Forensic Science International 206 (2011) 111118

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locate/forsciint

Relationship of spermatoscopy, prostatic acid phosphatase activity and prostate-specic antigen (p30) assays with further DNA typing in forensic samples from rape cases
Lydia Romero-Montoya a, Hugo Martnez-Rodrguez b, Miguel Antonio Perez b, Raul Arguello-Garca c,*
a b c

xico, 50090 Toluca, Mexico Laboratorio de Qumica Forense, Instituto de Servicios Periciales, Procuradura General de Justicia del Estado de Me Laboratorio de Qumica Forense, Subprocuradura General de Justicia de Texcoco, 56101 Texcoco, Mexico tica y Biologa Molecular, Centro de Investigacio y de Estudios Avanzados-IPN, 07360 Mexico City, Mexico n Departamento de Gene

A R T I C L E I N F O

A B S T R A C T

Article history: Received 23 February 2009 Received in revised form 28 June 2010 Accepted 7 July 2010 Available online 9 August 2010 Keywords: Rape analysis Vaginal swabs Acid phosphatase Sperm cytology Prostate specic antigen p30 Y-STRs Male DNA prole

In the forensic laboratory the biological analyses for rape investigation commonly include vaginal swabs as sample material combined to biochemical tests including sperm cytology (SC) and detection of acid phosphatase activity (AP) and prostate-specic antigen (PSA, p30) for the conclusive identication of semen components. Most reports comparing these tests relied on analysis of semen samples or donor swabs taken under controlled conditions; however their individual or combined efcacy under real live sampling conditions in different laboratories is largely unknown. We carried out SC, APA and PSA analyses in vaginal swabs collected from casework rapes submitted to Mexican Forensic Laboratories at Texcoco and Toluca. On the basis of positive and negative results from each assay and sample, data were classied into eight categories (IVIII) and compared with those obtained in the two only similar studies reported in Toronto, Canada and Hong Kong, China. SC and APA assays had the higher overall positivity in Toluca and Texcoco samples respectively and otherwise PSA had a lower but very similar positivity between these two laboratories. When compared to the previous studies some similarities were found, namely similar frequencies (at a ratio of approximately 1 out of 3) of samples being positive or negative by all techniques (Categories I and VI respectively) and a comparable overall positivity of APA and SC but higher than that of PSA. Indeed the combined results of using SC, APA and PSA tests was considered as conclusive for semen detection from approximately 1 out of 3 cases (Category I) to approximately 1 out of 2 cases in a scenario where at least SC is positive, strongly presumptive in 2 out of 3 cases (with at least one test positive) and the remainder 1 out of 3 cases (Category VI) suggested absence of semen. By determining Y-STR polymorphisms (12-loci) in additional samples obtained at Toluca laboratory, complete DNA proles were determined from all Category I samples, none marker was detected from all Category VI samples and mostly partial proles were obtained from samples of other categories. These observations give an overview on the variability in efcacy of each test performed at different laboratories and provide a general notion about the in praxis contribution of SC, APA and PSA tests for further DNA typing in the forensic analysis of rape. 2010 Elsevier Ireland Ltd. All rights reserved.

1. Introduction In the forensic investigation of rape, the conclusive identication of semen is required to corroborate the alleged sexual assault as it is a usually unwitnessed crime. The medical examination of the complainant and the laboratory analyses of biological samples pursuit both detecting assailant semen and supplying legal resources for judicial consignation. Provided that women are mostly by far the victims [1,2] and vaginal swabs are the more

* Corresponding author. Tel.: +52 55 5747 3800x5326; fax: +52 55 5747 3392. E-mail address: raularguellogarcia@yahoo.com (R. Arguello-Garca). 0379-0738/$ see front matter 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2010.07.012

reliable sample device [3], a number of sperm or seminal markers have been evaluated in these samples for forensic application. These include biochemical methods based either on detection of metabolites (free choline, Zinc, spermine, prostaglandin E) or enzyme activity (acid phosphatase [AP], gamma-glutamyl transpeptidase) as well as specialized immunological methods for specic proteins as are spermatozoid wall-specic antigens, seminal vesicle-specic antigens (semenogelins) and prostatespecic antigen (PSA, also called p30) [46]. Methodologies of more recent development based on detection of DNA (autosomal and Y-chromosome short tandem repeats [AS- and Y-STRs]) or RNA (semi-quantitative RT-PCR, Real Time PCR and microarrays) have been described for forensic purposes [7,8] and offer not only the

112

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118 2. Materials and methods 2.1. Specimens, fabric swabs and validation tests for PSA and APA The semen samples of reference were obtained from the following donors: one normospermic (sperm counts: 65.5 106 cells/mL), one oligospermic (23.5 106 cells/mL), one azoospermic and one vasectomized donor. To determine the detection limit for PSA assay it was used one semen sample from a normospermic individual with known PSA concentration as determined with the commercial kit VITEK ImmunoDiagnostic Assay System for total PSA [VIDASTPSATM] following manufacturers instructions. Samples of other bodily uids and materials for validation of biochemical tests included: human urine (from males aged 5, 9, 10, 13, 14, 15, 31 and 80 years old and two additional samples from adults), sweat (from seven males), breast milk (from four women), blood (from two adults), fecal materials (from two men and one woman), vaginal secretions (from one woman that was sampled three times within a 2-day period) and saliva (from two men and two women). Fabric swabs were prepared to validate sensitivity, specicity and interference by bodily uids/materials in PSA and APA assays. In these, 200 mL of either diluted semen (sensitivity assays) or other bodily uid/material (specicity assays), or a mixture of 200 mL of concentrate uid/material plus 200 mL of a 1:10 dilution of semen (interference assays) were directly applied to intact swabs. These latter were processed in the same manner as described below for casework swabs. 2.2. Vaginal swabs Two groups of samples were initially analyzed: the rst composed by onehundred and the second by forty-eight vaginal swabs collected by conventional technique [3] from the same number of alleged rape cases submitted to Laboratories of Forensic Chemistry of the Instituto de Servicios Periciales with venue at Toluca or Texcoco, in the State of Mexico, Mexico during FebruaryJune 2006 (Texcoco) and April 2006March 2007 (Toluca). For DNA typing studies, 27 additional casework swabs were collected during a further 3-months period. Most swabs were analyzed on the same day of collection or after 48 h of storage. In few cases (n = 8) two swabs were collected simultaneously from the same case and tested separately, i.e. on the same day or after 5-day storage; this delaying in analysis did not affect results as assessed by the similar results (positive or negative) obtained in SC, APA and PSA assays carried out in the two swabs of all these cases. The same swab from each case was the initial material for detection of sperm, APA and PSA after a sequential processing. Briey swabs were humidied with 750 mL SS (0.85% NaCl) within 15-mL centrifuge tubes and centrifuged at 650 g for 8 min. For PSA detection, 200 mL of supernatant were added to immunochromatographic membranes; for spermatoscopy the pellet was resuspended in a minimal volume of spent SS (approximately 30 mL) and 10 mL were deposited and dried on glass slides; in some cases the remaining volume of pellet suspension was used for Y-STR typing as described afterwards. For AP determinations the remaining pellet suspension was applied to the whole swab then this was used as described below. 2.3. Sperm cytology This was carried out using the Christmas Tree stain technique, reported as the most useful test when compared to hematoxylineosin and alkaline fuchsin [18]. Smears from each pellet suspension were prepared in glass slides, heat-dried, xed in alcohol and ether, and stained with the nuclear fast red solution for 15 min in a humidied chamber. Slides were washed with deionized water, stained with picroindigocarmine solution for 30 s then bleached with ethanol and air dried. Preparations were microscopically screened for spermatozoa with an 100 magnication and the total cell number was scored. The sample was considered as positive when at least 1 sperm (usually head) was unequivocally observed. 2.4. PSA (p30) detection For this technique, a commercially available immunochromatographic membrane test assay was used. It was the One Step ABA card PSATM (Abacus Diagnostics, West Hills, CA) with a reported sensitivity of detection for PSA as low as 4 ng/mL. The test consists of an all-inclusive single test device. Two hundred microliters of supernatant were placed in the sample window of the device and reaction was allowed to proceed for 10 min. The presence of a line formed by an antigen antibodydye sandwich in the reaction zone indicated a positive reaction. The device displays also an internal control line of reaction for monitoring device quality. This test has proven to be useful in previous studies [12,19]. 2.5. Acid phosphatase activity In this case the widely used qualitative technique based on the a-naphthyl phosphate (a-NP) substrate [20] was employed at Texcoco and Toluca laboratories with some differences. In brief, one whole swab was placed between two small pieces of lter paper then over a glass slide. Afterwards 150 mL of substrate solution (prepared in concentrations of 0.8% and 0.5%, w/v, at Texcoco and Toluca laboratories respectively) were added to the swab and it was left at room

potential benet of higher sensitivity and specicity than biochemical ones but the possibility of assailant identication. To date, RNA-based protocols are still considered as supplementary to DNA-based techniques on the basis of the higher stability of DNA over RNA in casework material. In an increasing number of forensic laboratories DNA-based assays have already been introduced. In particular Y-STRs have earned acceptance over AS-STRs since for the latter ones the female DNA might mask the male DNA prole in samples containing gender-mixed materials and when more than a male DNA is involved AS-STR data may be inconclusive [9]. Likewise the use of multiplexed Y-STRs and specic loci sets for forensic purposes has been supported and recommended by the DNA Commission of The International Society of Forensic Genetics (ISFG) [10]. However in laboratories with limited support, e.g. in the developing world and emergent economies, this issue might deserve a longer time to be done. Moreover the conclusive demonstration of semen materials in rape samples by biochemical tests preceding a reliable malespecic DNA proling still requires evaluations under real live sampling conditions. The most recommended biochemical techniques for the routine forensic analysis of rape include sperm cytology (SC), AP activity (APA) and recently PSA detection [11]. SC is a gold standard or conrmatory test; in cases of oligospermic, azoospermic or vasectomized individuals APA is a screening (presumptive) test because it can be found at lower concentrations in normal vaginal secretions [12] while PSA provides a more specic marker that has been detected with immunological techniques of increasing sensitivity [13]. Noteworthy, these markers have shown distinct stability when tested in vaginal uid after intercourse: spermatozoa may be found up to four days [14] while PSA (as detected by ELISA) show a mean decay time of 27 h and for quantitative APA it is 14 h-post coitus [15]. As expected, a delay in sampling vaginal swabs (e.g. !16 h) adversely affect detecting these markers albeit their relative stability is kept the same [4,11]. In this context, a vast majority of reports on the comparison of SC, APA and PSA have relied on analysis of semen samples or donor swabs taken after known times after intercourse. However it was clear from comparative studies reported by Poyntz and Martin [16] that the efcacy of any given marker, e.g. PSA (as detected by crossed-over immunoelectrophoresis) is signicantly higher in donor swabs (45 out of 52 cases) as compared to casework swabs (11 out of 59 cases). Several factors such as possible sampling errors, variable delays before sampling, environmental factors, artifact-derived sperm destruction and different handlings of casework samples from a laboratory to another account for and inuence these results. Indeed the individual and collective contribution of these systematic factors to biased conclusions is largely unknown due to a lack of information on data obtained under real live sampling conditions. To date, only two detailed studies carried out at Toronto, Canada and Hong Kong, China compared the detection rates of these three markers in forensic casework samples and data were categorized on the basis of a positive result in one, two or the three tests performed in 54 and 144 vaginal swabs respectively [4,17]. The aim of this work was to carry out SC, APA and PSA tests in vaginal swabs collected from casework rapes submitted to two different Forensic Laboratories in Mexico. Data were grouped into eight categories and compared with those obtained and grouped similarly in the two previous studies aforementioned. Furthermore Y-STR proling was carried out in representative samples from most of these categories. The present work gives an overview on the variability in efcacy of each test performed at different laboratories and provides a general notion about the individual and combined efcacy of these tests and its possible contribution for suitable DNA proling in forensic analysis of rape.

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118

113

Table 1 Sensitivity of SC, prostate-specic antigen PSA and acid phosphatase activity APA assays in swabs containing dilutions of semen from two donors with distinct sperm counts. Dilutions SCa Donor 1 1:10 1:100 1:500 1:1,000 1:10,000 1:50,000 1:100,000 1:200,000 1:400,000
a b c d e b

(PSA)a Donor 2 + + + +
c

(APA-1)a,d Donor 2 + + + + + + + Donor 1 + + + + Donor 2 + + + + +

(APA-2)a,e Donor 1 + + + + Donor 2 + + + +

Donor 1 + + + + + + +

+ + + + +

+ and for positive and negative SC, PSA and APA results. Sperm count: 65.5 106 cells/mL. Sperm count: 23 106 cells/mL. With APA technique as used at Texcoco laboratory. With APA technique as used at Toluca laboratory.

temperature for up to 1 min (Toluca laboratory) or 5 min (Texcoco laboratory). The presence of a deep violet color in the swab was indicative of a positive reaction. 2.6. DNA extraction and proling Pellet suspensions recovered from 27 representative casework swabs were subjected to Proteinase K treatment followed by differential extraction of sperm and epithelial fractions using the DifferexTM System kit (Promega Corp., Madison, WI) following manufacturers instructions. In fabric swabs this latter was not used except for samples from the azoospermic donor. DNA was extracted from the sperm and epithelial fractions using the DNA IQTM System (Promega) and 10 mL of these fractions were processed to quantify male and human (male + female) DNA by Real Time PCR using the QuantilerTM Duo kit (Applied Biosystems, Foster City, CA) which is capable to detect as low as $6 pg/mL of human DNA. In this protocol, the absolute DNA quantication was carried out using a standard curve (500.023 ng DNA) and the corresponding amplication plots were analyzed using the 7500 System SDS software which allowed to determine the cycle threshold (Ct) values for male and human DNA components (dyes FAMTM and VICTM respectively) and for the corresponding Internal Positive Control (IPC, dye NEDTM). This latter parameter served to determine the presence or absence of PCR inhibitors in rape casework samples. The equivalent to at least 0.5 ng DNA (when not possible, a maximum of 14.2 mL of DNA extracts) was amplied using the PowerPlexTM Y System (Promega) following suppliers indications in 25 mL total reaction volume. This kit allows the co-amplication and three-color detection of 12 Y-STR loci including the minimal (DYS19, DYS385I/II, DYS389I/II, DYS390, DYS391, DYS392 and DYS393) and extended (plus DYS437, DYS438 and DYS439) haplotypes [10] comprising a ladder of up to 102 alleles detecting amplicons <335 bp. The manufacturer species that this kit is able to amplify full proles from less than 250 pg of male DNA even in the presence of >100-fold excess of female DNA. Y-STR typing was carried out on the ABI Prism 310 Genetic Analyzer (Applied Biosystems) and the obtained GeneScan electropherograms were analyzed using the GeneMapper ID v3.2 software (Applied Biosystems). 2.7. Statistics Mean and standard deviations (S.D.) were obtained to estimate the variability in the frequencies of positive reactions scored in the distinct studies compared. In some cases two-tailed Students t-test was used to compare mean S.D. values and a level of condence of 95% was considered as signicant. The overall comparison of all categories in which SC, APA and PSA results were classied was performed by calculating the variation coefcient (S.D./mean).

3. Results Forensic analysts with similar training performed the mentioned SC, APA and PSA techniques at Texcoco and Toluca laboratories. An initial step was to determine the sensitivity of these protocols using dilutions of semen samples from one normospermic (donor 1) and one oligospermic (donor 2) donors (Table 1). In this case, APA assays were tested as established in these two laboratories and were thus considered as two different protocols. As shown by these data, the PSA technique displayed the higher sensitivity (positive at dilutions up to 1:100,000 for both donors) while SC and APA assays were always negative at a

1:50,000 dilution and in some cases even at a 1:10,000 dilution. As additional data, detection limits in PSA assays using semen samples from the azoospermic and the vasectomized donors were at 1:100,000 and 1:50,000 respectively and with APA technique these were 1:1000 and 1:500 (as from Tolucas laboratory). Considering sperm counts from donors, the detection limit of SC was of 1300 (normospermic) and 4600 (oligospermic) spermatozoa/swab in these samples. In the case of APA test carried out as in Texcoco or Toluca laboratories, for the latter the detection limit was at a 1:1000 dilution in both donors while for the former it was at a 1:1000 (donor 1) or 1:10,000 (donor 2) dilution suggesting a likely increased sensitivity of the protocol used in Texcoco laboratory. To obtain a precise value for the detection limit in PSA assays, the semen from a normospermic donor with known total PSA concentration (determined by VIDAS-TPSATM) was used. It was positive at a maximal dilution of 1:350,000 that corresponded to 1.72 ng/mL. This value is slightly better than that stated by the manufacturer (4 ng/mL) and is equivalent to that obtained in other studies using seminal uid standards [12]. In the case of APA and PSA assays, further analyses on specicity and interference by bodily uids/materials were carried out. Regarding specicity, from all the bodily uids and fecal materials mentioned in Section 2.1, as expected only vaginal secretions gave a false positive result in the APA assay as used in Texcoco laboratory albeit these samples were negative by the protocol used in Toluca laboratory. In the case of PSA assays performed for all these samples, only male urine from individuals aged 13 years old and older gave a false positive result even when diluted 1:100. On the other hand, none interference was observed with all bodily uids and fecal materials mixed with semen diluted at 1:10 (i.e. the reaction was always positive by PSA and by the two variants of APA technique). We next carried out SC, PSA and APA analyses in casework swabs after a maximum delay of 5 days between sampling and analysis in forensic laboratories at Toluca or Texcoco. As mentioned, this fact did not lead to biased results. In 39 out of 100 Toluca samples (39%) and 11 out of 48 Texcoco samples (22.9%) all assays were negative. Positive results in one, two or the three techniques were analyzed as total frequencies. In Toluca samples SC had the highest overall positivity (59/100 or 59%) followed by PSA (40%) while APA was less sensitive (34%). However SC and PSA showed similar overall positivities in Texcoco samples (21/48 [43.7%] and 20/48 [41.6%] respectively) which were signicantly lower than that of APA (36/48 [75%]). By analyzing the overall positivity in two out of the three assays, in both laboratories the frequencies were comparable: APASC in 34% and 41.6%, SCPSA in 38% and 33.3%, and PSAAPA in 29% and 39.5% for

114

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118

Table 2 Category-based results of acid phosphatase activity, spermatozoa and PSA detection in 100 forensic casework samples from Toluca and 48 forensic casework samples from Texcoco. Category AP activitya Sperm cytologya PSA detectiona Toluca No. I II III IV V VI VII VIII
a

Texcoco % 29.0 5.0 9.0 0 0 39.0 16.0 2.0 No. 15 5 1 12 4 11 0 0 % 31.2 10.4 2.0 25.0 8.3 22.9 0 0

+ + + +

+ + + +

+ + + +

29 5 9 0 0 39 16 2

+ and indicate positive and negative APA result; presence or absence of spermatozoa; positive or negative detection of PSA.

Toluca and Texcoco samples respectively. Thus the overall positivity in two of three assays was 33.6 3.6% in Toluca samples and 38.13 3.52% in Texcoco samples (non signicantly different, P > 0.05). The same case was for the overall positivity in all three assays: 29/100 (29%) in Toluca samples and 15/48 (31.2%) in Texcoco samples. From the above data it was apparent that the major differences between the two laboratories were derived from the overall positivity of the most sensitive assays in each case, namely SC in Toluca samples and APA in Texcoco samples. In the particular case of SC, the presence of few spermatozoa (under our conditions: one or less spermatozoa/eld in average or less than 10 spermatozoa/ smear) along to higher amounts of epithelial cells (likely from victims) was a common feature in 25 samples from Toluca and 16 samples from Texcoco. In spite that vaginal components might contribute to give false positive results in APA assays, there was not an obvious relation between positivity and negativity in paired SC APA data from the two laboratories. In order to obtain more detailed information on the inuence of the results from individual assays on these inter-laboratory differences, data were divided into eight categories corresponding to the possible combinations of results of SC, PSA and APA assays in each sample. As shown in Table 2, in Toluca samples there were no positive results only by APA or only negative by SC while in Texcoco samples none was positive only by either SC or PSA. Of note, there were 16% of Toluca samples positive only by SC (Category VII) and 25% of Texcoco samples positive only by APA (Category IV) that opposed to an absence of samples with similar results in the other laboratory. In the meanwhile as few as two Toluca samples (2%) and none of Texcoco samples were positive only by PSA (Category VIII). When the other assays were positive, PSA gave negative results in 5% of Toluca samples and 10.4% of Texcoco samples (Category II), APA had negative result in 9% of Toluca samples and in only one (2%) of Texcoco samples (Category III) while SC was negative in only four (8.3%) of Texcoco samples (Category V). These results conrmed a direct inuence of SC and

APA as the most sensitive techniques on the patterns of positivity in Toluca and Texcoco samples respectively. In order to extrapolate the analyses of the present study, it was of further interest to compare our Category-based results with the two other previously reported works providing data on the positivity of SC, PSA and APA assays in forensic casework samples. The former was carried out more than two decades ago at Toronto, Canada in which SC was carried out using the Christmas Tree stain, PSA was determined by crossed immunoelectrophoresis and APA with a-naphthyl phosphate as substrate in 144 vaginal swabs [4]. In the other, recently reported from Hong Kong, China [17], 54 samples were analyzed for SC as done by hematoxylineosin stain, PSA was detected by the One Step ABA card PSATM and APA was determined using a-naphthyl phosphate substrate (Benjamin C.M. Pang, personal communication). As a reference, in Toronto samples the overall positivity of APA and SC was similar (104/144 [72.2%] and 96/144 [66.7%] respectively) but higher than that of PSA (62/ 144 [43.1%]) and in Hong Kong samples the overall sensitivity of APA, SC and PSA were equivalent (29/54 [53.5%], 28/54 [51.6%] and 25/54 [46.1%] respectively). Also, there were 46/144 [31.9%] and 21/54 [38.8%] samples that were negative by all techniques and likewise 57/144 [39.6%] and 21/54 [38.8%] of samples were positive by all techniques in Toronto and Hong Kong collections respectively. Table 3 shows the comparisons among data from the four laboratories that were used to calculate global frequencies. Since the global variation coefcient was lesser to 0.2 for Category I (positive result in SC, PSA and APA) and Category VI (negative result in SC, PSA and APA) it was inferred that the results from the four laboratories reected a similar pattern in these cases. Indeed the global positivity in one or two out of the three techniques showed more variability (variation coefcient higher than 0.3). As occurred with data from Mexican laboratories, the most striking variations in frequency were observed in two cases: (1) in samples positive only by SC (Category VII) because these were high in Toluca and Toronto samples (16.0% and 18.1%) but low in Texcoco

Table 3 Category-based results of acid phosphatase activity, spermatozoa and PSA detection in forensic casework samples from Toluca and Texcoco (this study) and from two previous studies. Results are given as percentage values. Category I II III IV V VI VII VIII
a b c

Toluca (n = 100) 29.0 5.0 9.0 0 0 39.0 16.0 2.0

Texcoco (n = 48) 31.2 10.4 2.0 25.0 8.3 22.9 0 0

Torontoa (n = 144) 39.6 6.9 2.1 25.0 0.7 31.9 18.1 0.7

Hong Kongb (n = 54) 38.8 5.5 5.5 7.4 1.8 38.8 1.8 0

Global mean S.D. 34.6 4.6 6.9 2.1 4.65 2.8 14.3 10.9 2.7 3.2 33.1 6.5 8.9 8.1 0.6 0.8

Variation coefcientc 0.132 0.304 0.602 0.762 1.185 0.196 0.910 1.33

Data taken from Stubbings and Newall [4]. Data taken from Pang and Cheung [17]. Calculated dividing the S.D. value by the mean value.

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118 Table 4 Category-based results of Y-STR proles from sperm fractions of 27 representative forensic casework samples from Toluca laboratory. N.D.: not determined. Category I Number of samples (n) 13 Frequency (%) 48.1 DNA typing (loci amplied/total)a [n] 12/12 [13] Overall typing (%) 100 Spermatozoa/ smear (s) or eld (f) 672/s, 114/f

115

Male DNA input (ng)c [n] N.D. [5] 0.185.26 [8] N.D. [2] 0.046 [1]d 0.00.333 [2]d 0.0 [3] 0.032 [1]d 0.0 [1] 1.05d N.D. N.D. 0.12

III

7.4

11/12 [1] 2/12 [1] 6/12 [1] 0/12 [2] 0/12 [4]

54.1

6/s 1/s b

11.1

16.6

VI

14.8

0.0

VII VIII

1 4

3.7 14.8

0/12 (1] 3/12 [1] 4/12 [1] 10/12 [1] 12/12 [1]

0.0 60.4

1/s

a b c d

DNA typing was performed using the PowerPlexTM Y-System in sperm fractions unless otherwise specied. : Spermatozoa not detected. Male DNA input was quantied using the QuantilerTM Duo kit. DNA typing performed using epithelial fractions because male DNA was absent in sperm fractions.

and Hong Kong samples (0% and 1.8%), and (2) in samples positive only by APA (Category IV) where Texcoco and Toronto samples showed high frequencies (25.0% each) whilst Toluca and Hong Kong samples had low ones (0% and 7.4%). For the remainder cases (Categories II, III and V where one of the three techniques was negative, and Category VIII where only PSA was positive) their individual contribution was rather minor (<7% of total cases in each Category). The global overall positivity for APA and SC was similar (58.6 16.4% and 55.2 8.54% respectively) and higher than that for PSA (42.7 2.2%) but without statistic signicance (P > 0.05). In addition, there was a range of 61.0% (Toluca) to 77.1% (Texcoco) of samples being positive in at least one of the techniques used (i.e. including all but Category VI) with Toronto and Hong Kong samples having values of 68.1% and 61.1% respectively. Data derived from this inter-laboratory analysis rendered similar frequencies (approximately one-third each) for samples displaying either positivity or negativity by all biochemical techniques (Categories I and VI respectively) with the remaining one-third composed by variable frequencies derived from analyses using one or two techniques (Categories IIV, VIIVIII). As an approach to address if the quality of DNA proles from forensic swabs might kept a relation with previous results from AP, SC and PSA assays, 27 additional samples obtained at Tolucas laboratory were processed for DNA extraction from sperm and epithelial fractions and amplication and Y-STR proles were determined in the former extracts (Table 4). On one hand, the 12-loci PowerPlexTM Y-System has been used in other studies following detection of seminal markers as PSA and semenogelin [5]. On the other, in our validation study complete DNA proles could be obtained from sperm fraction in samples of the Donor 1 (normospermic) at <1:1000 dilutions and from epithelial fraction of the azoospermic individual at 1:10 dilution. Based on this, it was preferred to analyze both sperm and epithelial fractions from the additional casework swabs for presence of male DNA before its proling. In this manner, samples with apparently enough male DNA input (!0.25 ng) but with absent or incomplete Y-STR proles would imply the presence of PCR inhibitors. During sample fractioning, in some sperm or epithelial fractions male or female DNA was respectively the only component. In gender-mixed fractions, in spite that male DNA present in sperm or epithelial fractions was often at lower concentrations than human (male + female) DNA, the excess of female DNA was relatively low (<10-fold; see Supplementary Table 1) and allowed obtaining

complete or partial Y-STR proles from the 27 additional samples analyzed depending on their categorization by AP, SC and PSA assays (Table 4). In these determinations complete DNA proles (12 loci) were obtained from sperm fractions of Category I samples (n = 13) while none marker was amplied from Category VI samples using either fraction (n = 4). In contrast, only partial DNA proles (211 out of 12 loci) were obtained from the remainder samples (n = 10) belonging to Categories III, V, VII and VIII (Table 4). Of note, all samples containing at least 6 spermatozoa/smear (n = 13, all of Category I) rendered complete DNA proles with the exception of one sample without detectable spermatozoa (Category VIII) that rendered full DNA prole. Of samples where none marker was amplied (n = 7), 6 did not have detectable spermatozoa (Categories VVI) and the one remainder (Category VII) only had one sperm per smear. In another sample from Category III having also one sperm per smear, 2 out of 12 markers were identied. In general, partial Y-STR proles (211 out of 12 markers identied) were obtained in 6 samples belonging to Categories III (2/2), V (1/3) and VIII (3/4) and containing 06 spermatozoa/smear (all PSA-positive). In total, 20 out of 22 PSApositive samples rendered partial or full Y-STR proles. As far as the quantity of male DNA loaded in Y-STR typing assays was concerned, its quantication was carried out using suitable standard curves (Ct = 3.12[log(Qty)] + 29.4601, R2 = 0.9897; see Supplementary Table 3) for 18 samples because material from 9 samples was exhausted during proling tests (14.2 mL per assay) at the expense of DNA quantication. From these exhausted samples, in 5 cases complete proles were obtained, in 2 cases 1011 markers were identied and in the remainder 2 cases only 24 markers were resolved. Therefore in only 4 samples the precise cause of obtaining partial proles could not be further addressed. Of interest, in all 18 samples with quantied DNA input the presence of PCR inhibitors was ruled out because the IPC data for these samples indicated adequate amplication and Ct values were lower than 31 as stated by the kit manufacturer (Supplementary Table 3). In the 9 samples with known male DNA input where complete DNA proles were determined (8 from Category I and one from Category VIII), a range of 0.125.26 ng male DNA were used. In samples where male DNA input was determined and none marker was identied (n = 7), in 5 cases no male DNA was applied while in the 2 remainder cases the same result occurred in spite of applying 0.032 and 0.333 ng from epithelial fractions (Table 4). Finally, two samples rendered partial proles (3 and 6 out of 12 loci

116

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118

amplied) where 1.05 and 0.046 ng male DNA from the corresponding epithelial fractions were loaded. Regarding the frequencies of cases in which individual markers were identied, these ranged from 14 out of 27 (51.8%) for DYS19 to 19 out of 27 (70.3%) for DYS437 with the remainder 10 markers with intermediate frequencies (see Supplementary Table 2). In summary, complete Y-STR proles were obtained from 14 (51.8%) of casework samples. An absent prole due to absence of male DNA was observed in 5 cases (18.5%) while in the remainder 4 samples with absent or partial DNA proles were obtained the causes included low DNA input (0.032 and 0.046 ng) in 2 samples (7.4%) or other undetermined effects in the other 2 samples (7.4%) where the DNA input was apparently enough (0.333 and 1.05 ng) (Table 4). 4. Discussion In forensic practice, detecting semen material in casework vaginal swabs by reactivity of a single component (i.e. AP and PSA) in biochemical assays could give a mistaken conclusion (i.e. false positive results by vaginal secretions and urine respectively) [21]. Likewise if the SC used as a standard gold is often carried out by analysts with different training and skill indeed false-negative results may be obtained. In this context, the present work was aimed to determine the individual and combined frequencies of detection of spermatozoa, PSA and AP in vaginal swab samples and these data were compared using casework data obtained from different forensic laboratories: two at Mexico and two from previous reports carried out at Toronto and Hong Kong [4,17]. This very likely implies systematic sources of variation derived on a hand from the use of different techniques, some with different sensitivities as in the case of PSA (determined by rapid immunochromatography with the commercially available One Step ABA card PSATM device in this work and in Hong Kong samples or by crossed immunoelectrophoresis in Toronto samples), others with slightly different efcacy as in sperm detection (performed by Christmas Tree stain used in this work and in Toronto samples, or by hematoxylineosin stain in Hong Kong samples). In the case of APA determinations, these were performed using a similar protocol in the four laboratories (i.e. a-NP substrate). In Mexican laboratories the techniques reportedly considered as the most sensitive for SC and PSA assays were used [18,19] while the activity of AP determined by colorimetric assay is of extended use albeit it may be of lower efcacy than detection of AP molecule by immunoelectrophoresis [21]. On the other hand, additional sources of variation in casework material come from logistic information since complainant report is not necessarily infallible and during the crime the initial quantity of semen is unknown or even rape might not involve ejaculation. Also, using distinct sampling and handling procedures may cause a differential rate of degradation of semen markers [4,11,14,15]. By analyzing results from Mexican laboratories involved in the present study and where the same techniques were used with known differences, it was clear that SC and APA assays had the higher overall positivity in Toluca and Texcoco samples respectively. It is worth noting to mention that the relatively low positivity of APA technique (34%) in Toluca samples as compared with a 75% positivity in Texcoco samples are mostly explained by the use of a more diluted a-NP solution in the Toluca laboratory (0.5%) instead of the 0.8% a-NP solution routinely used in the Texcoco laboratory combined to the longer reaction time (5 min) in Texcoco laboratory as compared to 1 min in Toluca laboratory. Otherwise PSA had a lower but very similar positivity between these two laboratories (40% and 41.6% respectively). This latter observation is noteworthy since PSA was performed with a commercially available kit thus certain causes of systematic variation could be minimized. Thus PSA behaved as a more

consistent marker in Mexican samples in spite of the higher positivity of APA or SC tests. Some of the aforementioned observations could be extrapolated when compared to the previous studies reported [4,17]. For instance, APA again exhibited the highest overall positivity in Toronto and Hong Kong samples. Also, PSA positivity was almost the same in these two laboratories (43.1% and 46.1% respectively). These data t nicely with the notion of using APA as a test of screening or presumptive value and that PSA is a specic marker of lower but consistent efcacy in forensic practice even at the expense of an expected degree of variation in sperm detection frequencies (from 43.7% in Texcoco samples to 66.7% in Toronto samples). It was indeed obvious the inuence of environmental and/or handling factors on the potential degradation of PSA to concentrations below the sensitivity limit of immunochromatographic or immunoelectrophoretic methods used under eld conditions. This is in close agreement with previous observations [16] albeit the use of currently available kits has likely improved the efciency of PSA detection in rape samples (from 18.6% in [16] to an overall 42.7% observed in this work). As far as results on the efcacy of SC, APA and PSA assays in casework samples are concerned, it was clear that none was compelling by itself but some proposals to improve their combined use may be done. For example, the routine and generalized use of commercially available kits as are the rapid and qualitative staining kits for APA may extend the reliability of this marker from 14 to at least 24 h following rape [11]. In addition, these kits provide the advantages of internal controls and specialized skill for interpretation is not required. Furthermore detecting AP molecule [21] using standardized devices instead of its activity would improve frequencies of detection. Regarding PSA tests, in spite that its use in eld conditions reached positivity in 2 out of 5 cases, its high specicity may give sufcient support towards a conclusive identication of semen in these cases. Recent studies in vaginal samples from casework have also suggested the routine use of onestep PSA tests in forensic practice of rape [22]. In a general perspective, the combined results of using SC, APA and PSA tests could be considered as conclusive for semen presence from approximately 1 out of 3 cases (Category I) to approximately 1 out of 2 cases in a scenario where at least SC is positive (Categories I, II, III and VII combined), and strongly presumptive in 2 out of 3 cases (with at least one test positive, all but Category VI combined). Otherwise Category VI samples were strongly suggestive of semen absence in approximately one-third of rape cases. This notion suggests that results contained within Categories I and VI were of greater reliability (variation coefcient < 0.2) circumventing in part the inherent systematic, informational and environmental sources of variations among the distinct forensic laboratories. If semen presence/absence has to be conclusively determined to proceed/withdraw further DNA proling (i.e. for assailant identication) it is desirable to avoid as most as possible these variations independently of the possible lack of compelling logistic evidences (i.e. medical examinations and testimonies from victims, assailants and possible witnesses) in a given case of suspected rape. In good agreement with the data previously obtained and grouped from Mexican, Canadian and Chinese laboratories, the further Y-STR proling performed on 27 additional samples shed interesting observations. First, it was clear that all Category I samples (n = 13) gave complete Y-STR proles while all Category VI samples (n = 4) predicted well the absolute absence of DNA proling (Table 4). One exception was observed in one sample of Category VIII where a complete male DNA prole was obtained in the absence of sperm, but with detectable male DNA content (120 pg) in the sperm fraction (see Supplementary Table 1), a value above the sensitivity limit specied by the manufacturer ($6 pg).

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118

117

Secondly, it was observed that a PSA-positive result (Categories I, III, V and VIII) should be combined with a threshold sperm count (in our experience, >6 sperm/smear) to allow obtaining complete DNA proles. This is consistent with the use of the 12-loci PowerPlexTM Y-System in other studies following detection of seminal markers as PSA and semenogelin [5]. In this sense, multiplexed Y-STR proles have been obtained in the absence of spermatozoa using 3-loci [23], 7-loci [24] or even 10-loci [25] based protocols, albeit in the latter study all samples were APAand PSA-positive. It seems that the greater is the complexity of the Y-STR prole the greater is the difculty to obtain complete DNA proles but the greater is the certainty of assailant identication. Thirdly, during the present study the complete DNA prole from the azoospermic donor required the use of the epithelial fraction instead of the sperm one, suggesting the convenience of identifying and quantifying male DNA in both sperm and epithelial fractions, particularly in samples where spermatozoa are absent (Categories IVVI, VIII) before DNA proling. The informative value of sperm fractions was enough by itself in 22 out of the 27 samples subjected to DNA typing. Of the 5 samples where epithelial fractions had to be used to achieve DNA proling, in 4 no loci were amplied and only one sample from Category V showed amplication of 6/12 loci (see Supplementary Table 2). This suggests that analyzing epithelial fractions for assailant identication in casework swabs would be of help in some cases lacking data from sperm fractions. In addition, the selective quantication of male DNA gives insights on the likely causes (mostly the presence of PCR inhibitors) causing incomplete DNA proles. As mentioned before, the presence of PCR inhibitors was discarded in all samples analyzed with the use of internal controls (IPC) and Ct value determinations (Supplementary Table 3). In the present work either complete (n = 14), absent (n = 7) or partial (n = 6) Y-STR proles were obtained. In the group of absent Y-STR prole, in 5 out of 7 samples the cause was the absence of DNA, in one it was due to low DNA content and in the remainder one it was due to undetermined factors as the DNA input was enough. From the two cases analyzed in the group of partial Y-STR proles, in one sample this was related to low DNA content and in the other sample to undetermined factors. Based on the relative stability of the 12 loci amplied from the analyzed samples (Supplementary Table 2), the apparent lack of PCR inhibitors in vaginal swabs and the apparently enough DNA content in some samples where DNA proles were absent or incomplete (in this study, 2 out of 27), further studies are warranted to address if the primers used for STR detection might have distinct ability to anneal with target DNA or to be trapped or inactivated by agents particularly present in some casework samples as the result of assault conditions or swab sampling/handling. Taken together, the combined use of biochemical and DNA typing approaches kept a good relation, especially in samples positive for AP, sperm and PSA assays that will render complete YSTR proles; otherwise AP, sperm and PSA-negative samples will render absence of Y-STR proling. Future studies addressing the convenience of specic Y-STR proles regarding their complexity and marker stability in forensic samples would be useful to improve the routine and combined use of biochemical and molecular typing assays in the forensic analysis of rape. 5. Conclusions The observations derived from the present work give direct evidence about a representative and comparable global efcacy of the combined use of biochemical techniques (SC, APA and PSA) to detect seminal markers in casework vaginal swabs of rape even considering the inherent procedural variability among different forensic laboratories at distinct localities in the world. In particular, a positive result in all the techniques used reached a frequency of

about 1 of 3 cases (Category I) and was able to predict a complete YSTR proling that adds the benet of assailant identication [26] while a negative result by all these techniques (Category VI) strongly predicts the absence of both semen and male DNA proling although epithelial fractions would allow obtaining Y-STR proles. Nevertheless the generalized use of standardized PSA rapid tests coupled to Y-STR typing or other PCR-based protocols still deserves further research on identication of best practices based on inter-laboratory comparisons [27]. Also, this will require a greater nancial support from forensic authorities to the analytical infrastructure in a still signicant number of laboratories in the world. Acknowledgements This work was carried out according to current laws in Mexico and with the consent of the authorities from the Mexican forensic laboratories involved in this study. We thank Dr. Xochitl Adriana Felix-Lopez and Q. Alfonso Luna-Vazquez from Laboratory of Forensic Genetics, DGCSP-PGR for their valuable support performing Real Time PCR assays during validation studies of Y-STR typing and to XAFL for her valuable comments to this work. We also acknowledge Jean-Paul Anthony for critically reading this manu script and Mara de Lourdes Vega-Navarrete (Dist. Com. ZOGBI) for the kindly supply of reagents for validation of Y-STR typing studies. QFB Gabriela Chavez Marn (Laboratory of Forensic Chemistry, PGJEM) is acknowledged by her logistic support during DNA proling studies. Authors recognize the very detailed and interesting comments and suggestions of the Reviewers that contributed to improve the content of this work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.forsciint.2010.07.012. References
[1] C. Grossin, I. Sibille, G.L. de la Grandmaison, A. Banasr, F. Brion, M. Durigon, Analysis of 418 cases of sexual assault, Forensic Sci. Int. 131 (2003) 125130. [2] N. Riggs, D. Houry, G. Long, V. Markovcick, K.M. Feldhaus, Analysis of 1076 cases of sexual assault, Ann. Emerg. Med. 35 (2000) 358362. [3] M.L. Lawson, M. Maculuso, A. Bloom, G. Hortin, K.R. Hammond, R. Blackwell, Objective markers of condom failure, Sex. Trans. Dis. 25 (1998) 427432. [4] N.A. Stubbings, P.J. Newall, An evaluation of gamma-glutamyl transpeptidase (GGT) and p30 determinations for the identication of semen on postcoital vaginal swabs, J. Forensic Sci. 30 (1985) 604614. [5] I. Sato, F. Barni, M. Yoshiike, C. Rapoe, A. Berti, S. Nakaki, K. Yamazaki, F. Ishikawa, T. Iwamoto, Applicability of Nanotrap Sg as a semen detection kit before malespecic DNA proling in sexual assaults, Int. J. Legal Med. 121 (2007) 315319. [6] G. Sensabaugh, Isolation and characterization of a semen-specic protein from human seminal plasma: a potential new marker for semen identication, J. Forensic Sci. 23 (1978) 106115. [7] L. Roewer, Y chromosome STR typing in crime casework, Forensic Sci. Med. Pathol. 5 (2009) 7784. [8] M. Bauer, RNA in forensic science, Forensic Sci. Int. Genet. 1 (2007) 6974. [9] J.G. Shewale, H. Nasir, E. Schneida, A.M. Gross, B. Budowle, S.K. Sinha, Y-Chromosome STR system, Y-PlexTM 12, for forensic casework: development and validation, J. Forensic Sci. 49 (2004) 12781290. [10] L. Gusmao, J.M. Butler, A. Carracedo, P. Gill, M. Kayser, W.R. Mayr, N. Morling, M. Prinz, L. Roewer, C. Tyler-Smith, P.M. Schneider, DNA Commission of the International Society of Forensic Genetics (ISFG): an update of the recommendations on the use of Y-STRs in forensic analysis, Forensic Sci. Int. Genet. 1 (2007) 312. [11] N. Khaldi, A. Miras, K. Botti, L. Benali, S. Gromb, Evaluation of three rapid methods for the forensic identication of seminal uid in rape cases, J. Forensic Sci. 49 (2004) 15. [12] J.T. Chen, G.L. Hortin, Interferences with semen detection by an immunoassay for a seminal vesicle-specic antigen, J. Forensic Sci. 45 (2000) 234235. [13] C.M. Sturgeon, A.R. Ellis, Improving the comparability of immunoassays for prostate-specic antigen (PSA): progress and problems, Clin. Chim. Acta 381 (2007) 8592. [14] G.M. Willott, J.E. Allard, Spermatozoatheir persistence after sexual intercourse, Forensic Sci. Int. 19 (1982) 135154.

118

L. Romero-Montoya et al. / Forensic Science International 206 (2011) 111118 prostate specic antigen in vaginal specimens of raped women, J. Med. Assoc. Thai. 90 (2007) 26242629. ` I. Sibille, C. Duverneuil, G. Lorin de la Grandmaison, K. Guerrouache, F. Teissiere, M. Durigon, P. de Mazancourt, Y-STR DNA amplication as biological evidence in sexually assaulted female victims with no cytological detection of spermatozoa, Forensic Sci. Int. 125 (2002) 212216. J.A. Soares-Vieira, A.E. Correia-Billerbeck, E. Sadayo, M. Iwamura, R. Andrade Zampieri, G.J. Fgaro-Gattas, D. Romero-Munoz, J. Hallak, B. Bilhardinho Mendonca, A. Marmo-Lucon, Y-STR in forensic medicine: DNA analysis I semen samples of azoospermic individuals, J. Forensic Sci. 52 (2007) 664 670. C.L. Johnson, R.C. Giles, J.H. Warren, J.I. Floyd, R.W. Staub, Analysis of non-suspect samples lacking visually identiable sperm using a Y-STR 10-Plex, J. Forensic Sci. 50 (2005) 13. A. Hall, J. Ballantyne, Novel Y-STR typing strategies reveal the genetic prole of the semen donor in extended interval post-coital cervicovaginal samples, Forensic Sci. Int. 136 (2003) 5872. J.E. Allard, A. Baird, G. Davidson, S. Jones, J. Lewis, L. McKeena, C. Weston, D. Scrimger, G. Teppett, A comparison of methods used in the UK and Ireland for the extraction and detection of semen on swabs and cloth samples, Sci. Justice 47 (2007) 160167.

[15] H.C.B. Graves, G.F. Sensabaugh, D. Crim, E.T. Blake, Postcoital detection of a malespecic semen protein. Application to the investigation of rape, New Eng. J. Med. 312 (1985) 338343. [16] F.M. Poyntz, P.D. Martin, Comparison of p30 and acid phosphatase levels in postcoital vaginal swabs from donor and casework studies, Forensic Sci. Int. 24 (1984) 1725. [17] B.C.M. Pang, B.K.K. Cheung, Identication of human semenogelin in membrane strip test as an alternative method for the detection of semen, Forensic Sci. Int. 169 (2007) 2731. [18] J.P. Allery, N. Telmon, R. Mieusset, A. Blanc, D. Rouge, Cytological detection of spermatozoa: comparison of three staining methods, J. Forensic Sci. 46 (2001) 349351. [19] M.N. Hochmeister, B. Budowle, O. Rudin, C. Gehrig, U. Borer, M. Thali, R. Dirnhofer, Evaluation of prostate-specic antigen (PSA) membrane test assays for the forensic identication of seminal uid, J. Forensic Sci. 44 (1999) 10571060. [20] Metropolitan Police Forensic Science Laboratory. Biology Methods Manual. London, England, 1978. p. 3.18 (Chapter 3). [21] C.K. Bryson, A.M. Garlo, S.C. Piner, Vaginal swabs: endogenous and postcoital components, J. Forensic Sci. Soc. 29 (1989) 157171. [22] V. Peonim, T. Chirachariyave, K. Atamasirikul, J. Talthip, Comparable between rapid one step immunochromatographic assay and ELISA in the detection of

[23]

[24]

[25]

[26]

[27]