology for the initial genotyping of the M. tuberculosis complex at a populational level. Moreover, spoligotyping seems to be the most suitable method for ancient material, since only small amounts of an even highly fragmented mycobacterial aDNA can be retrieved in the samples under investigation. Other recent methods, such as IS6110 RFLP or VNTR require high molecular bacterial DNA and are therefore not applicable to ancient tissue material. Spoligotyping can be used for investigating evolutionary aspects of human tuberculosis and may help to clarify the origin and transmission of the disease in humans of various time periods and populations. In combination with other data it is suitable for constructing phylogenetic trees drawing back to the beginning of pathogenesis and spread of the disease in humans and animals.
Currently, there is still an open debate regarding the origins of tuberculosis in human and animal species. The classical hypothesis suggests that M. bovis is the probable ancestor which was transmitted from cattle to humans during domestication. This theory, however, is not supported by the as yet published spoligotyping results on ancient material from different time periods. In all studies, spoligotyping revealed no typical M. bovis type. In different studies on medieval bone samples [16,17], on bone samples ranging from 17 400 years of age to those of barely a hundred years and on naturally mummified Hungarians from the 18th to the 19th century, the patterns were found to be very similar to present day M. tuberculosis isolates. The as yet oldest successful spoligotyping pattern has been obtained from an isolated fossilized bison sample, dating back to ca. 17 000 years before present. Interestingly, in this study the spoligotyping pattern, which does not match with any known spoligotyping signatures, was different from M. bovis patterns. In our own recent study on ancient Egyptian material with molecularly proven tuberculosis, dating back to 4000 years before present, no M. bovis specific pattern was detected by spoligotyping. In contrast, M. africanum and M. tuberculosis specific patterns could be obtained while, quite interestingly, the M. africanum findings were detected in the oldest samples included in this study (Middle Kingdom, 2050–1650 BC). This supports the theory that a M. tuberculosis complex precursor evolved from M. africanum and that the present day M. tuberculosis and M. bovis may have developed in parallel. This theory is also substantiated by nucleotide sequence analyses of current M. tuberculosis isolates, revealing the absence of allelic variation. Therefore, the evolutionary origin of human pathogenic mycobacteria was suggested to be 15 000–20 000 years ago.
This clearly demonstrates that only the investigation of ancient biomaterial can provide a direct insight into the situation up to several thousand years ago. With this approach it will be possible to evaluate evolutionary models which are usually developed on data obtained from recent strains by comparing their frequencies, the occurrence of certain polymorphisms and their genetic differences. Since such statistical models are restricted to present day data and cannot take into account possible particular events of molecular evolution (such as bottlenecks, environmental disasters etc.), only the investigation of authentic historic material can gain access to the real story of molecular evolution and the development and spread of mycobacterial disease.
2.2Identification of other bacterial infections
The molecular identification of other pathogens such as M. leprae, Y. pestis, P. falciparum and others has also been performed successfully on ancient skeletal and mummified soft tissue material. They were, however, mainly based on isolated findings or small series. Nevertheless, these studies demonstrate the capability of molecular methods to detect pathogens even without having a morphological pendant. This is particularly true for the detection of Y. pestis in skeletons excavated from 16th and 18th century French mass graves of probable victims of the plague. The successful amplification of Y. pestis DNA in six out of 12 plague victim teeth confirmed the diagnosis of ancient septicemia, which had been assumed on the basis of historical data. This molecular approach represents the only possibility to prove the ancient infection, as manifestations of bone lesions usually do not occur in septicemic plague.
In addition to the application to specific primer sequences for the pointed detection of infectious organisms, the amplification of a segment of the eubacterial 16S rDNA with primer pairs recognizing conserved regions but flanking hypervariable species-specific regions allows the identification of unknown bacteria in a given sample (‘broad-range’ eubacterial PCR). This neat experimental strategy for pathogen identification is methodically restricted to the detection of predominant species within a complex mixture of bacterial organisms usually observed with ancient samples.
To overcome this limitation, PCR amplicons obtained with broad-range 16S rDNA primers from a mixture of bacterial species are cloned into appropriate vectors and at least 20 randomly selected clones from each experiment are sequenced. The comparison of the obtained sequences with all GenBank entries and other eubacterial 16S rDNA databases reveals scores of homology leading to a list of candidate bacteria (Fig. 4). This approach allows the determination of a variety of bacterial species simultaneously present in ancient tissue samples without targeting a single species. Thereby, pathogens or potential pathogenic bacteria may be detected in individuals without morphological evidence for a certain disease. Recently, we have applied this technique successfully in two case studies on an infant bone and a tooth sample of two different mummies from pharaonic Egypt. The analysis of the 16S rDNA revealed a probable septicemic infection by E. coli in the infant mummy and provided evidence for Corynebacterium spp. in the tooth sample, suggesting the presence of diphtheria in pharaonic Egypt.
Besides the identification of bacteria, these more general approaches can also be applied to detect protozoal infections. Likewise, malaria infections caused by one of the four different human pathogen Plasmodium species (P. falciparum, Plasmodium vivax, Plasmodium malarie and Plasmodium ovale) can be identified. This diagnosis may be observed particularly in those individuals with evidence for chronic anemia which in turn manifests itself in the skeleton as porotic hyperostosis of the orbital vault (‘cribra orbitalia’) and/or of the parietal skull bones. In fact, porotic hyperostosis as a bone reaction to chronic anemia may occur during many different pathological conditions, such as malnutrition, iron deficiency, other infections, etc. A clear diagnosis of malaria infections is therefore only possible by molecular detection of ancient plasmodial DNA. The existing studies already demonstrated the capability of PCR based identification of plasmodia, but PCR systems need to be improved in order to provide a sufficiently powerful approach to evaluate ancient malaria.
2.4Detection of viral infections
There exist only a few reports on the detection of viral infections in ancient tissue specimens. The most spectacular case was the molecular identification of a previously unknown strain of human T-cell lymphotropic virus type I (HTLV-I) provirus DNA in an Andean mummy dated approximately 1500 years old. Sequence comparison to those in contemporary Andeans and Japanese provided evidence that HTLV-I was carried into the New World during the ancestral migrations of humans from Asia across the Bering Strait. However, this conclusion was in contrast to the most molecular phylogenetic analyses that indicate a more recent origin for HTLV-I within South America and therefore led to a controversial scientific debate. Although there is still a lack of extended and reliable studies on ancient viral infections, the capability of this approach was demonstrated by the successful amplification of RNA from the 1918 ‘Spanish’ influenza virus. The influenza RNA was isolated from two paraffin-embedded and one permafrost lung tissue sample from victims of the influenza pandemic, which killed over 20 million people in 1918 and 1919. The complete sequence of the hemagglutinin gene of the 1918 virus could be detected and subsequently used for phylogenetic analyses. It can be hoped that further molecular studies on ancient viral infections will help to determine the genetic basis for such an exceptional virulence.
3Limitations of ancient pathogen retrieval
The work with ancient DNA bears some major limitations, which reduces the success rate of detecting ancient pathogens. One major point is the stability of the aDNA. Normally, DNA degrades rapidly after death. Several parameters such as high temperature, UV radiation, humidity and a low pH value significantly accelerate this DNA decay. On the other hand, a dry and cool climate or a rapid desiccation of an organism due to artificial or natural mummification processes may slow down the degradation and increases the probability of ancient DNA retrieval. In any case it has to be taken into account that the amount of authentic DNA in ancient tissue samples may be very low or even may have disappeared completely. In most instances, the aDNA molecules are significantly fragmented and destabilized by deamination and depurination. These factors can severely disturb the PCR amplification, leading to non-specific products or incorrect sequences. Therefore, an adequate primer design for ancient DNA purposes is necessary. This can lead to difficulties in finding the optimal PCR system, as in most instances only small DNA fragments of not more than 200 bp should be targeted and at the same time the amplified fragment should provide a high specificity for the wanted pathogen. Most approaches of molecular clinical diagnostics do not fulfill these criteria and cannot directly be used for ancient pathogen retrieval.
Another major limitation is the risk of contaminating the specimens with DNA of recent origin. This is a major problem when working with human DNA, since human DNA is prevalent in the environment and every person in contact with the material is a probable source of contamination. The danger of contamination with modern bacterial DNA is far less, since the researcher usually is not a contamination source and most pathogens are not present ubiquitously in the environment, but require living organisms or particular conditions to survive. Nevertheless, cross-contamination and overflow of modern microbial DNA have always to be taken into account and should strictly be avoided. Therefore, ancient DNA analyses should always be performed in laboratories which are exclusively used for these purposes. The identity of PCR amplification products should always be verified by additional molecular methods, e.g. determination of the sequence or restriction enzyme digestion of the PCR products. With regard to the properties of the ancient DNA, cloning of the PCR products may be very helpful to distinguish non-specific or nonsense sequences from the true microbial targets. Moreover, overlying sequences of e.g. fungal or modern bacterial DNA may thereby be excluded.
Despite the reduced susceptibility to contamination in detecting ancient DNA of pathogens, all authenticity criteria necessary to determine ancient DNA sequences should be considered, to avoid as far as possible any false positive results. It is important to produce reliable data which are comparable to other studies.
4Conclusions and outlook
In the beginning of the search for ancient microbial infections, much work was focused on the detection of pathogens which produce typical morphological alterations in bone or soft tissues. Some interesting findings on single individuals or small series have been published in which a presumed diagnosis was confirmed by molecular methods. Nevertheless, this already demonstrated the far-reaching capacity of the detection of ancient microbial DNA. The technical development for particularly adopted extraction and amplification procedures provided the successful detection of different pathogens in skeletal and mummified material up to several thousand years of age. Most molecular investigations have been made on the M. tuberculosis complex, which seems to be more appropriate for analysis due to the protective cell wall properties of the mycobacteria. The studies on ancient tuberculosis infections are no longer restricted to the diagnosis of the infectious disease, but have presented interesting data about the occurrence and frequency of tuberculosis in ancient populations. The recent introduction of the spoligotyping technique in this field offered the possibility to differentiate between the members of the M. tuberculosis complex and therefore allowed to investigate evolutionary aspects of mycobacterial infections. Along with these evaluations theories are emerging as to the origin and transmission of the disease. This represents an important contribution to modern microbiological research as the comparison of recent mycobacterial species can only lead to a theoretical or statistical model without direct evidence from the past.
Theoretically, any fungal, bacterial or viral infection can probably be detected, depending primarily on the preservation state of the specimens and the availability of suitable PCR systems. The most limiting factor is the susceptibility of DNA to degradation and modification in ancient tissue samples influenced by several factors such as temperature, humidity, pH value, etc.
There exists a lot of historical and archaeological evidence of other infectious diseases, such as leprosy, plague, malaria and others, which were already detected in ancient specimens by the successful amplification of the corresponding pathogens. However, there is still a lack of extended studies to gain more information about the frequency, transmission and probable evolution of these pathogens. Keeping in mind the clinical and populational significance of these infectious diseases, this holds particularly true for malaria and increasingly for tuberculosis. The ongoing research in this relatively new scientific field has the potential to contribute significantly to a better understanding of host–pathogen interaction, transmission and spread of infectious diseases in historic and modern times.