1
Assessing the suitability of miRNA-142-5p and miRNA-541 for bloodstain deposition
2
timing
3
4
Karolina Lech1, Katrin Ackermann1,2, Andreas Wollstein1,3, Victoria L. Revell4,Debra J Skene4
5
and Manfred Kayser1*
6
7
1
8
Rotterdam, Rotterdam, The Netherlands
9
2
-Department of Forensic Molecular Biology, Erasmus MC University Medical Centre
-Current affiliation: School of Chemistry and Centre of Magnetic Resonance, University of St
10
Andrews, St Andrews, Fife, UK
11
3
12
Munich LMU, Planegg-Martinsried, Germany
13
4
14
Guildford, Surrey, UK
-Current affiliation: Section of Evolutionary Biology, Department of Biology II, University of
-Centre for Chronobiology, Faculty of Health and Medical Sciences, University of Surrey,
15
16
*
17
Prof. Dr. Manfred Kayser, Department of Forensic Molecular Biology, Erasmus MC University
18
Medical Centre Rotterdam, Rotterdam, The Netherlands
19
Email: m.kayser@erasmusmc.nl
20
Tel: +31 10 703 80 73
21
Fax: +31 10 704 45 75
22
23
- Corresponding author:
24
Abstract
25
26
A recent proof-of-concept pilot study proposed using microRNA (miRNA) markers for
27
time of death determination. The markers - miRNA-142-5p and miRNA-541, were reported to
28
show considerable expression differences in vitreous humor between individuals who died
29
during the day or night. Here, we investigated whether these miRNA markers show the same
30
diurnal expression pattern in blood, which would make them useful for estimating bloodstain
31
deposition time to allow molecular alibi testing for forensic casework. We analyzed venous
32
blood samples collected from 12 healthy individuals every 4 hours during the 24 h day/night
33
period under controlled sleep-laboratory conditions. MiRNA-142-5p normalized against
34
miRNA-222 showed no statistically significant expression differences between blood samples
35
collected during daytime and nighttime (one-way ANOVA p=0.81), and also no statistically
36
significant rhythmicity during the 24h day/night period (cosine fit for all individuals p>0.05,
37
averaged data p=0.932). MicroRNA-541 amplification in blood was above the 34-cycle
38
threshold applied in the study, indicating too low quantities for obtaining reliable data. Overall,
39
we conclude that the two miRNA markers previously suggested for time of death determination
40
in vitreous humor are not suitable for estimating the deposition time of forensic bloodstains.
41
Future studies may find out if miRNA markers with significant diurnal expression pattern can be
42
identified and how useful they would be for forensic trace deposition timing.
43
44
45
46
47
Introduction
48
49
Circadian rhythms are entrainable, endogenously generated oscillations with a period of
50
approximately 24 hours (h), present in most organisms, including humans. In recent years it has
51
been shown that every cell contains its own molecular clock composed of a network of
52
oscillating mRNAs and their protein products [1,2]. Numerous biological processes exhibit
53
significant daily variations including activity and behavior, heart rate, hormone secretions, or
54
sleep-wake cycle amongst others [3,4]. Knowledge about the underlying molecular pathways of
55
these processes and their components could provide a plethora of potential rhythmic biomarkers
56
also for forensic applications. In principle, such rhythmic biomarkers can be useful for molecular
57
time estimations in two forensically relevant ways: estimating the time of death and estimating
58
the deposition time of biological traces found at a crime scene. The latter would provide a
59
molecular alibi test allowing DNA-identified sample donors to be linked with criminal acts,
60
which is highly appreciated for crime investigation [5].
61
The first approach to estimate trace deposition time by incorporating a chronobiological
62
aspect into the forensic field was published in 2010 [6]. In this proof-of-principle study, the 24
63
hour profiles of two circadian hormones, melatonin and cortisol, were reproduced in small blood
64
and saliva samples collected around the clock. The feasibility of applying this approach in a
65
forensic setting was also shown - it requires only minute amounts of samples and utilizes
66
commercial assays together with basic laboratory equipment. The time frame estimated with
67
those two hormones is, however, limited to either late night (melatonin) or early morning hours
68
(cortisol) [6], and both hormones are known to be influenced by external factors [6]. Hence, the
69
exploration for additional biomarkers with rhythmic changes in concentration that eventually
70
would prove invaluable for increasing the precision of biological trace deposition timing in
71
forensic applications, is encouraged.
72
Determining the time of death with biochemical markers, i.e., melatonin, has been
73
described earlier [7,8], and recently a proof-of-principle study proposed utilization of
74
microRNAs (miRNAs) for that purpose [9]. MicroRNAs are of great interest for various forensic
75
applications [5], due to their small size, tissue specific expression and higher resistance to
76
degradation compared to mRNA, and were previously proposed as suitable markers for forensic
77
body fluid identification [10,11,12]. In recent years various microRNAs have been implied in the
78
regulation of circadian clock [13,14,15], and some of them were shown to exhibit diurnal
79
expression changes in diverse tissues, such as liver or suprachiasmatic nucleus (SCN) [14,15].
80
In their study [9], the authors assessed the miRNA expression profile of vitreous humor,
81
and discovered that miRNA-142-5p and miRNA-541 exhibited considerable variations in their
82
expression levels between individuals who died during day- or nighttime, when normalized
83
against the reference marker miRNA-222. They also showed that there was no correlation
84
between miRNAs expression and the time elapsed after death, suggesting in-vitro stability of the
85
markers for at least 24 h, which is an important prerequisite for forensic applications.
86
In the present study, we investigated whether these two miRNA markers exhibit the same
87
diurnal changes in expression levels in other types of forensically relevant biological samples, in
88
this case blood, and if they do, to what degree their combined use together with melatonin and
89
cortisol could increase the precision of estimating blood trace deposition time. For this purpose,
90
we determined the temporal profiles of miRNA-142-5p and miRNA-541 in blood samples from
91
12 individuals collected every 4 hours during a 24h day/night cycle under controlled wake/sleep
92
laboratory conditions.
93
Materials and methods
94
95
1. Biological samples
96
Detailed characterization of sample collection as well as the sleep laboratory protocol used, is
97
described elsewhere [16]. In short, two-hourly blood samples were collected from 15 healthy
98
male volunteers aged 19 – 35 years (n = 15; mean age ± SD = 24 ± 5 years) participating in the
99
66 hour-long study (Clinical Research Centre, University of Surrey, UK) after various medical
100
assessments and completion of sleep questionnaires. The subjects underwent a week of regular
101
sleep-wake routine, before being admitted into the sleep laboratory. The sleep laboratory
102
protocol design controlled for food intake and lighting conditions and included an adaptation,
103
baseline and sleep deprivation night.
104
For the current study, 7 four-hourly blood samples (spanning a course of 24 hours) per each
105
participant (n = 12) were included. Three subjects were omitted from analysis due to having
106
either significantly shifted melatonin profiles or an abnormal white blood cells (WBC) count
107
[16,17]. The analyzed samples were collected during the baseline night (N2) at 22:00, 02:00,
108
06:00 h (n=36) during which the participants were not subjected to sleep deprivation (normal
109
sleep) and during a wake period (D3) mimicking a normal day, at 10:00, 14:00, 18:00 and 22:00
110
h (n=48).
111
112
2. RNA Isolation and quantitation
113
Blood samples were collected in EDTA tubes and stored at -80oC until use. Total RNA was
114
extracted from 400 µl of whole blood using the miRNeasyMiniKit (Qiagen, Hilden, Germany)
115
according to the manufacturer’s protocol with minor modifications, such that the samples after
116
homogenization were incubated on ice for 30 minutes and centrifuged at 4oC, speed 13 000 g for
117
5 minutes, and one volume of 70% ethanol (Merck, Darmstadt, Germany) was used for
118
precipitation. Purity of each sample was assessed with NanoDrop ND-2000 (NanoDrop
119
Technologies, Wilmington, DE, USA) and the quality and quantity were determined using the
120
Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany). Total RNA samples were
121
stored at -80oC until assayed.
122
123
3. Reverse Transcription Reaction
124
Reverse transcription (RT) reactions were performed in a final volume of 15 µl, using the
125
TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems, Life Technologies,
126
Carlsbad, CA, USA). Each reaction contained 100 ng of total RNA, 6 µl TaqMan MicroRNA
127
Assay RT Primer Pool (containing miRNA-142-5p, miRNA-222, miRNA-541), 0.2 µl of
128
100mMdNTPs, 3 µl of MultiScribe Reverse transcriptase enzyme (50 U/µL), 1.5 µl of reverse
129
transcription buffer (10x) and 0.19 µl of RNase Inhibitor (20 U/µL). The RT primer pool was
130
prepared according to the manufacturer’s instructions concerning custom primer pool preparation
131
(Applied Biosystems, Life Technologies, User Bulletin_4465407) with the exception of using
132
nuclease-free water instead of 1x TE buffer. Negative controls containing distilled water instead
133
of RNA were included. RT reactions were performed on MJ Research Thermal Cycler PTC-200
134
(GMI, Minnesota, USA) with the following program: 16oC for 30 minutes, 42oC for 30 minutes,
135
85oC for 5 minutes and 4oC on hold. Samples were kept at -20oC until assayed.
136
137
138
139
4. RT-qPCR reaction
140
MiRNA-142-5p and miRNA-222 expression levels were tested in 84 samples (12 subjects, 7
141
time points per subject). MiRNA-541 expression was tested in a subset of 14 samples (2
142
subjects, 7 time points per subject). The RT reaction contained 0.5 µl of appropriate TaqMan
143
MicroRNA Assay (Applied Biosystems, Life Technologies), 2 µl of two times diluted RT
144
product, 5µl of TaqMan Fast Universal PCR Master Mix (2x) NoAmpErase UNG (Applied
145
Biosystems, Life Technologies) and 2.5 µl of nuclease-free water in total volume of 10 µl. No-
146
template controls (NTC) with distilled water instead of RT reaction product were included. All
147
reactions were run in triplicate in 384-well plate format on Roche LightCycler 480 (Roche
148
Diagnostics, Mannheim, Germany). The program consisted of a 10 minutes pre-incubation step
149
in 95oC followed by 45 cycles of incubation and extension at 95oC for 15 seconds and at 60oC
150
for 60 seconds, respectively, and finished with 30 seconds at 40oC.
151
152
5. Data Analysis
153
Relative quantification of microRNA amplification was performed with the second derivative
154
maximum method (Roche Diagnostics) followed by the delta cycle threshold (CT) method [18].
155
All samples with CT values over 34 were considered as not expressing the tested microRNAs.
156
MiRNA-222 was used as the reference gene for the calculations, because it was shown to be
157
amplified stably and consistently in vitreous humor samples and was identified as the best single
158
candidate reference gene by the Genorm program [9]. Additionally, initial testing of miRNA-222
159
with the NormFinder [19] in a subset of blood samples revealed it to be a good candidate for
160
normalization purposes. To exclude the possibility of time-of-day dependent rhythmicity,
161
normalizer levels were further analyzed with nonlinear curve fitting and single cosinor methods.
162
The expression levels for miRNA-142-5p were estimated as fold change (FC), defined by the
163
equation FC=2dCT. Assessment of rhythmicity was performed with three independent methods.
164
First, a one-way ANOVA analysis was implemented to determine the difference of normalized
165
miRNA-142-5p expression levels between samples collected during the day and night. Secondly,
166
the relative expression values for each subject were analyzed with the single cosinor test with a
167
24-hour period to determine the presence or absence of diurnal expression pattern [17]. The
168
analysis was performed for each subject separately, as well as for averaged expression data from
169
all subjects.
170
Next, nonlinear curve fitting was performed using a cosine function with a 24-h period on
171
z-scored, miRNA-222-normalized expression data of miRNA-142-5p for all individuals
172
simultaneously, as described elsewhere [20], to obtain the estimates for amplitude and peak and
173
the significance of fit. The equation for nonlinear curve fitting is presented as
174
follows: normalized z − score = 𝛼𝑖 + 𝛽 ∙ cos(2𝜋 ∙
175
term for individual i, β is the amplitude and t is the peak time of the cosine function.
𝑇𝑃−𝑡
24
), where TP is the time point,𝛼𝑖 is the
176
177
Results
178
179
The expression levels of miRNA-142-5p (in the range of 31 – 33.6 CT) and miRNA-222
180
(between 22 – 25 CT) during the 24 h day/night cycle were tested in total RNA isolated from 84
181
blood samples of 12 subjects collected at 7 time points every 4 hours, and for miRNA-541 in a
182
subset of 14 samples from 2 participants. Testing of miRNA-541 was not performed in all
183
subjects due to initial data analysis indicating amplification of this particular miRNA above 34
184
CT (on average 37 CT), which was the detection threshold applied in this study. This result
185
indicates that the level of miRNA-541 found in these blood samples was too low for reliable data
186
generation. Before normalization we determined the significance of fit of a cosine curve on
187
miRNA-222 data, in order to exclude the possibility of using a rhythmic normalizer in our
188
analysis. With the single cosinor test, p-values > 0.05 were obtained in 11 out of 12 individuals,
189
and no significant time-of-day variation was found using nonlinear curve fitting for both
190
amplitude (p = 0.61) and peak (p = 0.35) estimates.
191
Calculated fold-changes of miRNA-142-5p, normalized against miRNA-222, throughout
192
the 24 h period are presented in figure 1. No significant differences between daytime and
193
nighttime expression levels were detected (one-way ANOVA p=0.81). Furthermore, rhythmicity
194
assessment using the cosinor method revealed no statistically significant results for any subject
195
individually (all p-values >0.05) and for data averaged across individuals (p = 0.932). The
196
expression of miRNA-142-5p for all subjects and all time-wise samples simultaneously was
197
additionally assessed with the nonlinear curve fitting method and the results are presented in
198
figure 2. The estimates for peak and amplitude obtained with this method are not statistically
199
significant (estimateamplitude = -0.05; pamplitude = 0.74; estimatepeak = -16.67; ppeak = 0.15). All the
200
applied methods showed that miRNA-142-5p levels do not display any rhythmic changes in the
201
blood samples we analyzed.
202
203
Discussion
204
205
In our study we aimed at establishing the 24 h day/night expression profiles of miRNA-
206
142-5p and miRNA-541 in blood, because both markers were recently proposed to be suitable
207
for time of death determination, due to observed diurnal differences in their expression levels in
208
vitreous humor [9]. In addition, previous microarray data suggested oscillation of miRNA-142-
209
5p in murine liver, however this finding was not confirmed by RT-qPCR [21] as it typically is
210
recommended in order to avoid technical artifacts. These findings motivated our study for testing
211
the suitability of both microRNA markers to estimate bloodstain deposition time for future
212
forensic applications.
213
Our data for miRNA-142-5p suggest that this miRNA marker is found in blood in
214
reliably measurable quantities, but also that its expression in this sample type does not oscillate
215
in regard to day- and nighttime. Although our data do not allow us to ultimately conclude
216
whether the changes in miRNA-142-5p expression that were reported in vitreous humor are
217
driven by the endogenous clock, or are a response to external stimuli or both, we suggest further
218
testing on miRNA-142-5p expression in vitreous humor and other human tissues before this
219
marker is applied for time of death determination. Because of its non-rhythmic profile in blood,
220
we do not regard miRNA-142-5p as suitable for any forensic time estimation from blood,
221
including trace deposition timing as well as death timing.
222
Our results for miRNA-541 suggest that it is found in blood at levels too low to provide
223
reliable results, at least when using RT-qPCR, which is currently the standard method for single-
224
marker expression analysis. MiRNA-541 is a member of the microRNA 379-410 cluster of
225
brain-specific microRNAs, and was shown to be abundantly expressed in distant axonal neuron
226
cells [22]. Hence, the low levels detected in blood samples in our study, may not be too
227
surprising and may possibly be a consequence of the tissue-specific expression of miRNA-541.
228
We conclude that, due to low quantities of miRNA-541 found in blood, it is not suitable for any
229
forensic time estimation from blood including trace deposition timing as well as death timing.
230
Although we could not confirm the previous findings from vitreous humor in blood, we
231
would like to emphasize the suitability of our protocol and samples to validate or de-novo
232
identify oscillating biomarkers potentially useful for trace deposition timing as well as for time
233
of death determination. The choice of sample type i.e., blood was based on its suitability in light
234
of both the forensic and chronobiological aspects. Blood represents the forensic trace
235
encountered commonly at crime scenes, typically allowing the occurrence of violent crime to be
236
confirmed, and is usually available with dead bodies. The blood samples we used in our analysis
237
were collected around the clock, spanning a full 24 h day/night period and allowing assessment
238
of temporal expression of a possibly rhythmic biomarker. Collection of these samples occurred
239
under strictly controlled conditions in a dedicated sleep laboratory specialized for human
240
chronobiology research, to assure elimination or minimization of exogenous factors that could
241
confound the rhythm of the internal clock. In particular, such a laboratory setting allows blood
242
sample collection during undisturbed sleep, which is otherwise impossible. Additionally, our
243
study design entailed stringent criteria of selection of participants – exclusion of extreme
244
chronotypes, restriction of the influence of sex and age on the phase of the clock [23,24] etc. The
245
conditions such as lighting intensity, meal composition and timing were controlled, yet still
246
resembled natural everyday settings, which is important when confronted with the real crime
247
case scenarios.
248
The ultimate goal of our study was to combine the time estimates obtained with
249
melatonin and cortisol together with miRNA-142-5p and miRNA-541 expression data to
250
increase the precision of bloodstain deposition timing. However, since miRNA-142-5p levels in
251
analyzed blood samples were found to be stable during the 24 h day/night period, and miRNA-
252
541 was not present in quantities high enough to allow reliable detection with the method used,
253
we could not consider these miRNAs as markers useful for our aim. Hence, until new suitable
254
biomarkers are identified in future studies, our previously suggested melatonin and cortisol
255
system [6] remains the only available molecular approach for a person’s alibi testing directly
256
from crime scene stains.
257
258
Acknowledgements
259
260
The authors thank the Surrey CRC medical, clinical and research teams for their help with the
261
sample collection. This study was supported by the Erasmus MC, a grant from the Netherlands
262
Genomics Initiative (NGI) / Netherlands Organization for Scientific Research (NWO) within the
263
framework of the Forensic Genomics Consortium Netherlands (FGCN), grant 727.011.001 by
264
the NWO Forensic Science Program, and by the EU 6th Framework project EUCLOCK
265
(018741). AW received additional funding by Volkswagen Foundation (ref 86042).
266
267
References
268
269
[1] Takahashi JS, Hong HK, Ko CH, McDearmon EL. The genetics of mammalian circadian
270
order and disorder: implications for physiology and disease. Nat Rev Genet. 2008;9:764-75.
271
[2] Mohawk JA, Green CB, Takahashi JS. Central and peripheral circadian clocks in
272
mammals. Annu Rev Neurosci. 2012;35:445-62.
273
[3] Partch CL, Green CB, Takahashi JS. Molecular architecture of the mammalian circadian
274
clock. Trends Cell Biol. 2014;24:90-9.
275
[4] Lowrey PL, Takahashi JS. Mammalian circadian biology: elucidating genome-wide
276
levels of temporal organization. Annu Rev Genomics Hum Genet. 2004;5:407-41.
277
[5] Kayser M, de Knijff P. Improving human forensics through advances in genetics,
278
genomics and molecular biology. Nat Rev Genet. 2011;12:179-92.
279
[6] Ackermann K, Ballantyne KN, Kayser M. Estimating trace deposition time with circadian
280
biomarkers: a prospective and versatile tool for crime scene reconstruction. Int J Legal Med.
281
2010;124:387-95.
282
[7] Ackermann K, Bux R, Rub U, Korf HW, Kauert G, Stehle JH (2006) Characterization of
283
human melatonin synthesis using autoptic pineal tissue. Endocrinology 147:3235–3242
284
[8] Mikami H, Terazawa K, Takatori T, Tokudome S, Tsukamoto T, Haga K (1994)
285
Estimation of time of death by quantification of melatonin in corpses. Int J Legal Med
286
107:42–51
287
[9] Odriozola A, Riancho JA, de la Vega R, Agudo G, Garcia-Blanco A, de Cos E, et al.
288
miRNA analysis in vitreous humor to determine the time of death: a proof-of-concept pilot
289
study. Int J Legal Med. 2013;127:573-8.
290
[10] Sood P, Krek A, Zavolan M, Macino G, Rajewsky N. Cell-type-specific signatures of
291
microRNAs on target mRNA expression. Proc Natl Acad Sci U S A. 2006;103:2746-51.
292
[11] Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell.
293
2004;116:281-97.
294
[12] Zubakov D, Boersma AWM, Choi Y, van Kuijk PF, Wiemer EAC, Kayser M.
295
MicroRNA markers for forensic body fluid identification obtained from microarray screening
296
and quantitative RT-PCR confirmation. Int J Legal Med. 2010;124:217-26.
297
[13] Chen RM, D'Alessandro M, Lee C. miRNAs Are Required for Generating a Time Delay
298
Critical for the Circadian Oscillator. Curr Biol. 2013;23:1959-68.
299
[14] Cheng HYM, Papp JW, Varlamova O, Dziema H, Russell B, Curfman JP, et al.
300
microRNA modulation of circadian-clock period and entrainment. Neuron. 2007;54:813-29.
301
[15] Mehta N, Cheng HYM. Micro-Managing the Circadian Clock: The role of microRNAs
302
in biological timekeeping. J Mol Biol. 2013;425:3609-24.
303
[16] Ackermann K, Revell VL, Lao O, Rombouts EJ, Skene DJ, Kayser M. Diurnal rhythms
304
in blood cell populations and the effect of acute sleep deprivation in healthy young men.
305
Sleep. 2012;35:933-40.
306
[17] Ackermann K, Plomp R, Lao O, Middleton B, Revell VL, Skene DJ, et al. Effect of
307
sleep deprivation on rhythms of clock gene expression and melatonin in humans. Chronobiol
308
Int. 2013;30:901-9.
309
[18] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time
310
quantitative PCR and the 2(T)(-Delta Delta C) method. Methods. 2001;25:402-8.
311
[19] Andersen CL, Jensen JL, Ørntoft TF. Normalization of real-time quantitative reverse
312
transcription-PCR data: a model-based variance estimation approach to identify genes suited
313
for normalization, applied to bladder and colon cancer data sets. Cancer Res. 2004 Aug
314
1;64(15):5245-50.
315
[20] Minors DS, Waterhouse JM. Mathematical and Statistical-Analysis of Circadian-
316
Rhythms. Psychoneuroendocrino. 1988;13:443-64.
317
[21] Na YJ, Sung JH, Lee SC, Lee YJ, Choi YJ, Park WY, et al. Comprehensive analysis of
318
microRNA-mRNA co-expression in circadian rhythm. Exp Mol Med. 2009;41:638-47.
319
[22] Zhang J, Zhang J, Liu LH, Zhou Y, Li YP, Shao ZH, et al. Effects of miR-541 on
320
neurite outgrowth during neuronal differentiation. Cell Biochem Funct. 2011;29:279-86.
321
[23] Dijk DJ, Duffy JF, Czeisler CA. Contribution of circadian physiology and sleep
322
homeostasis to age-related changes in human sleep. Chronobiol Int. 2000;17:285-311.
323
[24] Duffy JF, Czeisler CA. Age-related change in the relationship between circadian period,
324
circadian phase, and diurnal preference in humans. Neurosci Lett. 2002;318:117-20.
325
326
Figure Legends
327
328
Figure 1. MiRNA-142-5p levels, normalized against miRNA-222, in 12 healthy individuals
329
depicted in colored lines presented as fold change (FC) versus blood collection time within one
330
24 h day/night cycle. The black line represents the average fold change across all individuals
331
with the standard deviation bars. The high FC value of R07 at 22:00 h is probably due to partial
332
coagulation of this particular blood sample.
333
334
Figure 2. Z-scored, miR-222 normalized levels of miRNA-142-5p in each of the 12 individuals
335
depicted in colored lines with a superimposed cosine curve as calculated with the nonlinear curve
336
fitting method [19] depicted in black.
337
338