In this study, we describe an infant mummy from ancient Egypt that showed macromorphologic signs of chronic anemia and vitamin C deficiency. From this infant, we have obtained a sterile sample from a metatarsal bone to extract ancient bacterial DNA. Following polymerase chain reaction amplification and subcloning of the amplicons, the sequence of the 16S ribosomal DNA was determined in several resulting clones. The presence of pathogenic and apathogenic bacteria, such as Escherichia coli, are indicated by our result, providing evidence of bacteremia, which probably contributed to death due to septicemia. These findings suggest that the infant, who already had chronic anemia and vitamin C deficiency, acquired a gastrointestinal infection, which finally led to a systemic spread. To our knowledge, this is the first case identifying potentially septicemic bacterial dissemination in an ancient Egyptian mummy. Using our approach, we hope to investigate distinct paleomicrobiological aspects of ancient populations, which will potentially enlighten our understanding of the development and evolution of pathogenic bacteria.

Even in present-day populations, septicemia represents a serious complication of the neonatal period, resulting in a high incidence of lethality. Infections of the respiratory and gastrointestinal tracts are the 2 most common causes of septicemia. The most frequent pathogens are staphylococci, streptococci, and Escherichia coli.¹

The recent development of DNA amplification strategies using polymerase chain reaction (PCR) offers a unique opportunity to investigate the occurrence and frequency of infectious diseases of past populations. Thus, in previous populations, several infectious diseases have been analyzed successfully by identifying bacterial DNA, eg, tuberculosis,^2–5 leprosy,^6,7 malaria,⁸ Chagas disease,⁹ and plague (Yersinia pestis¹⁰).

Recently, with ancient DNA (aDNA) from bone and soft tissue samples of various ancient Egyptian mummies, we have used DNA amplification, followed by subcloning of the amplicons, to sequence species-specific segments of the bacterial 16S ribosomal DNA (rDNA). Using this approach, we were able to identify various bacterial species in bone and soft tissues, eg, Mycobacterium tuberculosis¹¹ and corynebacteria.¹²

In the present report, we describe our findings of an infant mummy, which showed signs of chronic anemia and vitamin C deficiency, indicating poor living conditions. In this case, the additional identification of bacterial DNA suggests septicemia, with pathogenic bacteria presumably originating from the gastrointestinal tract. To our knowledge, this is the first reported case of a paleomicrobiological investigation of ancient bacteremia in a child.

The infant mummy was found during an excavation of the Deutsche Archaeologische Institut, Cairo, Egypt, by Elina Grothe-Paulin, MA, in one of the Tombs of the Nobles from the necropolis of Thebes-West (designated TT-84). The burial site was built by the vizier Jamunedjeh during the reign of Pharaoh Amenophis II (1424–1398 BC). In the following periods, this tomb had been used for several intrusive burials. The completely wrapped mummified infant under investigation was detected in the upper level of the tomb chamber (TT-453). According to the archaeologic context and surrounding findings, the mummy could be dated to a period between 1000 and 700 BC (third intermediate period).

Several layers of linen bindings, which seemed to have been immersed only superficially by resin substances (Figure 1 ), covered the mummified infant. However, the skull was superficially destructed so that several skull bones were visible through the disrupted bindings. However, the other body parts were not visible. We decided to carefully unwrap the body under sterile conditions. During this procedure, a complete metatarsal bone, which was fully enveloped by the bindings and was not accessible directly from the superficial layers, was recovered and immediately transferred into a sterile cup. During the unwrapping procedure, a small sample of the linen bindings from the middle layer was also removed under sterile conditions. Similar samples were obtained from other mummy specimens found close to the infant. No soft tissues of the skin, musculature, or internal organs had survived; only osseous material was found, presenting a fairly complete infant skeleton (Figure 2 ).

Anthropologic examination revealed an infant approximately 18 months of age (±3 months) as evidenced by the developmental status of dentition.¹³ Sex could not be determined.

Careful inspection of the bones revealed distinct pathologic abnormalities. There was significant pitting of the orbital roofs as seen in the characteristic pattern of cribra orbitalia (Figure 3 ). This phenomenon is known to occur in chronic anemia.¹⁴ In addition, the inner table of the skull bones showed focal irregularities of the bony surface, with reactive new bone formation, which also suggests chronic anemic conditions (Figure 4 ). Furthermore, both distal femora and humeri revealed areas with minor superficial new bone formation as seen as the sequelae of subperiosteal bleeding.¹⁵

Using strictly sterile laboratory conditions, the surface of the bone sample was cleaned with 0.5% sodium hypochlorite solution and then mechanically removed to avoid surface contamination. The spongy bone material was then used for DNA extraction.

DNA purification was performed as follows. A 100-μL volume of 0.1-mm zirconium beads (Biospec Products, Bartlesville, Okla) in Tris-EDTA (pH 7.4) was added to the sample, and the residual supernatant fluid was removed after the beads settled. Then 100 μL of Tris-EDTA-NaCl (pH 8.0) and 50 μL of phenol-chloroform-isoamyl alcohol (25:24:1) were added to the tube. The tube was then shaken vigorously for 1 minute in a mechanical disrupter (MiniBead Beater model 3110, Biospec Products).

After centrifugation for 5 minutes, the aqueous phase was collected and mixed on a vortex mixer with an equal volume of chloroform-isoamyl alcohol (24:1). After another brief centrifugation, the upper phase was collected and boiled for 10 minutes, and DNA was precipitated with 0.1 volumes of 3 mol/L sodium acetate and 2 volumes of cold absolute ethanol. Finally, after centrifugation at 12000g for 20 minutes, the precipitated DNA was dried and resuspended in 10 μL of distilled water and used for PCR.

The DNA preparation was subjected to PCR amplification using different sets of well-established in-house consensus primer pairs. These recognize conserved regions of the eubacterial 16S ribosomal RNA gene but flank hypervariable, species-specific regions.¹⁶ The PCR reaction mixtures were electrophoretically separated and analyzed for the presence of specific amplification products.¹⁷

Applying broad-range bacterial primer pairs to the ancient samples, where the presence of different bacterial species can be expected, only the predominant species will be detected by sequence analysis of the amplicons. To receive further information on the bacterial composition, we decided to clone the PCR amplicon mixtures (TA cloning kit, Invitrogen BV, Groningen, The Netherlands) and to sequence 20 randomly selected clones from each experiment. Specifically, we wanted to detect and identify distinct species that are present in a small number within all other bacterial species.

The individual nucleotide sequences of cloned amplicons were determined by direct cycle sequencing of both strands of the plasmid preparations using an automated ABI 373A DNA sequencer. Computer-assisted comparison of the determined sequences with all GenBank entries and in-house eubacterial 16S rDNA databank revealed scores of homology that led to a reliable list of candidates.

Using our protocol of aDNA extraction and purification, we succeeded in receiving aDNA from the bone sample. Interestingly, the resulting aDNA obviously proved to be of very good condition, since we were able to amplify 369–base pair fragments of the 16S rDNA, which is well in line with previous observations that indicate that even larger fragments of aDNA may be intact.^4,6,11 The PCRs performed repeatedly revealed the same results. Thereby, several clones could be isolated from the bone sample containing bacterial DNA. The sequencing of the 16S rDNA allowed a distinction among different bacterial species on the basis of the hypervariable region, which is flanked from conserved eubacterial sequences. A comparison of the sequences obtained from the amplified material revealed a complete homology to modern bacterial sequences.

Thereby, the following bacteria could be identified: E coli (Figure 5 ), Frateuria auranta, and Halobacillus spp. Further tissue samples from mummies found adjacent to the infant mummy either did not show amplifiable aDNA (1 sample from an eye, 2 samples from retroperitoneal tissue of a mummy torso, probably kidney tissue) or provided apathogenic bacterial DNA (2 samples from gingival tissue, showing Halobacillus spp, Sporosarcina spp, and Bacillus panthothenicus). Test results of blank controls and the control sample from the linen bindings were negative.

There is little evidence from written sources that infectious diseases were present in ancient Egypt.¹⁸ Nevertheless, previous paleopathologic analyses provided data that indicated that bacterial infections have occurred in Thebes, the capital of ancient Egypt. Thus, to date, we have identified several individuals who had tuberculosis using molecular analysis of mycobacterial DNA.^4,11 Recently, we were successful in identifying ancient DNA from other bacteria, in particular from corynebacteria, suggesting, together with written evidence, that diphtheria was possibly present in ancient Egypt.¹² This latter study used the isolation of ancient DNA, its amplification by PCR, and subsequent bacterial selection using cloning strategies. Interestingly, several samples revealed fragments of aDNA that were large enough to allow successful sequencing of the 16S rDNA.^4,6,11,12 This novel strategy was now applied to a bone sample taken from an approximately 18-month-old mummy. The analysis of our data strongly suggests that our infant must have had a bacterial infection with systemic spread of the pathogenic bacteria. Recently, Fricker et al¹⁹ were able to identify E coli DNA in upper gut contents of the Lindow Man, an Iron Age bog body dating to approximately 300 BC. In addition, the presence of E coli DNA in our case, but not in several control cases buried nearby, merely excludes the possibility that postmortal spread of E coli may have occurred in those bodies.

Besides DNA from E coli, we identified 2 apathogenic bacteria that may be involved in the tissue decomposition during the decay process. Since our approach of cloning of amplified bacterial DNA excludes any quantitative measurement of the amounts of those DNAs, we cannot speculate on the amounts of E coli DNA versus the other bacterial aDNAs in our tissue sample. However, we have no data on any bacterial content of other samples or tissues or on the presence of bacteria on the bone surface. We, therefore, cannot exclude that the extensive decomposition of parosteal soft tissues associated with bacterial growth could have dislodged E coli into the bone, although there is no evidence that E coli is involved in tissue decay itself. The small sample removed from between the linen bindings (middle layer of bindings) did not contain amplifiable bacterial aDNA, which also excludes just a simple (postmortal) spread of bacterial DNA.

Further paleopathologic observations indicate that the infant had chronic anemia and vitamin C deficiency. These pathologic abnormalities were diagnosed from the presence of cribra orbitalia and porosis of the internal skull bones, which are typically, although not exclusively, seen in chronic anemia.¹⁴ In addition, the minor superficial new bone formation at the metaphyseal areas of the long bones can be regarded as the result of subperiosteal bleeding due to chronic vitamin C deficiency (scurvy¹⁵). These observations indicate that the child experienced poor living conditions.

Because of these reasons, it is conceivable that this child may have been easily susceptible to infectious diseases. The detection of bacterial DNA from E coli in a metatarsal bone clearly indicates that these bacteria must have spread out by bacteremia. Although we can only speculate on the final course of life of this infant, it is fair to assume that the spread of E coli by the blood stream may have been the ultimate cause of death from septicemic shock.

In this regard, it is critical to exclude contamination by recent bacterial DNA. Although contamination may principally have occurred during the excavation, handling of specimens, or laboratory analysis, in our case this risk was kept as low as possible. The mummy had been completely wrapped in linen bindings (except for the skull, where the linen bindings were superficially disrupted; however, the metatarsal bones were firmly packed in bindings), and we strictly applied sterile conditions during the removal of the small sample from the mummy before it was opened. In addition, a parallel analysis of further tissue samples from mummies near the infant mummy did not show either amplifiable bacterial aDNA or apathogenic bacterial DNA. Finally, all laboratory analyses were performed under sterile conditions, including several controls.

Recent data indicate that septicemia is mainly induced by staphylococci, streptococci, or E coli.¹ Most often these enter the blood stream by an infection of the (upper and lower) respiratory, gastrointestinal, or urogenital tracts. In this regard, it seems to be of particular interest that infant septicemia most commonly results from respiratory or gastrointestinal infections, and the bacteria identified in this study are also often seen in present-day cases of infant septicemia.^20–22

As a further result, we observed complete homology of the analyzed gene segment with the sequence known from recent E coli. This had initially been an unexpected observation, since the high replication rate of the bacteria and the long interval of approximately 3000 years would have suggested evolutionary sequence differences. However, since the gene segment chosen is highly conserved for the respective bacterial DNA, any significant change of the sequence may have lead to a new (and potentially nonpathogenic) bacterial strain or the affected bacterial strain would have been extinct. Thereby, the conservation of this gene sequence for thousands of years may have been a crucial point for the survival of the pathogen.

In summary, this is the first report, to our knowledge, that identifies E coli in an ancient Egyptian mummy. Furthermore, our data indicate that the affected infant, who already had chronic anemia and vitamin C deficiency and thus was more susceptible to infectious diseases, presumably underwent a gastrointestinal infection with final bacteremia and septicemia. We also provide circumstantial evidence that the paleomicrobiological analysis of ancient DNA may uncover diseases of historic populations and that this approach may open the way to investigate more closely the occurrence, frequency, and spread of infectious diseases of ancient times.

Click on thumbnail for full-sized image.

Click on thumbnail for full-sized image.