Auto-fluorescence emitted from the cell residues preserved in human tissues of medieval Korean mummies

Abstract

As a significant association has been established between residual ancient DNA (aDNA) and histological preservation, the morphological identification or confirmation of preserved cell residue in ancient tissues would greatly facilitate aDNA studies and enhance the definitiveness of their conclusions. However, morphological differentiation of cell residue from other tissue structures has always been difficult, even for experienced histologists, due to the severe degradation of cells over long burial durations. In the present study, using a fluorescence microscopy equipped with a specific type of filter set (excitation filter, 510–550 nm; dichroic mirror, 570 nm; emission filter, ∼590 nm), we found that certain structures in well-preserved mummified tissues emitted auto-fluorescence. Those structures were actually cell residues (e.g. fragmented DNA), laser capture microdissection and Quantifiler kit analysis having shown that preservation of nuclear DNA correlates with auto-fluorescence emission in laser capture microdissection-captured areas. Detection of auto-fluorescence could be an effective means of identifying cell residues in ancient tissue, enabling selection of the well-preserved samples necessary in successful aDNA studies.

Introduction

The authenticity of DNA obtained from ancient tissues nowadays is subject to serious challenge, as many ancient DNA (aDNA) results have been proven to be contaminated with modern DNA (Marota & Rollo, 2002). Therefore, various guidelines such as Authenticity Criteria to Determine Ancient DNA Sequences (Hofreiter et al. 2001) and Criteria of Authenticity (Willerslev & Cooper, 2005) have been formulated to ensure that aDNA results are authentic. Cell residue determination in ancient samples is also a significant tool for anthropologists hoping to confirm their aDNA results by morphological techniques. Since the first reports of a significant association between histological preservation and aDNA survival in ancient tissues (Colson et al. 1997; Haynes et al. 2002), histological study confirming the presence of nuclear residue in those tissues (Zink et al. 2005) has corroborated that argument. Nonetheless, discerning cell residue from other structures remains difficult, even for experienced histologists, because cell morphologies in ancient samples are typically severely degraded over long burial durations.

In the present study, tissue samples from well-preserved mummies discovered in medieval Korean tombs (Shin et al. 2003; Chang et al. 2006; Lee et al. 2009) were studied using microscopic techniques. Interestingly, the results revealed a few morphological signs that might be correlated with the preservation of cell residue. As this information could prove useful to researchers in the morphological selection of ancient samples for aDNA studies, we report our findings here.

Materials and methods

Among a number of medieval mummies recently found in Korea, two mummies (discovered in Gangneung and Yangju) were chosen as the subjects of this study (Fig. 1). They were selected for the excellent preservation of histological structures, as reported in two of our previous studies (Shin et al. 2003; Lee et al. 2009). In the case of the Yangju mummy, radiocarbon dating calibrated the most probable date range as 1411 ± 42 AD (Shin et al. 2003). Historical records (e.g. the lineage book of the clan, inscriptions on tombstones, etc.) meanwhile showed that the Gangneung mummy was born in 1561 and died in 1622 at the age of 61 (Lee et al. 2009).

Fig. 1.

The mummies from which the samples used in this study were collected. (A) and (B) Yangju mummy. (B) Right lung sample collected during endoscopic examination. (C) Gangneung mummy.

Samples were collected from the lung (Yangju), kidney (Yangju) and rib cartilage (Gangneung) by researchers wearing sterilized gowns and gloves and using sterilized or disposable instruments. The samples were stored in sterilized containers at 4 °C. No researchers or personnel could access the samples without permission.

The samples were subsequently rehydrated in Ruffer’s solution (distilled water : absolute ethanol : 5% sodium carbonate, 5 : 3 : 2) for 72 h (Ruffer, 1921). After being fixed in 4% paraformaldehyde in phosphate-buffered saline (pH 7.2) and stored at 4 °C overnight, they were washed with phosphate-buffered saline (pH 7.2) and then embedded in OCT compound (Sakura, Torrance, CA, USA). The samples were then cut into 5- or 12-μm sections on a cryostat (Leica, Nussloch, Germany). To rule out possible DNA contamination, the knife holder, storage shelf and specimen disc in the cryochamber of the cryostat were cleaned with 0.5% (w/v) sodium hypochlorite prior to their use (Richards et al. 1995). The glass slides used were pre-treated with bleaching solution for 20 min, after which they were heated again in a dry oven for 8 h at 200 °C.

The tissue sections, unstained by any fluorescent dyes, were observed using the U-MWG2 filter set (Olympus, Japan; 510–550-nm band-pass excitation filter; 570-nm dichroic mirror; 590-nm emission filter) with the BX51 fluorescence microscope (Olympus, Tokyo, Japan). For discernment of the distribution of cells within the tissues (Deitch, 1964; Šilha, 1966; Erenpreisa, 1977), the sections were stained with 0.05% methylene blue in accordance with the method of Bancroft & Stevens (1996). In addition, Masson trichrome staining was performed for detection of collagen fibers remaining in our samples (Sheehan & Hrapchak, 1980). To determine the presence of cellular nuclei, the tissue sections were also stained with 1 μg mL⁻¹ of 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen, Carlsbad, CA, USA) for 20 min at room temperature (20 to 25°C). Subsequently, the samples were observed with the XF06 filter set (Omega Optical, Brattleboro, VT, USA; 315–415-nm band-pass excitation filter; 400-nm dichroic mirror; 385–515-nm emission filter). Some sections were also observed under the confocal microscope LSM 510 META (Carl Zeiss, Jena, Germany). The auto-fluorescence, DAPI and differential interference contrast images obtained by confocal microscopy were merged using the Zeiss LSM Image Browser (Carl Zeiss).

Next, to show the correlation of the DNA quantities with the auto-fluorescence emission, laser capture microdissection (LCM) (model 2110; Arcturus, Mountain View, CA, USA) using Autopix 100 software (Ver. 1.6.5.3; Arcturus) was performed. First, we tried to locate the co-localization of methylene blue staining and the auto-fluorescence-emitting areas by merging the two pertinent images with Adobe Photoshop CS3 (Adobe Systems, San Jose, CA, USA). Next, the intense or weak methylene-blue-stained areas were captured separately (total area captured, 3 mm²) by means of Capsure Macro LCM Caps (Arcturus). The laser settings ranged from 60 to 100 mW (power) and from 1500 to 2000 μs (pulse) with 275 mV (intensity).

DNA extraction from the captured specimens was performed with the PicoPure DNA Extraction Kit (Arcturs) according to the manufacturer’s instructions. We used the Quantifiler human DNA quantification kit (Applied Biosystems, Foster City, CA, USA) developed for quantification of amplifiable human telomerase reverse transcriptase DNA (amplicon length, 62 bases). Using it, the differences in the quantities of amplifiable nuclear DNA remaining in each microdissected area could be compared with each other. After Quantifiler™ human primer mix (10.5 μL), Quantifiler™ polymerase chain reaction (PCR) reaction mix (12.5 μL) and extracted DNA (2 μL) were mixed in a reaction well and the plate was centrifuged for 3 min at 300 g, after which it was positioned in the instrumental thermal block of an ABI Prism 7000 Sequence Detector (Applied Biosystems). PCR amplification was performed as follows: initial incubation for 10 min at 95 °C, followed by 40 cycles of denaturation for 15 s at 95 °C and annealing/extension for 1 min at 60 °C. After the run was complete, the data were analyzed by 7000 system SDS software version 1.2.3 (Applied Biosystems). LCM capturing and a Quantifiler analysis were repeated three times.

A densitometric analysis of the methylene blue staining was conducted for each area subjected to LCM. The mean densities [i.e. the averages of the gray values (maximum level, 256)] of all of the pixels within the measured area could be obtained using the Image program (Scion Corporation, Frederick, MD, USA). The statistical significances of the densitometric data for the methylene blue staining and the Quantifiler-determined amplifiable DNA quantities were analyzed by Student’s t-test. P-values < 0.01 were considered statistically significant.

To minimize the probability of modern DNA contamination in the Quantifiler analysis, we followed the aforementioned guidelines of Hofreiter et al. (2001) and Willerslev & Cooper (2005). Briefly, we always wore protective gloves, masks, gowns and caps when doing post-sampling laboratory work. Our aDNA laboratory facilities, moreover, had been designed in accordance with Hofreiter et al. (2001); the rooms for aDNA extraction or PCR preparation were physically separated from our main PCR laboratory and the DNA extraction/PCR preparation rooms were equipped with night UV irradiation, isolated ventilation and a laminar flow hood (http://user.chol.com/~drdoogi/Lab_Facilities.tif).

Transmission (TEM) or scanning (SEM) electron microscopic studies were also performed on the samples, in accordance with the previously described methods (Hayat, 1970; Bozzola & Russell, 1992). In the TEM study, the samples were immersed in 2% paraformaldehyde/2.5% glutaraldehyde in neutral 0.1 m phosphate buffer for 1 h. They were post-fixed for 1 h in 1% (w/v) osmic acid, dehydrated in graded ethanol and embedded in Epon812 (EMS, Fort Washington, PA, USA). Ultrathin sections were cut for mounting on nickel grids coated with Formvar film preparatory to post-uranyl-lead counter-staining examination under an H-7600 transmission electron microscope (Hitachi, Tokyo, Japan). In the SEM study, each sample was pre-fixed by 4% paraformaldehyde/0.1% glutaraldehyde in neutral 0.1 m phosphate buffer. After the samples were post-fixed for 2 h in 1% (w/v) osmic acid, they were treated in a graded ethanol series and isoamyl acetate and dried in an SCP-II critical point dryer (Hitachi). The samples were then observed under a JSM-840A scanning electron microscope (JEOL, Tokyo, Japan) after having been coated with gold using the JFC-1100 ion coater (JEOL).

Results

When we observed the sections under a fluorescence microscope equipped with the U-MWG2 filter set, intense auto-fluorescence was emitted from certain tissue structures (Fig. 2). When the same sections were observed with the alternative XF06 filter set, however, no auto-fluorescence was identified. However, by methylene blue staining, we found cell-like residues in every tissue that we examined (Fig. 3), which corresponded well to the fluorescence-emitting structures (Fig. 2). Judging from the results of the methylene blue staining, the auto-fluorescent structures in cartilage had to be chondrocytes because most of them were localized within the chondral lacunae.

Fig. 2.

Fluorescence microscopic examination of sections of lung (left column), kidney (middle column) and cartilage (right column) with filter set U-MWG2. The figures in the lower row are magnified images. Intense fluorescence is emitted from cell-like residues. Auto-fluorescence-emitting residues in kidney are arranged in rows, which look like renal tubules. In cartilage, auto-fluorescence is only detected within the lacunae. Scale bars: 300 μm (upper row) and 100 μm (lower row) for lung and kidney; 200 μm for cartilage.

Fig. 3.

Lung (upper row), kidney (middle row) and cartilage (lower row) sections stained by methylene blue. The figures in the right column are magnified images. Cell-like residues are found in every microscopic field that we have examined. Scale bars: 200 μm for lung and kidney; 100 μm for cartilage.

To determine if the auto-fluorescence-emitting structures were actually cell residues (e.g. fragmented DNA), we measured quantities of amplifiable nuclear DNA in each tissue area exhibiting different intensities of auto-fluorescence. First of all, we confirmed the areas showing methylene blue staining or auto-fluorescence to be in perfect correspondence (Fig. 4). Next, by use of LCM capturing and Quantifiler kit analyses, the amounts of amplifiable DNA in the methylene-blue-intense areas were estimated and these amounts were further compared with those in the methylene-blue-weak areas. As nuclear DNA was very successfully amplified in the areas exhibiting intense methylene blue staining, we speculated that preservation of nuclear DNA might correlate with methylene blue staining intensity as well as with auto-fluorescence emission in each LCM-captured area (Fig. 5; Table 1; Supplementary Fig. S1). As for the failure of DNA amplification with the cartilage samples, we suspect that the cause might have been the smaller amounts of DNA preserved within the chondral lacunae, which could not be easily amplified for 40 cycles in the Quantifiler analysis.

Fig. 4.

Auto-fluorescence (AF) and methylene blue staining images of lung, kidney and cartilage. Lower row contains merged images. The areas exhibiting methylene blue and AF were in perfect correspondence. Scale bars: 200 μm for lung and kidney; 100 μm for cartilage.

Fig. 5.

Tissue areas before and after capturing with laser capture microdissection (LCM). LCM capturing was performed on intense and weak methylene blue staining areas, separately. Samples captured were subject to Quantifiler kit analysis. Scale bars: 100 μm.

Table 1.

Quantifiler analysis for laser capture microdissection capturing areas that were stained differently by methylene blue.

Sample	Methylene blue intensity (mean density)	Quantity (pg μL−1)
Lung	142.57 ± 10.09**	10.93 ± 3.01**
	110.89 ± 9.91	0
Kidney	124.14 ± 16.5**	22.90 ± 2.41**
	72.22 ± 12.12	0
Cartilage	95.85 ± 10.28**	0
	60.92 ± 10.8	0

^**P < 0.01.

In addition, the sections were stained with DAPI to confirm whether auto-fluorescence was being emitted from residual nuclear DNA. In Figs 6 and 7, all of the DAPI-stained residues emitted auto-fluorescence, although not every auto-fluorescence-emitting residue necessarily was DAPI-stained. Next, as collagen fibers, the most common residue found in ancient tissue (Chang et al. 2006), are known to be important fluorophores of the extracellular matrix (Blomfield & Farrar, 1969; Fujimoto, 1977; Monici, 2005), we performed Masson staining to determine whether the auto-fluorescence in our samples was also emitted from them. As seen in Fig. 8, the auto-fluorescence was emitted only by cell-like residues (Fig. 8) and, thus, collagen fibers could be excluded from the candidate structures for auto-fluorescence emission.

Fig. 6.

Lung, kidney and cartilage tissues observed with fluorescence microscope. Some auto-fluorescence (AF)-emitting cell residues are not stained by 4′,6-diamidino-2-phenylindole (DAPI) even if most of the cell residues stained by DAPI emit AF. Scale bars: 100 μm.

Fig. 8.

(A) Auto-fluorescence (AF) observed in kidney tissue. (B) Image of collagen fiber (green) stained by the Masson method. (C) Merged image of (A) and (B). AF is only emitted by the cell-like residues and not by collagen fibers. Scale bars: 100 μm.

Fig. 7.

Rib cartilage observed with confocal microscope. (A) Many chondrocytes showed co-localization between auto-fluorescence (AF) and 4′,6-diamidino-2-phenylindole (DAPI) staining. (B) A few chondrocytes emitted AF but were not stained by DAPI (indicated by arrows). DIC, differential interference contrast. Scale bars: 20 μm.

Scanning electron microscopy provided additional data that facilitated the identification of the auto-fluorescent cell residues. In SEM images of the rib cartilage, the cell residues showed various preservation patterns. Whereas most of them were relatively well preserved, the nuclei had disappeared in some others (Fig. 9). This pattern was confirmed by TEM imaging. Whereas some of the nuclei were well maintained, no traces of nuclei were observed in the other cell residues, possibly due to post-cell-death karyorrhexis (Fig. 10).

Fig. 9.

Scanning electron microscope images for rib cartilage. (A) Chondral lacunae (indicated by arrows) in cartilage. (B)–(E) Chondrocytes showing different preservation statuses were observed within lacunae. (B) Partly preserved chodrocytes. (C) Very well-preserved chondrocyte. (D) Chondrocyte without nucleus. Asterisk indicates vacant space where nucleus was situated. (E) In certain chondral lacuna, partly preserved chonrocyte (Ch) was observed, whereas the adjacent cell disappeared (double asterisks). Scale bars: 50 μm in (A); 10 μm in (B)–(E).

Fig. 10.

Transmission electron microscopic images of chondrocytes within lacunae (Lc). Different preservation statuses were observed in each chondrocyte. (A) The nucleus (Nu) was still maintained even if the morphology of the chondrocyte was not intact. (B) The nucleus was not found in the other cell residue. Scale bars: 2 μm.

Discussion

As aDNA or cytoplasmic materials in cells disappear or become difficult to detect after many years of burial, studies on biological molecules extant in ancient samples have sometimes met controversy over whether the ancient data obtained were authentic. For this reason, some researchers have speculated that proving the presence of cell residues by histological techniques can establish the preservation status of biomolecules in ancient samples, given that there is a significant association between histological preservation and survival of biomolecules (e.g. aDNA) in ancient samples (Colson et al. 1997; Haynes et al. 2002; Zink et al. 2005). Certainly, if a simplest means of morphological detection of cell residues in ancient tissues could be developed, it would be very helpful to researchers seeking ideal samples for studies on ancient biomolecules.

Judging from the morphological data obtained in this study, cell residues in ancient tissue samples seem to emit auto-fluorescence. The evidence supporting this conclusion can be summarized as follows. First and above all, the distribution of methylene blue staining, which is known to stain cellular components in various tissues (Deitch, 1964; Šilha, 1966; Erenpreisa, 1977), was co-localized with that of the auto-fluorescence. Second, combined use of LCM capturing and Quantifiler kit analysis showed that nuclear DNA could be very successfully amplified in areas where methylene blue was intensely stained. This means that preservation of nuclear DNA correlates well with intensities of methylene blue staining and even with auto-fluorescence emission. Third, emission of auto-fluorescence was detected in the areas where cells originally had been (i.e. chondrocytes within chondral lacunae or cell residues arranged in the pattern of renal tubules). Fourth, we could confirm that collagen fibers, well-known fluorophores in the extracellular matrix (Blomfield & Farrar, 1969; Fujimoto, 1977; Monici, 2005), did not emit auto-fluorescence. Fifth and finally, co-localization of DAPI staining with auto-fluorescence could also be considered evidence of the nature of the indicated structures. We can be certain that at least a portion of the auto-fluorescence effect is the result of residual DNA in ancient samples, although we had a few, exceptional cases of non-DAPI-stained residues emitting auto-fluorescence.

As for this absence of DAPI staining in some auto-fluorescent cell residues, scanning electron microscopy and transmission electron microscopy can help to interpret the phenomenon. For example, in our electron microscopic images, the cell residues showed various preservation statuses; whereas preserved nuclei were evident in some of them, some others offered no such traces. These latter were almost certainly auto-fluorescent, DAPI-negative cell residues. In this regard, materials extant in cytoplasmic compartments could also be considered a cause of auto-fluorescence in DAPI-negative cell residues. In any case, it cannot easily be denied that some cytoplasmic materials might be chemically changed into fluorophores over long burial durations. Of course, as we cannot rule out the possibility that nuclear DNA might spread throughout cells after disappearance of the nuclear membrane, more work is required to elucidate the exact causes of auto-fluorescence emission by DAPI-negative cell residues.

Even so, the advantages of auto-fluorescence detection for establishing the presence of ancient cell residues cannot easily be overstated. Above all, checking for auto-fluorescence emission in ancient tissues is a very simple way to determine if cell residues remain. We should also note that auto-fluorescence from cell residues can be detected only with one specific type of filter apparatus (i.e. U-MWG2) and not with others (e.g. XF06). In general, auto-fluorescence is of no benefit to histologists conducting immunofluorescence studies on ancient tissues. Once immunohistochemistry or in-situ hybridization with fluorescent-dye-labeled antibodies or nucleotide probes has been accomplished, auto-fluorescence emissions from cell residues must hamper the detection of specific signals. However, given the fact that, in the present study, auto-fluorescence from cell residues could be observed only with the U-MWG2 filter apparatus, there remains another option for successful application of immunofluorescence techniques to ancient tissue sections. If specific immunofluorescent signals are detected first with the alternative XF06 filter apparatus, after which the resultant images are merged with corresponding images from the U-MWG2 apparatus, we might readily establish whether or not such fluorescent signals are actually being emitted from cell residues.

However, for all of the advantages of auto-fluorescence in indicating cell residue preservation, it should be noted that the samples used in this study were taken from Korean mummies, which typically yield specimens showing an extraordinarily perfect preservation status. Applying conclusions from the current study to other ancient tissue cases must therefore be done very cautiously, if at all, until sufficient experience can be accumulated from similar cases around the world.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1 The raw data of Quantifiler analysis. This plot shows the relationship between the fluorescent signal vs. cycle number. In lung and kidney, all of the intensely-stained regions passed the threshold (C_T) but not the weakly-stained region. In the case of cartilage, none of the regions passed the C_T.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.