HCV RNA in peripheral blood cell subsets in HCV–HIV coinfected patients at the end of PegIFN/RBV treatment is associated with virologic relapse

Authors


Alejandro Vallejo, Laboratory of Molecular Virology, Hospital Universitario Virgen del Rocio, Manuel Siurot s/n. 41013 Seville, Spain. E-mail: avallejo@us.es

Abstract

Summary.  Extrahepatic replication may have important implications for the treatment of hepatitis C virus (HCV). Our aim was to analyse the association between the presence of positive/negative strand HCV RNA in different peripheral blood cell subsets at the end of PegIFN/RBV treatment, and treatment response in HIV-coinfected patients. Thirty-four HCV–HIV coinfected patients who concluded 48 weeks of PegIFN/RBV treatment were included in the present study. Positive/negative strand HCV RNA was detected by amplification of the 5′ untranslated region (5′ UTR) using high-temperature RT-PCR in immunomagnetic-isolated cell subsets. Twenty-three patients (67.6%) had sustained virologic response (SVR) while 11 patients (32.4%) relapsed. The frequency of positive/negative strand HCV RNA in any cell subsets was significantly lower in patients with SVR (8.6%) compared to relapsers (63.6%) (= 0.002). Baseline HCV viral load was statistically higher among patients who relapsed (= 0.008), while patients with SVR had very early virologic response more frequently (P = 0.003). Multivariate analysis showed, among these three variables, that only the presence of positive/negative strand HCV RNA was independently associated with relapse [= 0.024; OR 14 (14–137)]. In conclusion, the presence of positive/negative strand HCV RNA at the end of treatment is associated with relapse among HCV–HIV coinfected patients and might have important implications in the clinical practice.

Abbreviations:
HAART

highly active antiretroviral treatment

HCV

hepatitis C virus

IQR

interquartile range

PBMCs

peripheral blood mononuclear cells

PeglFN/RBV

pegylated-interferon-alfa and ribavirin

SVR

sustained virologic response

Introduction

With the introduction of highly active antiretroviral treatment (HAART), hepatitis C virus (HCV)–HIV coinfection emerged as the main cause of comorbidity [1], frequently associated with a faster progression of liver disease [2–4]. Hence, anti-HCV treatment, currently based on pegylated-interferon-alfa and ribavirin (PegIFN/RBV), in coinfected patients is of particular importance [5–10]. It has been reported an overall rate of 40% of sustained virologic response (SVR) among HCV–HIV coinfected patients [6], lower than the reported for HCV-monoinfected patients [6]. Different factors have been associated with this lower treatment response, such as higher serum HCV viral load at baseline, co-treatment with antiretroviral therapy, and higher degree of immune deterioration [3,11].

HCV is a positive strand RNA virus whose replication involves the synthesis of a negative strand RNA molecule. The virus mainly replicates in hepatocytes but it can also replicate in extrahepatic reservoirs, as peripheral blood mononuclear cells (PBMCs), particularly in patients coinfected with HIV [12–15]. Several studies have reported the presence of HCV RNA in PBMCs [13–16], dendritic cells, [17] monocytes/macrophages [18,19], and B lymphocytes [20], although these data have to be taken with caution because most of these studies utilized different technical approaches [21]. The detection of HCV RNA in such extrahepatic reservoirs might have important implications for transmission, disease progression, and effective treatment [4,11].

While different studies among HCV-monoinfected patients have reported the presence of HCV RNA in PBMCs in patients who had SVR [19,22], others have reported HCV RNA clearance in PBMCs as a predictor of response to antiviral therapy [23]. Nevertheless, the association between HCV RNA in isolated peripheral blood cell subsets at the end of anti-HCV treatment and the SVR has not been previously studied in HCV–HIV coinfected patients.

Hence, the objective of our study was to analyse the association between the presence of positive/negative strand HCV RNA in different isolated cell subsets at the end of treatment (48 weeks) and SVR in HCV–HIV coinfected patients.

Methods

Patients

From May 2004 to November 2005, 146 HCV–HIV co-infected Caucasian patients at Virgen del Rocio University Hospital, initiated anti-HCV therapy based on PegIFN-α-2a (180 μg/week) and RBV (1000–1200 mg/day, depending on the weight) for 48 weeks. For the purpose of the present study, only those who (1) concluded 48 weeks of treatment, (2) had undetectable plasma HCV RNA levels at week 48 of treatment, (3) had peripheral blood sample drawn at week 48 of treatment, and (4) virologic data 24 weeks after cessation of treatment, were eligible.

Thirty-four patients who fulfilled these criteria were studied. Written informed consent was obtained from all patients and the Ethical Committee of the Hospital approved the study.

Clinical, immunological and virological parameters were performed, analysed and recorded at baseline and during the PegIFN/RBV therapy on a 4-week basis until the end of therapy (week 48) and during the follow up after cessation (week 72).

Laboratory determinations

CD4 count was determined in fresh samples by flow cytometry. Plasma HCV RNA was detected by a commercially available PCR procedure (COBAS Amplicor; Roche Diagnostics, Barcelona, Spain) with a detection limit of 15 IU/mL. HCV genotype was determined using a reverse-hybridization assay (InnoLiPA HCV II; Innogenetics, Barcelona, Spain). Plasma HIV-1 RNA was measured by quantitative PCR assay (HIV Monitor™ Test Kit; Roche Molecular System, Hoffman-La Roche, Basel, Switzerland), according to the manufacturer’s instructions. This assay has a lower detection limit of 50 HIV-1 RNA copies/mL.

Isolation of peripheral blood cell subsets

Peripheral blood mononuclear cells were isolated from heparinized blood samples drawn at week 48 of treatment by Ficoll-Hypaque gradient centrifugation (Pharmacia Biotech, Uppsala, Sweden), and cryopreserved in liquid nitrogen with 10% DMSO. For this study, cell subsets were isolated by immunomagnetic separation technique (Dynabeads; Dynal, Paisley, UK) using monoclonal antibodies against specific surface antigens, including anti-CD19, anti-CD4 and anti-CD8 antibodies for positive cell isolation (Dynal, positive isolation kits), as well as an antibody mix with monoclonal antibodies toward CD3, CD14, CD36, CDw123, HLA Class II DR/DP and glycophorin A for negatively isolate NK cells (Dynal NK cell negative isolation kit), according to the manufacturer’s instructions. Each cell fraction contained at least 1–5 × 106 cells with purity above 98%. In addition, PBMCs were also available from a subgroup of 22 patients at initiation (baseline) of therapy and handled as above.

Specificity and sensitivity of the strand specific RT-PCR assay

Synthetic positive and negative strands of HCV were generated by in vitro transcription of pGEM-3Zf plasmid, which contains the complete cloned 5′UTR of the HCV genome generated by PCR using forward primer UTR-S-5′ TCGGAATTCGCGACACTCCACCATGAATC (position 21-40 HCV-1a H77) and reverse primer UTR-R-5′ GCGGATCCAGTAAACTCCACCAACGATCTG (position 459-436 HCV-1a H77), including EcoR1 and BamH1 restriction sites at 5′, respectively (italic). After plasmid linearization using EcoR1 or BamH1 (Promega, Madrid, Spain), the positive and negative strands were synthesized by transcription using T7 RNA polymerase (5′ to the DNA insert) or SP6 RNA polymerase (3′ to the DNA insert), respectively. The remaining DNA plasmid was then digested with DNase I (1,5 U/μg DNA, Promega) for 1 hour at 37 ºC, and the RNA was then extracted using QIAamp Viral RNA kit (Qiagen Diagnostics, Barcelona, Spain). The absence of residual DNA was confirmed by PCR. Concentrations of specific RNA strands were determined by spectophotometry (Nanodrop, ND-1000; Wilmington, DE, USA). Ten-fold serial dilutions of the specific RNA were made to assess the sensitivity and specificity of the RT-PCR. In addition, 0.5 μg RNA from an HCV-noninfected individual was added to each dilution.

Furthermore, ten HCV-positive plasma specimens were filtered (0.2 μm membrane filter; Millipore, Madrid, Spain) and HCV RNA was extracted using QIAamp Viral RNA kit (Qiagen Diagnostics). HCV RNA was tested for both strand specific RT-PCR assays.

Detection of strand-specific HCV RNA in isolated cell subsets and total PBMCs

Total RNA was extracted from total PBMCs and from each isolated cell subset at the end of the anti-HCV treatment (week 48) using QIAmp RNeasy (Qiagen Diagnostics) following the manufacturer′s instructions, and RNA concentration was determined by spectrophotometry. 5′UTR region of HCV RNA was amplified by high-temperature RT-PCR. Strand-specific cDNA was synthesized from 0.5 μg of RNA at 63 ºC for 30 min using a thermostable DNA polymerase (Transcriptor Reverse Transcriptase, Roche Molecular Biochemicals, Mannheim, Germany), and one of the following primers NR5 5′-TGCTCATGGTGCACGGTCTACGAGA (positions 348-323 HCV-1a H77) for the synthesis of the cDNA of positive polarity, and NF5 5′-GTGAGGAACTACTGTCTTCACGCA (positions 47-70 HCV-1a H77) for the synthesis of the cDNA of negative polarity. This high-temperature RT reaction prevents false priming of the incorrect strand. Once cDNA was synthesized, RT activity was inactivated by heating the reactions at 95 ºC for 30 min. For total PBMCs, both primers NR5 and NF5 were used together for cDNA synthesis.

For each cDNA, a first round of PCR was conducted during 35 cycles consisting of 30 s at 94 ºC, 30 s at 60 ºC and 40 s at 72 ºC. The reactions were carried out using 0.5 U Taq polymerase (Roche Molecular Biochemicals, Mannheim, Germany) in 50 μL buffer containing 10 mm Tris–HCl pH 8.8, 50 mm MgCl2, 0.1 Triton X-100, 33 μm each dNTP, and 0.5 μm each NR5 and NF5 primers.

The following nested-PCR was performed using 2 μL of the first PCR by two different methods: conventional PCR and real-time PCR. Conventional PCR was performed as above with forward primer 7700 5′TGGTCTGCGGAACCGGTGAGTACA (positions 143-167 HCV-1a H77) and reverse primer NR4 5′GGGTATCAGGCAGTACCACAAGG (positions 301-279 HCV-1a H77). Real-time PCR was performed in a mixture containing Brillant QPCR master mix (Stratagene, La Jolla, CA, USA), 4 mmol/L MgCl2, 5 pmol/L of primers SONY11 5′-GCGTCTAGCCATGGCGTTAGT (positions75-95 HCV-1a H77) and NAT12 5′-ACCCGGTCGTCCTGGCAATTC (positions 191-171 HCV-1a H77), and 15 pmol/mL of HCV-specific TaqMan-based probe FAM-5′TGAGTGTCGTGCAGCCTCCAGGACCCC-TAMRA (positions 97-123). The reaction consisted of 40 cycles of 15 s at 95 ºC, 20 s at 60 ºC and, 40 s at 72 ºC. All RT-PCR runs for every sample were performed in triplicate and an amplification of a housekeeping gene (glyceraldehyde 3 phosphate dehydrogenase GAPDH; sense 5′-ACCACAGTCCATGCCATCAC, and antisense 5′-TCCACCACCCTGTTGCTGTA) was included in parallel to assess RNA quality.

All samples were analysed blinded to clinical parameters and were subjected to three different cell isolation processes. Also, 0.5 μg RNA from PBMCs from five different HCV-non-infected subjects and from five different HCV-infected subjects (previously tested as PCR positive) were included in each run as control specimens.

Serum anti-IFN-α level quantification

Quantification of the levels of anti-IFN-α antibodies was determined in triplicate in serum samples at baseline and at weeks 12, 24, 36 and 48 during the treatment using high sensitivity colorimetric enzyme-linked immunosorbent assay (Human anti-IFN-α ELISA, PBL Biomedical Laboratories, New Brunswick, NJ, USA) with lower detection threshold of 1.38 ng/mL.

Statistical analysis

Continuous variables are shown as median [interquartile range (IQR)], and categorical variables as number of cases (percentage). Mann–Whitney U-test was used to analyse continuous variables between groups, while categorical variables were compared by chi-square test. All variables with univariate association level <0.1 were analysed with the multivariate logistic regression test to analyse the independent variables associated with SVR. The statistical analysis was performed using the Statistical Package for the Social Sciences software (spss 14.0; SPSS Inc., Chicago, IL, USA).

Results

Baseline characteristics of the 34 patients of this study are shown in Table 1. Of them, 23 patients (67.6%) had SVR, defined as undetectable plasma HCV RNA six months after completion of therapy, and subsequently for truly defining SVR. Among them, 11 were genotype 1, five were genotype 3, and seven were genotype 4. The other 11 patients (32.4%) relapsed, including eight genotype 1, two genotype 3, and one genotype 4. Among those who had SVR, three patients did not receive antiretroviral treatment, while five patients among those who relapsed did not received antiretroviral treatment. Only eight patients who had SVR had greater than six log plasma HCV RNA at baseline (34.7%), while eight patients who relapsed had greater than six log plasma HCV RNA at baseline (72.7%).

Table 1.   Characteristics of the 34 HCV–HIV coinfected patients at the beginning of PegIFN/RBV therapy
  1. *Previously treated with PegIFN-α-2a plus RBV.

  2. Mean viral load from the eight patients who did not receive antiretroviral treatment.

Age (years)43 (40–46)
Male gender28 (82.4%)
Injecting drug users25 (73%)
CD4+ count (cells/mm3)496 (366–695)
HCV infection
 Previously treated*4 (11.7%)
 HCV viral load (log10 IU/mL)5.9 (4.9–6.4)
 Genotype 119 (55.8%)
 Genotype 37 (20.62%)
 Genotype 48 (23.6%)
HIV infection
 Patients on HAART26 (76.5%)
 Time on HAART (years)6.9 (5.5–7.4)
 HIV viral load (log10 copies/mL)†3.2 (2.5–4.2)

The analysis of the specificity and sensitivity of our strand specific RT-PCR assay showed that the positive and the negative strand assay detected the HCV RNA positive and negative strand, respectively, at a limit of 1 fg. These assays did not detect HCV RNA negative and positive strand, respectively, using 0.1 μg. Thus, these assays completely discriminate between positive and negative HCV strands (Fig. 1). On the other hand, positive strand RT-PCR assay detected HCV positive strand from all the plasma specimens tested, while no negative HCV strand was detected using the negative strand assay (not shown).

Figure 1.

 Specificity and sensitivity of the positive and negative strand specific RT-PCR assays using different quantities of synthetic HCV-RNA mixed with 0.1 μg RNA from HCV-noninfected individual. WM, weight marker, 100 bp ladder; Neg, negative control; NS, negative strand; PS, positive strand.

The presence/absence of strand specific HCV RNA in isolated cell subsets and total PBMCs at the end of therapy (week 48) and before therapy (baseline) is summarized in Table 2. The same results were obtained either by conventional PCR or real-time PCR. At week 48, nine patients (26.4%) had either positive or negative strand specific HCV RNA in isolated cell subsets. Of these nine patients, two (22.2%) had SVR and the rest relapsed (77.7%). Positive strand HCV RNA was present in two patients with SVR at least in one cell subset (2/23, 8.6%), while it was present in seven patients (7/11, 63.6%) who relapsed (= 0.002). Negative strand HCV RNA was only found in one patient with SVR (4.3%), while it was found in five patients who relapsed (45%) (P = 0.01). None of the patients with SVR had HCV RNA in total PBMCs, while it was only found in three patients who relapsed (18.2%).

Table 2.   Positive and negative strand HCV RNA in cell subsets and PBMCs of HCV–HIV-coinfected patients at the end of anti-HCV treatment according to their treatment response
Week 48
 CodeGenotypeTotal PBMCPositive strandNegative strand
BCD4CD8NKBCD4CD8NK
  1. SVR, sustained virologic response.

 23A
SVR (n=23)1784C / 4D
88891A
91013A
2011A / 1B
2984C / 4D
3264C / 4D
3631A / 1B
1274C / 4D
513A
1301A
651A
3063A
8011A
8021A
8034C / D
8041A / B
8051B
8064C / D
8073A
8084C / D
2861B+
721B+++
 Total of patients0 (0%)0 (0%)1 (4.3%)1 (4.3%)1 (4.3%)0 (0%)1 (4.3%)0 (0%)0 (0%)
Relapse (n=11)531A
3363A
2071B
8094C / D
2301B++
2511A / 1B+++
2491B+++++++
353A+++++
1771A / 1B++++++
8101 A++++
8111A++++
 Total of patients3 (27%)5 (45%)3 (27%)6 (54%)6 (54%)2 (18%)0 (0%)3 (27%)3 (27%)
Baseline (total of patients)
SVR (n=14) 14 (100%)13 (92%)11 (78%)11 (78%)10 (71%)7 (50%)7 (50%)4 (28%)5 (35%)
Relapse (n=8) 8 (100%)7 (87%)8 (100%)7 (87%)7 (87%)6 (75%)4 (50%)5 (62%)6 75%)

At baseline, all the 22 patients with available samples had detectable HCV RNA in the total PBMCs, while positive strand HCV RNA was present in at least one cell subset in a high proportion of the patients (from 71% to 100%) regardless treatment response, as shown in Table 2. The negative strand HCV RNA was found from 28% to 50% of the patients who had SVR, and from 50% to 75% among patients who relapsed, although these differences were not statistically significant (= 0.9). The proportion of patients with negative strand HCV RNA in all the four cell subsets was significantly lower among patients with SVR (only one out of 14 patients) compared with patients who relapsed (three out of eight patients) (= 0.02).

Differences between patients with SVR and those who relapsed are shown in Table 3. Baseline HCV RNA viral load was statistically lower in patients with SVR (P = 0.008), very early virologic response was statistically higher in patients with SVR (= 0.003), and the presence of strand specific HCV RNA at week 48 of treatment was statistically lower in patients with SVR (P = 0.002). Multivariate logistic regression analysis of these three variables (only those with < 0.1 in the univariate analysis) showed that only the presence of strand specific HCV RNA at the end of treatment was independently associated to SVR (P = 0.024, OR 95% IC: 14 [1.4–137]).

Table 3.   Differences between patients with sustained virologic response and patients who relapsed
 SVR (n=23)Relapse (n=11)UnivariateMultivariate
PPOR (95% CI)
  1. Hazard ratios and 95% CI are shown only for independent predictive variables in the stepwise multivariate analysis. *Undetectable HCV viral load at week 4. Ns, not significant. Bold values represent statistical significance.

Age (years)41 (40–45)44 (42–46)0.123
Male gender18 (78.3%)10 (91%)0.637
Baseline CD4 count (cells / mm3)506 (360–710)410 (304–602)0.219
Baseline HCV viral load (log10 IU / mL)5.6 (4.6–6.2)6.2 (5.9–6.8)0.008ns
HCV genotype 111 (47.8%)8 (72.7%)0.237
Very early virologic response*15 (65.2%)1 (9%)0.003ns
Presence of strand specific HCV RNA at week 482 (8.6%)7 (63.6%)0.0020.02414 (1.4–137)

Since the proportion of patients with SVR with detectable HCV RNA in cell subsets at week 48 was lower compared to patients who relapsed, we analysed whether these patients had developed anti-IFN-α antibodies that could impair treatment efficacy. Nevertheless, none of the patients had detectable anti-IFN-α antibodies either at baseline or during the treatment (data not shown).

Discussion

This study shows for the first time that the presence of positive/negative strand HCV RNA in isolated cell subsets, B, CD4, CD8 and NK cells, at week 48 of PegIFN/RBV treatment could be directly associated with viral relapse in HCV–HIV coinfected patients.

Several previous studies have reported the presence of specific strand HCV RNA in PBMCs. Nevertheless, only one report associated this finding with the anti-HCV treatment response in HCV-monoinfected patients [23], while no studies have reported this association in HCV–HIV coinfected patients analysing different cell subsets at the end of treatment. In our study, a significant association between the presence of HCV RNA in cell subsets and viral relapse has been found. Nevertheless, since this is a small size study, larger cohorts are necessary to confirm our results. Of note, only two patients who relapsed had HCV RNA in total PBMCs in contrast to seven patients who had specific strand HCV RNA in cell subsets. This is of special importance because the absence of detection of HCV RNA in total PBMCs at the end of treatment could not reflect the real level of infection into the different isolated cell subsets. We have proved that the isolation of cell subsets enhance the amplification of HCV RNA when the viral load in cells is low. In this way, patients who relapsed had a significant higher proportion of negative strand HCV RNA in cell subsets at the end therapy compared to patients with SVR. In addition, the presence of HCV RNA in two patients with SVR in some cell subsets (although undetectable RNA load in PBMCs) is currently an unknown phenomenon that has to be further studied. However, these two patients were retested 12 weeks after the end of treatment for the presence of strand-specific HCV RNA. This was undetectable in both PBMCs and isolated cell subsets in both patients. On the other hand, four relapsers had no residual HCV RNA in either PBMCs or cell subsets. It is possible that HCV could be present at a very low level where was not detectable by our PCR assay. Also, HCV could have a residual replication at low level in the liver, as well not detectable in plasma, enough to rebound in the absence of treatment.

Other factors, as previously described, such as baseline HCV viral load and early virologic response are also associated with SVR [24]. In our study, the univariate analysis showed that these two variables along with the presence of strand specific HCV RNA in cell subsets at the end of therapy were significantly associated with SVR. Interestingly, only strand specific HCV RNA in cell subsets was independently associated with SVR in the multivariate analysis.

The positive predictive value for the presence of specific HCV RNA in cell subsets in our study was 82.6%, while the negative predictive value was 77.7%. A combination of the above three significant variables could increase their individual predictive value for treatment response. Hence, 13 patients who had baseline HCV RNA viral load lower than 6 log IU/mL, very early virologic response, and the absence of strand specific HCV RNA in cell subsets at the end of treatment, had SVR. In the same way, five patients who had baseline HCV viral load higher than 6 log IU/mL, no very early virologic response, and with the presence of strand specific HCV RNA in cell subsets at the end of treatment, relapsed. The positive predictive value for this combination of factors is 100%, and the negative predictive value is also 100%. To confirm these results, we are currently extending the number of patients.

The biological importance of finding negative strand HCV RNA in cells remains unclear; nevertheless it would be of importance in the clinical management of infected patients.

On the other hand, fluctuation of serum IFN-α levels during the treatment is unlike to occur since pegylation process makes interferon half-life steady from dose to dose (once per week) [11]. In addition, it is known that treatment with IFN-α could lead to the development of neutralizing anti-IFN-α antibodies in some patients [25,26]. Nevertheless, none of the patients in our study developed anti-IFN-α antibodies during the treatment. Despite all patients self-reported a complete adherence to the treatment during the 48 weeks, it has to be taken into account that some patients might have an irregular adherence that could decrease serum drug level by the end of the treatment. This eventual level decrease could have been enough to keep serum HCV RNA undetectable but insufficient to remove HCV RNA from extrahepatic reservoirs, such as cell subsets.

In conclusion, the presence of strand specific HCV RNA in different isolated cell subsets at the end of 48 weeks of therapy is associated with viral relapse in HCV–HIV coinfected patients. This could be of key importance for the clinical practice because it could modify the current management of HCV–HIV coinfected patients. Hence, a group of patients could be identified and proposed to prolong their treatment to 72 weeks. On the other side, this assay is expensive and not easy to be perform.

Acknowledgments

We thank the participants of this study for their cooperation. We also thank M.M. Rodriguez for her continuous technical support. This work was partially supported by Fondo de Investigaciones Sanitarias (FIS 04/0212), and Red de Investigaciones en Sida, Instituto de Salud Carlos III (RIS RD06/0006/0021).

Ancillary