Research article | Open | Open Peer Review | Published:
Transmission of Specific Genotype Streptomycin Resistant Strains of Mycobacterium tuberculosisin the Tokyo Metropolitan Area in Japan
BMC Infectious Diseasesvolume 9, Article number: 138 (2009)
The Erratum to this article has been published in BMC Infectious Diseases 2010 10:79
From 2003 through to 2004, an outbreak of tuberculosis was identified at a university campus in Yokohama City, located in the southern part of the Tokyo Metropolitan Area (TMA). All Mycobacterium tuberculosis (M. tuberculosis) strains detected with regards to this outbreak turned out to be Streptomycin resistant with matched patterns of 14 IS6110 bands of Restriction Fragment Length Polymorphism (RFLP). The M. tuberculosis bacilli, which had the matched IS6110 band patterns with resistance to Streptomycin to those of bacilli isolated in the outbreak, were also concurrently detected through either the population-based or the hospital-based DNA fingerprinting surveillance of M. tuberculosis either in Shinjuku City or in Kawasaki City respectively.
The aim of the present study is to describe the spread of the specific genotype strains of M. tuberculosis in the TMA as observed in the above incident, and to identify the possible transmission routes of the strains among people living in urban settings in Japan.
We applied Variable Numbers of Tandem Repeats (VNTR) analysis to all M. tuberculosis isolates which were resistant to Streptomycin with a matched IS6110-RFLP band pattern (M-strains). They were isolated either from cases related to the tuberculosis outbreak that happened at a university, or through DNA fingerprinting surveillance of M. tuberculosis both in Shinjuku City and in Kawasaki City. For VNTR analysis, 12MIRU loci, 4ETR loci, seven loci by Supply, four loci by Murase (QUB15, Mtub24, VNTR2372, VNTR3336) were selected.
Out of a total of 664 isolates collected during the study period, 46 isolates (6.9%) were identified as M-strains. There was a tendency that there was a higher proportion of those patients whose isolates belonged to M4-substrains, with four copies of tandem repeat at the ETR-C locus, to have visited some of the internet-cafés in the TMA than those whose isolates belonged to M5-substrains, with five copies at the ETR-C locus, although statistically not significant (38.1% vs. 10.0%, Exact p = 0.150).
Although firm conclusions could not be reached through the present study, it suggested that we have to take into consideration that tuberculosis can be transmitted in congregated facilities like internet cafés where tuberculosis high-risk people and general people share common spaces.
A population-based DNA fingerprinting surveillance of Mycobacterium tuberculosis (M. tuberculosis) using IS6110 Restriction Fragment Length Polymorphism (IS6110-RFLP) analysis has been conducted since 2002 in Shinjuku City, located in the centre of the Tokyo Metropolitan Area (TMA) with approximately 300 000 residents, in order to evaluate the transmission status of M. tuberculosis in an urban setting in Japan . In addition, since 2004 a hospital-based DNA fingerprinting surveillance of M. tuberculosis has been conducted at a public hospital in Kawasaki City, located in the south of the TMA with approximately 1.4 million residents . At the same time, from 2003 through to 2004, an outbreak of tuberculosis (TB) was identified at a university campus in Yokohama City, located just south of Kawasaki City and with a population of approximately 3.7 million . All M. tuberculosis strains detected with regards to this outbreak at the university campus turned out to be Streptomycin resistant and the IS6110-RFLP analysis showed matched patterns with 14 IS6110 bands. In addition M. tuberculosis bacilli, which had the matched IS6110 band patterns with resistance to Streptomycin to those of bacilli isolated in the outbreak, were also concurrently detected through either the population-based or the hospital-based DNA fingerprinting surveillance of M. tuberculosis either in Shinjuku City or in Kawasaki City respectively.
The aim of the present study is to describe the spread of the specific genotype strains of M. tuberculosis in the TMA as observed in the above incident, and to identify the possible transmission routes of the strains among people living in urban settings in Japan.
The design of the study is descriptive, to combine the epidemiological data of bacillary positive tuberculosis patients and DNA fingerprinting analysis results of M. tuberculosis isolated from each of the patients. We applied Variable Numbers of Tandem Repeats (VNTR) analysis to all M. tuberculosis isolates that were resistant to Streptomycin with a matched IS6110-RFLP band pattern (Figure 1). They were isolated either through the contact investigation of TB with regards to the TB outbreak that happened at a university campus in Yokohama City, from cases with possible links with the other outbreaks as suspected by public health centre (PHC) staff in the TMA, or through the population-based or hospital-based DNA fingerprinting surveillance of M. tuberculosis either in Shinjuku City or in Kawasaki City respectively. These two databases, which were independent of each other, were chosen because these were the only ones for us to investigate how these strains spread in the TMA. As the time of this study, there is no database covering the TMA in its entirety. In addition, the database of Kawasaki City, although its coverage among bacillary positive TB patients registered in the City was relatively low, approximately 21% (unpublished data), was chosen because we suspected there existed some link between the M-strains isolated from TB patients at the university campus in Yokohama City and TB bacilli in the Kawasaki City database for the university campus was located very close to the border between the two cities. The study period was set from September 2002 through to January 2008 because the outbreak incident at the university campus in Yokohama City, which prompted us to investigate the transmission situation of the M-strains in the TMA, seemed to have subsided by the end of 2006.
Drug susceptibility tests (DST) were performed at the laboratories where the strains of M. tuberculosis bacilli were initially isolated from the patients with TB. The DST results were collected by the PHC in the routine data collection process from doctors treating the patients. The DST was conducted with the proportion method with 1% Ogawa egg slant, with the critical concentration of drug at 0.2 mcg/ml for Isoniazid, 40 mcg/ml for Rifampicin, 10 mcg/ml for Streptomycin, and 2.5 mcg/ml for Ethambutol. The DST results were verified at the Research Institute of Tuberculosis (RIT), Tokyo, unless the laboratories where the DST was performed had been included in the network of quality assurance by RIT.
Once isolated at laboratories, cultures of M. tuberculosis from each patient were sent to RIT, for DNA fingerprinting analysis. The standardised method of DNA fingerprinting using IS6110 by the restriction fragment length polymorphism analysis was applied . The 27-locus VNTR analysis was performed using primers designed specifically to each of the loci selected. In the current VNTR analysis, 12 loci by Mazars et al. i.e., MIRU 2, 4, 10, 16, 20, 23, 24, 26, 27, 31, 39, 40 ; four loci by Frothingham et al. i.e., ETR A, B, C, F ; supplemented with seven loci recommended by Supply et al. i.e., QUB11b, QUB26, Mtub04, Mtub21, Mtub30, Mtub39, VNTR4156 ; and the other four loci recommended by Murase et al. i.e., QUB15, Mtub24, VNTR2372, VNTR3336  were selected. The last four loci were added in order to increase the discriminatory power in the context of high prevalence of Beijing-genotype M. tuberculosis strains in Japan [9, 10]. The VNTR profile of each of the analyzed M. tuberculosis was described as a set of figures of tandem repeat copy number according to this loci set order as mentioned above. The VNTR analysis was made following the previously described methods using Ex Taq with GC PCR buffer I (Takara Bio) [8, 11]. In brief, the PCR mixture was prepared in a 20 ml volume with 16GC PCR buffer I, 0.5 U Ex Taq, 200 mM each of four dNTPs, 0.5 mM each of the primer set and 10 ng template DNA. PCR was carried out for all loci under the following conditions: initial denaturation at 94 uC for 5 minutes, and then 35 cycles of 94 uC for 30 seconds, 63 uC for 30 seconds and 72 uC for 3 minutes, followed by a final extension at 72 uC for 7 minutes .
The M-strains are defined in this paper as M. tuberculosis strains which have a matched IS6110-RFLP band pattern including those with one band difference, if any (Figure 1); are resistant to Streptomycin; have any of the following VNTR profiles, either M4-substrain with four tandem repeat copies at the ETR C locus, i.e., 223315173533-4243-8844433-4337, or M5-substrain with five tandem repeat copies at the same locus including one copy number difference at the VNTR4156 locus, i.e., either 223315173533-4253-8844433-4337 or 223315173533-4253-8844434-4337 (Table 1)".
The epidemiological linkage, among those from whom M-strain was isolated, was investigated through routine or outbreak-related contact investigations by the PHC staff or by the infection control team at the above-mentioned university. The investigation was made mainly through interviews with the patients by the staff mentioned above. Most of the interviews had been done before the results of the molecular analyses became available.
An internet-café described in this paper is in general a type of cafeteria, being popular among youngsters for passing time and having a rest, which is commonly located in buildings close to a railway or an underground station in urban settings and is usually opened for 24 hours a day. At the place, they can spend some time by using a computer with Internet access. It is also commonly used as a temporary residence for homeless people to stay during night as its price is relatively inexpensive at night .
According to the Ethical Guidelines for Epidemiological Research by the government of Japan, informed consent does not necessarily need to be obtained from research subjects to conduct an observational study using the existing epidemiological data if the study does not use any human biological specimens . Therefore informed consent was not obtained for this study because the present study was an observational study using anonymous and unlinked data of TB patients, whose data sources were existing three databases that had already been collected and analyzed. The informed consent, however, had been routinely verbally obtained regarding the DNA fingerprinting surveillance of M. tuberculosis from each of the bacillary positive tuberculosis patients registered in the Shinjuku PHC by the PHC staff, and the documented consent forms were also routinely obtained from each of the TB patients by the hospital staff in Kawasaki City. The study protocol was approved by the Institutional Review Board Committee of the Research Institute of Tuberculosis with the reference number 21-1.
Chi-square test was applied for the comparison between two sample proportions and Student's t test was applied for the comparison between two sample means. The p-value of 0.05 was set as a statistically significant level.
Out of a total of 664 isolates collected during the study period; 464 isolates from the TB patients registered in Shinjuku City, 176 isolates from those in Kawasaki City, 20 isolates from those related to the outbreak at a university campus in Yokohama City, and the other four isolates from those related to the outbreak other than the university, 57 (57/664 = 8.6%) were detected to have a matched IS6110-RFLP band pattern including one band difference, if any. Among them, 55 isolates (55/57 = 96.5%) showed resistance to Streptomycin. The other two isolates were pan-susceptible to the four major anti-tuberculosis agents including Streptomycin. Among these 55 isolates, nine isolates were excluded from the analysis because their VNTR profiles significantly differed (seven strains showed VNTR profile as 223215173433-4243-8844433-4337, one strain as 223315173533-4244-8844433-4337, and the other as 213315173533-4243-8844433-4337). Consequently 46 isolates of M. tuberculosis (46/664 = 6.9%) were identified as M-strains, showing an identical or a very similar VNTR profile. All of them belonged to Beijing-genotype strains by the spoligotyping analysis (14, data not shown). Among them, only two isolates were resistant to Isoniazid as well (No.9 and 10 in Table 1). There was no significant difference in the proportion of gender distribution between the total 664 patients and the 46 patients with M-strains (78.2% vs. 89.1%, Chi-square = 3.11, p = 0.780), while the mean age of the total patients was significantly higher than that of the patients with M-strains (53.6 years old vs. 35.0 years old, Student's t = 4.527, p = 0.000).
These 46 isolates were further divided into two groups based on the VNTR analysis on the ETR-C locus. 26 M-strains had four copies of tandem repeat at the ETR-C locus (No.21 through 46), designated as M4-substrains, i.e., 223315173533-4243-8844433-4337. Among these M4-substrains, only one isolate showed six copies of tandem repeat at the QUB11b locus (No.42). The other 20 M-strains, which have five copies of tandem repeat at the ETR-C locus, designated as M5-substrains, i.e., either 223315173533-4253-8844433-4337 or 223315173533-4253-8844434-4337, eight isolates showed three copies and 12 isolates showed four copies of tandem repeat at the VNTR4156 locus. 18 isolates (No.1 through No.18) were obtained from the patients related to the outbreak-related contact investigations at the university campus in Yokohama City. It was therefore suspected that all of these 18 patients had been infected from the same index source related to the outbreak at the university (No.1). While the remaining two M5-substrains (No.19, 20), both of them detected from homeless people, were identified through the population-based DNA fingerprinting surveillance of M. tuberculosis in Shinjuku City, the epidemiological linkage of these two patients with other patients with M5-substrains was not confirmed.
Alternatively, among the 26 M4-substrains, two isolates (No.22, 23) were obtained from the patients concerning the above-mentioned outbreak of TB at the university campus, four isolates (No.24, 25, 32, 33) were from the other outbreak-related contacts in the TMA, six isolates (No.21, 29, 31, 39, 45, 46) were through the population-based DNA fingerprinting surveillance of M. tuberculosis in Shinjuku City, and the remaining 14 isolates were through the hospital-based DNA fingerprinting surveillance of M. tuberculosis in Kawasaki City. Epidemiological linkage among the patients with M4-substrains was confirmed only between No.30 and No.31 patients. The No.31 patient, who was 28 years old when he contracted TB, had been working at the front desk of an internet-café in the TMA where the No.30 patient, who was 47 years old homeless, stayed quite a long period, and was coughing whilst he was there. It was therefore confirmed that the No.30 patient was the source of infection to the No.31 patient. Both of the No.22 and No.23 patients were detected through the contact investigation related to the above-mentioned university outbreak of TB in Yokohama City but their TB strains exceptionally belonged to the M4-substrain. The No.22 patient had worked at an internet-café before he contracted TB and had no contact with the index case (No.1). Therefore it was suspected that the infection source of the No.22 patient might have been different from the index case. On the other hand, the infectious source of the No.23 patient was not confirmed to be different from the index case.
There was a tendency where patients whose M. tuberculosis isolates belonged to M4-substrains had higher proportion to have paid a visit to some internet-cafés in the TMA within the past two years prior to the TB registration than those whose M. tuberculosis isolates belonged to M5-substrains, although it is not statistically significant (30.8% vs. 10.0%, Exact p = 0.150).
The M-strains were sub-divided into M4-substrains and M5-substrains according to the copy number difference at the ETR C locus. The copy number at this locus showed relatively low diversity , furthermore its diversity was lower among Beijing-genotype strains than among non-Beijing-genotype strains . The sub-grouping of M-strains, M4-substrains and M5-substrains, according to the difference of tandem repeat copy numbers at the ETR C locus was supported by the epidemiological investigation results as described above.
As for M5-substrains, they were further divided into two strain groups, one was a group with three tandem repeat copies and the other with four copies at the VNTR4156 locus. The Hunter-Gaston Diversity Index at the VNTR4156 locus was relatively low among Beijing-genotype strains isolated in Japan as reported elsewhere . While some patients had very close contact with the index case (No.1) and some of them were acknowledged as members of the definite contact group with the index case, we could not epidemiologically differentiate the patients into two groups according to the copy number at the VNTR4156 locus. It was suggested therefore that one copy number difference at the VNTR4156 locus of M5-substrains had occurred by chance at some point on their transmission process.
Concerning M4-substrains, although it was not confirmed by patient interviews, it was suggested that transmission of the M4-substrains, to some extent, has a possible connection to the utilization of internet-cafés. The data of the population-based DNA fingerprinting surveillance of M. tuberculosis in Shinjuku City, which collected around 85% of all isolates of M. tuberculosis of bacillary positive tuberculosis patients city-wide, suggested that its prevalence might not be high enough to be considered as an endemic strain . It showed that only eight out of 464 isolates (1.7%) belonged to M-strains in Shinjuku City. In addition, the data from a nationwide drug resistance survey of M. tuberculosis in Japan in 2002 suggested that thus far, M-strains have not spread widely and nationwide . It showed that four among 325 randomly selected samples of M. tuberculosis from the survey indicated the matched IS6110-RFLP band pattern as that of M-strain, however none of them was identified as M-strain as described above (unpublished data).
The annual notification rate of bacillary TB patients in Japan has been declining from 14.6 in 1980 to 12.0 per 100 000 population in 2006. In accordance with the decline of the number of TB patients, the prevalence of TB infection was estimated to decline among all age groups especially among young age groups . Meanwhile, TB patients have been geographically concentrated mainly in urban settings in Japan where there are a number of people such as homeless people who have higher chances to be infected with and develop TB. For instance, the annual notification rates of bacillary TB patients per 100 000 population in mega-cities like Tokyo Metropolitan (15.2), Osaka City (35.4), Nagoya City (20.0), and Kawasaki City (16.1) were far above the national average (12.0) in 2006 . Considering the very low level of TB infection among young age groups, it is expected that a TB outbreak could be highly possible among young age groups once a contagious TB patient has a contact with them. For instance, Nakanishi et al. reported outbreak incidents of TB including homeless people and the workers at several public saunas in Shinjuku City where homeless patients often stayed with other homeless workers . They pointed out the possibility of transmission of TB in such places where quite a few homeless people stay for a long time as temporary residence.
We reported relatively high genotype clustering rate among homeless people (unadjusted Odds Ratio: 3.21, 95% Confidence Intervals 1.98–5.21) in Shinjuku City and many genotype clusters were a mixture of homeless and non-homeless patients, which may indicate transmission between these two population segments . In line with this, the present study suggested that specific Streptomycin resistant M4-substrains have been transmitting among both homeless people and contacts with them in internet-cafés. Internet-cafés in addition to public saunas are currently used as temporary residence for a considerable number of homeless people in urban areas in Japan. Furthermore, they are the sites where many youngsters stay concurrently. This implies that not only close contact among family members or friends but also casual contact like those who share a common floor or space in an internet-café could be a focus of transmission of TB. Bifani et al. reported a spread of multiple drug resistant M. tuberculosis widely across several states around New York City, and warned significant risk for TB control in the future . The present study similarly showed a spread of Streptomycin resistant specific genotype tuberculosis strains in urban settings in Japan.
The epidemiological investigation in the context of the present study has been limited to those whose strains sharing the closely related genotype caused the outbreak in a university campus, therefore this limits the ability to draw conclusions clearly as to the specific risk factors for acquisition of this strain. Nevertheless, the results of current molecular as well as epidemiological analysis successfully described the tip of the iceberg regarding the spread of specific Streptomycin resistant M. tuberculosis strains in the urban context in the TMA. The M-strains, which are resistant to Streptomycin, caused an outbreak of TB at a university involving numerous young age group people and also spread significantly in the urban area. All of the 46 TB patients infected with M-strains were not chronic or treatment failure cases but initial cases, which implies that the M-strains successfully transmit among people and induce development of active TB.
Although firm conclusions could not be reached through the present study, it suggested that we have to take into consideration that TB could transmit in congregated facilities like internet-cafés where TB high-risk people and general people share common spaces for relatively long time. The results of this study also imply that we further need to monitor the spread of this strain across the TMA and clarify the transmission routes precisely so that we will be able to control the transmission effectively in an urban context in Japan.
Ohkado A, Nagamine M, Murase Y, Uchimura K, Kaguraoka S, Tatsumi Y, Yamada N, Ohmori M, Maeda S, Maeda H, et al: Molecular epidemiology of Mycobacterium tuberculosis in an urban area in Japan, 2002–2006. Int J Tuberc Lung Dis. 2008, 12: 548-554.
Kinoshita S, Ohmori M, Tsukamoto K, Otuka G, Mashiko M, Fujiu M, Takahashi T, Hoshino H: [Outbreaks of Tuberculosis in Facilities Used by an Unspecified Number of People Near a Train Station – Problems Regarding Tuberculosis in Urban Areas -]. Kekkaku. 2007, 82: 749-757.
Funayama K, Tsujimoto A, Mori M, Yamamoto H, Fujiwara K, Nishimura T, Hasegawa N, Horiguchi I, Mori T, Marui E: [Usefulness of QuantiFERON TB-2G in Contact Investigation of a Tuberculosis Outbreak in a University]. Kekkaku. 2005, 80: 527-534.
Van Embden JDA, Cave MD, Crawford JT, Dale JW, Eisenach KD, Gicquel B, Hermans P, Martin C, McAdam R, Shinnick TM, et al: Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: Recommendations for a standardized methodology. J Clin Microbiol. 1993, 31: 406-409.
Mazars E, Lesjean S, Banuls AL, Gilbert M, Vincent V, Gicquel B, Tibayrenc M, Locht C, Supply P: High-resolution minisatellite-based typing as a portable approach to global analysis of Mycobacterium tuberculosis molecular epidemiology. PNAS. 2001, 98: 1901-1906. 10.1073/pnas.98.4.1901.
Frothingham R, Meeker-O'Connell WA: Genetic diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology. 1998, 144: 1189-1196. 10.1099/00221287-144-5-1189.
Supply P, Allix C, Lesjean S, Cardoso-Oelemann M, Rusch-Gerdes S, Willery E, Savine E, De Haas P, Van Deutekom H, Roring S, et al: Proposal for Standardization of Optimized Mycobacterial Interspersed Repetive Unit – Variable Number Tandem Repeat Typing of Mycobacterium tuberculosis. J Clin Microbiol. 2006, 44: 4498-4510. 10.1128/JCM.01392-06.
Murase Y, Mitarai S, Sugawara I, Kato S, Maeda S: Promising loci of variable numbers of tandem repeats for typing Beijing family Mycobacterium tuberculosis. J Med Microbiol. 2008, 57: 873-880. 10.1099/jmm.0.47564-0.
Yokoyama E, Kishida K, Ichinohe S: Improved Molecular Epidemiological Analysis of Mycobacterium tuberculosis Strains Using Multi-Locus Variable Numbers of Tandem Repeats Typing. Jpn J Infect Dis. 2007, 60: 235-236.
Iwamoto T, Yoshida S, Suzuki K, Tomita M, Fujiyama R, Tanaka N, Kawakami Y, Ito M: Hypervariable Loci that Enhance the Discriminatory Ability of Newly Proposed 15-loci and 24-loci Variable Number Tandem Repeat Typing Method on Mycobacterium tuberculosis Strains Predominated by the Beijing Family. FEMS Microbiol Lett. 2007, 270: 67-74. 10.1111/j.1574-6968.2007.00658.x.
Maeda S, Murase Y, Mitarai S, Sugawara I, Kato S: [Rapid, Simple Genotyping Method by the Variable Numbers of Tandem Repeats (VNTR) for Mycobacterium tuberculosis isolates in Japan – Analytical Procedure of JATA (12) – VNTR -]. Kekkaku. 2008, 83: 673-678.
De Witte Allard: The Cyber Homeless. Open Skies. 2008, 248: 157-161.
Ministry of Education, Culture, Sports, Science and Technology Ministry of Health, Labour and Welfare: Ethical Guidelines for Epidemiological Research. 2002, Tokyo, Ministry of Education, Culture, Sports, Science and Technology/Ministry of Health, Labour and Welfare
Bifani PJ, Mathema B, Kurepina NE, Kreiswirth BN: Global dissemination of the Mycobacterium tuberculosis W-Beijing family strains. TRENDS Microbiol. 2002, 10: 45-52. 10.1016/S0966-842X(01)02277-6.
Tuberculosis Research Committee (Ryoken): Drug-resistant Mycobacterium tuberculosis in Japan: a nationwide survey, 2002. Int J Tuberc Lung Dis. 2007, 11: 1129-1135.
Ohmori M, Ishikawa N, Yoshiyama T, Uchimura K, Aoki M, Mori T: Current epidemiological trend of tuberculosis in Japan. Int J Tuberc Lung Dis. 2002, 6: 415-423.
Japan Anti-Tuberculosis Association: [Statistics of TB 2007]. 2007, Tokyo, Japan Anti-Tuberculosis Association
Nakanishi Y, Oyama Y, Takahashi M, Mori T: [A Molecular Epidemiological Analysis of Several Outbreaks of Tuberculosis in Public Saunas A Problem of Tuberculosis Among Homeless People in the Metropolitan Area]. Japanese Journal of Public Health. 1997, 44: 769-778.
Bifani PJ, Plikaytis BB, Kapur V, Stockbauer K, Pan X, Lutfey ML, Moghazeh SL, Eisner W, Daniel TM, Kaplan MH, et al: Origin and Interstate Spread of a New York City Multi-Resistant Mycobacterium tuberculosis Clone Family. JAMA. 1996, 275: 452-457. 10.1001/jama.275.6.452.
The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2334/9/138/prepub
This study was conducted partly with funding from the research project "Emerging and Re-emerging Infectious Diseases" of the Ministry of Labour, Health, and Welfare, Japan (Principal investigators, Dr. Nobukatsu Ishikawa and Dr. Seiya Kato).
The authors declare that they have no competing interests.
All authors contributed to conduct the present study significantly. AO, MM, NH, GO, MN, HM, and MO collected data and analysed. YM and SM performed DNA fingerprinting analysis. AO wrote the manuscript and YM, NH, GO, KU, MO, NY, SK, TM, and NI edited it. All the 14 authors approved the contents of the manuscript.
An erratum to this article is available at http://dx.doi.org/10.1186/1471-2334-10-79.