Forensic Sci Med Pathol
DOI 10.1007/s12024-017-9869-2
ORIGINAL ARTICLE
Forensic aspects of gene expression signatures for age
determination in bruises as evaluated in an experimental
porcine model
Kristiane Barington 1 & Henrik Elvang Jensen 1 & Kerstin Skovgaard 2
Accepted: 31 March 2017
# Springer Science+Business Media New York 2017
Abstract Determining the age of bruises and the force used to
inflict the trauma is of crucial importance in both human and
veterinary forensic pathology. In the present study, the expres-
sion of more than 50 different genes in subcutaneous fat and
muscle tissue from experimental bruises in pigs was investi-
gated. The aim was to evaluate if expression signatures of
selected genes were capable of determining bruises according
to age and the force of impact. Eighteen experimental pigs
were anesthetized, and on each animal four blunt traumas
were inflicted on the back with a low, moderate or high force.
The pigs were euthanized from 1 to 10 h after infliction of the
trauma and subcutaneous fat and muscle tissues were sam-
pled. As control, subcutaneous fat and muscle tissues were
sampled from two un-injured pigs. Quantitative real-time po-
lymerase chain reaction was performed to evaluate mRNA
expression of genes involved in inflammation, tissue damage
and repair. Expression signatures of thirteen selected genes in
subcutaneous fat but not in muscle tissue reflected the age of
bruises with a precision of approximately ±2 h. Moreover, the
gene expression signature in the subcutaneous fat was to some
extend able to separate bruises inflicted with different forces.
Expression signatures of selected genes in the subcutaneous
Electronic supplementary material The online version of this article
(doi:10.1007/s12024-017-9869-2) contains supplementary material,
which is available to authorized users.
* Kristiane Barington
krisb@sund.ku.dk
1
Faculty of Health and Medical Sciences, University of Copenhagen,
Ridebanevej 3, DK-1870 Frederiksberg C, Denmark
2
Section for Immunology and Vaccinology, National Veterinary
Institute, Technical University of Denmark, Kemitorvet Kgs,
DK-2800 Lyngby, Denmark
fat will increase the precision of the age determination of
bruises in pigs. Further, due to the similarity of porcine and
human skin physiology and immunity, these results might also
provide valuable information in human forensic science.
Keywords Age estimation . Bruise . Force estimation . Gene
expression signature . qPCR
Introduction
The age assessment of bruises is of crucial importance in both
human and veterinary forensic pathology. All though various
methods for age estimation have been evaluated in animals
and humans, obtaining an accurate age of bruises continues to
be a challenge [1–6].
In forensic cases regarding bruises in pigs and humans,
lesions often have a tram-line pattern, characterized by two
parallel hemorrhages separated by a zone of undamaged skin
[7–9]. Histopathological evaluation of experimental bruises in
pigs is able to distinguish between bruises being more or less
than 4 h old [10]. However, more precise age estimation is
often requested as most bruises in pigs are assumed to be
inflicted within a timeframe of 8 h prior to slaughter, during
which period the pigs are handled by more people [8].
Moreover, histological changes in bruises not only depend
on the age but also on the force used to inflict the lesions
[11]. Due to the similarity between humans and pigs regarding
the appearance of bruises, skin anatomy, physiology and in-
flammatory reactivity (immune response), a porcine model of
bruises might also be highly relevant to improve prediction of
bruise age in humans [10, 13].
During recent years, the expression of mRNA in skin or
muscle tissue from experimental bruises in rats and mice has
been evaluated [14–16]. In these studies, the amount of

mRNA coding for tissue type plasminogen activator, skeletal
troponin 1 and sodium-coupled neutral amino acid transporter
showed a time dependent response [14–16]. Therefore, we
hypothesize that changes in mRNA expression profiles of
selected inflammatory and other genes involved in the innate
immune response to tissue damage and repair are likely to
further improve the assessment of the age of bruises and force
of impact in a porcine model.
In the present study, the expression of more than 50 differ-
ent genes involved in inflammation, tissue damage and repair
in subcutaneous fat and muscle tissue from experimental
bruises in pigs was investigated. The aim was to evaluate if
expression signatures of a small number of carefully selected
genes were capable of determining bruises according to age
and the force of impact.
Gene expression signatures in the subcutaneous fat but not
in muscle tissue were found to reflect the age of bruises with a
precision of approximately ±2 h. To some extent gene expres-
sion signatures were also able to separate bruises inflicted with
different forces.
Materials and methods
Animals
In total, 20 female Yorkshire-Landrace crossbred (18 experi-
mental and 2 controls) specific pathogen free pigs, with a
mean body weight of 31 kg (23–38 kg), were randomly divid-
ed into four groups (Table 1). The pigs were fed a commercial
pig diet (NAG, Helsinge, Denmark) twice a day and had free
access to tap water. They were acclimatized for one week and
housed in pens of two animals. The study and the experimen-
tal procedure were approved by the Danish Animal
Inspectorate (2013–15–2934 − 00849).
Experimental procedure
A full description of the experimental procedures has recently
been given [10, 11]. In summary, pigs were sedated with an
intramuscular injection of a mixture of tiletamin and
zolazepam 125 mg (Zoletil 50 Vet.; Virbac, Animal Health,
Carros, France), xylazin 20 mg/mL (Xysol vet.; ScanVet
Animal Health A/S, Fredensborg, Denmark), ketamine
100 mg/mL (Ketaminol Vet; Intervet International BV,
Table 1 Summary of force of
impact, number of pigs and age of Group Force of impact
bruises in the three experimental
groups and the control group. One 1 Low (2.14 N ± 0.22 N)
pig was euthanized at each of the 2 Moderate (4.24 N ± 0.16 N)
given time points 3 High (6.52 N ± 0.17 N)
Control No trauma
Forensic Sci Med Pathol
Holland) and butorphanoltartrat 10 mg/mL (Torbugesic Vet;
ScanVet Animal Health A/S, Fredensborg, Denmark). They
were maintained in anesthesia with inhalation of isoflurane
gas (Isoflurane, Vetflurane, Virbac, Animal Health, Carros,
France) and a continuous intravascular infusion of fentanyl
( F e n t a n y l 50 μ g / m L , F e nt an y l - H am el n; H a m e l n
Pharmaceuticals gmbh, Hameln, Germany) and midazolam
(Midazolam 5 mg/mL, Hameln Pharmaceuticals gmbh,
Hameln, Germany). While anesthetized, four blunt traumas
(area of impact Nos. 1, 2, 3 and 4) were inflicted on the back
of each of the 18 experimental pigs (Fig. 1). The four bruises
were inflicted over 3 to 4 min using a mechanical device [10,
11]. The pigs were divided into three groups according to the
force used to inflict bruises. In groups 1, 2 and 3, blunt
traumas were inflicted with a force ± standard deviation of
2.14 N ± 0.22 N (low force), 4.24 N ± 0.16 N (moderate
force), and 6.52 N ± 0.17 N (high force), respectively. After
infliction of the traumas, pigs were left in anesthesia and eu-
thanized after 1 to 10 h with an overdose of pentobarbital
(Glostrup Apotek, Glostrup, Denmark) given intravenously
(Table 1). The control pigs (without blunt trauma) were anes-
thetized for 6 h and then euthanized. Postmortem, subcutane-
ous fat and the underlying muscle tissue (0.5 × 0.5 × 0.5 cm)
from the bruises were sampled and preserved in 4.5 mL of
RNAlater [12], stored at 5 °C for 24 h. Samples were then
stored at −20 °C until RNA extraction. From the back of the
control pigs, subcutaneous fat and muscle tissues were sam-
pled from the corresponding areas (Nos. 1 to 4) of the exper-
imental pigs. In addition, uninjured subcutaneous fat and un-
derlying muscle tissues were sampled from the thighs of 8
experimental pigs and 1 control pig. The control tissues were
stored following the same protocol as was used for the tissue
underlying the bruises of the 18 experimental pigs.
Extraction, quantitation and quality assessment of RNA
Tissue samples of 100 mg were disrupted and homogenized in
QIAzol Lysis reagent (Qiagen, Hilden, Germany) using a
gentleMACS Dissociator (Miltenyi Biotec, Lund, Sweden).
RNA was extracted from the homogenized tissue using the
RNeasy Lipid Tissue Mini Kit (Qiagen, Hilden, Germany)
and all samples were treated with the RNAase-Free DNAase
set (Qiagen, Hilden, Germany) according to the manufac-
turer’s instructions. The purity of RNA was evaluated at opti-
cal density (OD) 260/280 and OD 260/230 ratios, and the
No. of pigs Age of bruises
4 2, 4, 6 and 8 h
4 2, 4, 6 and 8 h
10 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10 h
2
Forensic Sci Med Pathol
Fig. 1 Bruises on pig skin inflicted with the highest force. Bruises were
located on the back of the pig in area of impact Nos. 1 to 4. Inset: Cross-
section of a bruise. Hemorrhages are seen in the subcutis and muscle
tissue
concentration of RNA was quantified at OD 260 measured on
a Nanodrop ND-1000 spectrophotometer (Saveen and Werner
AB, Lindhamn, Sweden). The quality of RNA was assessed
by measuring the RNA integrity number (RIN) on an Agilent
2100 Bioanalyzer (Agilent Technologies, Glostrup, Denmark)
using the RNA 6000 Nano Kit (Agilent Technologies,
Glostrup, Denmark). RNA was stored at −80 °C until convert-
ed into cDNA.
cDNA synthesis
From each sample, duplicate cDNA syntheses were made
from 100 ng extracted RNA using the QuantiTect Reverse
Transcription Kit (Qiagen, Hilden, Germany) according to
the manufacturer’s instructions. In addition, non-reverse-
transcriptase controls were made to assess potential DNA con-
taminations. cDNA was diluted 1:9 in Low EDTA TE-Buffer
(VWR-Bie & Berntsen) prior to pre-amplification.
Primer design
Distribution and lengths of introns and exons within genes of
interest were found using e!Ensembl (http://www.ensembl.
org/Sus_scrofa/Info/Index), and primers were designed to
span intron/exon boundaries using Primer3 version 0.4.0
(http://bioinfo.ut.ee/primer3-0.4.0/) if possible and checked
for interspecies-variation using BLAST (http://blast.ncbi.
nlm.nih.gov/). Primers were synthesized at Sigma Aldrich.
Genes, primer sequences and amplicon length can be seen in
Online Resource 1.
Pre-amplification
Pre-amplification was carried out using 5 μL TaqMan PreAmp
Master Mix (Applied Biosystems, Foster City, CA, USA),
2.5 μL primer mix (200 nM) and 2.5 μL diluted cDNA.
Stocks of 200 nM primer pair mix were prepared combin-
ing equal amounts of all primers in low EDTA TE-buffer
(VWR- Bie & Berntsen). Primers, cDNA and master mix
were incubated in the Thermocycler at 95 °C for 10 min,
then 19 cycles of 95 °C for 15 s and 60 °C for 4 min. Pre-
amplified cDNA was incubated with 4 μL of 4 U/μL exo-
nuclease (New England Biolabs) 30 min at 37 °C followed
by 15 min at 80 °C before being diluted 1:9 in low EDTA
TE-buffer (VWR- Bie & Berntsen).
Quantitative PCR
Quantitative real-time polymerase chain reaction (qPCR)
was performed to evaluate mRNA expression of the genes
of interest and the reference genes listed in Online
Resource 1. qPCR of the preamplified cDNA was per-
formed in a Dynamic Array 96, 96 (Fluidigm, San
Francisco, CA, USA) combining 96 preamplified samples
with 96 primer sets for 9216 individual and simultaneous
qPCR reactions. Sample mix for each of the 96 reactions
consisted of 3 μL TaqMan Gene Expression Master Mix
(Applied Biosystems), 0.3 μL 20X DNA Binding Dye
Sample Loading Reagent (Fluidigm), 0.3 μL 20X
EvaGreen (Biotium, VWR – Bie & Berntsen), 0.9 μL
low EDTA TE-buffer (VWR – Bie & Berntsen) and
1.5 μL preamplified cDNA. The primer mixes consisted
of 3 μL 2X Assay loading reagent (Fluidigm) and 3 μL
of 20 μM forward and reverse primers. Control line fluid
was injected into the 96.96 Dynamic array according to the
manufacturer’s instructions (Fluidigm) and primed in the
HX IFC controller (Fluidigm). Thereafter, 4.9 μL sample
mix and 4.9 μL primer mix were dispensed into the appropri-
ate inlets in the 96.96 Dynamic array and then reloaded into
the HX IFC controller (Fluidigm) combining each of the 96
samples with the 96 primers in 9216 separate reaction cham-
bers. The 96.96 Dynamic array was subsequently placed in the
BioMark real-time PCR instrument (Fluidigm), and the fol-
lowing cycling parameters were used: 2 min at 50 °C, 30 min
at 80 °C for thermal mix, 2 min at 50 °C, 10 min at 95 °C,
followed by 35 cycles with denaturing for 15 s at 95 °C and
annealing/elongation for 1 min at 60 °C. Melting curves were
generated after each run (from 60 °C to 95 °C, increasing 1 °C/
3 s).
Data analysis and statistics
Preprocessing data: Three highly stable samples were used as
interplate calibrators to compensate for variation between the
dynamic chips used, and then data were corrected for PCR
efficiency for each primer individually. For the subcutaneous
tissue, the reference genes ACTB, B2M, HPRT1, PPIA,
RPL13A and TBP were the most stable, while for the muscle
tissue ACTB, B2M, GAPDH, HPRT1, RPL13A, TBP and
YWHAE were the most stable found using GeNorm [17]
and NormFinder [18]. The geometric mean of the reference

genes for each tissue was used to normalize all samples in
GenEx5 (MultiD Analyses AB, Sweden). The average of the
technical repeats of cDNA was calculated. Technical repeats
of cDNA deviating more than 1.5 Cq for more than 15% of the
samples were excluded from the data analysis. If data were
missing for more than 20% of the samples, the primer assay
was excluded from further analysis. Thereafter, the Cq values
were transformed to a linear scale (the relative expression for
all samples was calculated relative to the sample with the
lowest expression for each primer) and then log2 transformed
prior to statistical analysis.
If double primers (annealing at different sites at the mRNA
transcript) for a single mRNA sequence were analyzed, the
correlation between these was calculated. If the correlation
was non-significant, both primers were excluded (1 out of 43
double primers). For the remaining double primers (A or B),
the primer assay with most similar cDNA replicates, least miss-
ing values, and best PCR efficiency was included for further
analysis.
PCA and PLS: Before principal component analysis (PCA)
using LatentiX 2.12 (Frederiksberg, Denmark), the average
gene expression (log2 transformed values) in the four bruises
(area of impact 1 to 4) was calculated for the subcutaneous
tissue and the muscle tissue for each pig. In addition, the
average gene expressions in subcutaneous fat and muscle tis-
sue from the four corresponding areas in each of the two
control pigs were also calculated. Normal subcutaneous fat
and muscle tissues from the thighs were included as single
samples, i.e. no average calculated. Gene expression data were
autoscaled by multiplication with the inverse standard error
and mean centered before PCA. For each of the tissue types
(subcutis and muscle), a PCA including all pigs based on data
from all the genes of interest was made. Further a PCA in-
cluding samples of subcutaneous tissue from the high force of
impact (group 3) was prepared (Table 1). A model to predict
the age of bruises was made by partial least square regression
(PLS) analysis with full cross validation (LatentiX 2.12) of the
subcutaneous gene expression data from pigs in group 3
(Table 1). In the prediction-model, only genes with a regres-
sion coefficient above 0.04 or below −0.04 and with 95%
confidence limits, not including 0, were selected using
SIMCA (Malmö, Sweden). The expression of the selected
genes was used as independent variables, and the age of
bruises as the dependent variable.
Relative gene expression: Pigs with bruises inflicted with
the highest force (group 3) were further divided into three
intervals of age: 1 to 3 h, 4 to 6 h and 7 to 10 h, and the
average expression of the genes in the subcutaneous tissue
was calculated. Then the relative expression of genes in the
three age intervals was calculated relatively to the expression
in the control pigs scaled to one.
Comparison of gene expression according to area of im-
pact: Difference in gene expression between bruises inflicted
Forensic Sci Med Pathol
in area of impact 1 to 4 (Fig. 1) of all 18 experimental pigs was
analyzed by a one way ANOVA, followed by post-hoc Turkey
Kramer test in GenEx5. A p-value below 0.05 and a fold
change (relative quantities) of more than 2 between at least
two areas of impact were considered as a statistical and bio-
logical significant difference in gene expression, respectively.
Results
Subcutaneous tissue
The average gene expression (log2 transformed values) in the
subcutaneous fat from the four areas on the back of experi-
mental and control pigs are presented in Online Resource 2.
PCA including all pigs and expression data from 42 genes
Subcutaneous tissue underlying bruises clearly separated from
subcutaneous tissue from the back of control pigs and normal
subcutaneous tissue from the thighs of pigs, along the first
principal component. The subcutaneous tissue underlying
the bruises in the experimental pigs separated according to
force of impact, with the bruises inflicted with high force
being furthest away from the controls (Fig. 2). Genes of im-
portance (highest loadings) with regards to separation of
bruises inflicted with the highest force in the PCA of all ex-
perimental pigs included several proinflammatory cytokines
as IL1B, IL1RN, IL6, IL8, and IL18 but also CCL2, CFB,
HMOX1, and TGFB.
PCA including pigs from group 3 and expression data
from 42 genes
Subcutaneous tissue from pigs with bruises inflicted with the
highest force of impact (group 3) separated according to the
age of the bruises (Fig. 3). Bruises inflicted within 1 to 3 h
clearly separated from older bruises. Older bruises could fur-
ther be divided into two age intervals: 4 to 6 h and 7 to 10 h
(Fig. 3).
Relative gene expression
Table 2 shows the mean expression of all 42 genes for the
three age intervals (1 to 3 h, 4 to 6 h and 7 to 10 h) relative
to the mean expression in subcutaneous tissue from the control
pigs scaled to one. In Table 2 several types of expression
patterns can be seen, e.g. IL8, IL1RN and CCL2 were
highly upregulated during the entire experimental period
while PTGS2, FOS and SELE were highly upregulated
through the first 3 h and then rapidly declined.
Forensic Sci Med Pathol
Fig. 2 Principal component analysis (PCA) including all pigs and
expression data from 42 genes in the subcutaneous fat: control thigh
(light green), control back (green*), low force (blue), medium force
(yellow), high force (red). Each square represents the average of four
PLS including pigs from group 3 and 13 genes
The prediction-model was able to predict the age of bruises with a
root mean square error of 1.06 h, meaning that in 95% of bruises,
the prediction-model was able to determine the age with a preci-
sion of approximately ±2 h (Fig. 4). The model had a standard
error of precision of 1.11, a bias of 0.09 and a correlation
coefficient of 0.87. In the first PLS-component the explained
Fig. 3 Principal component analysis (PCA) including pigs from group 3
(high impact) and expression data from 42 genes, subcutaneous fat. Each
square represents the average of four bruises on the back of a pig. Bruises
bruises on the back of a pig, with the exception of the light green
squares that represents single samples of normal subcutaneous fat from
the thighs of pigs
variance in Yand X was 90% and 78%, respectively. The model
was based on the mRNA expression of the following genes:
APOA1, CCL2, CFD, FOS, ICAM1, IFNA1, IL6, NFKB1,
PLAT, PTGS2, SELE, SELP and TNFAIP3. All genes included
in the model had negative regression coefficients (with 95%
confidence intervals not including 0) meaning that the mRNA
expression was up-regulated at first and then declined during the
following 10 h after impact.
separated according to bruise age into three intervals: 1 to 3 h (blue
circle), 4 to 6 h (green circle), 7 to 10 h (red circle)
 Forensic Sci Med Pathol

Table 2 The mean expression of

42 genes after high force impact Control (n = 2) 1 to 3 h (n = 3) 4 to 6 h (n = 3) 7 to 10 h (n = 4)
for the three age intervals 1 to 3 h,
4 to 6 h and 7 to 10 h relative to Gene RE SEM RE SEM RE SEM RE SEM
the mean expression in APOA1 1.00 0.04 0.86 0.06 0.64 0.08 0.43 0.02
C3 1.00 0.07 1.53 0.57 1.20 0.46 0.48 0.05
subcutaneous tissue from the
CASP1 1.00 0.09 0.63 0.19 1.46 0.37 0.69 0.16
control pigs scaled to one. RE, CCL2(B) 1.00 0.00 21.26 2.50 17.80 3.97 9.81 2.11
relative expression; SEM, CCL3 1.00 0.05 3.02 0.86 1.89 0.41 1.56 0.40
standard error of the mean CCL5 1.00 0.16 0.61 0.28 0.67 0.05 0.70 0.20
CD163 1.00 0.02 1.73 0.46 1.63 0.36 1.55 0.68
CFB 1.00 0.38 1.30 0.23 1.62 0.46 0.89 0.06
CFD 1.00 0.02 0.85 0.11 0.50 0.02 0.40 0.03
CLDN1 1.00 0.27 0.87 0.13 1.05 0.61 1.82 0.82
CXCL10 1.00 0.36 0.44 0.18 0.28 0.07 0.11 0.01
FGF2 (B) 1.00 0.27 1.87 0.59 0.65 0.06 0.62 0.13
FN1 (B) 1.00 0.17 1.01 0.19 0.74 0.12 0.51 0.07
FOS (A) 1.00 0.37 10.94 6.68 1.08 0.08 0.61 0.10
HMOX1 (B) 1.00 0.27 7.34 2.53 5.55 0.96 2.60 0.20
ICAM-1 (A) 1.00 0.06 8.71 2.38 1.67 0.27 1.13 0.10
IFNA1 (A) 1.00 0.01 0.46 0.01 0.32 0.01 0.23 0.02
IL18 1.00 0.05 1.25 0.42 2.15 0.17 1.65 0.65
IL1B (A) 1.00 0.06 16.62 3.13 24.44 11.89 6.00 3.45
IL1RN 1.00 0.02 26.79 3.56 44.17 8.80 40.66 15.46
IL33 1.00 0.21 0.83 0.11 0.76 0.08 0.69 0.10
IL6 (B) 1.00 0.31 45.40 16.44 12.67 1.76 5.50 1.21
IL8 1.00 0.41 18.33 10.09 16.69 7.25 23.21 5.04
MAPK9 (A) 1.00 0.11 0.95 0.19 0.51 0.02 0.59 0.04
MMP2 (B) 1.00 0.21 0.88 0.12 0.50 0.05 0.42 0.04
NFKB1 1.00 0.07 1.49 0.19 0.97 0.04 0.77 0.00
NFKBIA 1.00 0.06 3.72 0.84 1.35 0.12 1.66 0.10
OCLN 1.00 0.19 0.80 0.23 0.40 0.12 0.52 0.05
PLAT (B) 1.00 0.12 1.47 0.07 0.60 0.06 0.58 0.07
PTGS2 (B) 1.00 0.02 9.25 1.99 2.07 0.34 1.08 0.32
SELE 1.00 0.34 76.59 52.36 7.31 4.39 2.79 2.01
SELP (A) 1.00 0.19 5.91 0.62 1.82 0.44 0.99 0.30
TF 1.00 0.19 1.16 0.44 0.52 0.12 0.94 0.18
TGFB (B) 1.00 0.02 2.38 0.43 2.65 0.11 1.63 0.04
TJP1 (A) 1.00 0.21 0.61 0.03 0.33 0.02 0.34 0.02
TNF (A) 1.00 0.12 2.28 0.51 1.92 0.59 0.81 0.33
TNFAIP3 1.00 0.02 1.37 0.28 0.77 0.04 0.50 0.04
TNNC2 (A) 1.00 0.43 62.86 43.25 1.62 0.67 20.14 18.35
TNNI2 (A) 1.00 0.16 35.78 23.75 1.58 0.47 13.07 11.37
VCAM1 (A) 1.00 0.15 0.96 0.16 1.02 0.09 0.71 0.09
VEGFA 1.00 0.18 1.78 0.38 0.59 0.01 0.69 0.15
VEGFB 1.00 0.16 0.99 0.12 0.44 0.01 0.49 0.01

Comparison of gene expression according to area
of impact
No significant difference in mRNA expression between bruises
inflicted in area of impact 1 to 4 was seen in subcutaneous tissue.
the highest force and muscle tissue from control pigs and
normal muscle tissue from the thighs of pigs was seen along
the first principal component (Fig. 5).
PCA including group 3 and 43 genes

Muscle tissue
The average gene expression (log2 transformed values) in the
muscle tissue from the four areas on the back of experimental
and control pigs are presented in Online Resource 3.
PCA including all pigs and 43 genes
Muscle tissue from the majority of pigs with bruises inflicted
with the moderate and minimum force of impact could not be
separated from the controls (Fig. 5). Some degree of separa-
tion between muscle tissue underlying bruises inflicted with
In pigs with bruises inflicted with the highest force of impact
(group 3) no separation according to bruise age was seen
(Online Resource 4).
Comparison of gene expression according to the four
areas of impact
Difference in mRNA expression between bruises inflicted in
area of impact 1 to 4 (Fig. 1) was seen for some genes
expressed in the muscle tissue: mRNA coding for TNNC1
(A), TNNT1 (B), TNNI1 (A) was significantly up-regulated
in area of impact no. 1 compared to the remaining three areas
Forensic Sci Med Pathol
Fig. 4 Partial least square regression analysis (PLS) including pigs from
group 3 (high impact) and 13 genes, subcutaneous fat. Each square
represents the average of four bruises on the back of a pig. The vertical
in the experimental pigs. This difference of gene expression in
areas of impact inflicted within 3–4 min was not seen in the
corresponding areas in the control pigs (Fig. 6). The amount of
mRNA coding for FOS (A) increased significantly from area
of impact 1 to 4. However, a tendency to this was also seen in
the control pigs (Fig. 6).
Fig. 5 Principal component analysis (PCA) including all pigs and 43
genes in the muscle tissue: control thigh (light green), control back
(green*), low force (blue), medium force (yellow), high force (red).
axis is the actual age (h) of the bruises and the horizontal axis is the age (h)
predicted by the model
Discussion
The prediction-model was able to predict the age of bruises
with a precision of approximately ±2 h. Moreover, PCA in-
cluding all genes was able to divide pigs inflicted with the
highest force into three age intervals (1 to 3 h, 4 to 6 h and 7
Each square represents the average of four bruises on the back of a pig,
with the exception of the light green squares that represents single
samples of normal muscle tissue from the thighs of pigs

Fig. 6 Comparison of gene
expression in muscle tissue
according to the four areas of
impact. Relative expression of
TNNC1, TNNT1, TNNI1 and
FOS in area Nos. 1 to 4 in pigs
with bruises (bruise) and control
pigs (control). The expression of
each gene in each area (Nos. 1 to
4) was calculated relative to the
expression in area No. 4 in the
control pigs scaled to one
to 10 h). Histological evaluation of the same bruises was able
to divide bruises as being more or less than 4 h [10].
Therefore, combining histological evaluation and gene ex-
pression data will increase the precision of age estimations
considerably. This is highly relevant as the majority of bruises
in pigs are assumed to be inflicted within a timeframe of 8 h
before slaughter [8].
The prediction-model was based on the mRNA expres-
sion data of selected acute-phase proteins, chemokines,
(pro-inflammatory) cytokines, cell adhesion molecules
and other proteins involved in the inflammatory response.
Grounding the prediction-model on the most relevant
genes not only reduces noise but also lower the costs and
time required to use gene expression signatures for ageing
bruises in forensic cases. As bruises inflicted with high
force had close resemblance to the gross appearance of
bruises in forensic cases of both pigs and humans we be-
lieve that the prediction model is likely to be valid for
estimating bruise age in a wide range of veterinary and
potentially also human forensic cases [11].
The mRNA expression profile (42 genes) in the subcu-
taneous fat was, to some extent, able to differentiate
bruises inflicted with varying force. Generally, bruises
inflicted with the low force were more similar to the con-
trols compared to bruises inflicted with moderate and high
force, respectively. This is in accordance with the histolog-
ical evaluation in which the inflammatory response is most
pronounced in bruises inflicted with the highest force [11].
Well-known proinflammatory cytokines; IL1B, IL1RN,
IL6, IL8, and IL18 were also found to be of high impor-
tance in separation of subcutaneous fat samples with regard
to force of impact. This predominant proinflammatory re-
sponse seen already after a few hours is in agreement with
previously reported histological evaluations of inflamma-
tion in the same tissue [11].
Forensic Sci Med Pathol
In the muscle tissue the gene expression signature (43
genes) could not clearly separate bruises according to the
force of impact or age. These unpredictable expression
data in the muscle tissue might result from variation in
thickness of the overlaying, protecting subcutaneous fat
tissue.
Interestingly, muscle tissue expression profiles of a few
genes (TNNC1, TNNT1, TNNI1 and FOS) differed between
areas of impact 1 to 4 of the same pig, inflicted with the same
force within a timeframe of 3 to 4 min. For the troponins,
which are proteins involved in the contraction of skeletal mus-
cle, differences between the four areas were only seen in the
experimental pigs and not in the two controls, indicating that
this difference is likely to be explained by the sequence of
impact rather than regional variations in the muscle. In area
of impact No. 1 all three types of troponin were clearly in-
creased compared to the control pigs. The basic for this pattern
is unknown. However, a hypothesis could be that the first
blunt trauma inflicted on the back of the pigs somehow has
exhausted the muscle tissues capability to express troponins
by saturating upstream receptors or transcription factor of tro-
ponin. The expression pattern in pigs is in direct contrast to a
study of experimental bruises in rats in which mRNA expres-
sion of troponin type I was significantly decreased compared
to normal muscle tissue [15]. It is tempting to speculate that
the difference in troponin expression could depend on the
species, the anatomical location of impact, and the force used
to inflict the injury. In favor of these speculations, human
serum levels of troponin were found to be higher in bone
injury patients compared to soft tissue injury patients and con-
trols [19]. Future studies will reveal which model resembles
human circumstances, but in favor of the porcine model, pre-
vious studies of genes involved in immunity and inflammation
show a higher degree of similarity between humans and pigs
than between humans and rats [13].
Forensic Sci Med Pathol
No differences in gene expression were seen between con-
trol tissue taken from the thighs of the experimental pigs and
the back of the control pigs, meaning that bruises on the back
did not induce a systemic reaction leading to differential
mRNA expression in the subcutaneous fat on the thighs.
Therefore, normal fat tissue from the experimental pigs can
serve as internal controls eliminating the control pigs in future
experiments using this model. This is in accordance with ob-
servations in rats with experimental bruises, in which normal
muscle tissue from the same animals was also found suitable
as internal control [15].
Conclusions
Gene expression signatures in the subcutaneous fat but
not in muscle tissue reflected the age of bruises with a
precision of approximately ±2 h. Moreover, the gene
expression signature in the subcutaneous fat was to
some extend able to separate bruises inflicted with dif-
ferent force. The data are strongly suggestive that a
combination of the expression signatures of selected
genes and histological evaluation will increase the pre-
cision of the age estimations of bruises in veterinary
and possibly also human forensic cases in the future.
Key points
1. The expression of more than 50 different genes in subcu-
taneous fat and muscle tissue from experimental bruises in
pigs was investigated. The aim was to evaluate if expres-
sion signatures of selected genes were capable of deter-
mining bruises according to age and the force of impact.
2. Quantitative real-time polymerase chain reaction was per-
formed to evaluate mRNA expression of genes involved
in inflammation, tissue damage and repair.
3. Expression signatures of thirteen selected genes in subcu-
taneous fat, but not in muscle tissue, reflected the age of
bruises with a precision of approximately ±2 h.
4. The gene expression signature in the subcutaneous fat was
able to separate bruises inflicted with different forces to
some extent.
5. Expression signatures of selected genes in the subcutane-
ous fat will increase the precision of the age determina-
tions of bruises in pigs and possibly also in humans due to
the similarities in skin anatomy, physiology and immune
reactivity.
Acknowledgements The authors wish to thank Karin Tarp at
Section for Immunology and Vaccinology, National Veterinary Institute,
Technical University of Denmark, for skilled technical assistance with
RNA extraction and qPCR analysis.
Compliance with ethical standards
Funding The project was funded by University of Copenhagen,
Denmark and National Veterinary Institute, Technical University of
Denmark. The funding source had no involvement in the experimental
design, analysis and interpretation of the results.
Conflict of interest The authors declare that they have no conflict of
interest.
Ethical approval All applicable international, national and/or institu-
tional guidelines for the care and use of animals were followed. The
experimental procedure was approved by the Danish Animal
Inspectorate (2013–15–2934 − 00849).
This article does not contain any studies with human participants
performed by any of the authors.
References
1. Barington K, Jensen HE. Experimental animal models of bruises in
forensic medicine – a review. Scand J Lab Anim Sci. 2015;41:14.
2. Stephenson T, Bialas Y. Estimation of the age of bruising. Arch Dis
Child. 1996;74:53–5.
3. Hughes VK, Ellis PS, Langlois NEI. Alternative light source
(polilight) illumination with digital image analysis does not assist
in determining the age of bruises. Forensic Sci Int. 2006;158:104–
7.
4. Byard RW, Wick R, Gilbert JD, Donald T. Histologic dating of
bruises in moribund infants and young children. Forensic Sci Med
Pathol. 2008;4:187–92.
5. Stam B, van Gemert JC, van Leeuwen TG, Teeuw AH, van der Wal
AC, Aalders MCG. Can color inhomogeneity of bruises be used to
establish their age? J Biophotonics. 2011;4:759–67.
6. Hughes VK, Langlois NEI. Visual and spectrophotometric obser-
vations related to histology in a small sample of bruises from ca-
davers. Forensic Sci Med Pathol. 2011;7:253–6.
7. Armstrong EJ. Distinctive patterned injuries caused by an expand-
able baton. Am J Forensic Med Pathol. 2005;26:186–8.
8. Barington K, Jensen HE. Forensic cases of bruises in pigs. Vet Rec.
2013;173:526–30.
9. Saukko P, Knight B. The pathology of wounds. In: Saukko P,
Knight B, editors. Knight’s forensic pathology. London: Arnold;
2004. p. 136–73.
10. Barington K, Jensen HE. A novel, comprehensive and reproducible
porcine model for the timing of bruises in forensic pathology.
Forensic Sci Med Pathol. 2016;12:58–67.
11. Barington K, Jensen HE. The impact of force on the timing of
bruises evaluated in a porcine model. J Forensic Legal Med.
2016;40:61–6.
12. Lader ES, Ambion Inc. Example 2 preparation of an exemplary
RNAlaterTM RNA preservation medium. In: Methods and reagents
for preserving rna in cell and tissue samples. 2012. http://www.
google.ch/patents/CA2298841C?hl=de&cl=en. Accessed 20 Mar
2017.
13. Dawson HD. A comparative assessment of the pig, mouse and
human genomes. In: McAnulty PA, Dayan AD, Ganderup NC,
Hastings KL, editors. The minipig in biomedical research. Boca
Raton: CRC Press; 2012. p. 323–42.
14. Takamiya M, Saigusa K, Kumagai R, Nakayashiki N, Aoki Y.
Studies on mRNA expression of tissue-type plasminogen activator
in bruises for wound age estimation. Int J Legal Med. 2005;119:16–
21.

15. Sun J, Wang Y, Zhang L, Gao C, Zhang LZ, Guo Z. Time-
dependent expression of skeletal muscle troponin I mRNA in the
contused skeletal muscle of rats: a possible marker for wound age
estimation. Int J Legal Med. 2010;124:27–33.
16. Du Q, Sun J, Zhang L, Liang X, Guo X, Gao C, Wang Y. Time-
dependent expression of SNAT2 mRNA in contused skeletal mus-
cle of rats: a possible marker for wound age estimation. Forensic Sci
Med Pathol. 2013;9:528–33.
17. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De
Paepe A, et al. Accurate normalization of real-time quantitative RT-
Forensic Sci Med Pathol
PCR data by geometric averaging of multiple internal control genes.
Genome Biol. 2002;3:RESEARCH 0034.
18. Andersen CL, Jensen JL, Ørntoft TF. Normalization of real-time
quantitative reverse transcription-PCR data: a model-based vari-
ance estimation approach to identify genes suited for normalization,
applied to bladder and colon cancer data sets. Cancer Res. 2004;64:
5245–50.
19. Onuoha GN, Alpar EK, Dean B, Tidman J, Rama D, Laprade M,
Pau B. Skeletal troponin-I release in orthopedic and soft tissue
injuries. J Orthop Sci. 2001;6:11–5.