Human Bone Proteomes before and after Decomposition: Investigating the Effects of Biological Variation and Taphonomic Alteration on Bone Protein Profiles and the Implications for Forensic Proteomics

Abstract

Bone proteomic studies using animal proxies and skeletonized human remains have delivered encouraging results in the search for potential biomarkers for precise and accurate post-mortem interval (PMI) and the age-at-death (AAD) estimation in medico-legal investigations. The development of forensic proteomics for PMI and AAD estimation is in critical need of research on human remains throughout decomposition, as currently the effects of both inter-individual biological differences and taphonomic alteration on the survival of human bone protein profiles are unclear. This study investigated the human bone proteome in four human body donors studied throughout decomposition outdoors. The effects of ageing phenomena (in vivo and post-mortem) and intrinsic and extrinsic variables on the variety and abundancy of the bone proteome were assessed. Results indicate that taphonomic and biological variables play a significant role in the survival of proteins in bone. Our findings suggest that inter-individual and inter-skeletal differences in bone mineral density (BMD) are important variables affecting the survival of proteins. Specific proteins survive better within the mineral matrix due to their mineral-binding properties. The mineral matrix likely also protects these proteins by restricting the movement of decomposer microbes. New potential biomarkers for PMI estimation and AAD estimation were identified. Future development of forensic bone proteomics should include standard measurement of BMD and target a combination of different biomarkers.

Keywords: age-at-death estimation; bone mineral density; forensic microbiology; forensic proteomics; forensic taphonomy; human decomposition; post-mortem interval estimation.

Figure 1

STRING protein network of the quantifiable proteins extracted from all samples. Immunoglobulin proteins (gene names IGHA1, IGHG2, IGHG3, IGKC, and IGLC2) were not found with STRING and are not represented in the figure. The light orange ellipse represents collagenous and collagen-binding proteins, the red one represents plasma and bone-related proteins, the yellow-green one at the top on the left side represents ubiquitous proteins, and the light green one at the top on the right side represents some muscle proteins. Other smaller clusters represent other types of proteins interacting less with the major clusters identified.

Figure 2

(A) Number of proteins extracted from each sample. Samples were grouped according to the bone type. All bone types were significantly different from each other (post-hoc pairwise Wilcoxon-test with corrections for multiple testing). Outliers are represented as pointed dots in the plot [two outliers identified here, one for iliac fresh (sample NP14, see Supporting Information, Table 2 for details) and one for tibia iliac skeletonized group (sample NP45, see Supporting Information, Table 2 for details)]. (B) Number of proteins extracted from fresh samples. Samples were grouped according to the donor. None of the donors resulted in being significantly different from each other (post-hoc pairwise Wilcoxon-test with corrections for multiple testing, p value >0.05).

Figure 3

Venn diagram to represent STRING protein networks of proteins significantly more abundant in fresh iliac samples (left) and fresh tibia samples (right) than in their skeletonized counterparts. Proteins shared between the two categories are represented in the middle. Immunoglobulin proteins (gene names IGHA1, IGHG2, IGHG3, IGKC, and IGLC2) were not found with STRING and are not represented in the figure. In the iliac crest category, red cluster represents plasma proteins, yellow cluster represents collagens and bone-related proteins, and green cluster represents ubiquitous proteins. No obvious clusters were identified for the shared proteins and for the ones belonging to the tibia category.

Figure 4

Number of proteins extracted from skeletonized samples, grouped by (A) depositional environment or (B) placement season. No significant differences were detected (Wilcoxon and Kruskal–Wallis p value >0.05).

Figure 5

Abundance of (A) CO3A1, (B) CO9, (C) COBA2, (D) MGP, (E) PGS2, and (F) TTHY protein in iliac crest-skeletonized samples and of (G) CO3 in tibia-skeletonized samples with increasing PMIs. Groups are significantly different from each other (ANOVA p value <0.05).

Figure 6

Relative abundance of fetuin-A in (A) fresh tibia, (B) skeletonized tibia, (C) fresh iliac crest, and (D) skeletonized iliac crest samples, of albumin in (E) fresh tibia, (F) skeletonized tibia, (G) fresh iliac crest, and (H) skeletonized iliac crest samples and of (I) olfactomedin like-3 in skeletonized iliac crest samples, arranged by the chronological age of the donors. ANOVA p value was reported for each plot. Only (A,D,G,H,I) resulted in being statistically significant.