- Email: [email protected]
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control
Accepted Manuscript Title: Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control Author: Alessandra Stampella Sabrina Di Marco Daniela Pirri Xavier de la Torre Francesco Botr`e Francesco Donati PII: DOI: Reference:
S0379-0738(16)30173-6 http://dx.doi.org/doi:10.1016/j.forsciint.2016.04.021 FSI 8437
To appear in:
FSI
Received date: Revised date: Accepted date:
15-11-2015 6-4-2016 15-4-2016
Please cite this article as: A. Stampella, S. Di Marco, D. Pirri, X. de la Torre, F. Botrgravee, F. Donati, Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control, Forensic Science International (2016), http://dx.doi.org/10.1016/j.forsciint.2016.04.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ce pt
ed
M
an
us
cr
ip t
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control
Homologous blood transfusions, a form of “blood doping” in sport, can be detected by forensic genetic techniques based on DNA typing. The proposed procedure, that tested effective also on mixed samples of red blood cell concentrates, allows to correctly identify also those samples that
Ac
would have given a false-negative result if assayed by the reference cytofluorimetric technique.
Page 1 of 22
Highlights
Forensic genetics techniques are proposed to detect homologous blood transfusions. The proposed method, based on DNA typing, tested effective on ex vivo blood mixtures
Mixed blood samples are revealed by triplets and quadruplets at one or more loci
The same procedure can be applied also to red blood cell concentrates
Ac
ce pt
ed
M
an
us
cr
ip t
Page 2 of 22
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control Alessandra Stampella1, Sabrina Di Marco1, Daniela Pirri1, Xavier de la Torre1, Francesco Botrè1, 2*,
ip t
Francesco Donati1
1: Laboratorio Antidoping, Federazione Medico Sportiva Italiana, Largo Giulio Onesti 1, 00197, Rome, Italy
cr
2: Dipartimento di Medicina Sperimentale, “Sapienza” Università di Roma, Viale Regina Elena 324, 00161
us
Rome, Italy
*Corresponding author
an
Prof. Francesco Botrè
ed
ce pt
phone: +39-06-87973500 fax: +39-06-8078971 email: [email protected]
M
Laboratorio Antidoping Federazione Medico Sportiva Italiana Largo Giulio Onesti 1 00197 Rome Italy
Keywords: Homologous blood transfusion; blood doping; doping analysis; red blood cells
Ac
concentrates; DNA typing; short tandem repeats.
Page 3 of 22
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control
ip t
ABSTRACT
cr
In this work we present the application of a method for the identification of homologous blood transfusions using forensic genetic techniques based on DNA typing. Ex vivo mixtures of human
us
blood samples - either whole blood or red blood cell concentrates - simulating homologous blood transfusions at different percentages of the donor were typed for a panel of 16 highly variable DNA short tandem repeats (STR). Tested samples included also mixtures, which gave false-
an
negative results if assayed by the reference flow cytofluorimetric method, which is based on the recognition of target antigens located on the membrane of the red blood cell. The recognition of
M
triplets and quadruplets at various loci gave information of the presence of cells belonging to different individuals, as it is the case for homologous blood transfusions. Specificity and sensitivity
d
of the method were assessed in the validation study. The method proved to be unequivocally specific since it was able to recognize all single profiles of each individual, clearly discriminating
te
them from mixtures. Sensitivity resulted as a consequence of the percentage of the donor aliquot in the total volume of the mixture. Although the source of DNA in a blood sample is represented
Ac ce p
only by nucleated white blood cells, the same procedure resulted effective also in detecting mixtures of red blood cell concentrates (RBCC) from leukodepletion procedure: DNA of the donor from the residual white blood cells resulted still detectable, even if with an expected loss of sensitivity. The proposed approach may contribute to reduce the risk of false-negative results, which may occur using the reference cytofluorimetric method.
Keywords: Homologous blood transfusion; blood doping; doping analysis; red blood cells concentrates; DNA typing; short tandem repeats.
1 Page 4 of 22
INTRODUCTION In sport doping, the administration or reintroduction into the circulatory system of any quantity of autologous, allogenic (homologous) or heterologous blood or of any red blood cell products, of
ip t
any origin, is prohibited by the World Anti-Doping Agency (WADA) [1]. Specifically, homologous blood transfusion (HBT) is a blood manipulation practice, which can be attractive for cheating athletes because of the enhancement of oxygen carrying to the tissues [2]. HBT is based on the
cr
infusion of whole blood (WB) or red blood cell concentrates (RBCC) from a donor to a recipient
us
subject whose blood is compatible for the matching of the major blood group antigens (ABO and Rh factor).
HBT was widely used in the late 70s-80s until it was banned in sport, first by the
an
International Olympic Committee and then by the WADA. In the 90s, Erythropoietin (EPO) was considered by athletes a safer and more effective method to illicitly boost the oxygen transport
M
capacity in blood; later on, following the implementation of a direct method to detect the intake of EPO and analogues in urine [3-4], cheating athletes reconsidered the practice of blood transfusion. The implementation of the first method for the detection of HBT dates back to 2004
d
and was first used on the occasion of the anti-doping test for the Athens Summer Olympic Games:
te
in the same year the first adverse analytical finding for HBT was also reported [5]. The direct method for the detection of the abuse of homologous blood transfusion (HBT) is
Ac ce p
still in use by the WADA-accredited laboratories and is based on the recognition of different population of erythrocytes, both of the donor and of the receiver, in the same blood sample. This method is based on the identification, by flow cytofluorimetry, of antigens of minor human blood groups. The test is indeed based on the analysis of 8 different antigens belonging to four different human blood groups (namely big C, small c and big E of the Rh group, Jka and Jkb of the Kidd group, Fya and Fyb of the duffy group and big S of the MNS group) [6–7]. Despite the proven performances of this method in terms of sensitivity, it may happen that
two individuals express an identical antigenic profile, leading to the possibility of false-negative results [8-9]. This can be considered the most significant limitation of the flow cytofluorimetric method for HBT. In our laboratory, we first tried to solve this limitation by increasing the panel of whole blood antigens to be analyzed (extending it from 8 antigens of the basic panels to 13 antigens by an “extended” panel). Surely the discrimination power of the method was increased,
2 Page 5 of 22
but clearly false negative results, although with a much lower statistical frequency, could still occur. For this reason we have recently explored the utility of DNA based forensic analysis, presenting a preliminary analytical approach based on DNA genotypic differences and no longer on phenotypic
ip t
differences between the donor and the receiver [9]. In parallel, Manokhina et al. proposed another DNA based-approach for the detection of homologous blood transfusion as an illicit doping practice [10]. Their work focused on the possibility to discriminate donor blood from the
cr
receiver by a genotyping and sensitive strategy with detection on quantitative PCR, thus
us
confirming utility and versality of DNA-based investigative approaches for the detection of homologous blood transfusion.
We are here proposing to follow a forensic human DNA typing strategy, based on the analysis of
an
little portions of DNA called Short Tandem Repeats (STRs). Each locus STR may show a wide number of alleles, which differs for the number of the repeated sequences. 15 high variable
M
autosomal STRs plus the amelogenin locus (used for sex determination) are normally tested in human identification typing [11]. The probability that two different individuals share the same haplotype pattern is virtually zero. DNA extracted from different sources (e.g. blood, urine, buccal
d
swab) is amplified by PCR into a master mix containing the primer set of all of the 16 loci of the
te
identification panel (in brief D8S1179, D21S11, D7S820, CSF1PO, D3S1358, TH01, D13S317, D16S539, D2S1338, D19S433, vWA, TPOX, D18S51, D5S818, FGA and amelogenin). Each primer is
Ac ce p
specific for a certain locus. PCR generates a very high number of fragments containing the alleles of interest. The amplified sample is then analyzed in a capillary electrophoresis system to separate the fragments locus by locus. Detection of different alleles is made by a five-color detection system. It is possible to identify an individual for the high inter-individual variability of alleles at each STR locus that depends on the number of repetitions of the repeated sequence. Human genome is diploid and the profile of a single (non-transfused) individual is unique and can present one or two allelic peaks for each locus, indicating homozygosity or heterozygosity respectively for that locus. Our preliminary results showed that DNA typing, thanks to the very high discrimination power of DNA analysis, is an alternative/complementary analytical technique, which enhances the specificity of the classical flow cytofluorimetry-based method. In case of whole blood or RBCC transfusions with a sufficient residual population of nucleated cells (e.g. white blood cells), the profile of a mixture of blood originated from two different subjects can be recognized by the presence of more than two allelic peaks (triplets and quadruplets) at various loci [12-13]. In a two3 Page 6 of 22
person mixture the interpretation becomes difficult when the minor profile is less than one-third of the response of the major profile, since it could be difficult to distinguish true alleles from stutter peaks [14-15]. In this work, the forensic DNA method based on STRs is proposed as an
ip t
alternative/complementary way to detect different forms of HBT (either of whole blood or of RBCC), also to overcome the limitation due to the risk of false negative results that may occur by applying the current reference cytofluorimetric method, based on the recognition of target
Ac ce p
te
d
M
an
us
cr
antigens located on the membrane of the red blood cell.
4 Page 7 of 22
MATERIAL AND METHODS Blood Samples Whole blood samples of several different individuals were used to prepare 20 different mixed
ip t
samples with four different proportions of the donor (10, 5, 2.5, 1, 0.5%) for a total number of 100 mixes to be analyzed. Prior to mixing, AB0 group and Rh factor were determined by a hemagglutination assay (DiaClon BIO-RAD antibodies anti-A, anti-B and anti-D) and minor antigens
cr
by flow cytofluorimetry were analyzed on each single sample. Samples were mixed in order to
us
simulate a false-negative result coming from a flow cytofluorimetric screening, which means that samples were mixed only when not more than one antigenic difference was first detected. The same analytical approach was applied on mixed samples obtained after leukodepletion of the
an
donor using Ficoll-Paque PLUS (GE Healthcare). Mixtures at < 10% of the donor were produced from red blood cell concentrates. Residual white blood cells were counted by Sysmex XE-2100
M
Automated Hematology Blood Analyser to evaluate whether their number was comparable to
Flow Cytofluorimetry Analyses
d
those admitted by law in a normal transfusion routine practice.
Ac ce p
by Giraud et al. 2008 [6].
te
Phenotyping of whole blood samples prior the generation of the mixes was performed as detailed
Preparation of concentrated RBC
Ficoll-Paque™ PLUS (GE Healthcare) was used to obtain concentrated RBC from 1 mL of whole blood sample. The preparation was performed according to the manufacturer’s recommendations. White blood cell (WBC) count
White blood cells residue number was estimated using Sysmex XE-2100 hematologic analyser (Dasit). For the determination of the number of total white blood cell and the number of each type of blood cell present in the concentrated samples, the method adopted is flow cytofluorimetric with the use of a semiconductor laser. DNA Genomic DNA was isolated from 2µL of sample (either whole blood or RBCC) using the PrepFiler™ 5 Page 8 of 22
Forensic DNA Extraction Kit (Applied Biosystems®). All DNA extracts were quantified in triplicate by real time PCR using the Quantifiler® Human DNA Quantification Kit (Applied Biosystems®) and the Applied Biosystems 7500 Fast real-time PCR system. Quantified samples were then amplified with PCR using 10 µL of 0.1ng/µL of the final extract; any dilution was performed using the elution
ip t
buffer of PrepFiler™ Forensic DNA Extraction Kit (Applied Biosystems®). This amplification step was performed on the GeneAmp® PCR System 9700 thermocycler using the AmpFlSTR ® Identifiler® PCR Amplification kit or AmpFlSTR® Identifiler® Plus PCR Amplification with the
cr
recommended amplification conditions (a total of 28 cycles, annealing temperature 59°C,
us
extension temperature 72°C as advised by manufacturer). DNA typing was executed with the ABI PRISM 3100 Genetic Analyser (Applied Biosystems®). In brief, 1.5 µL of amplified DNA was diluted in 12 μL of form amide and 1µL of GeneScan™-500 Size Standard (Applied Biosystems®); POP-4
an
polymer was used and 5s injections were applied. All data were analyzed using the GeneMapper™ ID Software (Applied Biosystems®), which performs the allele calling by the help of a reference
Ac ce p
te
d
M
DNA Allelic Ladder containing all the alleles of each locus analyzed.
6 Page 9 of 22
RESULTS AND DISCUSSION Specificity of the DNA-based method Specificity was assessed by evaluating the ability of the analytical method to accurately obtain a
ip t
DNA profile from individual samples in the presence of interfering and not confusing them as mixtures. The aim was to verify whether it was possible the identification of single individual gene profiles without any triplet and/or quadruplet, typical of mixed samples. The method resulted
cr
unequivocally specific, since no samples belonging to single individual were identified as mixtures.
us
For each locus only one or two peaks have been identified, indicative respectively of homozygosity
M
an
and heterozygosity for that locus (Figure 1).
Figure 1. Example of a four loci DNA profile for an individual carrying two heterozygous loci (double peaks D19S433 and VWA) and
te
d
two homozygous loci (single peaks: TPOX and D18S51).
Sensitivity: identification of mixtures
Ac ce p
From a pool of whole blood samples, 20 different mixtures were produced. Blood major groups antigenic profile was first estimated by hemagglutination assay in order to produce mixtures where donor and receiver were compatible for ABO and Rh factor (positive + or negative according to the presence of D antigen or less). Five A+ mixes, one B+ mix, three 0+ mixes and one 0- mix were prepared. Ten mixes with compatible donor/receiving have been prepared as follows: three 0+/A+, two 0+/B+; one 0+/AB+; one 0-/A-; one 0-/A+; one B+/AB+ and one A+/AB+. 32 single samples have been first analyzed by the flow-cytofluorimetric method; all 20 mixtures we used for the following DNA analysis resulted as "false negatives" with the flow-cytofluorimetric method (an example of such a case is shown in Figure 2). Note that a false-negative result refers to the result of a flow cytofluorimetric assay where no more than one antigenic difference is recorded. For each pair of samples 5 different mixtures, with decreasing concentrations of the donor (from 10% down to 0.5%, for a total of 100 final mixtures), were considered.
7 Page 10 of 22
ip t cr
us
Figure 2. Analysis of a HBT sample in which donor and receiver share identical single haplotype for 13 different antigens of whole blood minor groups. Flow cytofluorimetric analysis cannot discriminate between the contributions given by each sample (donor or receiver) to the resulting flow cytofluorimetric pattern, so that no double populations of erythrocytes can be recorded, leading to a
an
false-negative result.
The highest percentage of the donor (10%) has been chosen to simulate the situation of a recent
M
transfusion of a whole blood bag (450 mL). The other percentages of donor (5%, 2.5%, 1% and 0.5%) simulate a blood sample collection at various times after the transfusion of blood to the
d
receiver, to consider the period of blood replacement after the transfusion. The low concentration cytofluorimetric method.
te
of the donor at 0.5% was considered because quite close to the limit of sensitivity of the flow
Ac ce p
The presence of triplets or quadruplets at one or more loci is the essential information for the identification of a mixture; it might occur that no triplets and quadruplets were detected in a mixture, but the probability of not having any informative locus out of the 15 target loci is considered to be extremely low, close to the statistical zero. Table 1 Sensitivity in percentage (%) of the DNA typing method for the five different mixtures with THR=50 and THR=200. Whole blood donor (%)
Sensitivity (%) THR 50
Sensitivity (%) THR 200
10
100
100
5
100
95
2.5
95
80
1
95
70
0.5
85
65
8 Page 11 of 22
The choice of an appropriate analytical threshold (THR) is crucial for the interpretation of the mixtures. In this work, thresholds at 50 and 200 relative fluorescent units (RFU) were considered. A relative low threshold, such as 50 RFU, takes both advantages and disadvantages; in fact, more information and identification power of very small peak alleles could be obtained, but on the
ip t
contrary, background noise peaks or "stutters" peaks may be erroneously evaluated as true peaks. The use of a high threshold (e.g. 200 RFU) allows avoiding most of the interpretation problems, but, at the same time, may not lead to an allelic call of the peak. The choice of the right THR
cr
should be carefully evaluated and may be determined to reach a sort of “compromise” between
us
the necessity to identify a peak and the parallel need to discard any not significant signal. As shown by Tab.1, sensibility of 100% was achieved with mixtures at 10% of the donor independently by the chosen THR. Mixtures at 5% of the donor shown sensitivity of 100% only
an
with THR=50, while sensitivity slightly decreases at 95% when THR = 200 was applied. As expected, with the reduction of donor concentration, the percentage of identified mixtures decrease, but
M
anyway in the worst situation (mixes at 0.5% of the donor) more than 50% of the samples were correctly recognized as mixes (Tables 2 and 3).
te
THR=5 0
d
Table 2. Mixes at 5 different percentages of the donor at THR 50
Identified mixes
n° total triplets + quadruplets
Mean triplets + quadruplets
n° total triplets
Mean triplets
Minimummaximum triplets
n° total quadruplets
Mean quadruplets
Minimummaximum quadruplets
Mix 10%
20
20
191
9.5
126
6.3
3-11
65
3.25
0-7
Mix 5%
Ac ce p
Total mixes
20
20
193
9.6
126
6.3
2-10
67
3.35
0-7
Mix 2,5%
20
19
86
4.3
115
5.75
1-10
39
1.95
0-5
Mix 1%
20
19
116
5.8
93
4.65
1-9
23
1.15
0-4
20
17
86
4.3
68
3.4
1-9
18
0.9
0-3
Mix 0,5%
Table 3. Mixes at 5 different percentages of the donor at THR 200
THR=20 0
Total mixes
Identified mixes
n° total triplets + quadruplets
Mean triplets + quadruplets
n° total triplets
Mean triplets
Minimummaximum triplets
n° total quadruplets
Mean quadruplets
Minimummaximum quadruplets
Mix 10%
20
20
111
6.2
126
6.2
1-11
44
2.4
0-6
6Mix 5%
20
19
89
4.9
126
4.9
1-8
29
2.2
0-4
Mix 2,5%
20
16
52
3.2
115
3.2
0-7
13
2.1
0-4
Mix 1%
20
14
36
2.6
93
2.6
0-7
12
1.7
0-3
Mix 0,5%
20
13
23
2.5
68
2.5
0-7
7
1.6
0-3
9 Page 12 of 22
A second step of data analysis was the evaluation of the number of triplets or quadruplets (which are informative loci for detecting mixtures of two individuals) at each locus (Table 4). As expected, total triplets and quadruplets numbers in all mixtures at different concentrations gradually
ip t
decrease with the decrement of the donor. Table 4. Number of loci with triplets and quadruplets with different THR (50, 200)
10%
5%
2.5%
1%
0.5%
THR 200
50
200
50
200
D8S1179
16
8
16
7
15
2
D21S11
17
13
15
11
13
5
D7S820
13
8
13
3
10
1
CSF1P0
12
6
11
3
5
D3S1358
11
11
11
11
TH01
13
13
13
D13S317
16
12
16
D16S539
13
12
13
D2S1338
12
10
12
D19S433
10
8
10
vWA
14
12
10
15
50
200
50
200
2
8
3
7
1
6
-
4
1
2
-
3
-
3
-
10
5
7
3
5
1
11
an
4
13
8
12
5
7
4
10
11
5
12
4
5
3
9
8
4
6
5
3
2
10
8
2
6
1
2
1
8
9
6
8
4
8
4
14
12
14
10
13
10
13
10
7
9
6
7
2
8
2
6
2
12
14
10
10
3
7
2
5
3
D18S51
d
te
TPOX
M
11
Ac ce p
us
50 Loci
cr
Mixture (%donor)
D5S818
9
7
11
6
7
4
7
2
5
1
FGA
14
7
13
5
8
1
3
2
3
2
Analyzing all mixtures with both THR 50 and 200, the more informative loci resulted to be vWA, D8S1179, D13S317, D21S11 and TH01, while the less informative one resulted the locus CSF1P0. Figure 3 shows the informative power of each locus at different analytical thresholds. As expected, the informative power of all loci decreases at the increasing of the applied threshold. CSF1PO confirmed to be the less informative locus at both thresholds.
10 Page 13 of 22
Figure 3. Comparison of the percentages of loci with triplets and quadruplets for the mixtures tested with
M
an
us
cr
ip t
THR 50 and THR 200
d
Assays on red blood cells concentrates (RBCC)
te
As previously detailed, RBCC deprived of the buffy-coat were produced from whole blood samples with a leukodepletion procedure based on density gradient centrifugation steps. In a non-
Ac ce p
leukodepleted whole blood transfusion bag of 450 mL, the number of leukocytes can range among 4,500-10,000 per µL. A bag of RBCC usually has a capacity of 250 mL and the residual leukocyte content cannot exceed the amount of about 2700 per µL [16]. Based on these data, it follows that the extraction of DNA from a blood bag coming from a leukodepletion procedure is also technically possible.
In our experiments with leukodepleted blood, we worked with five samples containing fewer WBCs per μL than the amount permitted by law [17]. The number of leukocytes in our RBCC ranged indeed from 60 to 350 per µL. Despite the very low number of WBC, amounts of 0.5 to 6.5 ng of DNA for 50 μL of eluate (0.010.13 ng/µL) was extracted from the leukodepleted samples. Given the low concentrations of extracted DNA, we preferred to perform an amplification of STR loci using the PCR kit-PLUS (Life Technologies), which allows to amplify DNA samples with more sensitivity so improving the
11 Page 14 of 22
performance with respect to the classical PCR used to amplify the extracted DNA from whole
an
us
cr
ip t
blood samples. For each sample, complete profiles have been obtained. (Figure 4)
Figure 4. Example of a RBCC DNA typing
M
Prior to be used for the preparation of mixes, both whole blood samples and RBCC, used as donors, were typed to know in advance their genotype. Five mixtures have been prepared mixing 9 parts of whole blood samples and 1 part of RBCC (10% of the donor).
d
These mixtures present a further imbalance in the number of WBCs of recipient and the
te
donor with respect to mixtures prepared exclusively with whole blood. Considering the numbers of WBCs, the mixtures with RBCC correspond to mixtures of whole blood with a
Ac ce p
range from 0.15% to 0.88% of the donor contribute. Only THR=50 was applied in data analyses; both the ability of the method to recognize mixes based on the presence of triplets and quadruplets and the level of information of each locus have been assessed. In the data analysis, particular attention has been paid to the possibility of confounding donor allelic peaks with stutter peaks. For indeed, the data analysis software usually identifies a stutter peak from an allele when the ratio of the height of the stutter does not exceed the 15% of the reference peak height; in this case the software does not automatically label the stutter. In our mixed samples, the relative height of the allelic peaks of the donors is significantly lower than those of the receivers extracted from whole blood samples, so, a very low allelic peak of the donor, when is placed “near” the main peak (± 1 bp) could be mistaken for a stutter peak (see representative data reported in Table 5).
12 Page 15 of 22
Table 5. Number of informative loci for each mixture, considering possible stutters as peaks (Informative loci) or eliminating them
Loci w/o stutters
1
11
7
2
9
6
3
4
3
4
3
0
5
0
cr
Informativ e loci
us
Mix
ip t
from the analysis (Loci w/o stutters).
an
0
Knowing in advance the typing of all the single samples, it is possible to distinguish
M
between donor allelic peaks and stutters. We moved forward by eliminating each peak that could be confused with a stutter from the analysis in order to verify the applicability of the method also with this kind of mixes. In only one mixture (mixes #5) neither triplets nor
d
quadruplets have been detected, so, that mixture cannot be recognized. In the other four
te
cases, from 3 to 11 loci with triplets or quadruplets have been identified, so that these samples were clearly recognized as mixtures. When the analysis proceeded eliminating
Ac ce p
the peaks that could be confused with stutter peaks (main peak ±1), many loci lost their level of information; despite this, three samples still could be identified as mixes for the presence of 3-7 informative loci. Only mix 4 is no longer recognized as a mixture in fact, referring to the number of WBCs, it could be compared with a mixture of whole blood samples where the donor is 0.5%.
Figure 5 shows the typing of Mix1 loci labeled with NED™; here different circumstances can arise. For the locus D19S433 both the donor and the recipient are heterozygous with one allele in common (namely 14.2); in the mixture only the alleles of the receiver have been called probably due to the low amount of DNA extracted from the donor (allele 12 from the donor and not in common with the receiver, is not detectable). For the locus vWA, the initial condition is the same of D19A433; but in the mixture all of the three alleles have been called (allele 18 from the donor, and not in common with the receiver, is detectable). The donor allele 18 cannot be confused with a stutter peak and it is informative for the 13 Page 16 of 22
recognition of the mixture. For the locus TPOX donor and receiver were both heterozygous with no alleles in common, so allele quadruplet was expected in the mixture (8, 9, 10, 11), but the donor allele 9 does not appear in the mixture as a result of drop-out; anyway the presence of a triplet makes this locus informative. Locus D18S51 presents a quadruplet in
ip t
the profile of the mixture and this clearly indicates the presence of a mixture. Figure 5. Typing of loci labeled with NED™ of mixture1. (A) The RBCC donor sample, (B) the whole blood receiver
Ac ce p
te
d
M
an
us
cr
sample and (C) their mixture.
Besides the presence of triplets and quadruplets, additional information for the identification of mixes can be obtained by an accurate study of the electropherograms. One of them can be represented by the strong imbalance of heights of the peaks detected at one locus. Usually, the presence of only two allelic peaks of similar height in a locus indicates the heterozygosity for that locus (Figure 6A). However, the height ratio between the two peaks can be altered by several factors, including degradation processes and/or other events that could lead to heterozygous unbalance [18]. Mixtures prepared with RBCC resulted with several loci showing apparently unbalanced doublets (see Figure 6B). Such a pattern is typical of mixes in which the receiver is homozygous and the donor is an heterozygous with one common allele or homozygous with a different allele from the receiver. In all cases in which the occurrence of an allelic dropout can be excluded, such circumstances lead to the evidence of a mixed sample.
14 Page 17 of 22
ip t cr
Figure 6. Examples of different allelic doublets. (A) an individual heterozygous; (B) a mixture with recipient and donor homozygous
us
for different alleles.
These signals, combined with the presence of triplets and quadruplets in other loci of the
an
same sample, corroborate the recognition of mixtures, also allowing to understand which
Ac ce p
te
d
M
allele belongs to the receiver and which one to the donor.
15 Page 18 of 22
CONCLUSIONS
Doping control laboratories have the main objective of detecting prohibited substances and methods in athletes’ samples. This work confirms that DNA typing analysis can be applied
ip t
as an alternative, sensitive and specific technology to detect homologous blood transfusion in sport doping. More specifically, DNA typing can overcome the main limitation of the reference flow-cytofluorimetry based method, which is the risk of reporting
cr
false-negative results for samples in which the donor and the receiver show the same profile in terms of target surface antigens of the red blood cell. DNA typing can instead
us
recognize mixed samples by the presence of more than two alleles in some locus, in particular of triplets and quadruplets in two individual mixtures. The sensitivity to recognize
an
mixtures is strictly connected to the percent of donor blood in the receiver and to the choice of the analytical threshold; anyway sensitivity is maximal (100%) with both thresholds (200 and 50 RFU) when donor and receiver ratio is at least 1:9 (donor at 10%).
M
Sensitivity and the number of information decrease with the reduction of the donor percentage; nevertheless, also with the smaller percentage of the donor (0.5%), mixtures were successfully typed and correctly identified in 65% of the cases. Some loci resulted
d
always more informative than others with both THRs, e.g. vWA, while on the contrary,
te
others were the less informative in each mix (CSF1PO). The choice of the right THR resulted critical to avoid the losing of valuable information and to avoid the
Ac ce p
misinterpretation of background signals as allelic peaks. This analytical method is also applicable to mixtures of whole blood and RBCC, despite the strongly reduced number of nucleated cells of the donor present in the latter.
16 Page 19 of 22
REFERENCES The World Anti-Doping Code. International Standard. Prohibited List. The World AntiDoping Agency, Montreal, Canada, September 2015. Available on line at: https://wadamain-prod.s3.amazonaws.com/resources/files/wada-2016-prohibited-list-en.pdf (last accessed: October 10th, 2015).
[2]
B. Ekblom, A.N. Goldbarg, B. Gullbring, Response to exercise after blood loss and reinfusion., J. Appl. Physiol. 33 (1972) 175–80.
[3]
F. Lasne, J. De Ceaurriz, Recombinant erythropoietin in urine, Nature 405 (2000) 635.
[4]
J. Segura, J.A. Pascual, R. Gutiérrez-Gallego, Procedures for monitoring recombinant erythropoietin and analogues in doping control., Anal. Bioanal. Chem. 388 (2007) 1521–9. doi:10.1007/s00216-007-1316-x.
[5]
J. Mørkeberg, Blood manipulation: current challenges from an anti-doping perspective., Hematology Am. Soc. Hematol. Educ. Program. 2013 (2013) 627–31. doi:10.1182/asheducation-2013.1.627.
[6]
S. Giraud, N. Robinson, P. Mangin, M. Saugy, Scientific and forensic standards for homologous blood transfusion anti-doping analyses. Forensic Sci. Int. 179 (2008) 23–33. doi:10.1016/j.forsciint.2008.04.007.
[7]
M. Nelson, H. Popp, K. Sharpe, M. Ashenden, Proof of homologous blood transfusion through quantification of blood group antigens., Haematologica. 88 (2003) 1284–95. http://www.ncbi.nlm.nih.gov/pubmed/14607758 (accessed September 17, 2015).
[8]
P.A. Arndt, B.M. Kumpel, Blood doping in athletes--detection of allogeneic blood transfusions by flow cytofluorimetry., Am. J. Hematol. 83 (2008) 657–67. doi:10.1002/ajh.21196.
[9]
F. Donati, A. Stampella, X. de la Torre, F. Botrè, Investigation on the application of DNA forensic human identification techniques to detect homologous blood transfusions in doping control, Talanta 110 (2013) 28–31. doi:10.1016/j.talanta.2013.02.042.
[10]
I. Manokhina, J.L. Rupert. A DNA-based method for detecting homologous blood doping. Anal Bioanal Chem. 2013 Dec;405(30):9693-701
Ac ce p
te
d
M
an
us
cr
ip t
[1]
[11] J.M. Butler, Short tandem repeat analysis for human identity testing., Curr. Protoc. Hum. Genet. Chapter 14 (2004) Unit 14.8. doi:10.1002/0471142905.hg1408s41. [12]
P. Gill, R. Sparkes, R. Pinchin, T. Clayton, J. Whitaker, J. Buckleton, Interpreting simple STR mixtures using allele peak areas., Forensic Sci. Int. 91 (1998) 41–53.
[13]
P. Gill, C.H. Brenner, J.S. Buckleton, A. Carracedo, M. Krawczak, W.R. Mayr, et al., DNA commission of the International Society of Forensic Genetics: Recommendations on the interpretation of mixtures., Forensic Sci. Int. 160 (2006) 90–101. doi:10.1016/j.forsciint.2006.04.009.
[14]
R. Puch-Solis, L. Rodgers, A. Mazumder, S. Pope, I. Evett, J. Curran, et al., Evaluating forensic DNA profiles using peak heights, allowing for multiple donors, allelic dropout and stutters., Forensic Sci. Int. Genet. 7 (2013) 555–63. doi:10.1016/j.fsigen.2013.05.009.
[15]
H. Haned, C.C.G. Benschop, P.D. Gill, T. Sijen, Complex DNA mixture analysis in a forensic context: evaluating the probative value using a likelihood ratio model., Forensic Sci. Int. Genet. 16 (2015) 17–25. doi:10.1016/j.fsigen.2014.11.014. 17 Page 20 of 22
European Directorate for the Quality of Medicines & HealthCare, Guide to the Preparation, Use and Quality Assurance of Blood Components, Recommendation No. R (95) 15, (n.d.). http://www.avis.it/repository/cont_schedemm/4877_documento.pdf (accessed September 16, 2015).
[17]
G. Liumbruno, F. Bennardello, A. Lattanzio, P. Piccoli, G. Rossetti, Recommendations for the transfusion of red blood cells., Blood Transfus. 7 (2009) 49–64. doi:10.2450/2008.0020-08.
[18]
A. Debernardi, E. Suzanne, A. Formant, L. Pène, A.B. Dufour, J.R. Lobry, One year variability of peak heights, heterozygous balance and inter-locus balance for the DNA positive control of AmpFℓSTR© Identifiler© STR kit., Forensic Sci. Int. Genet. 5 (2011) 43–9.
Ac ce p
te
d
M
an
us
cr
ip t
[16]
18 Page 21 of 22
ACKNOWLEDGEMENT
This project has been supported in part by the Italian National Anti-Doping Commission
Ac
ce pt
ed
M
an
us
cr
ip t
(Commissione per la Vigilanza sul Doping) of the Italian Ministry of Health.
Page 22 of 22
S0379-0738(16)30173-6 http://dx.doi.org/doi:10.1016/j.forsciint.2016.04.021 FSI 8437
To appear in:
FSI
Received date: Revised date: Accepted date:
15-11-2015 6-4-2016 15-4-2016
Please cite this article as: A. Stampella, S. Di Marco, D. Pirri, X. de la Torre, F. Botrgravee, F. Donati, Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control, Forensic Science International (2016), http://dx.doi.org/10.1016/j.forsciint.2016.04.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ce pt
ed
M
an
us
cr
ip t
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control
Homologous blood transfusions, a form of “blood doping” in sport, can be detected by forensic genetic techniques based on DNA typing. The proposed procedure, that tested effective also on mixed samples of red blood cell concentrates, allows to correctly identify also those samples that
Ac
would have given a false-negative result if assayed by the reference cytofluorimetric technique.
Page 1 of 22
Highlights
Forensic genetics techniques are proposed to detect homologous blood transfusions. The proposed method, based on DNA typing, tested effective on ex vivo blood mixtures
Mixed blood samples are revealed by triplets and quadruplets at one or more loci
The same procedure can be applied also to red blood cell concentrates
Ac
ce pt
ed
M
an
us
cr
ip t
Page 2 of 22
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control Alessandra Stampella1, Sabrina Di Marco1, Daniela Pirri1, Xavier de la Torre1, Francesco Botrè1, 2*,
ip t
Francesco Donati1
1: Laboratorio Antidoping, Federazione Medico Sportiva Italiana, Largo Giulio Onesti 1, 00197, Rome, Italy
cr
2: Dipartimento di Medicina Sperimentale, “Sapienza” Università di Roma, Viale Regina Elena 324, 00161
us
Rome, Italy
*Corresponding author
an
Prof. Francesco Botrè
ed
ce pt
phone: +39-06-87973500 fax: +39-06-8078971 email: [email protected]
M
Laboratorio Antidoping Federazione Medico Sportiva Italiana Largo Giulio Onesti 1 00197 Rome Italy
Keywords: Homologous blood transfusion; blood doping; doping analysis; red blood cells
Ac
concentrates; DNA typing; short tandem repeats.
Page 3 of 22
Application of DNA-based forensic analysis for the detection of homologous transfusion of whole blood and of red blood cell concentrates in doping control
ip t
ABSTRACT
cr
In this work we present the application of a method for the identification of homologous blood transfusions using forensic genetic techniques based on DNA typing. Ex vivo mixtures of human
us
blood samples - either whole blood or red blood cell concentrates - simulating homologous blood transfusions at different percentages of the donor were typed for a panel of 16 highly variable DNA short tandem repeats (STR). Tested samples included also mixtures, which gave false-
an
negative results if assayed by the reference flow cytofluorimetric method, which is based on the recognition of target antigens located on the membrane of the red blood cell. The recognition of
M
triplets and quadruplets at various loci gave information of the presence of cells belonging to different individuals, as it is the case for homologous blood transfusions. Specificity and sensitivity
d
of the method were assessed in the validation study. The method proved to be unequivocally specific since it was able to recognize all single profiles of each individual, clearly discriminating
te
them from mixtures. Sensitivity resulted as a consequence of the percentage of the donor aliquot in the total volume of the mixture. Although the source of DNA in a blood sample is represented
Ac ce p
only by nucleated white blood cells, the same procedure resulted effective also in detecting mixtures of red blood cell concentrates (RBCC) from leukodepletion procedure: DNA of the donor from the residual white blood cells resulted still detectable, even if with an expected loss of sensitivity. The proposed approach may contribute to reduce the risk of false-negative results, which may occur using the reference cytofluorimetric method.
Keywords: Homologous blood transfusion; blood doping; doping analysis; red blood cells concentrates; DNA typing; short tandem repeats.
1 Page 4 of 22
INTRODUCTION In sport doping, the administration or reintroduction into the circulatory system of any quantity of autologous, allogenic (homologous) or heterologous blood or of any red blood cell products, of
ip t
any origin, is prohibited by the World Anti-Doping Agency (WADA) [1]. Specifically, homologous blood transfusion (HBT) is a blood manipulation practice, which can be attractive for cheating athletes because of the enhancement of oxygen carrying to the tissues [2]. HBT is based on the
cr
infusion of whole blood (WB) or red blood cell concentrates (RBCC) from a donor to a recipient
us
subject whose blood is compatible for the matching of the major blood group antigens (ABO and Rh factor).
HBT was widely used in the late 70s-80s until it was banned in sport, first by the
an
International Olympic Committee and then by the WADA. In the 90s, Erythropoietin (EPO) was considered by athletes a safer and more effective method to illicitly boost the oxygen transport
M
capacity in blood; later on, following the implementation of a direct method to detect the intake of EPO and analogues in urine [3-4], cheating athletes reconsidered the practice of blood transfusion. The implementation of the first method for the detection of HBT dates back to 2004
d
and was first used on the occasion of the anti-doping test for the Athens Summer Olympic Games:
te
in the same year the first adverse analytical finding for HBT was also reported [5]. The direct method for the detection of the abuse of homologous blood transfusion (HBT) is
Ac ce p
still in use by the WADA-accredited laboratories and is based on the recognition of different population of erythrocytes, both of the donor and of the receiver, in the same blood sample. This method is based on the identification, by flow cytofluorimetry, of antigens of minor human blood groups. The test is indeed based on the analysis of 8 different antigens belonging to four different human blood groups (namely big C, small c and big E of the Rh group, Jka and Jkb of the Kidd group, Fya and Fyb of the duffy group and big S of the MNS group) [6–7]. Despite the proven performances of this method in terms of sensitivity, it may happen that
two individuals express an identical antigenic profile, leading to the possibility of false-negative results [8-9]. This can be considered the most significant limitation of the flow cytofluorimetric method for HBT. In our laboratory, we first tried to solve this limitation by increasing the panel of whole blood antigens to be analyzed (extending it from 8 antigens of the basic panels to 13 antigens by an “extended” panel). Surely the discrimination power of the method was increased,
2 Page 5 of 22
but clearly false negative results, although with a much lower statistical frequency, could still occur. For this reason we have recently explored the utility of DNA based forensic analysis, presenting a preliminary analytical approach based on DNA genotypic differences and no longer on phenotypic
ip t
differences between the donor and the receiver [9]. In parallel, Manokhina et al. proposed another DNA based-approach for the detection of homologous blood transfusion as an illicit doping practice [10]. Their work focused on the possibility to discriminate donor blood from the
cr
receiver by a genotyping and sensitive strategy with detection on quantitative PCR, thus
us
confirming utility and versality of DNA-based investigative approaches for the detection of homologous blood transfusion.
We are here proposing to follow a forensic human DNA typing strategy, based on the analysis of
an
little portions of DNA called Short Tandem Repeats (STRs). Each locus STR may show a wide number of alleles, which differs for the number of the repeated sequences. 15 high variable
M
autosomal STRs plus the amelogenin locus (used for sex determination) are normally tested in human identification typing [11]. The probability that two different individuals share the same haplotype pattern is virtually zero. DNA extracted from different sources (e.g. blood, urine, buccal
d
swab) is amplified by PCR into a master mix containing the primer set of all of the 16 loci of the
te
identification panel (in brief D8S1179, D21S11, D7S820, CSF1PO, D3S1358, TH01, D13S317, D16S539, D2S1338, D19S433, vWA, TPOX, D18S51, D5S818, FGA and amelogenin). Each primer is
Ac ce p
specific for a certain locus. PCR generates a very high number of fragments containing the alleles of interest. The amplified sample is then analyzed in a capillary electrophoresis system to separate the fragments locus by locus. Detection of different alleles is made by a five-color detection system. It is possible to identify an individual for the high inter-individual variability of alleles at each STR locus that depends on the number of repetitions of the repeated sequence. Human genome is diploid and the profile of a single (non-transfused) individual is unique and can present one or two allelic peaks for each locus, indicating homozygosity or heterozygosity respectively for that locus. Our preliminary results showed that DNA typing, thanks to the very high discrimination power of DNA analysis, is an alternative/complementary analytical technique, which enhances the specificity of the classical flow cytofluorimetry-based method. In case of whole blood or RBCC transfusions with a sufficient residual population of nucleated cells (e.g. white blood cells), the profile of a mixture of blood originated from two different subjects can be recognized by the presence of more than two allelic peaks (triplets and quadruplets) at various loci [12-13]. In a two3 Page 6 of 22
person mixture the interpretation becomes difficult when the minor profile is less than one-third of the response of the major profile, since it could be difficult to distinguish true alleles from stutter peaks [14-15]. In this work, the forensic DNA method based on STRs is proposed as an
ip t
alternative/complementary way to detect different forms of HBT (either of whole blood or of RBCC), also to overcome the limitation due to the risk of false negative results that may occur by applying the current reference cytofluorimetric method, based on the recognition of target
Ac ce p
te
d
M
an
us
cr
antigens located on the membrane of the red blood cell.
4 Page 7 of 22
MATERIAL AND METHODS Blood Samples Whole blood samples of several different individuals were used to prepare 20 different mixed
ip t
samples with four different proportions of the donor (10, 5, 2.5, 1, 0.5%) for a total number of 100 mixes to be analyzed. Prior to mixing, AB0 group and Rh factor were determined by a hemagglutination assay (DiaClon BIO-RAD antibodies anti-A, anti-B and anti-D) and minor antigens
cr
by flow cytofluorimetry were analyzed on each single sample. Samples were mixed in order to
us
simulate a false-negative result coming from a flow cytofluorimetric screening, which means that samples were mixed only when not more than one antigenic difference was first detected. The same analytical approach was applied on mixed samples obtained after leukodepletion of the
an
donor using Ficoll-Paque PLUS (GE Healthcare). Mixtures at < 10% of the donor were produced from red blood cell concentrates. Residual white blood cells were counted by Sysmex XE-2100
M
Automated Hematology Blood Analyser to evaluate whether their number was comparable to
Flow Cytofluorimetry Analyses
d
those admitted by law in a normal transfusion routine practice.
Ac ce p
by Giraud et al. 2008 [6].
te
Phenotyping of whole blood samples prior the generation of the mixes was performed as detailed
Preparation of concentrated RBC
Ficoll-Paque™ PLUS (GE Healthcare) was used to obtain concentrated RBC from 1 mL of whole blood sample. The preparation was performed according to the manufacturer’s recommendations. White blood cell (WBC) count
White blood cells residue number was estimated using Sysmex XE-2100 hematologic analyser (Dasit). For the determination of the number of total white blood cell and the number of each type of blood cell present in the concentrated samples, the method adopted is flow cytofluorimetric with the use of a semiconductor laser. DNA Genomic DNA was isolated from 2µL of sample (either whole blood or RBCC) using the PrepFiler™ 5 Page 8 of 22
Forensic DNA Extraction Kit (Applied Biosystems®). All DNA extracts were quantified in triplicate by real time PCR using the Quantifiler® Human DNA Quantification Kit (Applied Biosystems®) and the Applied Biosystems 7500 Fast real-time PCR system. Quantified samples were then amplified with PCR using 10 µL of 0.1ng/µL of the final extract; any dilution was performed using the elution
ip t
buffer of PrepFiler™ Forensic DNA Extraction Kit (Applied Biosystems®). This amplification step was performed on the GeneAmp® PCR System 9700 thermocycler using the AmpFlSTR ® Identifiler® PCR Amplification kit or AmpFlSTR® Identifiler® Plus PCR Amplification with the
cr
recommended amplification conditions (a total of 28 cycles, annealing temperature 59°C,
us
extension temperature 72°C as advised by manufacturer). DNA typing was executed with the ABI PRISM 3100 Genetic Analyser (Applied Biosystems®). In brief, 1.5 µL of amplified DNA was diluted in 12 μL of form amide and 1µL of GeneScan™-500 Size Standard (Applied Biosystems®); POP-4
an
polymer was used and 5s injections were applied. All data were analyzed using the GeneMapper™ ID Software (Applied Biosystems®), which performs the allele calling by the help of a reference
Ac ce p
te
d
M
DNA Allelic Ladder containing all the alleles of each locus analyzed.
6 Page 9 of 22
RESULTS AND DISCUSSION Specificity of the DNA-based method Specificity was assessed by evaluating the ability of the analytical method to accurately obtain a
ip t
DNA profile from individual samples in the presence of interfering and not confusing them as mixtures. The aim was to verify whether it was possible the identification of single individual gene profiles without any triplet and/or quadruplet, typical of mixed samples. The method resulted
cr
unequivocally specific, since no samples belonging to single individual were identified as mixtures.
us
For each locus only one or two peaks have been identified, indicative respectively of homozygosity
M
an
and heterozygosity for that locus (Figure 1).
Figure 1. Example of a four loci DNA profile for an individual carrying two heterozygous loci (double peaks D19S433 and VWA) and
te
d
two homozygous loci (single peaks: TPOX and D18S51).
Sensitivity: identification of mixtures
Ac ce p
From a pool of whole blood samples, 20 different mixtures were produced. Blood major groups antigenic profile was first estimated by hemagglutination assay in order to produce mixtures where donor and receiver were compatible for ABO and Rh factor (positive + or negative according to the presence of D antigen or less). Five A+ mixes, one B+ mix, three 0+ mixes and one 0- mix were prepared. Ten mixes with compatible donor/receiving have been prepared as follows: three 0+/A+, two 0+/B+; one 0+/AB+; one 0-/A-; one 0-/A+; one B+/AB+ and one A+/AB+. 32 single samples have been first analyzed by the flow-cytofluorimetric method; all 20 mixtures we used for the following DNA analysis resulted as "false negatives" with the flow-cytofluorimetric method (an example of such a case is shown in Figure 2). Note that a false-negative result refers to the result of a flow cytofluorimetric assay where no more than one antigenic difference is recorded. For each pair of samples 5 different mixtures, with decreasing concentrations of the donor (from 10% down to 0.5%, for a total of 100 final mixtures), were considered.
7 Page 10 of 22
ip t cr
us
Figure 2. Analysis of a HBT sample in which donor and receiver share identical single haplotype for 13 different antigens of whole blood minor groups. Flow cytofluorimetric analysis cannot discriminate between the contributions given by each sample (donor or receiver) to the resulting flow cytofluorimetric pattern, so that no double populations of erythrocytes can be recorded, leading to a
an
false-negative result.
The highest percentage of the donor (10%) has been chosen to simulate the situation of a recent
M
transfusion of a whole blood bag (450 mL). The other percentages of donor (5%, 2.5%, 1% and 0.5%) simulate a blood sample collection at various times after the transfusion of blood to the
d
receiver, to consider the period of blood replacement after the transfusion. The low concentration cytofluorimetric method.
te
of the donor at 0.5% was considered because quite close to the limit of sensitivity of the flow
Ac ce p
The presence of triplets or quadruplets at one or more loci is the essential information for the identification of a mixture; it might occur that no triplets and quadruplets were detected in a mixture, but the probability of not having any informative locus out of the 15 target loci is considered to be extremely low, close to the statistical zero. Table 1 Sensitivity in percentage (%) of the DNA typing method for the five different mixtures with THR=50 and THR=200. Whole blood donor (%)
Sensitivity (%) THR 50
Sensitivity (%) THR 200
10
100
100
5
100
95
2.5
95
80
1
95
70
0.5
85
65
8 Page 11 of 22
The choice of an appropriate analytical threshold (THR) is crucial for the interpretation of the mixtures. In this work, thresholds at 50 and 200 relative fluorescent units (RFU) were considered. A relative low threshold, such as 50 RFU, takes both advantages and disadvantages; in fact, more information and identification power of very small peak alleles could be obtained, but on the
ip t
contrary, background noise peaks or "stutters" peaks may be erroneously evaluated as true peaks. The use of a high threshold (e.g. 200 RFU) allows avoiding most of the interpretation problems, but, at the same time, may not lead to an allelic call of the peak. The choice of the right THR
cr
should be carefully evaluated and may be determined to reach a sort of “compromise” between
us
the necessity to identify a peak and the parallel need to discard any not significant signal. As shown by Tab.1, sensibility of 100% was achieved with mixtures at 10% of the donor independently by the chosen THR. Mixtures at 5% of the donor shown sensitivity of 100% only
an
with THR=50, while sensitivity slightly decreases at 95% when THR = 200 was applied. As expected, with the reduction of donor concentration, the percentage of identified mixtures decrease, but
M
anyway in the worst situation (mixes at 0.5% of the donor) more than 50% of the samples were correctly recognized as mixes (Tables 2 and 3).
te
THR=5 0
d
Table 2. Mixes at 5 different percentages of the donor at THR 50
Identified mixes
n° total triplets + quadruplets
Mean triplets + quadruplets
n° total triplets
Mean triplets
Minimummaximum triplets
n° total quadruplets
Mean quadruplets
Minimummaximum quadruplets
Mix 10%
20
20
191
9.5
126
6.3
3-11
65
3.25
0-7
Mix 5%
Ac ce p
Total mixes
20
20
193
9.6
126
6.3
2-10
67
3.35
0-7
Mix 2,5%
20
19
86
4.3
115
5.75
1-10
39
1.95
0-5
Mix 1%
20
19
116
5.8
93
4.65
1-9
23
1.15
0-4
20
17
86
4.3
68
3.4
1-9
18
0.9
0-3
Mix 0,5%
Table 3. Mixes at 5 different percentages of the donor at THR 200
THR=20 0
Total mixes
Identified mixes
n° total triplets + quadruplets
Mean triplets + quadruplets
n° total triplets
Mean triplets
Minimummaximum triplets
n° total quadruplets
Mean quadruplets
Minimummaximum quadruplets
Mix 10%
20
20
111
6.2
126
6.2
1-11
44
2.4
0-6
6Mix 5%
20
19
89
4.9
126
4.9
1-8
29
2.2
0-4
Mix 2,5%
20
16
52
3.2
115
3.2
0-7
13
2.1
0-4
Mix 1%
20
14
36
2.6
93
2.6
0-7
12
1.7
0-3
Mix 0,5%
20
13
23
2.5
68
2.5
0-7
7
1.6
0-3
9 Page 12 of 22
A second step of data analysis was the evaluation of the number of triplets or quadruplets (which are informative loci for detecting mixtures of two individuals) at each locus (Table 4). As expected, total triplets and quadruplets numbers in all mixtures at different concentrations gradually
ip t
decrease with the decrement of the donor. Table 4. Number of loci with triplets and quadruplets with different THR (50, 200)
10%
5%
2.5%
1%
0.5%
THR 200
50
200
50
200
D8S1179
16
8
16
7
15
2
D21S11
17
13
15
11
13
5
D7S820
13
8
13
3
10
1
CSF1P0
12
6
11
3
5
D3S1358
11
11
11
11
TH01
13
13
13
D13S317
16
12
16
D16S539
13
12
13
D2S1338
12
10
12
D19S433
10
8
10
vWA
14
12
10
15
50
200
50
200
2
8
3
7
1
6
-
4
1
2
-
3
-
3
-
10
5
7
3
5
1
11
an
4
13
8
12
5
7
4
10
11
5
12
4
5
3
9
8
4
6
5
3
2
10
8
2
6
1
2
1
8
9
6
8
4
8
4
14
12
14
10
13
10
13
10
7
9
6
7
2
8
2
6
2
12
14
10
10
3
7
2
5
3
D18S51
d
te
TPOX
M
11
Ac ce p
us
50 Loci
cr
Mixture (%donor)
D5S818
9
7
11
6
7
4
7
2
5
1
FGA
14
7
13
5
8
1
3
2
3
2
Analyzing all mixtures with both THR 50 and 200, the more informative loci resulted to be vWA, D8S1179, D13S317, D21S11 and TH01, while the less informative one resulted the locus CSF1P0. Figure 3 shows the informative power of each locus at different analytical thresholds. As expected, the informative power of all loci decreases at the increasing of the applied threshold. CSF1PO confirmed to be the less informative locus at both thresholds.
10 Page 13 of 22
Figure 3. Comparison of the percentages of loci with triplets and quadruplets for the mixtures tested with
M
an
us
cr
ip t
THR 50 and THR 200
d
Assays on red blood cells concentrates (RBCC)
te
As previously detailed, RBCC deprived of the buffy-coat were produced from whole blood samples with a leukodepletion procedure based on density gradient centrifugation steps. In a non-
Ac ce p
leukodepleted whole blood transfusion bag of 450 mL, the number of leukocytes can range among 4,500-10,000 per µL. A bag of RBCC usually has a capacity of 250 mL and the residual leukocyte content cannot exceed the amount of about 2700 per µL [16]. Based on these data, it follows that the extraction of DNA from a blood bag coming from a leukodepletion procedure is also technically possible.
In our experiments with leukodepleted blood, we worked with five samples containing fewer WBCs per μL than the amount permitted by law [17]. The number of leukocytes in our RBCC ranged indeed from 60 to 350 per µL. Despite the very low number of WBC, amounts of 0.5 to 6.5 ng of DNA for 50 μL of eluate (0.010.13 ng/µL) was extracted from the leukodepleted samples. Given the low concentrations of extracted DNA, we preferred to perform an amplification of STR loci using the PCR kit-PLUS (Life Technologies), which allows to amplify DNA samples with more sensitivity so improving the
11 Page 14 of 22
performance with respect to the classical PCR used to amplify the extracted DNA from whole
an
us
cr
ip t
blood samples. For each sample, complete profiles have been obtained. (Figure 4)
Figure 4. Example of a RBCC DNA typing
M
Prior to be used for the preparation of mixes, both whole blood samples and RBCC, used as donors, were typed to know in advance their genotype. Five mixtures have been prepared mixing 9 parts of whole blood samples and 1 part of RBCC (10% of the donor).
d
These mixtures present a further imbalance in the number of WBCs of recipient and the
te
donor with respect to mixtures prepared exclusively with whole blood. Considering the numbers of WBCs, the mixtures with RBCC correspond to mixtures of whole blood with a
Ac ce p
range from 0.15% to 0.88% of the donor contribute. Only THR=50 was applied in data analyses; both the ability of the method to recognize mixes based on the presence of triplets and quadruplets and the level of information of each locus have been assessed. In the data analysis, particular attention has been paid to the possibility of confounding donor allelic peaks with stutter peaks. For indeed, the data analysis software usually identifies a stutter peak from an allele when the ratio of the height of the stutter does not exceed the 15% of the reference peak height; in this case the software does not automatically label the stutter. In our mixed samples, the relative height of the allelic peaks of the donors is significantly lower than those of the receivers extracted from whole blood samples, so, a very low allelic peak of the donor, when is placed “near” the main peak (± 1 bp) could be mistaken for a stutter peak (see representative data reported in Table 5).
12 Page 15 of 22
Table 5. Number of informative loci for each mixture, considering possible stutters as peaks (Informative loci) or eliminating them
Loci w/o stutters
1
11
7
2
9
6
3
4
3
4
3
0
5
0
cr
Informativ e loci
us
Mix
ip t
from the analysis (Loci w/o stutters).
an
0
Knowing in advance the typing of all the single samples, it is possible to distinguish
M
between donor allelic peaks and stutters. We moved forward by eliminating each peak that could be confused with a stutter from the analysis in order to verify the applicability of the method also with this kind of mixes. In only one mixture (mixes #5) neither triplets nor
d
quadruplets have been detected, so, that mixture cannot be recognized. In the other four
te
cases, from 3 to 11 loci with triplets or quadruplets have been identified, so that these samples were clearly recognized as mixtures. When the analysis proceeded eliminating
Ac ce p
the peaks that could be confused with stutter peaks (main peak ±1), many loci lost their level of information; despite this, three samples still could be identified as mixes for the presence of 3-7 informative loci. Only mix 4 is no longer recognized as a mixture in fact, referring to the number of WBCs, it could be compared with a mixture of whole blood samples where the donor is 0.5%.
Figure 5 shows the typing of Mix1 loci labeled with NED™; here different circumstances can arise. For the locus D19S433 both the donor and the recipient are heterozygous with one allele in common (namely 14.2); in the mixture only the alleles of the receiver have been called probably due to the low amount of DNA extracted from the donor (allele 12 from the donor and not in common with the receiver, is not detectable). For the locus vWA, the initial condition is the same of D19A433; but in the mixture all of the three alleles have been called (allele 18 from the donor, and not in common with the receiver, is detectable). The donor allele 18 cannot be confused with a stutter peak and it is informative for the 13 Page 16 of 22
recognition of the mixture. For the locus TPOX donor and receiver were both heterozygous with no alleles in common, so allele quadruplet was expected in the mixture (8, 9, 10, 11), but the donor allele 9 does not appear in the mixture as a result of drop-out; anyway the presence of a triplet makes this locus informative. Locus D18S51 presents a quadruplet in
ip t
the profile of the mixture and this clearly indicates the presence of a mixture. Figure 5. Typing of loci labeled with NED™ of mixture1. (A) The RBCC donor sample, (B) the whole blood receiver
Ac ce p
te
d
M
an
us
cr
sample and (C) their mixture.
Besides the presence of triplets and quadruplets, additional information for the identification of mixes can be obtained by an accurate study of the electropherograms. One of them can be represented by the strong imbalance of heights of the peaks detected at one locus. Usually, the presence of only two allelic peaks of similar height in a locus indicates the heterozygosity for that locus (Figure 6A). However, the height ratio between the two peaks can be altered by several factors, including degradation processes and/or other events that could lead to heterozygous unbalance [18]. Mixtures prepared with RBCC resulted with several loci showing apparently unbalanced doublets (see Figure 6B). Such a pattern is typical of mixes in which the receiver is homozygous and the donor is an heterozygous with one common allele or homozygous with a different allele from the receiver. In all cases in which the occurrence of an allelic dropout can be excluded, such circumstances lead to the evidence of a mixed sample.
14 Page 17 of 22
ip t cr
Figure 6. Examples of different allelic doublets. (A) an individual heterozygous; (B) a mixture with recipient and donor homozygous
us
for different alleles.
These signals, combined with the presence of triplets and quadruplets in other loci of the
an
same sample, corroborate the recognition of mixtures, also allowing to understand which
Ac ce p
te
d
M
allele belongs to the receiver and which one to the donor.
15 Page 18 of 22
CONCLUSIONS
Doping control laboratories have the main objective of detecting prohibited substances and methods in athletes’ samples. This work confirms that DNA typing analysis can be applied
ip t
as an alternative, sensitive and specific technology to detect homologous blood transfusion in sport doping. More specifically, DNA typing can overcome the main limitation of the reference flow-cytofluorimetry based method, which is the risk of reporting
cr
false-negative results for samples in which the donor and the receiver show the same profile in terms of target surface antigens of the red blood cell. DNA typing can instead
us
recognize mixed samples by the presence of more than two alleles in some locus, in particular of triplets and quadruplets in two individual mixtures. The sensitivity to recognize
an
mixtures is strictly connected to the percent of donor blood in the receiver and to the choice of the analytical threshold; anyway sensitivity is maximal (100%) with both thresholds (200 and 50 RFU) when donor and receiver ratio is at least 1:9 (donor at 10%).
M
Sensitivity and the number of information decrease with the reduction of the donor percentage; nevertheless, also with the smaller percentage of the donor (0.5%), mixtures were successfully typed and correctly identified in 65% of the cases. Some loci resulted
d
always more informative than others with both THRs, e.g. vWA, while on the contrary,
te
others were the less informative in each mix (CSF1PO). The choice of the right THR resulted critical to avoid the losing of valuable information and to avoid the
Ac ce p
misinterpretation of background signals as allelic peaks. This analytical method is also applicable to mixtures of whole blood and RBCC, despite the strongly reduced number of nucleated cells of the donor present in the latter.
16 Page 19 of 22
REFERENCES The World Anti-Doping Code. International Standard. Prohibited List. The World AntiDoping Agency, Montreal, Canada, September 2015. Available on line at: https://wadamain-prod.s3.amazonaws.com/resources/files/wada-2016-prohibited-list-en.pdf (last accessed: October 10th, 2015).
[2]
B. Ekblom, A.N. Goldbarg, B. Gullbring, Response to exercise after blood loss and reinfusion., J. Appl. Physiol. 33 (1972) 175–80.
[3]
F. Lasne, J. De Ceaurriz, Recombinant erythropoietin in urine, Nature 405 (2000) 635.
[4]
J. Segura, J.A. Pascual, R. Gutiérrez-Gallego, Procedures for monitoring recombinant erythropoietin and analogues in doping control., Anal. Bioanal. Chem. 388 (2007) 1521–9. doi:10.1007/s00216-007-1316-x.
[5]
J. Mørkeberg, Blood manipulation: current challenges from an anti-doping perspective., Hematology Am. Soc. Hematol. Educ. Program. 2013 (2013) 627–31. doi:10.1182/asheducation-2013.1.627.
[6]
S. Giraud, N. Robinson, P. Mangin, M. Saugy, Scientific and forensic standards for homologous blood transfusion anti-doping analyses. Forensic Sci. Int. 179 (2008) 23–33. doi:10.1016/j.forsciint.2008.04.007.
[7]
M. Nelson, H. Popp, K. Sharpe, M. Ashenden, Proof of homologous blood transfusion through quantification of blood group antigens., Haematologica. 88 (2003) 1284–95. http://www.ncbi.nlm.nih.gov/pubmed/14607758 (accessed September 17, 2015).
[8]
P.A. Arndt, B.M. Kumpel, Blood doping in athletes--detection of allogeneic blood transfusions by flow cytofluorimetry., Am. J. Hematol. 83 (2008) 657–67. doi:10.1002/ajh.21196.
[9]
F. Donati, A. Stampella, X. de la Torre, F. Botrè, Investigation on the application of DNA forensic human identification techniques to detect homologous blood transfusions in doping control, Talanta 110 (2013) 28–31. doi:10.1016/j.talanta.2013.02.042.
[10]
I. Manokhina, J.L. Rupert. A DNA-based method for detecting homologous blood doping. Anal Bioanal Chem. 2013 Dec;405(30):9693-701
Ac ce p
te
d
M
an
us
cr
ip t
[1]
[11] J.M. Butler, Short tandem repeat analysis for human identity testing., Curr. Protoc. Hum. Genet. Chapter 14 (2004) Unit 14.8. doi:10.1002/0471142905.hg1408s41. [12]
P. Gill, R. Sparkes, R. Pinchin, T. Clayton, J. Whitaker, J. Buckleton, Interpreting simple STR mixtures using allele peak areas., Forensic Sci. Int. 91 (1998) 41–53.
[13]
P. Gill, C.H. Brenner, J.S. Buckleton, A. Carracedo, M. Krawczak, W.R. Mayr, et al., DNA commission of the International Society of Forensic Genetics: Recommendations on the interpretation of mixtures., Forensic Sci. Int. 160 (2006) 90–101. doi:10.1016/j.forsciint.2006.04.009.
[14]
R. Puch-Solis, L. Rodgers, A. Mazumder, S. Pope, I. Evett, J. Curran, et al., Evaluating forensic DNA profiles using peak heights, allowing for multiple donors, allelic dropout and stutters., Forensic Sci. Int. Genet. 7 (2013) 555–63. doi:10.1016/j.fsigen.2013.05.009.
[15]
H. Haned, C.C.G. Benschop, P.D. Gill, T. Sijen, Complex DNA mixture analysis in a forensic context: evaluating the probative value using a likelihood ratio model., Forensic Sci. Int. Genet. 16 (2015) 17–25. doi:10.1016/j.fsigen.2014.11.014. 17 Page 20 of 22
European Directorate for the Quality of Medicines & HealthCare, Guide to the Preparation, Use and Quality Assurance of Blood Components, Recommendation No. R (95) 15, (n.d.). http://www.avis.it/repository/cont_schedemm/4877_documento.pdf (accessed September 16, 2015).
[17]
G. Liumbruno, F. Bennardello, A. Lattanzio, P. Piccoli, G. Rossetti, Recommendations for the transfusion of red blood cells., Blood Transfus. 7 (2009) 49–64. doi:10.2450/2008.0020-08.
[18]
A. Debernardi, E. Suzanne, A. Formant, L. Pène, A.B. Dufour, J.R. Lobry, One year variability of peak heights, heterozygous balance and inter-locus balance for the DNA positive control of AmpFℓSTR© Identifiler© STR kit., Forensic Sci. Int. Genet. 5 (2011) 43–9.
Ac ce p
te
d
M
an
us
cr
ip t
[16]
18 Page 21 of 22
ACKNOWLEDGEMENT
This project has been supported in part by the Italian National Anti-Doping Commission
Ac
ce pt
ed
M
an
us
cr
ip t
(Commissione per la Vigilanza sul Doping) of the Italian Ministry of Health.
Page 22 of 22