Short tandem repeat methodology for genotypic identification of single-person versus multi-person use of syringes

Objective: To develop laboratory methods to differentiate between single- versus multi-person use of syringes by injection drug users.

Cited Here...: Forensic short tandem repeat (STR) genetic analysis was undertaken to cross-validate a test panel of trace blood contents from syringes representing single- versus multi-person syringe use. Laboratory-simulated scenarios of needle sharing generated 34 syringe washes that were blinded for evaluation. Polymerase chain reaction was used to amplify the polymorphic STR locus D6S502 from blood trace contents in used syringes. Alleles were sized and quantified using a commercial gene sequencer. A statistical algorithm was developed to determine the number of alleles present in the amplified DNA fragments. Syringes with more than two expected alleles were considered to represent multi-person syringe use. Sensitivity, specificity and the kappa coefficient were calculated.

Cited Here...: Allelic matrix-based analysis of alleles from the single STR successfully characterized single-use (n = 12) and multiple-use (n = 22) syringes with 68% sensitivity and 100% specificity upon re-analysis. The extent of agreement over and above chance (kappa = 0.6;P < 0.0001) indicated good agreement for differentiating single- versus multi-person syringe use.

Conclusions: These findings suggest that improved genotypic STR analysis of syringe material could be an adjunct to methods for validating self-reported needle sharing, conducting behavioral surveillance of needle-sharing behaviors, and evaluating interventions such as needle-exchange programs. Assays based on multiple STR loci will undoubtedly improve upon the promising results obtained from laboratory simulations of needle sharing.

Introduction

Needle sharing among injection drug users (IDUs) is a major source of transmission of blood-borne pathogens such as HIV, hepatitis B virus (HBV) and hepatitis C virus (HCV) ^[1]. Injection drug use accounts for approximately one-third of all reported AIDS cases, and one-half of all new HIV infections in the United States, either directly through sharing of injection equipment, or indirectly through heterosexual and perinatal transmission ^[2,3]. Strategies to reduce multi-person syringe use include education and increasing access to sterile injection equipment for IDUs who cannot or will not cease injecting ^[1,4-6].

Most evaluations of behavioral interventions among IDUs have relied on self-reports of needle sharing behaviors. Several studies have shown that needle-exchange attendance is associated with declines in self-reported needle borrowing and lending among HIV-uninfected and HIV-infected IDUs ^[5-8]. However, concerns have been raised regarding the validity and reliability of self-reported syringe sharing ^[9-11]. IDUs may under-report risky behaviors such as needle borrowing and lending. IDUs participating in research studies may be subject to socially desirable responding due to ethically mandated risk-reduction counseling and familiarity with detailed questionnaires pertaining to HIV risks. In addition, IDUs often fail to recognize risks associated with injection practices such as 'back-loading' or 'front-loading', which involve syringe-mediated sharing of injection equipment ^[1,12,13]. The occurrence of 'immaculate infections' (i.e., HIV seroconversion with no reported risk factors) ^[5,14] indicates the need for standardized laboratory methods to cross-validate self-reported rates of multi-person syringe use.

Obtaining a valid biological marker of multi-person syringe use that does not rely on self-reports would represent a considerable scientific advance. Such data would be extremely helpful for validating self-reported needle-sharing behaviors, evaluating HIV interventions and conducting individual and population level behavioral surveillance among IDUs. Until recently, no methods were available to systematically validate multi-person use of syringes. Herein, we report upon a novel method using a genetic examination of the number of distinct genomes present in syringe exudates to validate self-reports of syringe sharing. Our methods were based on genotypic analysis of one tetra-nucleotide (TCTA)_n short tandem repeat (STR) locus, D6S502, which has been previously used in forensic studies ^[15,16]. D6S502 has eight alleles and a heterozygosity of 0.78 in Caucasians ^[16]. Using polymerase chain reaction (PCR) to amplify this STR from DNA in blood traces of used syringes, the capability to differentiate between syringes that have been used by single versus multiple persons was demonstrated. To our knowledge, this is the first report of an application that is capable of obtaining a valid biological measure of needle sharing.

Materials and methods

Laboratory simulations of needle-sharing scenarios

To develop methods for differentiating between single- and multi-person use of syringes, laboratory conditions were created to simulate needle-sharing scenarios. Blood samples from three different individuals were used as references. Mixtures of these references were prepared by simulating the multiple-use syringes by different individuals as follows. Registering (i.e., the process whereby IDUs pull back a small amount of blood into the syringe to confirm venous access), was simulated by drawing blood into the 1 cm³ syringe until the barrel was three-quarters full and then expelling back into the blood container. To simulate 'booting' (i.e., the process by which IDUs pull back a small amount of blood into the syringe a second time to ensure that the remainder of drug is expelled), 1 cm³ of blood was drawn into the syringe and left in it for a few seconds before expulsion. The syringe was then stored at room temperature for 2 days and either washed as outlined below or processed further to simulate sharing. The syringe was then exposed to additional rounds of the procedure outlined above with a different reference blood sample. To preserve blinding of the study samples, two separate laboratories took part in the study. The first laboratory prepared the syringe washes and the second conducted DNA extraction, PCR and analysis.

Syringe washing

A 50 μl volume of wash buffer solution consisting of 10 mM Tris HCl pH 8.3, 50 mM KCl, 2.5 mM MgCl2, Tween 20 and NP40 was pipetted into a microfuge tube. The tip of the syringe needle was placed into the wash solution at the bottom of the microfuge tube, and then the plunger was drawn back to fill the syringe with the wash-buffer and then slowly expelled for a total of three times. These needle washes were stored at -70°C before extracting the DNA.

DNA extraction

The syringe washes were treated with RNAse and Proteinase K (Qiagen, Hilden, Germany) and the DNA was carefully extracted by column chromatography using the QiaAmp Blood Kit as described in the user manual. The DNA was eluted in 30 μl water and one-third of the sample was used for PCR amplification.

STR amplification

STRs are tandemly repeated simple sequences of 2-7 bases in length that vary in the number of repeat units among alleles from different individuals. They are abundant and well distributed across the human genome, characterized by Mendelian inheritance and provide a rich source of polymorphic markers for individual identifications ^[17-19]. Analyses of minute biological materials left at crime scenes with STRs is a standard procedure at crime laboratories that can successfully distinguish between individuals with a probability of 1 × 10^-8[20,21].

The PCR was carried out in a volume of 50 μl, containing 10 μl of the extracted DNA samples from the syringes and 40 μl of the PCR mix which consisted 0.5 mM of each D6S502 primers, 250 μM of each dNTP, 10 mM Tris-HCl (pH8.3), 50 mM KCl, 2 mM MgCl2 and 3.5 units of AmpliTaq Gold (Perkin Elmer Biosystems, Foster City, California, USA). Oligonucleotide primers for D6S502 were obtained from Genosys (Woodlands, Texas, USA) with a fluorescent dye (tet) attached to the 5′ end of the reverse primer. The primer sequences were forward: TTTTTGTATT TCATGTGTACATTCG and reverse: (tet)-CGTAGC TCTAATTAGTTCATTTTC. The PCR amplifications were performed in GeneAmp PCR system 9600 (Perkin Elmer Biosystems) with a program consisting of 9 min denaturing at 94°C, followed by 30 cycles of 50 s denaturing at 94°C, 60 s annealing at 54°C and 60 s extension at 72°C, with a final extension of 10 min at 72°C.

Genotypic analysis

Genotypic analysis was carried out using commercial hardware and software (P E Biosystems). Electrophoresis was performed in an ABI PRISM 377 DNA Sequencer (PE Biosystems) using 0.2 mm × 20 cm × 34 cm denaturing gel containing 5% Long-Ranger matrix acrylamide (FMC, Rockland, Maine, USA), 10% urea and 1 × Tris-borate-EDTA. The PCR product (3 μl) was mixed with loading buffer consisting of 1 μl internal lane standard (Genescan-Tamra 350; Perkin Elmer Biosystems) and 2 μl of formamide and denatured at 94°C for 5 min. The samples were then immediately cooled to 4°C. Samples (∼2 μl) were loaded on every other lane of the gel to minimize and avoid cross-contamination. The analyses of the gel and the sizing of DNA peaks on each lane were performed using Genescan 3.0 software (Perkin Elmer Biosystems) as described in the user manual. Electropherogram analysis with Genotyper 2.1 was used to call and edit peaks corresponding to sized PCR products of D6S502 alleles whose labels represent their observed size in base pairs. Interpretation of electropherograms was conducted by two of the authors (S.S. and M.W.S.), who were blinded to the needle-sharing status of the syringe-wash specimens.

Statistical analysis

During PCR, products are sometimes generated by slippage of Taq DNA polymerase resulting in minor products that were either one-repeat unit smaller, or less commonly one repeat larger than the true alleles ^[22]. To minimize the possibility of misclassification, the following analytic strategy was designed to distinguish slippage products from true additional alleles in the samples of mixed origin. A statistical algorithm was developed based on the ratio of the height of the slippage bands relative to the height of adjacent alleles from the three reference samples. The mean and the standard deviation of the ratio of these slippage products relative to true alleles were calculated for both directions (i.e. upstream and downstream of the true alleles) from the known reference samples.

Electropherograms from syringe-wash samples were examined by first locating the alleles from a single individual. The two highest peaks were assigned as the true alleles from the single user, with the second peak being of a height no less than 33% of the largest to account for possible homozygotes. Peaks of less than 3% of the largest peak height or 25 fluorescence units were considered to represent background noise. The remaining peaks were considered subsequently by decreasing the order of their height to minimize the possibility that slippage during PCR amplification could contribute to misclassification. Slippage loss and gain ratios were compared to their respective means plus five standard deviations for each potential allele. PCR products of less than these cutoffs were assumed to originate from slippage, and those greater to represent an additional D6S502 allele in the sample under analysis. The set of mixed sample alleles was also matched to the alleles in the individual reference samples used to generate the syringes in a determination of the origins of these mixtures. Any sample with more than two alleles was assumed to represent multi-person syringe use.

The validity of these laboratory methods in distinguishing single- versus multi-person syringe use was determined by calculating sensitivity and specificity, treating the known contents of each syringe specimen as the gold standard. Sensitivity measured the ability to identify correctly the multi-use syringe specimens, whereas specificity measured the ability to correctly identify the single-use syringe specimens. The kappa statistic was calculated to measure the extent of agreement beyond chance ^[23].

Results

Typical examples of STR genotypic analysis of syringe exudates that illustrate single- versus multi-person use from the test panel are depicted in Figure 1. Electropherograms are labeled as reference samples along with single use syringes and multiple use syringes whose distinctly sized products were identified as alleles, marked with asterisks, according to the statistical criteria as described above. These electropherograms show that reference 'A' has distinct A175 and A183 alleles. Reference 'B' has A179 and A195 and reference 'C' has A171 and A187 alleles.

Fig. 1

Before unblinding the results of the test panel that contained known numbers of needle sharers in a given syringe, the following conclusions were made for these examples. From the genotypes observed in the single usage syringes, it was concluded that single-use syringe 'D' was the same as reference 'C', single-use syringe 'E' = reference 'A' and single-use syringe 'F = reference 'B'. Mixed syringe washes were also determined in the analysis. Multiple-usage syringe 'G' was found to have the alleles A175, A179, A183 and A195, indicating a mixture of the samples of reference 'A' (A175 and A183) and reference 'B' (A179 and A195). Multiple-use syringe 'H' had the alleles A171, A179, A187 and A195 from reference 'B' and 'C' whereas multiple-use syringe 'I' had the alleles 171, 175, 183 and 187 from reference 'A' and reference 'C'. After the experiment was unblinded these observations were found to be consistent with the known origins of the samples.

In our study, 39 simulated syringe washes were tested in blinded genotypic diagnosis. Using the initial allelic matrix-based analysis procedure described above, we characterized single-use and multi-use syringes as summarized in Table 1. Two-by-two contingency table analysis for the gold standard (unblinded samples) versus the classification of single-use versus multi-person use of syringes (blinded samples), yielded a sensitivity and specificity of 74% (20 of 27) and 75% (9 of 12), respectively. The kappa coefficient, which is an indication of the extent of agreement over and above chance, was 0.45, reflecting 'good' agreement (P < 0.0001) ^[24]

Table 1

In an attempt to maximize specificity, we re-examined these data by re-adjusting the background noise cutoff to 50 fluorescent units, and removing one sample with low overall signal and four with poorly-resolved fragments. Among 34 of the above samples, the repeated sensitivity and specificity calculations were 68% (15 of 22) and 100% (12 of 12). The kappa statistic was 0.6 (P < 0.0001).

Discussion

The present study showed that it is possible to differentiate single- versus multi-person use of syringes using STR genotypic analysis. To our knowledge, this is the first report of an application that can be used to validate self-reported needle-sharing behaviors among IDUs. In our simulated experiment using known syringe contents in needle-sharing scenarios, data from one STR, D6S502, was sufficient to yield a sensitivity and specificity of 68 and 100%, respectively, after refining the cutoff representing background noise. Our study lends confidence to the notion that injection practices such as 'registering' and 'booting' needle sharing and other syringe-mediated practices such as 'backloading' or 'frontloading' leave sufficient quantities of trace DNA to be efficiently amplified by PCR, as previously reported by Heimer ^[25]. Consistent with our own experience, previous forensic case studies have indicated that STR genotypic analysis can be conducted using as little as 1 ng of trace DNA ^[26,27].

In previous studies, HIV-1 antibodies as well as HIV RNA and integrated HIV DNA have been detected using enzyme-linked immunosorbent assay and PCR to examine trace amounts of biological material in syringe exudate from injection drug users ^[25,28-31]. For example, in New Haven, Connecticut, detection of HIV DNA through PCR, coupled with syringe-tracking methods using bar-coded needles was used to evaluate a needle-exchange program [28]. Using these techniques, it was estimated that needle-exchange program attendance was associated with a 33% reduction in HIV incidence. The application of STR genotypic analysis to identify needle sharing provides an exciting tool for extending this earlier research. For the first time, a biological measure of syringe sharing can be obtained as a gold standard, without the need to rely on self-reports of questionable validity and reliability.

A limitation of the STR genotyping technique is the phenomenon that occurs due to Taq DNA polymerase 'slipping' on the tetranucleotide repeat during PCR amplification creating products of predominantly minus one repeat unit ^[22,32]. In forensic studies, slippage products are observed adjacent to allelic peaks ^[33-35]. We used a tetranucleotide repeat locus, D6S502 with minimal slippage and applied a carefully designed algorithm to distinguish slippage bands from true alleles in our samples. Indeed, the presence of these slippage bands and the need to ignore them contributed to the inability to identify some mixed samples. Some misclassification of multiple-use syringes as single-use was due to the presence of small amounts of the second genome (< 5 to 10%) which could not be confidently detected due to the background of slippage. Development of more appropriate STR markers and more stringent criteria for identifying the presence or absence of alleles above background should significantly improve sensitivity without sacrificing specificity in differentiating between trace DNA from unique individuals. Fortunately, the forensic community has focused on identification of STR markers with minimal slippage and a high level of informativeness ^[27,33-35]. We have begun to investigate the usefulness of forensic applications using pentanucleotide repeats at CD4, NOS2A and GSTP1; our unpublished data support the published reports that slippage can be reduced to 1-2%^[35]. Hopefully a set of pentanucleotide markers can be developed with very low slippage, small allele sizes (to minimize degradation), and high informativeness for differentiating between individuals.

Using one STR marker, D6S502, we conservatively limited our interpretation of needle-sharing scenarios to single- versus multi-person use of syringes, without consideration of the number of persons who had shared a particular syringe. Using multiple markers, forensic analysis of mixed biologicals can identify a perpetrator's DNA from crime scenes even when the genetic material from the victim predominates ^[36,37]. With the detection of multiple loci, it may be possible to estimate the minimum number of people who have used a given syringe, their sex and other characteristics in future analyses; however, this was beyond the scope of the present study. It is very likely that examination of multiple loci will make it easier to determine multi-person syringe use, since favorable combinations of allelic sizes for detection of more than one genome are more likely to be observed.

We demonstrated the application of a forensic and genetic method to the novel problem of differentiating between single- versus multi-person use of syringes among IDUs. Further refinement of these methods, based on panels of known syringe contents and multiple loci, is expected to increase the sensitivity beyond 90% as well as achieving 100% specificity. When applied to epidemiological studies of IDU populations, these forensic genetic methods will be invaluable for validating self-reported needle sharing, conducting behavioral surveillance, evaluating interventions such as needle-exchange programs, and the identification of individuals who are resistant to viral infection.

Acknowledgements

We gratefully acknowledge the technical expertise of Victor David, Melissa Levasseur, Beth Marosy, Melissa Marx, Joe Bareta and Marianne Subleski along with helpful discussions with Dr Hyoung Doo Shin, Dr Marilyn Raymond and Dr Stephen J. O'Brien.

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

This project has been funded in whole or in part with Federal funds from the National Cancer Institute under Contract No. NO1-CO-56000 and the National Institute on Drug Abuse under Grant #DA12568 of the National Institutes of Health.