The misuse of somatic gene therapy for the purpose of enhancing athletic performance is perceived as a coming threat to the world of sports and categorized as ‘gene doping’. This article describes a direct detection approach for gene doping that gives a clear yes-or-no answer based on the presence or absence of transgenic DNA in peripheral blood samples. By exploiting a priming strategy to specifically amplify intronless DNA sequences, we developed PCR protocols allowing the detection of very small amounts of transgenic DNA in genomic DNA samples to screen for six prime candidate genes. Our detection strategy was verified in a mouse model, giving positive signals from minute amounts (20 μl) of blood samples for up to 56 days following intramuscular adeno-associated virus-mediated gene transfer, one of the most likely candidate vector systems to be misused for gene doping. To make our detection strategy amenable for routine testing, we implemented a robust sample preparation and processing protocol that allows cost-efficient analysis of small human blood volumes (200 μl) with high specificity and reproducibility. The practicability and reliability of our detection strategy was validated by a screening approach including 327 blood samples taken from professional and recreational athletes under field conditions.
The long and unbroken history of doping plainly shows us that some athletes are willing to try anything to get an edge, and therefore do not hesitate to adapt even poorly evaluated drugs and methods. Stunning results reported from experimental animal studies have raised concerns about the illicit use of gene transfer technologies, that is gene doping, for athletic enhancement purposes.1, 2, 3 Currently there is no clear evidence that gene doping has already found its way to the sports ground; however it is definitely only a question of time when first athletes will start a try. The World Anti-Doping Agency is taking the probability of such malpractice seriously, and, in 2003, the term ‘gene doping’ was enumerated in the official list of banned substances and methods.4 Since then, significant attention and resources have been dedicated by the World Anti-Doping Agency to research in the detection of gene doping.
As gene doping is based on the intuitive concept of administering genetic information, thus enabling the athlete's cells to produce its own doping substances—which in most cases would be indistinguishable from endogenous proteins—it is commonly believed that gene doping detection strategies should aim at metabolic, transcriptomic or proteomic changes as a consequence of the delivery and/or expression of the transgene.5 However, it remains an open issue if incontrovertible doping signatures can be established that are independent of any confounding variables such as age, gender, physical or ethnic background. We have to keep in mind that elite athletes are ‘exceptional people’ with ‘exceptional skills’ and ‘exceptional physical abilities’ that will also be mirrored in ‘exceptional signature patterns’, for example, because of distinct genetic polymorphisms, methylation patterns or histone modifications.
Recently, our group proposed a direct PCR-based detection strategy that would give a clear yes-or-no answer based on the presence or absence of transgenic DNA in peripheral blood samples.6 The procedure is based on the specific detection of intronless DNA by using a single-copy PCR protocol using a priming strategy that is specific for DNA sequences that could be abused for gene doping purposes (see Figure 1 for details). Our priming strategy is targeting the transgene itself, as this sequence part is the minimal prerequisite to achieve a doping effect. In contrast to this, other non-human sequence parts delivered by gene transfer, such as viral sequences or promoter sequences, may vary depending on the technique used. In this study, we provide experimental evidence that transgenic DNA can be detected for several weeks in the peripheral circulation of mice that received intramuscular adeno-associated virus (AAV)-mediated gene transfer, representing one of the most likely candidate vector systems to be used in gene doping practice. Transgenic DNA could be detected in minute amounts of blood samples, thus allowing minimally invasive sample retrieval, a fundamental prerequisite to legitimate doping testing. To make our detection strategy amenable for routine testing, we implemented a robust sample preparation and processing protocol that allows specific and reproducible analysis of human blood samples to screen for six prime gene doping candidate genes: erythropoietin (EPO), insulin-like growth factor 1 (IGF-1), vascular endothelial growth factors A and D (VEGF-A, -D), human growth hormone 1 (GH-1) and follistatin (FST).1, 2, 3
A specific, highly sensitive nested PCR setup to detect transgenic DNA in a high genomic background
Our priming strategy for the detection of transgenic VEGF-A is illustrated in Figure 1. VEGF-A exists in multiple isoforms of variable exon content and strikingly contrasting properties and expression patterns. To cover all relevant splice variants, splice site-spanning primer pairs were located in the canonical region comprising exons 1–5. To achieve maximum specificity and sensitivity, we developed a nested PCR assay with both rounds carried out across the splice site junctions of the transgenic cDNA sequence. By this approach, we are able to detect very small amounts of VEGF-A cDNA in a high background of 750 ng of genomic DNA without producing any detectable false priming artifacts (Figure 2). Reliability, sensitivity and specificity of the established purification and detection protocol were confirmed by spiking plasma samples with defined copy numbers of cDNA standards and by analyzing whole-blood samples spiked with transduced cells at known cell numbers (data not shown, see also Beiter et al.6). In this manner, we established protocols for the detection of six prime gene doping candidates, namely, EPO, IGF-1, VEGF-A, VEGF-D, GH-1 and FST (Figure 2).
To reduce the expense of reagents, as well as the preparation time and required amounts of sample materials, we attempted to develop a multiplex PCR for the simultaneous detection of multiple transgenes. For the selected candidates established in single nested PCRs, multiple combinations with various primer concentrations and cycling conditions were tested to establish a multiplex PCR that gives clean signals and accurate sensitivity. By combining a first-round multiplex PCR with target-specific separate single second-round nested PCRs, we managed to include four candidate genes (GH-1, VEGF-A, VEGF-D and EPO) in the first-round multiplex PCR. Reliability of this multiplex approach could be shown for PCR setups containing up to 1.25 μg genomic DNA without producing detectable false priming artifacts (Figure 3).
Detection of transgenic DNA in the blood of mice after intramuscular gene transfer
To validate our detection method, we determined the time course of detectable transgenic DNA in the blood of mice after in vivo intramuscular gene transfer, representing the most realistic ‘gene-doping scenario’. To deliver the transgene, we chose a ‘naked’ plasmid vector expressing human VEGF-A (hVEGF-A) under the control of the cytomegalovirus promoter (pCR3.1-hVEGF-A) as the easiest accessible but also the least effective modality, as well as a recombinant AAV vector pseudotyped with viral capsids from serotype 1 (rAAV2/1-hVEGF-A), representing the most suitable approach for in vivo gene transfer, achieving high specificity for skeletal muscle tissue. C57BL/6N mice received intramuscular injection of 10 μg of pCR3.1-hVEGF-A (n=3) or 3.3 × 1011 viral genomes of rAAV2/1-hVEGF-A (n=6), respectively, in the left hind limb quadriceps muscle. At indicated time points after injection, 20 μl of tail-tip blood samples was analyzed for the presence of transgenic DNA. VEGF-A transgenic DNA was detectable in all rAAV-treated animals till day 28, and even on day 56 after gene transfer, four out of six animals were tested positive (Figure 4). Transgenic DNA was not detectable in samples from plasmid-treated animals, with the exception of one animal at day 2, but plasmid gene transfer also did not succeed to induce expression of hVEGF-A, as shown by quantitative RT-PCR expression analysis of RNA extracted at day 56 from skeletal muscle tissue (Figure 5). Animals treated with rAAV demonstrated a pronounced expression of hVEGF-A in the injected muscle, indicating successful gene transfer. It is noteworthy that minor transgene expression could also be detected in the contralateral limb site. Expression of endogenous murine VEGF-A was not affected by any treatment.
Validation of sample processing and detection protocols
To make our detection strategy amenable for routine testing, we implemented a robust sample preparation and processing protocol that was validated by analyzing human control blood samples taken before and after intensive exercise from professional and recreational athletes who did not undergo any gene transfer treatment. For each time point, two samples (sample parts A and B) were taken and stored at −20 °C for further analysis. Batches of 14 blood samples (part A) were processed and amplified at the same time under the same conditions, using the same PCR master mix and in the same thermocycler. Three PCR-positive controls containing a control DNA sample spiked with cDNA standards (10 copies each), as well as genomic amplification controls, were included to verify that the PCR master mix was prepared correctly and that the DNA isolation procedure was performed properly. In total, 327 blood samples were analyzed, representing 1962 second-round PCRs. No false priming artifacts could be observed. However, in two sample batches we obtained faint false-positive signals (14 samples for follistatin, 1 sample for growth hormone-1), which had to be attributed to a cross-contamination from positive control samples in the second-round PCR. Repetition of PCR analyses from these sample batches was negative, as well as a second DNA preparation of the original part A blood samples. A third reevaluation by processing samples of part B confirmed the negative result.
In its 2010 Prohibited List, the World Anti-Doping Agency includes the transfer of genetic material with the potential to enhance athletic performance into the section ‘M3. Gene Doping’.7 Somatic gene therapy carries immense potential as a therapeutic modality for the treatment of inherited or acquired diseases by supplying functional copies in substitution of mutated genes, and by improving the body's natural ability to cope with diseases and infections. Although apparently simple in concept, the practical realities of translating gene therapy strategies into clinical practice have proven to be tremendously challenging. However, previous experiences have shown that even drugs that are still in the experimental phases of research may rapidly find their way into the athletic world.8 Several gene therapy trials aimed at the treatment of anemia, muscular dystrophy and peripheral vascular diseases have been initiated or are currently on the way. Moreover, impressive results from preclinical animal trials might entice some athletes to try these drugs, especially as no detection method is yet at hand. Dramatically improved muscle mass and muscle performance has been shown by rAAV-mediated intramuscular transfer of insulin-like growth factor 1 and follistatin in mice.9, 10 Growth hormone serum levels in rodents could be significantly increased by salivary gland and intraperitoneal administration of rAAV vector systems.11, 12 Intramuscular transfer of rAAV-VEGF has been shown to improve muscle survival and regeneration following muscle damage in mice.13 Using a tetracycline-inducible rAAV-based vector system, sustained functional neovascularization could be achieved after intramuscular gene transfer of VEGF in a murine ischemia model.14 By exploiting a similar Tet-On system, precise regulation of blood hemoglobin levels could be demonstrated after intramuscular electro-mediated gene transfer of erythropoietin in mice.15
Among the various application strategies and vector constructs that have been established in gene therapy trials, attempts to transfer gene therapy into doping practice would most likely primarily focus on intramuscular gene transfer using AAV-derived vectors as gene delivery systems. As current non-viral vectors have proven to be highly inefficient in humans for in vivo applications, recombinant AAV would be most suitable to deliver the transgene of choice in a ‘gene doping scenario’. This is because of its low immunogenicity, stable support of gene expression in slowly dividing or postmitotic cells, and the possibility of improving tissue specificity by pseudotyping the transgene cassette into alternative AAV vector capsids.16 Production of these vectors is by now a routine task in many molecular biology labs, and custom-made viral vectors are offered by commercial suppliers as ready-to-use stocks. Considering application strategies, in vivo intramuscular gene transfer would be the most practicable and promising approach, as skeletal muscle is easily accessible for vector administration, is sufficiently vascularized and enables prolonged expression of the transgene because of its postmitotic state.17
Several detection strategies for gene doping are currently under investigation.2, 17 As expression of the transgene cassette is uncoupled from intrinsic regulatory pathways, most detection strategies aim at deciphering biomarkers and expression profiles (transcriptomic, proteomic and metabolomic approaches) that are indicative of ‘unnatural’ gene expression. It remains an open issue whether such indirect methods can be sufficiently validated to exclude all natural causes for aberrant signature patterns. Because in sports jurisdiction a positive doping test alone establishes comfortable satisfaction of guilt, athletes charged with doping offenses are functionally presumed guilty until proven innocent. But how can an athlete provide evidence for an exceptional genetic or epigenetic predisposition when current knowledge of the complex orchestration of genomic, epigenomic, transcriptomic and proteomic regulatory networks is spotty at best? Undoubtedly, a direct detection method that unambiguously identifies the doping agent should always be preferable. In this study, we provide evidence that gene doping can be detected by a direct approach based on the presence of traces of transgenic DNA in blood circulation after somatic gene transfer. Biodistribution studies show that entry of a significant portion of the vector into the blood stream is unavoidable even after direct injection of the vector into the target tissue.18, 19, 20, 21 Using whole-blood samples for DNA purification, our detection strategy comprises all extracellular and intracellular sources of transgenic DNA, including free circulating vector particles, transduced blood cells and free circulating DNA derived from transduced apoptotic/necrotic target tissue. Owing to the high sensitivity of our PCR setup, we were able to detect traces of transgenic DNA in very small blood sample volumes (20 μl) in a mouse model after intramuscular rAAV gene transfer for up to 56 days. Transduction efficiency in mice is usually much higher than the one that can be currently achieved in humans because of smaller tissue mass and lack of preexisting immunity to AAV capsid antigens. Therefore, disproportionally higher viral loads have to be administered in humans to achieve sufficient transgene expression in the target tissue. Our detection method allows 10-fold higher amounts of sample volumes in humans than has been analyzed in the mouse model. Thus, we are confident that our approach will also be applicable to convict ‘gene dopers’.
So far, we developed target-specific PCRs for the detection of the most important gene doping candidates VEGF-A and -D, erythropoietin, GH-1, insulin-like growth factor 1 and follistatin. Sample preparation and PCR setup procedures were established to ensure straightforward, cost-effective, minimally invasive clinical testing with high specificity and sensitivity. Of course, the benefit of our PCR-based detection method, that is, its great sensitivity, is also a drawback, as even the smallest amount of contaminating cDNA can lead to false-positive results. Therefore, the accuracy of our testing system depends on awareness of risk factors and the proper use of procedures for the prevention of nucleic acid contamination. The test laboratory should be designed and operated in a way that prevents contamination of reactions with amplified products from previous assays and cross-contamination between samples, both of which can lead to false-positive results. Unidirectional workflow, separation of workspace and appropriate control procedures are fundamentally crucial. As a matter of course, these prerequisites cannot be completely fulfilled in a small molecular biology lab that is routinely dealing with a bulk of cDNA preparations. Substitution of positive control cDNAs by recombinant controls that are distinguishable from transgenic DNA will further reduce the risk of cross-contamination, thus making our procedure amenable to high-throughput screening in accredited laboratories.
In our opinion, it is crucial to develop efficient gene doping detection methods not only to expose unscrupulous cheaters but also to establish a daunting scenario to prevent gene doping even before it becomes reality, thereby keeping ahead of a new doping wave that might change sports as we know it today forever.
Materials and methods
Selection of primers within the coding sequence of a candidate gene was performed with respect to the gene-specific exon–intron structure. Primers were located within canonical and conserved exonic sequence regions to cover all mRNA variants of a candidate gene that are of relevance for physical performance (Table 1). If possible, primers were designed to selectively bind at the exon junctions of cDNA sequences (Figure 1). If not possible, primers were placed in exonic sequences located upstream of long genomic introns. Regions with known single-nucleotide polymorphisms were excluded using the database information given in dbSNP at the NCBI (http://www.ncbi.nlm.nih.gov). All primers were purchased from Eurofins MWG Operon (Ebersberg, Germany).
Nested PCR setup
First-round (outer) and second-round (inner) PCR runs were performed with Promega GoTaq Green Master Mix (Promega GmbH, Mannheim, Germany). First-round PCR cycling conditions included an initial denaturation step at 95 °C for 3 min 30 s, a touchdown protocol at 94 °C for 20 s, 62 °C for 25 s (decreasing 0.3 °C for 10 cycles), 72 °C for 35 s, followed by 30 cycles of 94 °C for 20 s, 59 °C for 25 s, 72 °C for 35 s and a final elongation step of 72 °C for 7 min. The first-round PCR product was diluted at a ratio of 1:50 and 3 μl was submitted to second-round (inner) PCR. Second-round PCR was performed in a total volume of 50 μl as follows: initial denaturation at 94 °C for 3 min 30 s, followed by 33 cycles of 94 °C for 20 s, 59 °C for 25 s and 72 °C for 30 s, and a final extension for 7 min at 72 °C. To combine the advantages of a multiplex approach (simultaneous detection of multiple transgenes from the same starting material) with the particular sensitivity of the nested PCR protocol, we combined a first-round multiplex PCR containing all outer primer pairs of selected candidate genes with separate second-round nested PCRs specific for each target template. Final PCR products were analyzed on a 1.6% (W/V) agarose gel and visualized by UV illumination after staining with GelRed (Biotium, Hayward, CA, USA).
Preparations of transgenic DNA standards
The effectiveness of the PCR protocols was tested on different preparations of 750 ng of total genomic DNA spiked with known copy numbers of respective cDNAs as positive controls. Unspiked DNA samples represented negative controls. All 1 kb cDNA standards were constructed by target-specific PCR from cDNA libraries. Concentrations of the respective standard cDNA stocks were determined photometrically (NanoDrop 1000, Thermo Fisher Scientific, Waltham, MA, USA) and by photodensitometry from serial dilutions run on 1.2% agarose gels using Quantity One 1-D Analysis Software (Bio-Rad, Muenchen, Germany). Copy numbers were calculated using Oligo Calculator (http://www.basic.northwestern.edu/biotools/oligocalc.html) and standards with defined copy numbers were prepared by serial dilutions.
DNA extraction from human blood samples
Prerequisite for DNA extraction was to establish a suitable sample processing method that should yield reproducible amounts of high-quality DNA from minimized sample volumes with minimal risk of sample cross-contamination. For human blood samples, we used the QIAamp DNA Blood Mini Kit (Qiagen GmbH, Hilden, Germany), followed by a purification/concentration step, using Microcon YM-30 centrifugal filter devices (Millipore Corporation, Billerica, MA, USA). Briefly, DNA was extracted from 200 μl of EDTA whole-blood samples according to the ‘blood and body fluid protocol’ provided by the manufacturer, with a final elution volume of 100 μl. DNA eluate was transferred to the sample reservoir of the Microcon unit and centrifuged for 12 min at 14 000 g. Before inverting the filter, 20 μl of nuclease-free water (Promega Corporation, Madison, WI, USA) was added and incubated for 3 min. After inversion of the filter, DNA was recovered by centrifugation at 1000 g for 4 min. This exceptional short and easy to handle protocol produced high-quality DNA preparations from 200 μl of EDTA whole-blood samples with an average yield of 5 μg at concentrations of 250 ng μl−1.
A 5.6 kb pCR3.1-hVEGF-A plasmid was used as a source of human VEGF-A cDNA. The EcoRI restriction fragment was cloned into the commercial pAAV-MCS vector using the Stratagene rAAV cloning system (Stratagene, La Jolla, CA, USA) to obtain the AAV cis-plasmid (pAAV-hVEGF-A) containing humanVEGF165 under the control of a cytomegalovirus promoter. Using this pAAV-hVEGF-A construct, pseudotyped recombinant rAAV2/1-hVEGF-A virus encoding the human transgene VEGF-A was generated by Vector Biolabs (Philadelphia, PA, USA) and purchased as a ready-to-use virus stock. This hybrid rAAV vector (rAAV2/1) was based on recombinant AAV2 with cap gene derived from AAV1, which has been shown to have superior transduction efficiencies for skeletal muscle compared with AAV serotype 2.22
Twelve-week-old male C57BL/6N mice were obtained from Charles River (Sulzfeld, Germany). Mice received intramuscular injections using a volume of 30 μl per injection. The vector suspension containing 3.3 × 1011 viral genomes of viral vector rAAV2/1-hVEGF-A or 10 μg of naked plasmid vector pCR3.1-hVEGF-A was injected into the left hind limb quadriceps muscle. In total, 12 animals were distributed among three groups, of which three animals received NaCl as control injection, three received pCR3.1-hVEGF-A and six received rAAV2/1-hVEGF-A. In total, 20 μl of blood samples was collected from the tail vein with an EDTA-coated capillary (Sarstedt AG, Nümbrecht, Germany) on days −1, 2, 7, 14, 28 and 56. Animal experiments were approved under the German federal regulation rules and carried out in accordance with institutional guidelines. Total DNA from 20 μl murine blood samples was extracted with a QIAamp DNA Micro Kit (Qiagen) using a modified protocol adapted for mice blood samples, with the following changes to the manufacturer's protocol. Incubation time of cell lysis was extended to 30 min, and the initial column centrifugation step was increased to 10 000 g. After the two standard washing steps, an additional washing step with 500 μl ethanol (85%) was included (6000 g, 1 min). DNA was recovered with a final elution volume of 100 μl, and the eluate was further concentrated/purified using Microcon YM-30 filter devices as described above. On day 56, all animals were killed and tissues from the injected and the contralateral uninjected muscle were excised and stored in RNAlater (Applied Biosystems, Foster City, CA, USA) at −80 °C for later quantitative analysis.
Hind limb skeletal muscle samples were mechanically homogenized using Qiagen TissueLyser, and total RNA was extracted with the RNeasy Fibrous Tissue Mini Kit (Qiagen) according to the manufacturer's instructions. Total RNA (500 ng) was reverse transcribed using SuperScript III reverse transcriptase (Invitrogen GmbH, Karlsruhe, Germany) with random hexamer primers according to the manufacturer's instructions. To distinguish murine (GenBank accession NM_009505) and human (GenBank accession NM_001025368) VEGF-A transcripts, primer pairs were located in species-specific regions in exons 2 and 3. Species specificity of murine (5′-CACGACAGAAGGAGAGCAGAAG-3′; 5′-GTTACAGCAGCCTGCACAGC-3′) and human (5′-GGCAGAATCATCACGAAGTGG-3′; 5′-GCATGGTGATGTTGGACTCCTC-3′) VEGF-A primers was confirmed by PCR on cDNA preparations of the respective species and vice versa. Real-time PCR and quantitative expression analysis were carried out as described previously.23, 24 Data are given as mean normalized expression values that reflect the relative expression of the target gene to the reference gene glyceraldehyde-3-phosphate dehydrogenase (GenBank accession NM_002046).
To validate our procedures for routine testing, 327 blood samples taken before and after intensive exercise were obtained from professional and recreational athletes. The sports activity of the athletes ranged from individual sports such as track and fields and gymnastics to basketball and soccer (professional league). Samples were taken before and after fitness training, resistance training, typical sports-specific training (soccer) and at a cross-country running competition. Each subject gave a written informed consent before participation in the study. The experimental protocols were approved by the Institute's Human Ethics Committee according to the principles set forth in the Declaration of Helsinki of the World Medical Association.
This project has been carried out with the support of WADA (research grants 06B7PS and 08C17PS). We thank Irina Smirnow and Andrea Schenk (Department of Gastroenterology & Hepatology, Medical Clinic, Eberhard-Karls-University of Tuebingen) for excellent technical assistance.
About this article
The University of Tübingen, Germany has a patent pending for the ‘Detection of transgenic DNA’(PCT/EP2007/003385; http://www.wipo.int/pctdb/en/wo.jsp?WO=2007124861) that relates to the detection of transgenic DNA in a living being and to a kit for performing such a method. A free use without charge of the patent pending procedure has been granted to the World Anti-Doping Agency for the purpose of doping analysis in sports.