Exploring new methods to
determine the PMI of modern
remains
Nienkemper J.M. (10458131)
European Credits: 36
Duration: 7 months (feb – aug) Date of submission: 17-08-2018
MSc in Forensic Science, University of Amsterdam 1st supervisor: Drs. ing. Tristan Krap
2nd supervisor: Dr. Annemieke van Dam Examiner: Prof. dr. Roelof-Jan Oostra
Department of Anatomy, Embryology and Physiology & Biomedical Engineering and Physics at the Amsterdam University Medical Centres
ABSTRACT 3
Introduction 3
Materials & Methods 6
Study Material 6
Part I: General experiments 7
Preparation of the cross-sections 7
Histological index 7
UV-fluorescence 8
Part II: Staining experiments 8
Sirius Red/ Fast Green 8
Staining with Sirius Red/ Fast Green 8
Measuring Sirius Red/ Fast Green 9
Absorption 9
Eosin 9
Staining with Eosin 9
Measuring Eosin 10
Absorption 10
Fluorescence 11
Nile Red 11
Staining with Nile Red 11
Measuring Nile Red 11
Fluorescence 12
Results 13
Histological index 13
UV-fluorescence 14
Influence of mass and thickness 15
Sirius Red/ Fast Green 15
Absorption 16 Eosin 16 Absorption 17 Fluorescence 17 Nile Red 19 Fluorescence 19 Discussion 20 Conclusion 26 List of References 27 APPENDIX I 30 APPENDIX II 31 APPENDIX III 32
ABSTRACT
An indication of the post-mortem interval (PMI) of skeletal remains can be of tremendous help to forensic investigators. However, at present, few methods exist which are capable of estimating the PMI of modern remains. The present study will explore new methods which can potentially contribute to estimating the PMI. This is achieved by applying multiple different dyes to pig bones which have been exposed to the environment up to twenty months in an open field. Sirius Red and Fast Green stain collagen pink and non-collagenous proteins green, Eosin stains proteins pink, collagen being its main target, while Nile Red is known to fluoresce red when bound to lipids. Fluorescence and absorption spectrometry was used to quantify the bound dyes. Sirius Red/ Fast Green staining showed a significant correlation with PMI, yet results contradict with previous research. Through absorption measurements a significant correlation was found between Eosin-quantity and PMI, yet, fluorescence measurements were not able to confirm this relationship. Lastly, Nile Red staining was not proven to correlate to PMI statistically, but observations suggest an interrelation between lipid-quantity and PMI.
KEYWORDS: post-mortem interval, modern skeletal remains, Sirius Red/ Fast Green, Eosin,
Nile Red, spectrophotometry
Introduction
When a skeletonised body is found an indication of the time since death could be of tremendous help to the forensic investigators; it could contribute to identifying the deceased, the arrest of a suspect and solving a crime. However, one of the first questions in need of an answer is whether the remains are from before 1920 or after. 1920 is when the Dutch missing person’s register was founded and any findings from before this date are considered to be of archaeological interest, while remains from after 1920 are forensically relevant [1]. The time since death is also known as the post-mortem interval (PMI). An estimation of the PMI can be based on the decomposition of a body, which starts shortly after death and occurs in different stages. Initially, soft tissue is still present, enabling a pathologist to measure the core temperature of a body and correlate this to PMI. Settling of the blood and stiffening of the limbs could also be used to get an indication of the time since death. However, when all soft tissue is degraded and only skeletal remains are left, far less material is present from which a PMI can be deduced.
At present, few methods exist which are capable of estimating the post-mortem interval of skeletal remains accurately. Studying morphological features such as discolouration of the bone is sometimes used to distinguish between the more recent remains, which are forensically interesting, and the archaeological ones [2]. Bone discoloration can be induced by sun exposure, contact with soil or algae or decomposition of the soft tissue [3]. The environment has a substantial influence on the discoloration of bone and bone degradation in general, therefore, simply observing a change in morphological appearance is not sufficient for an accurate estimation of time since death. Another commonly applied technique is that of luminol, which reacts with hemoglobin in blood, resulting in chemiluminescence [4]. A decrease in chemiluminescence is observed with increasing PMI and after a period of approximately 80 years, no more chemiluminescence can be detected [5]. However, since the usage of luminol is an observer-based technique, it is susceptible to classification errors; modern remains can be wrongly classified as ancient and vice versa. A different option is to
study the autofluorescence of bone by exposing it to UV-light. Fresh bone is found to fluoresce blue, while historical specimens show yellow to no fluorescence [6]. Proteins are thought to be the source of the blue fluorescence, with collagen being the main contributor. A decrease in collagen, as well as modifications of the mineral phase could be the cause of the yellow fluorescence seen in historical specimens. Some cross-sections however, are found to have zones of yellow or reduced fluorescence along the outer -, and sometimes inner, rim of the section; the so-called sandwich effect. It is argued by Ramsthaler et al. [7] that these zones of reduced fluorescence are a result of inhibition of the emission of UV-light due to remaining body fat, yet no other research supports this hypothesis. Just as the luminol method, the UV-fluorescence technique can only be used to distinguish between forensically relevant and historical bone. Yet again, the technique is observer based and inconsistencies exist in the scoring method, leaving the ability of UV-fluorescence to be used to distinguish between forensic and historical bone to be disputable [6]. Several studies have attempted to measure the change in nitrogen-, amino acid-, collagen-, citrate- or overall protein-levels of bone and correlate this to PMI, however, none of these experiments show reliable results capable of accurate dating [8–10].
The only technique currently available which is more precise, is that of radiocarbon dating. This method is based on the finding that the atmosphere contains radiocarbon (14C). This radiocarbon is incorporated into plants, which are subsequently being consumed by animals. During life, certain C14- levels are maintained, but when consumption ceases at death, the levels of radiocarbon start to decrease [11]. When measuring the amount of radiocarbon present in a deceased animal, the time of death can be calculated. However, since the date of death is calculated with a 3- to 5-year interval, an estimation of the PMI of very recent skeletal material remains problematic [12].
Each of the aforementioned dating techniques relies on the structural degradation of bone and its constituents. The main component present in bone is fibrous collagen, which forms a close structural connection with hydroxyapatite, a crystalline matrix of minerals [8] (Figure 1). Collagen fibres provide flexibility to the bone and a structure on which the mineral crystals can grow, while minerals add strength. Because of the small size of the minerals, they easily dissolve in water and could percolate from the bone, yet, because of the close framework with collagen, the apatite is protected [13]. In turn, the apatite protects the collagen mineralised within from degradation by microbial enzymes [8]. However, certain microorganisms are capable of dissolving the apatite by excreting acids, which results in tunnels formed in the bone. The increased porosity leads to exposure of collagen, which is then reduced to peptides, and eventually amino acids, by bacterial collagenases. The amino acids easily wash away in the groundwater, which is enhanced by elevated water permeability of the soil [14]. Which bacteria are active is dependent on the pH-value of the surrounding soil. Non-collagenous proteins are thought to persist longer in bone than collagen because of a higher affinity to hydroxyapatite [15]. Apart from microbial attack, collagen can also degrade through hydrolysis of the peptide bonds, fragmenting the protein, however, this process occurs less prominent and is much slower [8,15,16]. Lipids too are incorporated in the mineralized bone matrix and are present in bone cells, mostly osteocytes [17]. Bone marrow, the soft inner part of bone, also contains high amounts of fat; approximately 5 to 15 times more than the mineralized tissue. It is thought that, in their liquid phase, lipids can seep out of the bone, but no literature can be found on the behaviour of lipids in degrading bone.
Figure 1. The structure of human bone (edited from Zimmermann et al., [18]).
Histological stains can be used to visualize some of the aforementioned elements. Dyes are applied to a specimen to highlight a particular feature or to add contrast to the tissue, facilitating microscopic observations. Degradation of a certain element is expected to result in a reduction of bound dye, therefore, histological staining could be an interesting technique to estimate PMI. Boaks et al. [19] implemented this concept on cross-sections of pig bones which have been deposited in a field and exposed to the environment for up to a year. The study group used the dyes Sirius Red and Fast Green, which colour all types of collagen (Co) pink and all other non-collagenous proteins (NCo) green. The dyes were subsequently eluted from the samples to calculate a Co/NCo-ratio. A moderate, yet significant correlation was found between the degradation of Co/NCo protein concentrations and time. Jellinghaus and colleagues [8] tried to reproduce these results with buried pig remains but found an overall lower Co/NCo-ratio in fresh bone samples than Boaks et al. [19] and could only partly confirm the correlation between the ratio and PMI. Contradicting results indicate the need for more research.
The present study is going to reproduce the experiment performed by Boaks et al. [19] with slight alterations and will apply two more dyes to pig bone cross-sections. Pig bones (sus
scrofa domesticus) are chosen as a substitute for human material, because of ethical
constraints. The focus of the current study will be on the degradation of modern remains (up to twenty months old) which have aged above ground. Whether a deceased body is buried or deposited on the surface has a substantial influence on the process of decomposition. For example, above ground a corpse will endure higher temperatures compared to underground, accelerating the process of skeletisation [8]. While underground, the microbial and mineral content of soil has a greater influence on decomposition as compared to above surface. Capella et al. [2] presented a list of 24 articles of which they think are the most important researches concerning techniques to estimate the PMI of remains; only one of these 24 investigations also focused on non-buried remains, indicating the need to investigate the degradation of bone exposed to the open air more thoroughly.
Next to Sirius Red/Fast Green, the second dye applied is Eosin (E) which is used to give collagen and other positively charged proteins a pink colour [20]. This dye is semi-quantified with the use of fluorescence and absorption spectrometry. Thirdly, Nile Red (NR) is used to stain the lipids present in the osteocytes in bone, which fluoresces red when seen
from under a fluorescent microscope [21]. This dye is, again, semi-quantified with the use of fluorescence spectrometry. It is assumed that the amount of observed dye is proportional to the element they stain. The results obtained from the spectrophotometric techniques are plotted against time, to investigate whether there is a correlation between the quantity of bound dye (and thus proteins/lipids) and PMI. Resulting from this, the main question of the present research is: Can Sirius Red/Fast Green, Eosin and Nile Red be used to estimate the PMI
of modern skeletal remains? Yet first, a sub question needs to be answered; Can Sirius
Red/Fast Green, Eosin and Nile Red be used to stain cross sections of modern remains of piglet bone? Additionally, to get insight into the level of preservation of the cross-sections of varying PMIs, the Oxford Histological Index will be used to score each sample. Structural alterations to bone caused by fungi and bacteria can take place within the first few post-mortem months or even days [16,22], therefore, studying histological changes with the use of the histological index could potentially provide valuable information regarding PMI. Moreover, UV-light will be applied to the specimens to study the presence of the mentioned sandwich effect, which can possibly be correlated to protein or lipid content.
The present study provides a foundation for future research into a more accurate method to determine the PMI of modern skeletal remains.
Materials & Methods
Study Material
Long bones of domestic pigs (sus scrofa domesticus) were used to study the degradation of bone. Eight piglet femora which have aged for different periods of time were collected from the Forensic Anthropological Outdoor Research Facility in Den Ham (NL) and taken to the Amsterdam University Medical Centres (AUMC) in plastic containers. These pig cadavers have been placed on the surface of a field in metal cages by other researchers for different observation studies. At the time of collection, the pig bones were deposited 10, 13, 17 and 20 months ago. This resulted in four moments of measurement within a period of 20 months of degradation. From each of the 8 bones, 6 cross-sections were created from the intermediate part of the diaphyses. With each of these 6 cross-sections a different part of the experiment was performed, in duplo (Table 1). Additionally, a fresh femur of an adult pig was obtained from a butcher, which was used for optimization experiments.
Table 1. Overview of number of samples.
Condition Age of the bone Number of pigs Number of femora Number of cross-sections Bones collected from Den Ham 10 months 2 2 12
13 months 2 2 12 17 months 2 2 12 20 months 2 2 12
Part I: General experiments
Preparation of the cross-sections
Preparation of the specimens was performed according to the protocol of Maat et al. [23] with minor adjustments. A mitre box was placed on a glass plate with a wet piece of tork paper in between, and under the glass plate, to stabilize the setup. A piece of tinfoil was wrapped around the outer end of the bone, to enlarge the surface area for better grip. A small hacksaw was used to saw six sections of approximately 2 mm wide from the diaphysis, which took place under the fume hood to avoid inhalation of bone particles. The sections were temporarily stored in small plastic containers, separately. The three dye conditions were assigned to the six cross sections in a different order for each bone (Table 2).
Sandpaper with grit size P220 was attached to a second glass plate using Vaseline. With the use of the ‘gripping device’ described by Maat et al. [23], a specimen was sanded until it reached a thickness of around 0,3 mm. To minimize the groove-size caused by the sanding, the sample was sanded with fine sandpaper (grit size P1200) for 10 seconds on both sides. The final thickness of the section was to be between 0,2 mm and 0,3 mm, which was ensured with the use of a calliper. Lastly, each section was weighted with an analytical scale. The finished products were kept in small plastic containers, with added water to prevent dehydration.
Table 2. The different conditions assigned to the cross-sections, in duplo. R/G = Sirius Red and Fast Green, E =
Eosin, NR = Nile Red.
Bone # PMI (in months) 1 2 3 4 5 6 L1 10 R/G1 R/G2 E1 E2 NR1 NR2 26 10 E1 E2 NR1 NR2 R/G1 R/G2 28 13 NR1 NR2 R/G1 R/G2 E1 E2 29 13 R/G1 R/G2 E1 E2 NR1 NR2 32 17 E1 E2 NR1 NR2 R/G1 R/G2 33 17 NR1 NR2 R/G1 R/G2 E1 E2 37 20 E1 E2 NR1 NR2 R/G1 R/G2 39 20 NR1 NR2 R/G1 R/G2 E1 E2
Histological index
Once the sections were created the first step of the experiment was to observe each cross-section microscopically, to score histological features according to the Oxford histological
preservation index [24]. With this index, each slide can be classified into one of six stages, based on the degree of preservation of the bone (Table 3). When observing the sections, attention was paid to the presence of tunnelling, often as a result of microbial attack, and destructions of the lamellar structure. Each section was observed with a Leica DM1000 microscopeat a magnification of 40x (4x10x) and a set light intensity.
Table 3. 6 Oxford histological index values which can be assigned to bone cross-sections to indicate the degree
of preservation (based on Hedges & Millard [24]). Index Approx. % of
intact bone
Description
0 <5 No original features identifiable, other than Haversian canals
1 <15 Small areas of well-preserved bone present, or some lamellar structure preserved by pattern of destructive foci
2 <33 Clear lamellate structure preserved between destructive foci 3 <67 Clear preservation of some osteocyte lacunae
4 <85 Only minor amounts of destructive foci, otherwise generally well preserved 5 >95 Very well preserved, virtually indistinguishable from fresh bone
UV-fluorescence
Subsequently, the sections were studied under UV-light to observe the potential presence of the previously mentioned sandwich effect. A crime light with a wavelength ranging from 350 nm to 380 nm was focused on the sample from a height of 15 cm. A Canon EOS 40D camera was set in a fixed position directly above the sample at a height of 50 cm. Camera settings were as follows: aperture = 2.8, shutter speed = 30”. A dark background was put under the sample which was placed on a microscope glass. The photos were taken in duplo, in a dark room. Based on the photos it was determined whether the samples showed the presence of a sandwich effect ‘yes’ or ‘no’. Additionally, two photos were taken with a white background and white light.
Part II: Staining experiments
Sirius Red/ Fast Green
Staining with Sirius Red/ Fast Green
To stain the cross-sections with Sirius Red and Fast Green, the staining kit ‘Sirius Red/Fast Green Collagen Staining Kit’ from Chondrex was used and handled according to the manufacturer's protocol. After testing with pig femora obtained from a butcher, minor adaptations were made, which resulted in the following protocol being applied to the bones retrieved from Den Ham: each section was placed in a ‘dipping device’, a 10ml plastic tube with holes poked in the bottom, so that the samples could be transferred from one solution to
another without touching it. First, the section was fully submerged in MiliQ for 2 minutes, in triplo, to wash the sample. Afterwards, the section was transferred into PBS for another 5 minutes, to maintain the pH within the sample. Next, the section was placed in a petri dish with a lid, on a wet piece of filter paper, to prevent dehydration. 0,2 Mm of Dye Solution was pipetted onto the section, which was incubated for 30 minutes. The section was turned over half way to ensure coverage of the dye on both sides. After incubation, the specimen was rinsed with the use MiliQ, which was pipetted over the slide until the water ran clear. The section was placed on a microscope glass with a drop of MiliQ after which it could be photographed under a Leica DM1000 microscope with camera (magnification 40x).
Measuring Sirius Red/ Fast Green
Absorption
For semi-quantification of the dyes, the dyes were eluted from the sample. The section was placed in a petri dish where 1 ml of Extraction buffer was pipetted on the sample and incubated for 10 minutes. After incubation, the liquid was pipetted into a cuvette and the section was mounted onto a microscope glass with neomount. The cuvette was placed in the Lambda Bio+ spectrometer and with the use of Lambda Bio XLS Optical Density-values were obtained for the wavelengths 540 nm and 605 nm, in triplo. 540 Nm is the absorption maximum for Sirius Red, which binds collagen, and 605 nm is that of Fast Green, which binds non-collagenous proteins. A cuvette filled with 1 ml of Extraction buffer was used as a reference sample. Corrections were applied to these values according to the Chondrex protocol. With the corrected values a ratio was calculated which is proportional to the ratio of collagen and non-collagenous proteins (Co/NCo-ratio). After a test of normality, a Pearson’s correlation was calculated between the obtained ratios and PMI.
Eosin
Staining with Eosin
With cross-sections created from an adult pig femur obtained from a butcher, the protocol of de Boer et al. [23] was tested for its ability to stain with Hematoxylin and Eosin. Since the protocol was designed to stain archaeological human bone, optimization was needed so that it could be applied to modern pig bone. Optimization experiments lead to the following protocol which was applied to the samples from Den Ham; each cross-section was placed in the ‘dipping-device’ and submerged in MiliQ for 2 minutes, in triplo. Afterwards, the specimen was placed in PBS (1x solution) for 5 minutes. Next, the section was put in Eosin G ready made for 5 minutes. It was chosen not to use Hematoxylin, since it coloured all of the sections a dark purple and seemed to suppress the Eosin. After Eosin, the section was placed in 70% ethanol, 90% ethanol and 100% ethanol for 3 minutes each, to dehydrate the sample. The extracted water needed to be replaced, therefore the sample was submerged in Xylene for 5 minutes. With the use of neomount the sample was fixed onto a microscope glass. A Leica DM1000 microscope with camera was used to photograph the section (magnification 40x).
Measuring Eosin
To quantify the bound Eosin, it was attempted to eliminate the Eosin from the sample so that it could be analysed with the use of spectrometry. Ali et al. [25] suggested that a pH below 2 would transform anionic Eosin into an unbound neutral state, which would result in a decrease of protein-dye binding. To investigate this hypothesis, Hydrogen Chloride (HCl) was added to test-cross-sections, to lower the pH of the samples below 2. This addition resulted in a pH<2 and an initial disappearance of the pink colour of Eosin, however a subsequent increase of pH by adding MiliQ showed a reappearance of pink. What exactly caused this reaction to occur is not known, but it showed that Eosin is not washed out from the sample with the use of this method. Therefore, it was decided to measure the fluorescence and absorption of Eosin while still bound to the bone sample and mounted to a microscope glass.
Absorption
The USB4000-UV-VIS absorption spectrometer (range 200 nm - 850 nm) was used to measure the absorption intensity of Eosin and SpectraSuite served to produce the corresponding graphs. First, a spectralon was placed directly under the fibre at approximately 1 cm for a background measurement. Secondly, all cross-sections assigned to Eosin were measured prior to staining at the widest part of the cross-section (Figure 2). This measurement served as a baseline measurement. After staining with Eosin, the section was measured at four locations shown in Fig. 2.
Microsoft Excel 2018 was used to determine the intensity at 530 nm and
970 nm (Graph 1). A ratio was calculated between the two obtained values. After testing for normality, a Pearson’s correlation was calculated between the ratio’s and PMI, using IBM SPSS Statistics. A paired sample
T-test was used to determine whether there was a significant mean difference between the ratio’s obtained in the ‘before staining’ and ‘after staining’ condition.
Graph 1. The blue line is the baseline measurement (pre-stained bone) and the red line is the stained bone. The intensity is determined at the location of the dashed lines; 530 nm and 970 nm.
Figure 2. A cross-section
of a piglet femur. The green circle is the location of the baseline measure-ment. 1, 2, 3 and 4 are the locations of post-staining measurements.
Fluorescence
The Perkin Elmer LS-55 fluorescence spectrometer was used to measure the fluorescence intensity of two different elements; hydroxyapatite (HA) and Eosin (E). After optimization the used setting were: hydroxyapatite: excitation wavelength = 375 nm and emission wavelength range = 400 nm - 600 nm [26], and Eosin: excitation wavelength = 500 nm and emission wavelength range = 525 nm - 700 nm [25]. SpectraSuite visualised the corresponding graphs. Each sample assigned to Eosin was measured pre- and post-staining at the four locations indicated in Fig. 2. The samples were placed in a black box under the optic fibre of the spectrometer at a distance of 0,5 cm. Microsoft Excel 2018 served to calculate the AUC for all obtained data. The AUC for hydroxyapatite was calculated 420 nm and 510 nm and for Eosin between 540 nm and 600 nm. A ratio was calculated between the AUC for Hydroxyapatite and the AUC for Eosin. After testing for normality, a Kendall’s tau correlation between the obtained ratio’s and PMI was calculated. A Wilcoxon signed rank test was used to determine whether there was a significant mean difference between the ratio’s obtained in the ‘before staining’ and ‘after staining’ condition.
Nile Red
Staining with Nile Red
Test-staining on fresh pig femora obtained from a butcher, gave the following protocol which was applied to the bone-sections from Den Ham: a stock solution of Nile Red (NR) was prepared by adding 5 mg of Nile Red to 10 ml acetone, which was stored in the dark. A working solution was created by taking 100 μl of stock solution and adding that to 10 ml acetone. A section was submerged in MiliQ with the help of the dipping-device for 2 minutes, in triplo. The second step was to fully immerse the section in PBS for 5 minutes, after which the specimen was placed in a petri dish with a lid, on a wet piece of filter paper. 0,2 ml of stock solution was pipetted onto the sample, which was incubated for 40 seconds. The section was turned over after 20 seconds to ensure complete coverage of the dye. The petri dish was covered in tin foil to keep out the light and a lid covered the petri dish to ensure no/less stock solution would evaporate. After a quick dry, the section was mounted onto a microscope glass with the use of neomount. A 100x magnified image was captured using a Nikon Eclipse E600 fluorescence microscope emitting green light and a build-in DS-Fi2 camera with a red filter.
Measuring Nile Red
To quantify the amount of Nile Red bound to the lipids in the sample, it was attempted to remove Nile Red from the sample. Trichloroethylene is a common chemical used to degrease metal and is used to eliminate fat from bone [27]. Therefore, is was experimented whether samples treated with Nile Red would lose its fluorescent property after treatment with trichloroethylene + ethanol (1:2). Even after an incubation time of 25 hours, Nile Red was still present in the sample. Moreover, the solution evaporated rather quickly, making it difficult to collect and measure the liquid. Consequently, the fluorescence intensity of Nile Red was measured when still bound to the bone.
Fluorescence
To measure the fluorescence intensity of collagen, hydroxyapatite and Nile Red, the Perkin
Elmer LS-55 fluorescence spectrometer was used. After optimization the used setting were:
hydroxyapatite: excitation wavelength = 375 nm and emission wavelength range = 400 nm - 600 nm, and Nile Red: excitation wavelength = 550 nm and emission wavelength range = 580 nm - 700 nm [28]. SpectraSuite visualised the corresponding graphs. All of the samples assigned to Nile Red were measured pre- and post-staining at the four locations indicated in Fig. 2. The samples were placed in a black box under the optic fibre of the spectrometer at a distance of 0,5 cm. Microsoft Excel 2018 served to calculate the AUC for all obtained data. The AUC for hydroxyapatite was calculated between 420 nm and 510 nm and for Nile Red between 600 nm and 660 nm. A ratio was calculated between the AUC of Hydroxyapatite and the AUC of Nile Red. After testing for normality, a Kendall’s tau correlation between the obtained ratio’s and PMI was calculated. A Wilcoxon signed rank test was used to determine whether there was a significant mean difference between the ratio’s obtained in the ‘before staining’ and ‘after staining’ condition.
Results
Histological index
Based on the Oxford histological preservation index [24] each bone section was assigned a score of 5 (very well preserved). No obvious tunnelling or other microstructural destructions seemed to be present in any of the specimens. Fig. 3 shows that each PMI resulted in a similar level of preservation.
Figure 3. One cross-section per bone visualized with corresponding microscopic image (magnification 100x).
Each time period has an equal degree of preservation. The sample with a PMI of 0 had no photo of the cross-section available.
UV-fluorescence
The UV-reflection experiment showed that all bone samples collected from Den Ham, except for bone 39 (PMI: 10 months), displayed the so-called sandwich effect (Figure 4). A fresh sample obtained from a butcher showed an even blue fluorescence.
Figure 4. PMI is in months. The left side of the image is a bone section excited by UV-light and the right side is the
Influence of mass and thickness
In order to determine whether mass and thickness influenced the AUC-values and Co/NCo-ratios, a correlation analysis was conducted. First, a Shapiro-Wilk test revealed that the data of the categories ‘Nile Red Fluorescence’ and ‘Eosin fluorescence’ was not normally distributed (NR: p = 0,001**, E: p = 0,011*). Therefore, Kendall’s tau was used to calculate the correlations concerning the categories ‘Nile Red fluorescence’ and ‘Eosin fluorescence’, while a Pearson’s correlation was used for all other normally distributed data. The previous applies to all subsequent correlation analyses performed. Table 4 shows the results of these analyses. Additionally, mass and thickness were found to have a very strong correlation of r = 0,930 (p = 0,001**). The strength of the correlation is verbally described according to guidelines suggested by Evans [29]; an r between 0 and 0,19 is considered ‘very weak’, r = 0,20-0,30 is ‘weak’, r = 0,40-0,59 is ‘moderate’, r = 0,60-0,79 is ‘strong’ and r = 0,80-1 is ‘very strong’.
Table 4. The influence of mass and thickness on AUC-values and ratios. Values marked light grey show a very
weak correlation, and values marked dark grey show a weak correlation. None of the correlations is significant. Eosin Nile Red Sirius Red/Fast
Green Correlation between: Absorption
(AUC) Fluorescence (AUC) Fluorescence (AUC) Absorption (ratio) Mass r = -0,310 (p = 0,095) r = 0,176 (p = 0,344) r = 0,276 (p = 0,137) r = -0,107 (p = 0,693) Thickness r = -0,060 (p = 0,751) r = 0,180 (p = 0,341) r = 0,026 (p = 0,892) r = 0,202 (p = 0,452) ** The p-value is significant for = 0,001
* The p-value is significant for = 0,05
Sirius Red/ Fast Green
Fig. 5 is an example of the result of staining with Sirius Red/Fast Green. Sirius Red stains collagen pink and Fast Green gives the non-collagenous proteins a green colour. Fresh samples contain high amounts of collagen and therefore mainly colour pink, while with increasing PMI collagen degrades and a transition to green can be seen. Appendix I shows all stained sections; here it can be seen that samples with a PMI of 0 show considerably more pink dye than samples with a longer PMI. Samples with a PMI ranging from 10 to 20 months seem to be in transition from predominantly pink to green.
Figure 5. The left imagae shows specimen 39 (PMI = 10 months) pre-staining and the right image shows 39 after
Absorption
With absorption spectrometry the eluted dyes Sirius Red and Fast Green were measured. With the obtained values a Collagen/Non-Collagenous proteins-ratio was calculated, over which a correction was applied according to the Chondrex protocol. These ratios are plotted against PMI in Graph 2. A strong correlation of 0,737 was found between the Co/NCo-ratios and PMI, with a significant p-value of 0,001**. The fresh adult bone obtained from the butcher gave a Co/NCo-ratio of 0,050 0,014. Specimens with a PMI of 10 and 13 months show Co/NCo-ratios within a similar range, and so do specimens with a PMI of 17 and 20 months. This distribution in values could not be explained by level of browning of the bone, location in the field in Den Ham, thickness or mass.
Graph 2. Co/NCo-ratios determined for the Sirius Red/Fast Green-condition with the use of absorption
spectrometry, plotted against PMI.
Eosin
Fig. 6 shows an example of the result of Eosin staining. Eosin stains proteins pink, collagen being the main contributor. By simply observing all specimens with the naked eye, no apparent increase or decrease of Eosin seems to occur over time. See Appendix 2 for all stained sections.
Figure 6. The left image shows specimen 26 (PMI = 20 months) before staining and on the right image is 26 shown
Absorption
Graph 3 shows the result of absorption spectrometry performed on samples stained with Eosin. Ratios were calculated between the intensity at 530 nm and 970 nm, which are plotted against PMI. With the Pearson’s correlation test a strong correlation of -0,631 was found between intensity-ratios and PMI, with a significant p-value of 0,009*. The fresh adult bone sample gave an intensity-ratio of 2,712 0,471.
The intensity measured at 970nm was relatively stable over time and across samples, therefore, any change in ratio was predominantly caused by a change in intensity measured at 530 nm; the absorption maximum of Eosin bound to proteins [30]. Hence, a decrease in intensity-ratio correlates to a decrease in intensity measured at 530 nm. Yet, a lower intensity corresponds to higher levels of absorption. Higher levels of absorption are in turn thought to correlate to a higher quantity of Eosin. In summary, a decrease in intensity-ratio is correlated to an increase in Eosin quantity. Thus, when PMI increases, Eosin seems to show an increase too.
Graph 3. Intensity-ratios determined for the Eosin-condition with the use of absorption spectrometry, plotted
against PMI.
A paired sample T-test showed that the mean intensity-ratios calculated in the ‘Before staining’ and ‘After staining’ conditions do not differ significantly (p = 0,940, t = 0,076).
Fluorescence
Results of measuring the Eosin stained samples with fluorescence spectrometry are depicted in Graph 4. Ratios were calculated between the AUC obtained for hydroxyapatite (HA) and the AUC for Eosin (E), which are plotted against PMI. A Kendall’s tau-test revealed a weak correlation of 0,373 between the HA/E-ratio and PMI, with a p-value of 0,062. Fresh adult bone gave an HA/E-ratio of 0,024 0,038. Samples with a PMI of 13 and 20 months show a relatively large variance, yet there is no clear explanation for this observation.
Graph 4. HA/E-ratios determined for the Eosin-condition with the use of fluorescence spectrometry, plotted against
PMI.
Since the data was not normally distributed, a Wilcoxon signed rand test was used to calculate whether there was a significant mean difference between the AUCs for Eosin, before and after staining. This test showed that the mean AUCs are significantly different (p = 0,001**). Graph 5 visualizes the difference in obtained signal before and after staining with Eosin when a sample with a PMI of 10 months is excited at 500 nm.
Graph 5. shows the signals obtained by exciting sample 39 (PMI =
Nile Red
Fig. 7 shows the result of staining with Nile Red, which gives lipids a red colour when seen from under a fluorescent microscope. See Appendix III for all stained sections. When observing the specimens from under microscope, a decrease of stained lipid spots seems to occur with increasing PMI.
Figure 7. The left image shows specimen 39 (PMI = 10 months) under white light, while the right image is taken
excited by green light with a red filter. The red dots within the lamellar structure are the lipids captured in osteocyte lacunae.
Fluorescence
Fluorescence spectrometry was used to measure samples stained with Nile Red. From the obtained signals AUCs were obtained. A ratio between the AUC of hydroxyapatite and the AUC of Nile Red (NR) was calculated, which are plotted against PMI in Graph 6. Since the data was not normally distributed, a Kendall's tau test was used to reveal a weak correlation of 0,243 between the ratios and PMI (p = 0,224). A fresh adult bone gave an HA/NR-ratio of -1,850 8,320.
Graph 6. HA/E-ratios determined for the Nile Red-condition with the use of fluorescence spectrometry, plotted
against PMI.
A Wilcoxon signed rank test showed that the mean AUCs calculated in the ‘Before staining’ and ‘After staining’ conditions do differ significantly (p = 0,002*).
Discussion
The samples used for the experiments came from an outdoor research facility in the Netherlands. The piglets were exposed to the weather, which has a substantial influence on the degradation of the cadavers. Ambient temperature, humidity, pH, soil moisture, and the presence of bacteria and fungi all affect decomposition. However, it was outside the bounds of the present project to control environmental conditions or to include the influence of the weather on the degradation of skeletal remains as a covariate.
A fresh femur of an adult pig was used to perform optimization experiments and simultaneously served as a t=0 measurement. However, since the pig femur was obtained from a butcher, the time passed since death was unknown and therefore t=0 is not exact. Moreover, the histological structure of adult bone differs significantly from that of piglets. Bone tissue of infant pigs was observed to have more circumferential lamellae and less osteons than adult bone and a higher number of osteocytes, which is in agreement with findings of Feng & Jasiuk [31]. Moreover, the concentric lamellae, which are part of the osteons, do not seem to show the distinct circular pattern as can be seen in adult bone, but rather a more disorderly pattern. Adult bone has an increased mineral content and lipid concentration compared to young pigs and a more mature collagen matrix with more collagen cross-links [17,31,32]. Moreover, because of the larger surface area of adult bone cross-sections, more dye could have bound to the specimens. It is thought that adult bone which has not been exposed to the environment or experienced decomposition should not be compared to piglet femora. Therefore, the values obtained by measuring the adult bones are mentioned to give an indication of the possible t=0 values but are not included in the (statistical) analysis.
Not only do infant and adult pig bones differ, there is also a considerable difference between porcine and human bone (Figure 8). Human bone has higher collagen levels, but lower non-collagenous protein levels than pigs [33]. Humans have thicker cortical bone, larger osteons and larger Haversian canals compared to animals [34]. Moreover, animal bones are found to be less affected by bacterial attack than human bones [35]. These dissimilarities complicate the usage of the Oxford Histological Index to score infant pig bone, since the index is based on adult human remains. The index would need to score different aspects for it to be able to be used to score infant pig skeletons. For example, browning of the bone was not included in the index, while this was the only visible change over time in the present study. It is possible that the microstructural destructions or other alterations scored with the histological index were present, but simply not observed with bright-field microscopy. It is advised to use a polarized light microscope, which aids the observation of the anisotropy of bone and therefore facilitates the detection of small alterations [36], however, this was beyond the scope of the current project. Moreover, certain types of tunnelling (linear longitudinal and budded tunnelling) might only be detected when a thicker section is observed. In conclusion, it is thought that the histological index is not suited for scoring thin sections of piglet bones.
Figure 8. Left depicts a microscopic image of a fully developed human bone, showing very distinct circular osteons
containing the Haversian canal in the centre (edited from Todt [37]). Right displays a microscopic image of infant pig bone, with a more disorderly lamellar structure and less well-defined osteons.
After attributing scores according to the histological index, the sections where analysed for the presence of the sandwich effect with the use of UV-light. All samples, except for sample 39, showed the presence of the sandwich effect. Ramsthaler and colleagues [7] opted that the reduced zones of fluorescence seen in the sandwich effect might be due to inhibition of emission by remaining body fat. However, observations made in the present study resulted in a contradicting view. It was seen that yellow fluorescence (reduced fluorescence) was present in those areas which showed brown discoloration when seen with the naked eye (Figure 9). When studied under a fluorescence microscope, these brown areas showed reduced Nile Red spots, which corresponds to reduced osteocyte lacunae containing lipids. It is therefore thought that regions of reduced fluorescence correlate to reduced lipid content, however, more research is needed to support this hypothesis.
Figure 9. The left picture shows sample 32 (PMI = 13 months) under UV-light, where faded yellow fluorescence
can be seen along the outer rim of the section. The middle picture depicts sample 32 excited by white light, showing moderate brown discoloration along the outer rim (bottom). The right picture shows this same sample excited by green light and a red filter. Here it can be seen that the outer rim contains less lipid spots stained by Nile Red.
Not only were the specimens assessed through observations, they were also measured with the help of spectrophotometric techniques. Absorption spectrometry resulted in Co/NCo-ratios when performed on sections stained with Sirius Red/ Fast Green, and in intensity-ratios for sections stained with Eosin. Fluorescence spectrometry gave AUC-ratios for both Nile Red and Eosin stained specimens. Weak to very weak correlations were found between mass and thickness of a section and the different ratios, though non-significant. Mass and thickness were found to have a near perfect correlation (r= 0,930 & p=0,001**), yet separately they influence the ratios differently. Therefore, it is thought that a third factor,
surface area, is of importance. A low mass often corresponded to a very small surface area and as a result the width of certain sections was sometimes more narrow than the beam of the excitation light of the fluorescence spectrometer. When sections were too narrow, part of the black background was included in the measurement, resulting in lower peaks and more fluctuation of the signal, which in turn resulted in smaller AUCs. With absorption spectrometry, the non-absorbing white background was partly included in the measurement, which possibly altered the results. Thinner sections let through more light and consequently reflected less, which resulted in fluctuations of the peak and an overall lower signal. Moreover, thin sections become more porous, giving rise to scattering of the light. This too resulted in lower peaks and fluctuations in the signal. Thicker or larger sections have more surface to which a dye can bind, so when eluting the dye, this could result in higher Optical Density values, yet, a weak, non-significant correlation (Mass; r = -0,107 & p = 0,693 and thickness; r = 0,202 & p = 0,452) shows this was not the case. In conclusion, thicker or heavier sections do not necessarily contain more dye but seem to give a better signal.
The Co/NCo-ratios acquired from the Sirius Red/Fast Green staining showed a significant and strong correlation with PMI (r = 0,737 & p=0,001). However, these findings are not in line with the results from Boaks et al. [19] nor Jellinghaus et al. [8]. Due to different time intervals, only the 10-month and 13-month old specimens of the present study can be compared to the results of Boaks and her colleagues [19]. At 10 months, Boaks and colleagues [19] found a ratio of 1,22 ± 0,04, which is higher by a factor of nearly 50 compared to the ratio of 0,025 ± 0,004 obtained in the current study. The same applies to the ratios obtained at 13 months. Jellinghaus et al. [8] found ratios in a similar range as our results, however, since her oldest specimen was 3 months old, the ratios cannot be directly compared. The most striking difference is that Jellinghaus and Boaks and colleagues found an overall negative correlation with PMI, while the current experiments showed a positive correlation between Co/NCo-ratios and PMI. Explaining this difference is made difficult by the fact that only ratios are presented in their articles and no OD-values. One reason could be differences in protocol regarding the extraction of dyes. During optimization experiments it was observed that the green dye would not completely dissolve into the extraction buffer. It was decided to incubate the sections in the extraction buffer for 10 minutes, despite the dye not being completely extracted after this period of time. Jellinghaus et al. [8] noted that they too had difficulties removing the dye and therefore modified the washing procedure, but since they did not specify how, dissimilarities could have occurred in this part of protocol. Their explanation for the not correctly functioning of the extraction buffer, was that it was not cooled properly during transport from the supplier in the USA. However, the extraction buffer, NaOH-methanol, is advised to be stored dry and at room temperature.
Another conclusion which could be drawn is that it is not time, but other factors such as soil composition, temperature, bacteria and fungi which have a greater influence on protein degradation. In independent studies, these factors have been found to impact protein diagenesis substantially [35,38–40]. Differing environmental conditions could have caused the inconsistent results. With a more controlled environment in an experimental setup, Sirius Red/ Fast Green could still prove to be a good estimator of PMI. Such a situation might not represent the conditions of an actual crime scene, but it could help understand the process of bone degradation and assist in improving methods for estimating PMI.
No significant correlation was found between PMI and HA/Eosin-ratios, when measured with fluorescence spectrometry. A downside to Eosin is that its specificity is low since it binds to
certain amino acid residues which are present in more than proteins alone. Moreover, the level of protein-dye-binding is pH dependent, where a low pH leads to less protein-dye-binding [25]. However, since each specimen was handled identically, they should all have had an equal pH and show the same amount of non-specific binding. Peaks obtained with fluorescence spectrometry varied greatly in intensity when comparing the four different locations measured, which might be explained by differences in bone structure within the section. To minimize this variety in intensity, the four locations were averaged into one value and a ratio was calculated with the AUC of hydroxyapatite. It is thought that when a sample is excited at 375 nm, the obtained emission peak might not be the result of hydroxyapatite alone, but rather a combination of collagen and hydroxyapatite as a result of the tight knit structure of the two elements. Yet, the uncertainty of what exactly is measured should not pose a problem, since this uncertainty is equally present in each of the samples.
Despite the non-significant correlation, the fact that Eosin bound to bone can be semi-quantified with the use of fluorescence spectrometry is innovating. When exciting a sample stained with Eosin at a wavelength of 500 nm, this resulted in an emission peak at 566 nm. Non-stained sections did not show this emission peak when excited at the same wavelength. A Wilcoxon signed rank test showed that the AUCs calculated before and after staining were significantly different (p = 0,001**). Moreover, the obtained emission peak lies outside the spectrofluorometric profile of porcine bone and can therefore easily be distinguished from peaks that are a result of light emitted by intrinsic bone components (Graph 7).
Using fluorescence spectrometry to quantify Eosin shows to be a promising technique, and therefore it is suggested to further investigate the usage of this method in different settings. The environment in which the bones age should be controlled and a larger sample set should be used. In addition, it would be preferred to use human long bones for the adjusted experiment to investigate whether this will lead to different results compared to pig bones. Ali
et al. [25] demonstrated that Eosin Y fluorescence can be used to quantify the severity of
certain liver diseases. An injured liver has denatured proteins, which have more exposed amino acids residues to which Eosin can bind. Accordingly, more dye-binding occurs with damaged proteins, which increases the fluorescence. This research shows that Eosin fluorescence could be used for quantification, however, it also indicates that it is possible that degradation of collagen might not result in an expected corresponding decrease of Eosin fluorescence, but rather an increase because of more exposed amino acid residues. This might explain the observed increase of Eosin over time when measured with absorption spectrometry; where fluorescence spectrometry gave no significant results, absorption spectrometry resulted in a strong negative correlation between Eosin intensity-ratios and PMI. A lower intensity correlates to higher levels of absorbance. Thus, an increase in PMI corresponds to an increase of absorbance at the wavelength of protein-bound Eosin (530 nm). However, when comparing Intensity-ratios from before and after staining with Eosin, there was no significant mean difference. This would imply that the addition of Eosin has no effect on the level of absorption of a cross-section. Yet, when visually comparing the obtained graphs, a distinct dip (at 530 nm) was only found to be present in the samples stained with Eosin. Why this difference would not show when comparing the intensity-ratios might be explained by the fact that, while measuring, any slight movement of the specimen would result in an increase/decrease of the intensity. Therefore, for future research it is advised to ensure that
each sample is placed under the fibre in such a way that the maximum intensity is equal in all samples.
Graph 7. shows the spectrofluorometric profile of porcine bone in an Excitation-Emission-Map (EEM). This map
demonstrates that at an excitation of 500 nm no fluorescence occurs in non-stained pig bone (Edited from van der Laan [41]).
Statistically, Nile Red quantity was not found to correlate to PMI. However, observations would suggest that specimens with a lower PMI which are not affected by brown discolouration show a stronger fluorescence of lipids when observed with a fluorescence microscope. In turn, stronger fluorescence was observed to correspond to a peak with a higher intensity (when measured with fluorescence spectrometry). Fig. 10 depicts a microscopic image of a sample with a PMI of 20 months, which shows a limited number of stained lipid droplets. Fluorescence measurements of this same sample resulted in very low intensity peaks. A sample with a PMI of 10 months is shown to have a considerably higher number of stained lipid spots. Measurements of this sample led to a more defined peak with a higher intensity. Yet, with the current setup, the relationship between the observed fluorescence and PMI cannot be proven statistically. Optimization experiments showed that a short incubation time (40 seconds) of Nile Red was preferred, since otherwise saturation would occur. However, Alemán-Nava and colleagues [42] suggest an incubation time of 10 minutes results in maximum fluorescence (in algae). A longer incubation time might have resulted in a higher intensity of the signal. Moreover, fluorescence of hydrophobic Nile Red is found to be quenched in water [43]. Perhaps, the water in which the sections were stored partly remained in the sections and interfered with the fluorescence. Another possibility is that most of the lipids have been extracted from the bone within the first ten months. This would explain why samples with a PMI of 13 months or higher barely showed a peak when measured with fluorescence spectrometry, and hardly showed the presence of lipid droplets when seen from under the fluorescence microscope. Closer examination of samples with a PMI shorter than 10 months might give conclusive results. If lipids are found to be dissolved within approximately 10
months, this element might proof itself to be useful for dating specimens with a post-mortem interval which is expected to be very short; e.g. a few months old.
A longer incubation time, drying of the sample, shorter time intervals and a larger sample size are advised for future research. Additionally, it is suggested to control the level of discoloration of the bone, to investigate whether the hypothesized relationship between discoloration and lipid-quantity truly exists.
Figure 10. The top left image shows specimen 26 (PMI = 20 months) excited by green light with a red filter. The
top right graph shows the emission graph when specimen 26 is excited at 550 nm at four locations indicated in Fig. 2. The bottom left image depicts sample 39 (PMI = 10 months) excited by green light with a red filter. The bottom right graph shows the corresponding emission graph when specimen 39 is excited at 550 nm at four locations.
Conclusion
The aim of the present study was to explore new methods to determine the PMI of skeletal remains, which was achieved by applying three different dyes to porcine bone samples with varying PMIs. The main research question was whether Sirius Red/ Fast Green, Eosin and Nile Red could be used to estimate the PMI of modern skeletal remains. Prior to answering this question, it was established that each of the dyes is successful at staining cross-sections of infant porcine bone.
Results obtained by staining with Sirius Red/ Fast Green gave a significant strong correlation with PMI but were in contradiction with previous studies. Yet, under experimental conditions Sirius Red/ Fast Green staining could still prove to be an informative method.
Absorption measurements of specimens stained with Eosin resulted in a strong correlation between Eosin abundance and PMI, yet, fluorescence measurements were not able to confirm this correlation. Nonetheless, the present study identifies Eosin staining measured with fluorescence spectrometry as a method with potential for the (semi-) quantification of proteins.
Nile Red staining is observed to correlate to PMI, however, this finding is not statistically supported. Moreover, there seems to be a relationship between lipid-quantity, discoloration of bone and UV-fluorescence, yet again, this is not supported by statistics.
Additionally, the Oxford Histological Index was found to be unfit for scoring thin sections of infant pig bones.
For future research, it is advised to reproduce the performed experiments with the use of a larger sample size and more/shorter time intervals, to investigate whether the found correlations and observations are also existent when measured at a larger scale. Moreover, one should focus on controlling environmental conditions, to minimize the potential influence thereof. Further research will aid in the understanding of bone degradation and assist in the creation of more accurate methods to estimate the PMI of modern remains.
List of References
[1] T.K. van den Hövel, investigating the forensic relevance of skeletal human remains by fluorescence spectroscopy ; pilot study for a portable device, unpublished work. [2] A. Cappella, D. Gibelli, E. Muccino, V. Scarpulla, E. Cerutti, V. Caruso, E. Sguazza,
D. Mazzarelli, C. Cattaneo, The comparative performance of PMI estimation in
skeletal remains by three methods (C-14, luminol test and OHI): analysis of 20 cases, Int. J. Legal Med. 132 (2018) 1215–1224. doi:10.1007/s00414-015-1152-z.
[3] C.A. Junod, Subaerial bone weathering and other taphonomic changes in a temperate climate, (2013) 114.
[4] G. McLaughlin, I.K. Lednev, Potential application of Raman spectroscopy for determining burial duration of skeletal remains, Anal. Bioanal. Chem. 401 (2011) 2511–2518. doi:10.1007/s00216-011-5338-z.
[5] F. Introna, G. Di Vella, C.P. Campobasso, Determination of postmortem interval from old skeletal remains by image analysis of luminol test results., J. Forensic Sci. 44 (1999) 535–8. doi:10.1520/jfs14505j.
[6] N. Hoke, A. Grigat, G. Grupe, M. Harbeck, Reconsideration of bone postmortem interval estimation by UV-induced autofluorescence, Forensic Sci. Int. 228 (2013) 176.e1-176.e6. doi:10.1016/j.forsciint.2013.03.013.
[7] F. Ramsthaler, S.C. Ebach, C.G. Birngruber, M.A. Verhoff, Postmortem interval of skeletal remains through the detection of intraosseal hemin traces. A comparison of UV-fluorescence, luminol, Hexagon-OBTI®, and Combur® tests, Forensic Sci. Int. 209 (2011) 59–63. doi:10.1016/j.forsciint.2010.12.011.
[8] K. Jellinghaus, C. Hachmann, K. Hoeland, M. Bohnert, U. Wittwer-Backofen, Correction to: Collagen degradation as a possibility to determine the post-mortem interval (PMI) of animal bones: a validation study referring to an original study of Boaks et al. (2014) (International Journal of Legal Medicine, (2018), 132, 3, (753-763), 10.1007/s00414-017-1747-7), Int. J. Legal Med. 132 (2018) 765.
doi:10.1007/s00414-018-1782-z.
[9] B. Knight, I. Lauder, Practical methods of dating skeletal remains: a preliminary study., Med. Sci. Law. 7 (1967) 205–208.
[10] H.P. Schwarcz, K. Agur, L.M. Jantz, A new method for determination of postmortem interval: Citrate content of bone, J. Forensic Sci. 55 (2010) 1516–1522.
doi:10.1111/j.1556-4029.2010.01511.x.
[11] R.E. Taylor, J.M. Suchey, L. a Payen, P.J. Slota, The use of radiocarbon (14C) to identify human skeletal materials of forensic science interest., J. Forensic Sci. 34 (1989) 1196–1205.
[12] Y. Fernández-Jalvo, P. Andrews, D. Pesquero, C. Smith, D. Marín-Monfort, B. Sánchez, E.M. Geigl, A. Alonso, Early bone diagenesis in temperate environments. Part I: Surface features and histology, Palaeogeogr. Palaeoclimatol. Palaeoecol. 288 (2010) 62–81. doi:10.1016/j.palaeo.2009.12.016.
[13] C.N. Trueman, D.M. Martill, The long-term survival of bone: The role of bioerosion, Archaeometry. 44 (2002) 371–382. doi:10.1111/1475-4754.t01-1-00070.
[14] B.B. Dent, S.L. Forbes, B.H. Stuart, Review of human decomposition processes in soil, Environ. Geol. 45 (2004) 576–585. doi:10.1007/s00254-003-0913-z.
[15] M. Buckley, C. Wadsworth, Proteome degradation in ancient bone: Diagenesis and phylogenetic potential, Palaeogeogr. Palaeoclimatol. Palaeoecol. 416 (2014) 69–79. doi:10.1016/j.palaeo.2014.06.026.
[16] M.J. collins, C.M. nielsen-marsh, J. hiller, C.I. smith, J.P. roberts, R. V PRIGODICH, T.J. wess, J. csapò, A.R. millard, G. turner-walker, the Survival of Organic Matter in Bone: a Review, Archaeometry. 44 (2002) 383–394.
[17] A. During, G. Penel, P. Hardouin, Understanding the local actions of lipids in bone physiology, Prog. Lipid Res. 59 (2015) 126–146. doi:10.1016/j.plipres.2015.06.002.
[18] E.A. Zimmermann, E. Schaible, B. Gludovatz, F.N. Schmidt, C. Riedel, M. Krause, E. Vettorazzi, C. Acevedo, M. Hahn, K. Puschel, S. Tang, M. Amling, R.O. Ritchie, B. Busse, Intrinsic mechanical behavior of femoral cortical bone in young, osteoporotic and bisphosphonate-treated individuals in low-and high energy fracture conditions, Sci. Rep. 6 (2016) 1–12. doi:10.1038/srep21072.
[19] A. Boaks, D. Siwek, F. Mortazavi, The temporal degradation of bone collagen: A histochemical approach, Forensic Sci. Int. 240 (2014) 104–110.
doi:10.1016/j.forsciint.2014.04.008.
[20] Z.R. Fischer AH, Jacobson KA, Rose J, Hematoxylin and eosin staining of tissue and cell sections., Cold Spring Harb Protoc. (2008).
[21] Z.J. Storms, E. Cameron, H. de la Hoz Siegler, W.C. McCaffrey, A Simple and Rapid Protocol for Measuring Neutral Lipids in Algal Cells Using Fluorescence, J. Vis. Exp. (2014) 1–7. doi:10.3791/51441.
[22] V. Marchiafava, A. Ascenzi, Fungal ostcoclasia: a model of dead bone resorption, Calcif. Tissue Res. 14 (1974) 195–210.
[23] H.H. de Boer, M.J. Aarents, G.J.R. Maat, Staining ground sections of natural dry bone tissue for microscopy, Int. J. Osteoarchaeol. 22 (2012) 379–386.
doi:10.1002/oa.1208.
[24] V. Hedges, R., Millard, A. y Pike, Measurement and relationship of diagenetic alteration of bone from three archaeological sites., J. Archaeol. Sci. 22 (1995) 201– 209. doi:10.1006/jasc.1995.0022.
[25] H. Ali, S. Ali, M. Mazhar, A. Ali, A. Jahan, A. Ali, Eosin fluorescence: A diagnostic tool for quantification of liver injury, Photodiagnosis Photodyn. Ther. 19 (2017) 37–44. doi:10.1016/j.pdpdt.2017.03.016.
[26] C.H. Bachman, Fluorescence of bone, Nature. 208 (1965) 239–241.
[27] J.W. Kuiper, C. Van Kuijk, J.L. Grashuis, A.G.H. Ederveen, H.E. Schütte, Accuracy and the influence of marrow fat on quantitative CT and dual-energy X-ray
absorptiometry measurements of the femoral neck in vitro, Osteoporos. Int. 6 (1996) 25–30. doi:10.1007/BF01626534.
[28] P. Greenspan, E.P. Mayer, S.D. Fowler, Nile red: A selective fluorescent stain for intracellular lipid droplets, J. Cell Biol. 100 (1985) 965–973.
doi:10.1083/jcb.100.3.965.
[29] J.D. Evans, Straightforward Statistics for the Behavioral Sciences, Brooks/ Cole Publ. Pacific Grove. (1996).
[30] A.A. Waheed, K.S. Rao, P.D. Gupta, Mechanism of dye binding in the protein assay using eosin dyes, Anal. Biochem. 287 (2000) 73–79. doi:10.1006/abio.2000.4793. [31] L. Feng, I. Jasiuk, Multi-scale characterization of swine femoral cortical bone, J.
Biomech. 44 (2011) 313–320. doi:10.1016/j.jbiomech.2010.10.011. [32] A.& S.R. Smith, S.Y., Varela, Bone toxicology, Springer Int. Publ. (2017).
[33] J. Aerssens, S. Boonen, G. Lowet, J. Dequeker, Interspecies Differences in Bone Composition, Density, and Quality: Potential Implications for in Vivo Bone Research 1,
Endocrinology. 139 (1998) 663–670. doi:10.1210/endo.139.2.5751.
[34] F.M. Nor, R.F. Pastor, H. Schutkowski, Histological study to differentiate between human and non-human long bone, Int. Med. J. 22 (2015) 413–417.
[35] M.M.E. Jans, C.M. Nielsen-Marsh, C.I. Smith, M.J. Collins, H. Kars, Characterisation of microbial attack on archaeological bone, J. Archaeol. Sci. 31 (2004) 87–95. doi:10.1016/j.jas.2003.07.007.
[36] I. Kontopoulos, P. Nystrom, L. White, Experimental taphonomy: post-mortem microstructural modifications in Sus scrofa domesticus bone, Forensic Sci. Int. 266 (2016) 320–328. doi:10.1016/j.forsciint.2016.06.024.
[37] W.L. Todt, Bone histology, (2017).
http://wwwp3.cord.edu/faculty/todt/336/lab/bone/index.htm (accessed August 8, 2018).
[38] A.M. Child, A. Bousher, M. Chandra, R. Edyvean, Towards an understanding of the decomposition of bone in the archaeological environment, Biodeterior. Biodegrad. 9.
Leeds, UK. (32676BC) UK-10.
[39] L.S. Bell, M.F. Skinnerb, S.J. Jonesa, lforeilslc Science Illterll ~ tii The speed of post mortem change to the human skeleton and its taphonomic significance, 82 (1996) 129–140. doi:10.1016/0379-0738(96)01984-6.
[40] C.I. Smith, C.M. Nielsen-Marsh, M.M.E. Jans, M.J. Collins, Bone diagenesis in the European Holocene I: patterns and mechanisms, J. Archaeol. Sci. 34 (2007) 1485– 1493. doi:10.1016/j.jas.2006.11.006.
[41] L. Van Der Laan, Oxidatie van lipiden vanuit de botmatrix, unpublished work. [42] G.S. Alemán-Nava, S.P. Cuellar-Bermudez, M. Cuaresma, R. Bosma, K. Muylaert,
B.E. Ritmann, R. Parra, How to us Nile Red, a selective fluorescent stain for microalgal neutral lipids, J. Microbiol. Methods. 128 (2016) 74–79.
doi:10.1016/j.mimet.2016.07.011.
[43] P. Fowler, Standley D. and Greenspan, Application of Nile Red, a fluorescent hidrophobic probe, for the Detection of Neutral Lipid Deposits in Tissue Section., J. Histochem. Cytochem. 33 (1985) 833–836.
APPENDIX I
Below nine bone samples with varying PMI are shown. The samples are stained with Sirius Red/ Fast Green as described by the protocol in Methods. A Leica DM1000 microscope with camera was used to photograph the section (magnification 40x).
APPENDIX II
Below nine bone samples with varying PMI are shown. The samples are stained with Eosin as described by the protocol in Methods. A Leica DM1000 microscope with camera was used to photograph the sections (magnification 40x, apart from the sample with a PMI of 0; these sections were magnified 100x).
APPENDIX III
Below nine bone samples with varying PMI are shown. The samples are stained with Nile Red as described by the protocol in Methods. A 100x magnified image was captured using a Nikon
Eclipse E600 fluorescence microscope emitting green light and a build-in DS-Fi2 camera with a red filter.
