Patients, Materials, and Methods

Patients Eighty-five patients attending the outpatient clinic at the National Institute for Infectious Diseases “L. Spallanzani,” Rome, for human papillomavirus (HPV) monitoring were retrospectively selected, on the basis of a known positivity to HCV antibody, from a larger cohort of 350 HIV-positive women. The cervical cytobrush–Ayres spatula sample was used to prepare the Pap smears, immersed in a tube that contained 1 mL of PBS, and gently agitated. The liquid was then aliquoted in 2 tubes and stored at −80°C until use. One aliquot was used for the detection and typing of HPV, and the remaining aliquot was used for the present study. All patients provided signed, informed consent for the use of their samples for research purposes. The collection of personal data and behavioral information was obtained using a questionnaire. The study was authorized by the institutional ethics committee. HCV viremia data were available for 80 of the HCV antibody–positive women: 63 had measurable HCV RNA (median, 5.70 log₁₀ IU/mL; range, 2.87–6.83 IU/mL); 17 women had HCV RNA loads below the lower limit of detection of the quantification assay (i.e., <2.78 log₁₀ IU/mL). For the remaining 5 patients, viremia data were not available

Screening for the presence of HCV RNA in cervical cytobrush samples was performed for all 85 women from this subgroup, and 23 tested positive. The analysis of HCV quasi species in plasma and cytobrush samples was completed for 8 patients. Of these, 2 were infected with HCV genotype 1a, and 6 were infected with HCV genotype 3a

Virological assessment Screening for plasma HCV RNA was performed using qualitative HCV RNA detection by polymerase chain reaction (PCR; limit of detection, 50 IU/mL) (Amplicor; Roche Diagnostic Systems). The viral load in positive plasma samples was determined by quantitative reverse-transcription (RT)–PCR (HCV Amplicor Monitor) with a linear range of 6×10²–5×10⁵ IU/mL. When the specimen volume was sufficient, plasma samples with values exceeding the linear range of the assay were retested after a 1:100 dilution. HCV genotyping was done using a reverse-hybridization line probe assay (INNO-LiPA HCV; Innogenetics) based on the 5′ noncoding region (NCR)

RNA extraction and RT-PCR conditions RNA was extracted from plasma and cervical cytobrush samples using the method of Boom et al. [20]. The presence of inhibitors from the sample matrix was ruled out by testing the presence of β-globin DNA in all extracts from cytobrush samples

RT of total RNA was performed using a reverse transcriptase kit (Invitrogen) in accordance with the manufacturer’s instructions. Amplification of the 5′ NCR from cervical cytobrush samples was done using nested PCR protocols, including TaqGold DNA polymerase (Applied Biosystems). Ten microliters of cDNA was amplified in the first round of amplification using the primers 5′-AACTACTGTCTTCACGCAGAA-3′ (sense; positions 16–36, HCV clone MD2b10-2, GenBank accession number AY232749.1) and 5′-GATGCACGGTCTACGAGACCTC-3′ (antisense; positions 304–283, HCV clone MD2b10-2, GenBank accession number AY232749.1). Amplification conditions were as follows: denaturation for 15 min at 94°C, then 34 cycles of 30 s at 94°C, 30 s at 45°C, and 1 min at 72°C. The amplicon size for the primary reaction was 290 bp. An aliquot (2.5 μL) of the first-round reaction was used for the second round of amplification, using the internal primers 5′-ATGGCGTTAGTATGAGTG-3′ (sense; positions 48–65, HCV clone MD2b10-2, GenBank accession number AY232749.1) and 5′-GCGACCCAACACTACTCGGCT-3′ (antisense; positions 202–223, HCV clone MD2b10-2, GenBank accession number AY232749.1). Amplification conditions were as follows: denaturation for 15 min at 94°C, then 24 cycles of 30 s at 94°C, 30 s at 45°C, and 30 s at 72°C. The nested amplicon size was 175 bp. Amplification of HVR-1 in RNA from plasma and vaginal fluids was performed with nested PCR protocols, using primers that targeted the aminoterminal region of the E2 gene for the different genotypes. PCR conditions for HVR-1 genotypes 1b and those for genotypes 1a, 2, and 3 were different and were derived from the literature [21, 22]

Positive and negative controls, which contained standardized viral RNA extracts and nuclease-free water, respectively, were included in each RT-PCR assay. The amplified products were analyzed by agarose gel electrophoresis and visualized by UV fluorescence after staining with ethidium bromide

HVR-1 cloning, sequencing, and quasi species assessment The HVR-1 region was cloned using nested PCR products. Amplified products were purified using the QIAquick PCR Purification kit (Qiagen), ligated into the pCR4-TOPO vector contained in the TOPO TA Cloning kit (Invitrogen Life Technologies), and transformed in competent Escherichia coli cells using the One Shot TOP10 system (Invitrogen Life Technologies). Plasmid DNA was extracted using the QIAprep Miniprep kit (Qiagen). A total of 160 HVR-1 clones (5–16 clones from each sample) were sequenced on ABI Prism 3100 using the BigDye Terminator cycle sequencing kit (Applied Biosystems)

Quasi species analysis was performed on an 81-bp nucleotide sequence in the E2 region of the HVR-1 gene (positions 1416–1496, according to HCV 1a GenBank accession number AY231582; positions 1489–1569, according to HCV 3a GenBank accession number AF046866). Nucleotide sequences were aligned and translated into amino acid sequences using BioEdit software (Ibis Therapeutics; version 5.0.9)

The HCV nucleotide and amino acid complexity of the HVR-1 region was determined by calculating Shannon entropy using the formula S=-Σ_i(p_i ln p_i), where p _i is the frequency of each sequence in the viral quasi species. Then, the normalized entropy, S _n, was calculated using the formula S_n=S/ln N, where N is the total number of sequences analyzed (in our study, 5–16 clones for each compartment for each patient). The proportions of synonymous substitutions per potential synonymous site (K _s ) and nonsynonymous substitutions per potential nonsynonymous site (K _a ) were calculated using the Nei-Gojobori method with the Jukes-Cantor correction

The mean genetic distance between nucleotide sequences was calculated using a Kimura 2-parameter distance matrix with a transition:transversion ratio of 2.0. To assess the amino acid diversity, the mean genetic distance was calculated using the p-distance model. Phylogenetic trees were constructed using the neighbor-joining method. Trees were rooted by including the nucleotide sequences of HVR-1 relating to the 1a and 3a genotypes (hvr1-1a and hvr1-3a; GenBank accession numbers AY231582 and AF046866, respectively). Bootstrap analysis with 500 replications was performed to assess the significance of the nodes; values >80% were considered to be significant. All of the algorithms used were included in MEGA software (version 2.1; available at: http://www.megasoftware.net)

Statistical analysis The results of continuous variables were expressed as means ± SEs, and comparisons were performed using either paired or unpaired Student’s t test, as indicated. Proportions were evaluated using the χ² test

Results

Determination of the presence of HCV in cytobrush samples from HIV-coinfected women To assess the presence of HCV in cytobrush samples from HCV/HIV-coinfected women, nested PCR specific for the 5′ NCR was performed for samples from 85 patients attending the outpatient clinic at the National Institute for Infectious Diseases “L. Spallanzani,” Rome. Of these, 23 (27%) had HCV sequences present in cytobrush samples. In particular, 22 (96%) of 23 cytobrush-positive women had measurable circulating HCV RNA (median, 5.70 log₁₀ IU/mL; range, 4.38–6.20 log₁₀ IU/mL), and only 1 had a circulating HCV RNA load <2.78 log₁₀ IU/mL. Of the 62 cytobrush-negative women, 41 (66%) had measurable HCV RNA in plasma (median, 5.70 log₁₀ IU/mL; range, 2.87–6.83 log₁₀ IU/mL), and 16 had HCV RNA loads <2.78 log₁₀ IU/mL (table 1). For the remaining 5 cytobrush-negative women, HCV viremia information was lacking. Table 1 shows that the proportion of HCV-positive cytobrush samples was significantly higher (34.9%) in women with measurable HCV viremia than in those with HCV RNA loads <2.78 log₁₀ IU/mL (5.9%) (P=.041). However, median HCV RNA loads did not differ between women with or without HCV RNA present in the cytobrush sample (5.7 log₁₀ IU/mL in both groups)

HVR-1 heterogeneity analysis of plasma and genital-tract HCV To establish the heterogeneity of HCV in the 2 body compartments, a subset of 8 women with HCV-positive cytobrush samples and plasma viral loads >2.78 log₁₀ IU/mL were selected, and their clinical features are described in table 2. Genetic heterogeneity of the HVR-1 of the E2 gene was investigated using a cloning and sequencing approach. A total of 98 plasma and 79 cytobrush clones were analyzed. The complexity, diversity, and proportion of K _s and K _a for each subject and for each compartment, as well as median values, are shown in table 3. The complexity, which reflects the number of different variants present in the analyzed compartment, was significantly lower in cytobrush samples than in plasma samples at both the nucleotide and amino acid level. Diversity, K _s, K _a, and the K _a:K _s ratio also tended to be lower in cytobrush samples than in plasma samples. This tendency was particularly evident in patients B, C, and D and was reflected in mean values, although the differences were not statistically significant (table 3)

HVR-1 quasi species compartmentalization in plasma and cytobrush samples With the aim of evaluating compartmentalized viral replication of HCV from cytobrush samples, compared with that from plasma samples, phylogenetic trees were constructed on the basis of nucleotide sequences. Figure 1 shows the phylogenetic tree of all clones from the 8 patients. As expected, the clone sequences from the different patients segregated independently from each other, indicating the absence of cross-contamination between the samples and then remaining associated with their respective genotype (1a and 3a). It should be noted that clones from patients G and F formed 2 very homogeneous, independent clusters. Information about the duration of HCV infection was available only for patient G (⩾4 years). The high homogeneity of clones from patients G and F was consistent with the corresponding values of heterogeneity parameters (table 3). This table shows that, in each patient, cytobrush and plasma samples varied in the composition of viral quasi species, with some clones represented in both and others represented in only 1 of the 2 body sites. In particular, 7 of 8 patients had clones in plasma samples that were not present in cytobrush samples, whereas 5 patients had clones in cytobrush samples that were not present in plasma samples, although bootstrap values at nodes showed significant clusters segregating only for plasma clones. Notably, 1 patient (patient F) harbored a homogeneous viral quasi species in both compartments. When we considered the amino acid sequences of the clones, the phylogenetic tree was virtually superimposable on that for nucleotide sequences (not shown)

The individual phylogenetic trees of the 5 patients with unique variants in cytobrush samples, together with the corresponding amino acid alignments, are shown in figure 2. As can be seen, unique variants could also be identified for the 2 body sites at the amino acid level, which supports the hypothesis that different viral quasi species were actually present in plasma and cytobrush samples. For example, in patient C, a very unusual amino acid substitution was detected at position 23 (G→T) that derived from a double-nucleotide replacement at the corresponding codon. However, it was observed in only 1 clone, and we could not rule out that this result derived from a certain level of infidelity by the polymerase used in the PCR

Discussion

In the present study, we evaluated the prevalence of HCV genome sequences in cervical cytobrush samples from HCV/HIV-coinfected women, and we analyzed the HVR-1 quasi species present in this body compartment, compared with those in plasma, to explore the possibility that female genital tract can harbor viral quasi species different from those in plasma

Approximately one-quarter of these women were found to harbor HCV in their genital tract. This prevalence was very similar to that recently reported in HCV/HIV-coinfected women [17] and to that from a previous report of HIV-negative women [18]

The frequency of HCV in cytobrush samples was significantly higher in women with measurable HCV RNA loads, but it was also detected in a cytobrush sample from a woman with an HCV RNA load below the limit of quantification of our assay, which indicates the possibility that women who are virtually without viremia may harbor HCV in other body compartments. These results are not unanticipated, given that previous reports have shown that, in patients negative for HCV RNA, HCV may be present in other body sites, such as the gastric mucosa [23]

We were unable to measure the HCV RNA loads in cytobrush samples using commercial methods, because of the scarcity of available material. However, the HCV genome was detected by an in-house method targeting the 5′ NCR, and the HVR-1 region could be successfully amplified as well. The availability of HVR-1 amplicons prompted the use of a cloning and sequencing approach based on a coding region, so that the corresponding amino acid sequences could be deduced. This region is of particular interest when studying HCV quasi species, because it is under the selective pressure of host immune response and is the main focus of interest when analyzing viral dynamics and evolution. In addition, we could determine not only the complexity and diversity but also K _a and K _s, whose ratio is informative with respect to the extent of immunologic pressure. In fact, it is generally assumed that a K _a:K _s ratio >1 indicates the existence of immunological pressure on the evolution of viral quasi species

Our results indicate that the heterogeneity of viral quasi species was indeed different in plasma and cytobrush samples at both the nucleotide and the amino acid levels. The mean complexity in cytobrush samples was significantly lower (table 3), and a similar trend was also observed for diversity, K _a, K _s, and K _a:K _s values, although the differences did not reach statistical significance. When considering the phylogenetic trees, viral variants were identified in the background of several common clones at both the nucleotide and amino acid level that were unique of either body compartment

Our findings extend the results of Nowicki et al. [17], who recently showed, on the basis of the sequence analysis of the 5′ NCR region, compartmentalization of HCV in the blood and genital tract of HCV/HIV-coinfected women, which suggests the possible localized replication of HCV in the genital tract. Our data show, to our knowledge for the first time, a different composition of viral quasi species based on the coding region that includes epitopes under immunologic pressure. For both fluids, the K _a:K _s ratio was >1, which confirms that the evolution of this region is driven by immunologic pressure in both compartments. In addition, because K _a:K _s values in plasma were higher than those in the genital tract, the extent of such pressure seemed to be higher in the circulation, which is possibly due to a higher exposure of HCV to the effector arms of the immune response. In fact, if HCV actually replicates in the genital mucosa, it is reasonable that immune pressure in this area is lower than that in circulation. However, further experiments are necessary to substantiate this hypothesis—possibly those based on in situ hybridization or RT-PCR of biopsy samples from the endocervical lining, which would be more suitable for obtaining information about local reservoirs of HCV in the genital tract. In this respect, the quasi species analysis of HCV present in body sites other than blood in the uncommon persons without detectable HCV RNA loads will be highly informative about the extent of immunologic pressure in local compartments and will be the focus of further studies

In addition, our data show that the qualitative composition of viral quasi species in individual patients is different in plasma and cytobrush samples, given that unique sequences were identified in the 2 body sites, with monophyletic segregating clusters observed only in plasma. In a recent study of the HVR-1 region in seminal fluid from HCV/HIV-coinfected patients, Briat et al. [24] described the presence of unique variants in blood and semen, which suggests that HCV that is shed into the male genital tract originates from a preferential transfer of viral particles from blood into genital secretions

On the basis of the our data, it is not possible to establish whether the different composition of viral quasi species in the genital tract of HCV/HIV-coinfected women could actually reflect compartmentalized viral replication or the existence of selective mechanism(s) controlling the transfer of plasma quasi species through the barrier separating the 2 compartments. It should be noted that the cytobrush samples for the quasi species analysis were evidently not contaminated by blood, so it is reasonable to assume that HCV in cytobrush samples did not derive merely from blood contamination of the specimen. Inflammation was present in 4 women (table 2), so it is possible that circulating mononuclear cells harboring HCV infection can cross the anatomical barrier between plasma and the female genital tract and, thus, act as HCV carriers. However, in 2 women with inflammation (patients B and C), cytobrush-specific variants were observed (figure 2). However, there is also the possibility that localized replication of HCV occurs in cell reservoirs resident in the genital mucosa, leading to an independent evolution of HCV quasi species. In fact, for HCV present in cerebrospinal fluid (CSF), a clear segregation of 5′ untranslated region variants has been observed, which is suggestive of local replication [25]. Moreover, on the basis of similarity between HCV variants from CSF and peripheral blood mononuclear cells (PBMCs), it has been postulated that PBMCs may carry the virus into the brain and that, thereafter, resident microglial cells maintain viral replication

In conclusion, independent of the mechanisms involved in the presence of HCV in the female genital mucosa, our data confirm and extend previous evidence that viral quasi species in this specific body compartment contain components different from those that circulate. This finding may have pathogenetic implications with respect to the possible perinatal and sexual transmission of distinct viral variants