Detection of sexually transmitted infection and human papillomavirus in negative cytology by multiplex-PCR

Sexually transmitted diseases (STDs) are fairly common in people of reproductive age; more than 50% of new STD infections occur in adolescents. However, only 50% to 80% of reportable STDs that occur in the United States are actually reported [1]. Testing for STDs can result in negative cytology, that is, laboratory test results that revealed reactive cellular change (RCC) or atypical squamous cells of undetermined significance (ASCUS). RCC or ASCUS is routinely screened for, but they are not associated with neoplasms or preneoplastic stages known as squamous intraepithelial lesions (SILs). RCCs are cytological alterations related to inflammation or irritation but, whether subtle or marked, are typically not associated with human papillomavirus (HPV) infection. Common features of RCCs include the formation of small perinuclear halos, mild nuclear enlargement or nuclear degeneration, hyperchromasia, cytoplasmic pseudokeratinization, cytoplasmic vacuolization, and notably a background of inflammation. The abnormal cytology of ASCUS can usually be distinguished from RCC, but does not meet all criteria of HPV-infected koilocytes [2].

With the exception of a few organisms that cause sexually transmitted infections (STIs) such as trichomoniasis or candidiasis, most inflammatory agents are only rarely recognized under the microscope and usually produce nonspecific cytopathic effects. For example, bacterial vaginosis (BV) is a leading cause of abnormal vaginal discharge as a sequel of gynecologic conditions, including endometritis, pelvic inflammatory disease (PID), post-surgical abortion, post-hysterectomy infection, and cervical intraepithelial neoplasia [3]. However, few attempts have been made to identify the etiologic organisms of BV, and diagnoses have relied only on the presence of common "clue cells" detected in cervical swabs [4, 5]. The prime manifestation of BV constitutes a massive micro-ecologic alteration of vaginal flora characterized by decreased Lactobacillus spp., logarithmically-increased Gardnerella vaginalis, and the presence of potentially pathogenic bacteria, including Bacteroides, Peptococcus, Mobiliuncus, Ureaplasma urealyticum, and Mycoplasma spp. Mycoplasma spp. and U. urealyticum are most frequently detected in the vaginal tract [6], and are strongly associated with infertility, intra-amniotic infection, postpartum infection, and pelvic inflammatory disease (PID) [1–5]. The multiple strains of Mycoplasma are the only cell wall-deficient, membrane-bound organisms capable of independent self-replication: they include Mycoplasma hominis, M. genitalium, M. primatum, M. spermatophilum, and M. penetrans [7, 8].

Despite the fact that many pathogens are not detected by Gram stain, Gram-stained vaginal secretions are still accepted as the diagnostic gold-standard, rather than the culture of secretions. With the advent of advanced techniques such as those based on liquid-based cytology [9–13], it is now feasible to accurately and efficiently identify microorganisms in cervicovaginal swab specimens of asymptomatic patients with multiplex PCR or DNA chip. Although detection of infectious organisms is not the primary goal of cervical cytology, it can be an important component of screening in asymptomatic patients or subtle disease. We aimed to estimate the incidence of STIs in healthy women of reproductive age with negative cytologic findings by multiplex PCR and find any correlation of HPV infection with specific pathogens.

High-throughput technology including hybrid capture 2 (HC2) and DNA microarray has been currently adopted to detect HPV [14]. HC2 is a relatively easy method with high screening power, but it is unable to genotype or excludes low-risk HPV types. DNA microarray can identify specific genotypes, and can also detect co-infection in a single sample, but is labor-intensive. Despite widely available high-throughput technology for the detection of microbes or HPV, few studies of STDs and HPV with high fidelity data exist due to lack of a gold standard. We therefore designed multiplex PCR assays to compare with the available methods of HC2 and HPV DNA microarray, and analyzed discordant results by direct sequencing.

A total of 235 married and sexually active women (mean age of 37.9 y; ages range 21 to 48 years) were recruited by the Department of cytopathology of Yonsei University College of Medicine for this study. Specimens collected from cervical liquid-based cytology medium were negative for intraepithelial lesions. Participants had no clinical STD symptoms that required a clinical visit, no history of STDs, and no abnormal Papanicolau tests (Pap smear) results. We also excluded patients with SIL or carcinoma, postmenopausal women, or those with previous operative or therapeutic history related to gynecologic disease. The study protocol was approved by the local institutional review board, and written informed consent was obtained from all participants.

Reports of the incidence of STIs in asymptomatic women are not currently available, because of few attempts to recognize STI species despite advanced techniques based on liquid-based cytology [9–13].

We found that the overwhelming majority of patients with negative cytologic findings were positive for at least one STI (97.9%); many were positive for multiple infectious agents (84.9%). These results are surprising in that we did not observe high rates of infection with microbes that are typically widespread, such as N. gonorrhea and C.trachomatis [10, 13]. In more than 25% of the samples, we detected microbes that cause asymptomatic, dormant non-gonococcal and non-chlamydial infections, such as G. vaginalis, U. urealyticum, Group β streptococcus, and M. hominis. Microbes detected in less than 2% of the specimens were CT-1, NG, and C. parapsilosis (CP) [13].

MH was often detected in the same sample with other microbes, such as UU, M. mulieris (MM), M. curtisii (MC), and CT. However, co-infection of multiple mycoplasma strains has rarely been reported [15].

Because HPV and STIs rely on the same route of infection, this increased risk of HPV infection in patients with STIs was expected. Studies of women visiting STD clinics have reported a greater prevalence of HPV infection in this group than in the general population [16, 17]. In one study, approximately half of the women evaluated at an STD clinic were infected with genital HPV, including HPV types-16, 83, 56, 52, and 59 (HR group) and 66 (LR group) [16]. In the present study of 235 asymptomatic women who had negative cytology results, we found that approximately 20% were infected with HPV. Using a genotyping microarray or grouping multiplex PCR, the most common HR types we detected were HPV types-16, 18, 51, 58, and 33, and the most common LR type was HPV-70. Epidemiologic data concerning HPV-types differ based on the method of detection used and the target samples. Multiple infections with more than two different HPV genotypes were found in 22% to 26% of the specimens, and HPV-16 was most detected (50%).

Specimens infected with LR-HPV were frequently co-infected with CT, whereas those with HR-HPV were coinfected with GBS. The fact that CT associates with LR-HPV rather than HR-HPV suggests that CT is less likely to contribute to cervical intraepithelial neoplasia or squamous cell carcinoma. Previous reports suggested that coinfection with CT and HPV is common in ASCUS and could contribute to the development of intraepithelial lesions [18–24]. It is possible that the modulation of cervical immune responses in response to CT may influence the clearance of HPV lesions [18, 19], and suggesting that CT infection with HPV persistence is not a coincidence [20]. Likewise, we are the first to document the prevalence of coinfection of GBS and HR-HPV in negative cytology (including RCC and ASCUS). However, it remains uncertain that GBS may be an independent factor or cofactor for HPV in the development of cervical intraepithelial neoplasm. For identification of its cofactor, novel immune evasion strategies and the analysis of their functions in the context of viral and bacterial infections should be required in further studies.

We compared high-throughput HPV detection methods in the present study, and found that discrepancies with direct sequencing were observed only in 10.6%. The microarray (DNA chip) method also resulted in a significant number of false negative results. Analysis of 25discordant samples showed that multiplex PCR-HPV method exhibited the highest sensitivity, positive predictive value, and accuracy. However, each HPV detection method covered different genotypes, and different typing or grouping methods were used to present the data. For example, the relatively low sensitivity of HC2 may be due to its inability to detect LR genotypes, and its lower positive predictive values and accuracy may be due to limitations in detecting individual genotypes, and thus, multiple infections.

In conclusion, asymptomatic women of reproductive age with negative cytologic findings had a high frequency of infection with GV, UU, GBS, and MH. Most HPV infections were not associated with any specific STI pathogen; however, LR-HPV infection correlated with CT infection, and HR-HPV with GBS infection. In terms of sensitivity and prediction value, multiplex PCR yielded data most similar to that of DNA microarray.

HSS and SMN summarized, analyzed data and equally drafted entire manuscript. ARP performed statistical analysis of given data. YNL & KSK participated in cytology screening and sample preparation with performance of HC2/DNA microarray. JKK & HJC performed multiplex PCR. NHC designed and participated in DNA direct sequencing, and revised manuscript. All authors read and approved the final manuscript.

Two hundred thirty-eight asymptomatic women of reproductive age with negative cytology had high frequency of sexually-transmitted infection (96.6%), and 24.4% had HPV.