827

Quantitative Analysis of Hepatitis C Virus in Peripheral Blood and Liver: Replication Detected Only in Liver Judie Boisvert,1 Xiao-Song He,2 Ramsey Cheung,2,3 Emmet B. Keeffe,2 Teresa Wright,4 and Harry B. Greenberg1,2,3,5

Departments of 1Microbiology and Immunology and 2Medicine, Stanford University School of Medicine, Stanford, 3Department of Veterans Affairs Palo Alto Health Care System, Palo Alto, 4 Department of Veterans Affairs Medical Center, San Francisco, and 5Aviron, Mountain View, California

Prior studies seeking evidence of viral replication in peripheral lymphocytes of hepatitis C virus (HCV)–infected patients have yielded conflicting results. This study sought to quantitatively determine whether a permissive HCV cell interaction could be detected in leukocytes from infected patients. Peripheral leukocytes from chronically infected patients were purified and were tested for HCV RNA. The results show that virus load is highest in B cells. Other subsets of peripheral leukocytes consistently had very low levels of viral RNA or were negative. Negative-strand HCV was found only in hepatocytes. To determine whether HCV replication could be induced by activation, B cells from HCV-infected patients were stimulated in vitro. No HCV replicating in peripheral leukocytes was detected by a highly sensitive assay. If HCV replication occurs in the leukocyte subsets analyzed here, it is at extremely low levels or occurs under alternate physiological conditions.

Hepatitis C virus (HCV) infects ∼3% of the world’s population and nearly 2% of the US population [1, 2]. Although primarily a hepatotropic virus, HCV is also associated with lymphoproliferative diseases, such as non-Hodgkin’s B cell lymphoma and cryoglobulinemia [3–7]. The role of HCV in the pathogenesis of these diseases is unknown. The association of HCV RNA with peripheral blood leukocytes has been documented since 1992 [8–13]. Many early studies described the detection of negative-strand HCV, the viral replicative intermediate, in peripheral blood mononuclear cells (PBMC) [8, 9, 11, 12]. However, the specificity of the methods used has been questioned [14–16]. More recent studies, which used an optimized negative strand–specific (NSS) reverse-transcriptase polymerase chain reaction (RT-PCR) assay, detected negative-strand HCV only in PBMC taken from posttransplant or human immunodeficiency virus–coinfected HCV patients, and not in PBMC from typical patients with chronic HCV infection [17–19]. Most of these studies analyzed PBMC and not purified leukocyte subpopulations. In studies in which PBMC were purified into subpopulations, either the strand Received 17 April 2001; revised 1 June 2001; electronically published 30 August 2001. Presented in part: 10th International Symposium on Viral Hepatitis and Liver Disease, Atlanta, April 2000 (abstract C127). Informed consent was obtained from patients, and human experimentation guidelines of the US Department of Health and Human Services and those of Stanford University were followed in the conduct of the research. Financial support: National Institutes of Health (AI-40034; AI-07328 to J.B.); Hutchinson Program in Translational Medicine at Stanford University. Reprints or correspondence: Dr. H. B. Greenberg, Aviron, 297 N. Bernardo Ave., Mountain View, CA 94043-5205 ([email protected]). The Journal of Infectious Diseases 2001; 184:827–35 䉷 2001 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2001/18407-0003$02.00

specificity of the NSS RT-PCR assays was not assessed or the overall sensitivity of the assay with respect to the number of purified cells analyzed was not addressed [9, 12, 20, 21]. An unanswered question is whether HCV associates with, and perhaps replicates in, a relatively rare cell type, resulting in dilution by nonpermissive cells and an undetected signal. Furthermore, most circulating leukocytes are in a resting state. Another B cell–tropic virus, Epstein-Barr virus, remains latent while the host cell is quiescent but is reactivated and enters a lytic replication phase once the host cell is activated (reviewed in [22]). The goals of this study were to determine the relative virus load and replication status of HCV in purified peripheral leukocyte subsets and hepatocytes of HCV-infected persons, to determine the detection limits of our assays, and to determine whether HCV levels were increased by in vitro activation of B cells.

Materials and Methods Leukocyte purification. About 50 mL of peripheral blood was collected from chronically HCV-infected patients into EDTA-containing Vacutainer tubes (Becton Dickinson). Polymorphonuclear leukocytes (PMNL) and red blood cells (RBCs) were separated from PBMC by density gradient centrifugation over ficoll (Amersham Pharmacia Biotech). RBCs were lysed by a hypotonic sodium chloride solution (0.2%). T cell–enriched and T cell–depleted populations were prepared from PBMC with use of magnetic beads, as described in the manufacturer’s instructions (CD2 CELLection kit; Dynal). Cells were stained with anti–human CD14–fluorescein isothiocyanate, anti–human CD19-phycoerythrin (PE), and anti–human CD5Cy5PE (BD PharMingen), were fixed with 0.5% paraformaldehyde in PBS for 5 min at room temperature, and were sorted by flow cytometry (FACS Vantage; Becton Dickinson). PMNL were sorted by gating on forward and side-scatter parameters. T cells were sorted

828

Boisvert et al.

by use of a lymphocyte forward and side-scatter gate and CD5⫹CD19⫺CD14⫺ fluorescence parameters. B cells were purified by using a lymphocyte forward and side-scatter gate and CD19⫹CD14⫺ fluorescence parameters. Purification of CD5⫹ and CD5⫺ B cells included the additional CD5 fluorescence parameter. Monocytes were purified by using a monocyte forward and sidescatter gate and CD14⫹CD19⫺CD5⫺ fluorescence parameters. After sorting, the cells were treated with 0.05% trypsin and 0.53 mM EDTA (Gibco BRL Life Technologies) for 15 min at 37⬚C. Cells were washed, counted, and aliquoted for either quantitative (Q) RT-PCR or NSS RT-PCR. Cell pellets designated for Q RTPCR were frozen at ⫺80⬚C and were analyzed within 1 month. From cells designated for NSS RT-PCR, RNA was extracted within 1 h (ToTALLY RNA kit; Ambion), aliquoted, and frozen in a 70% ethanol solution at ⫺80⬚C for future analysis. For the in vitro stimulation assay, B cells were enriched from whole blood (RosetteSep; Stem Cell Technologies), followed by ficoll density gradient centrifugation (Amersham Pharmacia Biotech) performed according to the manufacturers’ instructions. Purity was confirmed by fluorescence-activated cell sorting (FACS) analysis on either a FACS Vantage or a FACSCaliber machine (Becton Dickinson). Hepatocyte purification. Hepatocytes infected with HCV were prepared from HCV-positive explant liver tissue obtained at the time of transplantation and were processed immediately. The tissue was cut into small cubes and was washed with RPMI (BioWhittaker) containing 10% heat-inactivated fetal bovine serum (Gibco/BRL), to remove contaminating peripheral blood lymphocytes. Tissue chunks then were processed in a blender (Stomacher 400; Seward). Hepatocytes were isolated from the pellet of ficoll gradients and were cryopreserved in 10% dimethyl sulfoxide until analysis. Q RT-PCR. This analysis was performed on samples PB1PB5, PB9-11, and LV2 by use of Amplicor Monitor kits (Roche Molecular Systems). Patient samples PB6-PB8, PB12-PB17, LV1, and LV3-LV5 were analyzed by Amplicor Monitor COBAS kits (Roche), according to the manufacturer’s instructions. To control for contamination, we used 3 different rooms, as follows: specimen preparation and RT and PCR mix preparation, amplification, and detection [23]. Statistics. Statistical analyses were done with the StatView software program (SAS Institute) by first averaging the 2 values (where available) for the duplicate aliquots of each sample. For each cell type, after these values were averaged, SE was calculated, and statistical significance was evaluated by paired t test. Because the B cell value for patient PB2 was an outlier of much higher value than other patient samples, this number was omitted from statistical analyses. Transcripts and primers. In vitro transcribed RNA molecules were prepared as follows. Plasmid pTET/T7HCV(D)BglII/5corr containing HCV sequences corresponding to nucleotides 1–3236 and 8938 to the end of the 3 noncoding region (NCR) was provided by C. Rice (Rockefeller University, New York); an XbaI-BglII digestion product containing the HCV sequences was subcloned into a pBluescript II KS⫹ vector (Stratagene), which contains both T7 and T3 promoters (pBlue/HCV). Orientation was confirmed such that T7 transcription yielded positive-sense HCV transcripts and T3 transcription yielded negative-sense HCV transcripts. Competent Escherichia coli cells were transformed with pBlue/HCV. Plas-

JID 2001;184 (1 October)

mid DNA was isolated and linearized by restriction digest with XbaI or EcoRI for runoff transcription. Digestion products were gel purified before phenol-chloroform extraction. In vitro transcription was done by using T7 and T3 Megascript kits (Ambion), followed by DNase digestion and phenol-chloroform extraction. Stock and serial dilution aliquots were stored at ⫺80⬚C. Primer sequences were as follows: primer 49, GGC GAC ACT CCA CCA TGA ATC AC; primer 66, CAT GGT GCA CGG TCT ACG AGA CC; primer 50, GGA ACT ACT GTC TTC ACG CAG AA; and primer 56, TCG CAA GCA CCC TAT CAG GCA GT [24, 25]. All primers were synthesized by Integrated DNA Technologies. Strand-specific RT-PCR. We used a modified version of the NSS assay, developed by Lanford et al. [15], to detect antisense HCV [15]. Reverse transcription was performed as described. Forward primer (no. 49) was used for reverse transcription and reverse primer (no. 66) in the PCR reaction. Reverse transcription was done on a thermocycler (model 9600; PE Biosystems) at 95⬚C for 5 min, followed by 65⬚C for 30 min to 1 h. PCR was done as follows: 1 cycle at 95⬚C for 1 min; 50 cycles at 95⬚C for15 s, 58⬚C for 30 s, and 72⬚C for 30 s; and 1 cycle at 72⬚C for 5 min. PCR products (expected product size, 325 bp) were detected by Southern blotting by using a nonisotopic psoralen-biotin–labeled probe and chemiluminescence (Ambion). The hybridization probe was prepared by PCR amplification of plasmid DNA containing the HCV sequences, by using primers (50 and 56) internal to the NSS RT-PCR target region. PCR products of the expected size (260 bp) were gel purified and were labeled with psoralen-biotin with a labeling kit (BrightStar Psoralen-Biotin; Ambion), per the manufacturer’s instructions. Each experiment included the following controls: serial dilutions of in vitro–transcribed RNA molecules containing positive- or negative-sense HCV nucleotide sequences corresponding to the 5 one-third of the viral genome (5 NCR to N-terminal region of NS2). Positive sense transcript dilutions contained an estimated 1 ⫻ 10 6, 1 ⫻ 10 5, and 1 ⫻ 10 4 copies, respectively. Negative-sense transcript dilutions contained an estimated 1 ⫻ 10 4 , 1 ⫻ 10 3 , 1 ⫻ 10 2 , 10, and 1 copies, respectively. Copy numbers were estimated on the basis of the RNA concentration (as determined by spectrophotometry and agarose/ formaldehyde gel electrophoresis) and the molecular weight of the RNA transcript. Working dilutions were made in 100 mg/mL of total RNA extracted from an HCV-negative tissue culture cell line (Daudi; American Type Culture Collection). Negative controls included RNA extracted from Daudi cells, RNase-free water (Ambion), and 100 copies each of positive- and negative-sense in vitro RNA transcripts, respectively, from which the primer (no. 49) was omitted during reverse transcription and added during PCR. RNA extracted from HCV-positive explant hepatocytes was included in most experiments as a positive control. RNase protection assay. RNase protection was done as described elsewhere [26]. In brief, total cellular RNA was extracted from HCV-infected explant hepatocytes, allowed to hybridize in the absence of probe, and then followed by RNase treatment, to remove excess positive-sense HCV RNA. After proteinase K treatment, phenol/chloroform extraction, and ethanol precipitation, RNA was subjected to NSS RT-PCR, as described above. B cell stimulation assay. In vitro stimulation of peripheral B cells was done by coculturing 5 ⫻ 10 5 B cells/mL with 1 ⫻ 10 5 irradiated murine L cells constitutively expressing human CD40L

JID 2001;184 (1 October)

Analysis of HCV in Peripheral Leukocytes

(provided by S. Levy with permission from Y.-J. Liu, DNAX Research Institute, Palo Alto, CA). Culture medium contained Iscove’s modified Dulbecco’s medium (Gibco BRL), 2% fetal bovine serum, 0.5% bovine serum albumin (Sigma-Aldrich), 50 mg/mL human transferrin (Sigma), 5 mg/mL bovine insulin (Sigma), 15 mg/ mL gentamicin, and 2 ng/mL human interleukin (IL)–4 (R&D Systems).

Results Patient characteristics. Table 1 lists relevant clinical information for each of the 19 patients with HCV infection enrolled in this study: age, sex, disease status (chronic, transplant, or posttransplant), virus load in serum or plasma, genotype, risk factors, and disease duration. Hepatocytes purified from liver recipient (HCV-infected transplant) tissue samples were used as positive controls for detection of total HCV RNA and negativestrand HCV RNA. Purification of leukocyte subsets. To determine the relative virus load in different leukocyte subsets, peripheral blood was collected from HCV patients and was purified, as described in Materials and Methods, into the following leukocyte subpopulations: PMNL, T cells (CD5⫹CD19⫺CD14⫺), B cells (CD19⫹ CD14⫺), and monocytes (CD14⫹CD5⫺CD19⫺). Cell purity was assessed by FACS and was usually 1 95% (data not shown). Because these cells were fixed due to safety concerns and were trypsinized to remove or reduce cell-surface–adsorbed viTable 1.

rus particles, there was concern that fixation or trypsinization might reduce the sensitivity of the RT-PCR assays. To address this concern, duplicate PBMC samples from patients with HCV infection were treated either with trypsin or with paraformaldehyde, as described in Materials and Methods, or were left untreated and analyzed by Q RT-PCR. No statistically significant differences were found for treatment or no treatment samples (data not shown). Q RT-PCR of HCV RNA level in different cell populations. Duplicate aliquots of purified cells from patients PB1–PB13 were analyzed by Q RT-PCR (Roche Amplicor Monitor kits) to determine the relative virus load in each cell population. Analyzed subsets included unfractionated PBMC, PMNL, T and B cells, and monocytes. In addition, hepatocytes obtained from HCV-positive transplant liver tissue were analyzed as a positive biologic control. In addition to the internal negative controls included with the Monitor kits, 21 samples known to be HCV negative also were tested and were found to be negative by this assay (13 plasma or serum samples and 8 PBMC samples; data not shown). We show the results of our analysis as mean HCV RT-PCR copies per 1 ⫻ 10 4 cells for each subset (figure 1). The highest average virus load was found in B cells, followed by hepatocytes, PBMC, and PMNL; monocytes were usually positive, with very low HCV RNA levels. With the exception of 1 sample (of 22 analyzed), T cells were always negative. Of note, when

Characteristics of hepatitis C virus (HCV)–infected patients. HCV infection status

Transplant a date

Virus load

Serum ALT level, U/L

— —

NA NA

88 188

Chronic Chronic Chronic Posttransplant Posttransplant Posttransplant Chronic Chronic End-stage Chronic

— — — 1988, 1990 1996 1997 — — — —

NA 3.08 ⫻ 105 3.70 ⫻ 105 9.40 ⫻ 105 2.73 ⫻ 106 1.32 ⫻ 106 4.15 ⫻ 105 5.00 ⫻ 105 4.70 ⫻ 105 6.30 ⫻ 105

108 134 390 150 NA NA 121 96 53 246

Chronic Chronic Chronic Chronic Chronic Posttransplant Chronic Transplant Transplant Transplant Transplant Transplant

— — — — — 2/1999 — 7/1999 7/1999 7/1999 12/1999 1/2000

4.85 ⫻ 105 1.97 ⫻ 103 3.32 ⫻ 105 1.66 ⫻ 105 8.48 ⫻ 105 7.50 ⫻ 105 NA 9.4 ⫻ 104 1.5 ⫻ 103 NA NA 3.4 ⫻ 103

23 26 30 80 73 NA 108 48 232 86 43 89

Age, years

Sex

PB1 PB2

NA 43

Female Male

Chronic Chronic

PB3 PB4 PB5 PB6 PB7 PB8 PB9 PB10 PB11 PB12

44 53 47 57 52 53 54 35 53 49

Male Male Male Male Male Male Male Male Male Male

PB13 PB14 PB15 PB16 PB17 PB18 PB19 LV1 LV2 LV3 LV4 LV5

65 52 58 44 NA 62 52 49 42 46 44 53

Male Male Female Male Male Female Male Male Male Male Female Male

Patient

829

b

NOTE. ALT, alanine aminotransferase; IDU, injection drug use; NA, not available. a A dash (—) indicates not applicable. b HCV RNA copies/mL of serum or plasma.

Risk factor(s) Transfusion IDU, cocaine, ear piercing, transfusion IDU Tattoo Unknown Unknown Unknown Unknown None Cocaine use Unknown IDU, transfusion, ear piercing, tattoo HCV-positive wife Unknown Unknown None Tattoo, ear piercing Transfusion IDU Tattoos Transfusion Transfusion IDU Unknown

Disease duration, years 12 15–20 25 25–30 Unknown 9 (3d) 4 (2d) 3 (2d) Unknown 12–18 Unknown 20–24 Unknown Unknown Unknown Unknown 13–15 0.25 (2d) Unknown 30 5 46 28 Unknown

830

Boisvert et al.

compared with unfractionated PBMC, PMNL, T cells, or monocytes, B cells had a significantly higher mean HCV RNA level (P ! .002, paired t test). However, the difference in mean HCV RNA levels between CD5⫹ and CD5⫺ B cells was not statistically significant (P 1 .82 , paired t test), nor was the difference between B cells and hepatocytes. Because the manufacturer replaced the original Monitor kits with the newer COBAS kits during the study period, 8 patient samples were analyzed with the original Roche Amplicor Monitor kits and 9 were analyzed with the newer Roche Amplicor Monitor COBAS system. To determine the relative sensitivity of the 2 assays, 21 samples were tested by both assays in parallel and the results compared. Fourteen of these samples were within the linear range of the assay. Results obtained with the COBAS kits were consistently higher than those with the original monitor kit, by an average of 3.3-fold (data not shown). To adjust for this discrepancy, all Q RT-PCR results obtained with the COBAS test were divided by 3.3. Detection of HCV replication intermediate negative strand. Leukocyte samples from 15 patients (PB1–PB13, PB18, and PB19) tested negative for NSS HCV. Selected leukocyte samples (from patients PB1, PB3, PB4, PB11, and PB18) were tested twice in independent experiments. Figure 2 shows a represen-

JID 2001;184 (1 October)

tative blot for an experiment containing positive and negative controls and data for leukocyte subsets from patient PB6. To determine the sensitivity and strand specificity of our assay, serial dilutions of positive- and negative-sense in vitro– transcribed RNA molecules were included as controls in each experiment. The sensitivity of this assay is 1–10 copies, and the strand specificity is ⭓10,000-fold; that is, 1 ⫻ 10 5 to 1 ⫻ 10 6 positive-sense copies were required to yield a false-positive result in our assay (figure 2). In addition, hepatocytes obtained from HCV-positive posttransplant liver tissue were included as a positive biologic control; negative controls included water and RNA extracted from tissue culture cell lines (Daudi). To control for residual RT activity, samples containing 100 copies each of positive- and negative-strand in vitro transcripts from which primer was excluded during reverse transcription, but added for PCR, were also included (figure 2). Detection of the negative-strand RNA of positive-strand RNA viruses can be complicated by a large molar excess of positive-sense RNA, which in poliovirus, for example, is 30–70fold [26, 27]. To determine whether this might be the case for our NSS HCV assay, an RNase protection assay was performed on RNA extracted from HCV-positive explant hepatocytes obtained from posttransplant liver tissue in which excess positive-

Figure 1. Mean hepatitis C virus (HCV) RNA level of different leukocyte subsets isolated from 13 patients and analyzed in duplicate. Cell types analyzed were peripheral blood mononuclear cells (PBMC), polymorphonuclear leukocytes (PMNL), T cells, B cells, CD5⫹ B cells, CD5⫺ B cells, and monocytes. Mean HCV RNA levels in explant hepatocytes taken from posttransplant liver tissue of 5 patients with HCV infection (LV1–LV5) are shown for comparison. HCV RNA levels were significantly higher in B cells than in unfractionated PBMC, PMNL, T cells, and monocytes (P ! .002 for each comparison). HCV RNA levels in CD5⫹ vs. CD5⫺ B cells did not differ significantly (P 1 .82). HCV RNA levels in hepatocytes were not significantly different from those of peripheral B cells (P 1 .56 ). Bars, SE for each cell type. P values were calculated by paired t tests. *Comparison between B cells and PBMC, PMNL, T cells, and monocytes; **comparison between CD5⫹ and CD5⫺ B cells. RTPCR, reverse-transcriptase polymerase chain reaction.

JID 2001;184 (1 October)

Analysis of HCV in Peripheral Leukocytes

Figure 2. Detection of hepatitis C virus (HCV) replication intermediate negative strand. Southern blot shows polymerase chain reaction products from a representative experiment that assessed the presence of negative-strand HCV RNA in purified leukocyte subsets from HCV patient PB6. Also shown are positive (⫹sense) and negative (⫺sense) in vitro transcripts (IVTs) and hepatocytes from transplant liver tissue of 2 patients with HCV infection (LV1 and LV2). Negative controls included duplicate samples of RNA from HCV-negative tissue culture cell lines, duplicate samples with water as template, and 100 copies of each ⫹sense and ⫺sense HCV in vitro RNA transcripts from which primer was omitted during reverse transcription. All samples shown were processed in 1 experiment. B, B cells; monos, monocytes; PBMC, peripheral blood mononuclear cells; PMNL, polymorphonuclear leukocytes; T, T cells.

sense RNA molecules were removed by RNase digestion after allowing positive- and negative-sense RNA molecules to hybridize overnight. Control samples were subjected to the same conditions, except that RNases were omitted. Duplicate RNA aliquots were analyzed in 2 independent experiments: In both cases, RNase treatment did not increase the sensitivity for negative-strand detection, indicating that the inability to detect negative-sense RNA in the leukocyte samples is not due to excess positive-sense RNA (data not shown). Quantitative analysis of the NSS assay. In total, 47 leukocyte samples were positive for HCV RNA by the Q RT-PCR assay but negative for the HCV replicative intermediate by the NSS RT-PCR assay. To determine whether this assay was sensitive enough to detect negative-strand HCV in leukocyte subsets, we calculated the predicted number of negative-strand molecules for each sample. This calculation requires the assumption that negative-strand synthesis in leukocytes, if it occurs, would be at the same rate as in hepatocytes. The other assumption is an estimate of the ratio of total HCV RNA per cell to negative-strand HCV RNA in cells supporting viral replication, as determined by the Q RT-PCR and NSS RT-PCR assays used in this study, respectively. Duplicate aliquots of HCV-infected hepatocytes from 3 patients (LV1–LV3) were analyzed for the copy number of total HCV RNA and negativestrand HCV RNA, respectively. Total HCV RNA was determined by the Roche quantitative

831

assay described above. We analyzed negative-strand HCV RNA by limiting serial dilution in the NSS RT-PCR assay, as described. Serially diluted hepatocyte RNA samples and positive and negative in vitro–transcribed HCV RNA molecules were used as templates for NSS RT-PCR. The copy number of negative-strand HCV was determined to be equal to the copy number of the negative-strand transcript sample with the same endpoint titers. Next, we calculated the ratio of total HCV RNA to negative-strand HCV RNA for each of the 3 samples (14, 27, and 81), the average of which was 41. Figure 3 shows the predicted number of negative-strand molecules that would be expected to be present in each sample, on the basis of the hypothesized ratio of 41. Assuming that the NSS RT-PCR assay has a sensitivity of 10 copies, 30% (14/47 samples) of the leukocyte samples in question would be above the threshold of negative-strand HCV RNA required to be detected in our assay. Assuming a sensitivity of 1 copy, 66% (31/47) of these samples would have sufficient negative-strand HCV RNA above the threshold to be detected (figure 3). In vitro stimulation of B cells. To determine whether replication of HCV could be induced after in vitro B cell activation, we cultured purified B cells obtained from HCV patients (patients PB14–PB17), with or without B cell activation. B cells were enriched from peripheral blood by depletion of T cells, monocytes, NK cells, PMNL, and RBCs, by use of RosetteSep (Stem Cell Technologies). Cells were cultured either in medium alone or with mouse L cells expressing CD40L in the presence of IL-4. In addition, RNA was immediately extracted from about one-third of the B cell–enriched population before culturing. RNA was frozen for future analysis of HCV RNA. Activation of B cells cultured under stimulating conditions was confirmed by FACS analysis, and stimulated cells were compared with cells cultured in medium alone. Stimulated cells adopt a larger blastlike appearance and up-regulate expression of activation molecules CD69, CD54, and CD80, compared with unstimulated cells (figure 4). Figure 5 shows a marked decrease in the relative HCV RNA levels in each sample over time, with or without activation. When analyzed for the presence of negative-sense HCV RNA, all of these samples, including preculture B cells, tested negative (data not shown), indicating that HCV is not replicating at detectable levels under these conditions (data not shown).

Discussion The finding that the average HCV RNA level in hepatocytes is similar to that of B cells is not entirely surprising (figure 1). Previous studies that quantitatively analyzed HCV in hepatocytes by use of similar methods showed a wide range of virus load (1 ⫻ 10 2–1 ⫻ 10 8 genome equivalents/1 mg of RNA) [17, 19, 28–30]. Considering that 1 ⫻ 10 6 cells yield ∼10 mg of total RNA, our data are well within that range. In addition, the liver

832

Boisvert et al.

JID 2001;184 (1 October)

Figure 3. Quantitative analysis of negative strand–specific assay. Data show approximate no. of hepatitis C virus (HCV) negative-strand molecules per test predicted to be present if HCV were replicating in these cells at positive:negative strand HCV RNA ratios similar to those calculated for hepatocytes.

samples used in our study came from HCV-infected patients undergoing liver transplantation, and the virus load might be expected to be lower. Certainly, the serum HCV RNA levels in these patients appear to be substantially lower than those in our chronic infection study population (table 1). These studies show a preferential association of HCV with B cells. This is especially interesting since B cells comprise such a small percentage of PBMC (10%–15%), whereas T cells, which are a much larger fraction of PBMC (70–80%), are rarely positive for HCV. Passive adsorption of virus particles to the cell surface remains a possible explanation, although our data were obtained after treatment of the cells with trypsin. Another possibility is that free virions are taken up as antibody–virus particle complexes by B cells, via HCV-specific surface immunoglobulins, or by Fc receptors. One study found specific binding of HCV surface glycoprotein E2 with CD81, a nearly ubiquitous cell surface marker that may prove to be important for viral entry [31]. Although it is unlikely that CD81 functions as the sole receptor for HCV, since it is expressed on T cells, with which HCV appears not to associate, it may function as a binding and/or entry receptor in complex with other molecules. For example, CD81 associates with B cell–specific molecules, CD21 and CD19, and CD19 associates with surface immunoglobulin molecules. Perhaps the association of CD81 with these cell-specific molecules allows HCV cell entry. The association of HCV with monocytes

and PMNL can be explained by the phagocytic role these cells play in peripheral immunosurveillance. CD5 is a cell surface marker found on all T cells and on a subpopulation of B cells, also referred to as B1 B cells [32–34]. B1 B cells are thought to derive from a separate stem cell found in the fetal liver, not in bone marrow. B1 B cells are numerous during fetal and perinatal development but rare in adult life, and they appear to be self-renewing. Of note, B1 B cells have a high propensity to recognize autoantigens and may be responsible for a large fraction of the IgM found in serum. This is of special interest, considering the close association between HCV and cryoglobulinemia, a disease characterized by overproduction of mono- and polyclonal IgM and IgG autoantibodies (reviewed in [4]). The proportion of CD5⫹ B cells is significantly greater in chronically HCV-infected patients than in persons whose HCV infection has resolved or in uninfected volunteers [35]. However, little is known about CD5 function in vivo. Consequently, B cells from patients PB9–PB13 were purified into CD5⫹ and CD5⫺ fractions, to determine whether the virus preferentially associates with either B cell subpopulation. Our data show no significant difference in HCV RNA level in these 2 subpopulations. The detection of negative-strand HCV sequences suggests active replication of the virus. Any attempt to determine whether HCV replicates outside the liver must address a number of tech-

JID 2001;184 (1 October)

Analysis of HCV in Peripheral Leukocytes

833

Figure 4. In vitro stimulation of hepatitis C virus (HCV)–positive peripheral B cells. Shown is phenotypic profile of HCV-positive peripheral B cells cultured in vitro for 4 days without (A–C) or with (D–F) CD40L and interleukin-4 stimulation. A and D, Forward-scatter gate (FSC) vs. side-scatter (SSC) parameters show increase in cellular size and complexity after stimulation. B and E, CD69–fluorescein isothiocyanate cell surface expression. C and F, CD54-Cy5–phycoerythrin (PE) vs. CD80-PE cell surface expression.

nical obstacles. HCV is present in most infected patients at very low levels. Thus, its detection alone can be a technical challenge. Obtaining maximal sensitivity without resorting to nested RTPCR, which is prone to contamination, is also highly desirable. Ensuring and demonstrating strand-specificity in an NSS assay is especially important to control for false positives, which can be obtained as the result of false priming of the positive-sense genomic RNA during reverse transcription, residual RT activity after addition of the antisense primer, or self-priming hairpin

structures in the RNA. If no evidence of replication is found, it is important to determine whether the assay is sensitive enough to detect replication if it occurs. In calculating the predicted number of negative-strand HCV genome copies that might be expected to be present, if replication is occurring, 2 important considerations must be made. First is the assumption that replication in B cells (or other leukocyte subsets) yields positive:negative strand RNA ratios similar to those in hepatocytes. And, second, what is that ratio?

Figure 5. Hepatitis C virus (HCV) RNA in in vitro–stimulated B cells. Shown are relative HCV RNA levels in purified peripheral B cells from patients PB14–PB17 before stimulation or after 4 days of coculture in the presence of irradiated CD40L-expressing cells and interleukin-4 or without stimulation. Error bars, SEs for duplicate samples of each cell type. RT-PCR, reverse-transcriptase polymerase chain reaction.

834

Boisvert et al.

By using 3 explant liver samples, we calculated an average positive:negative HCV RNA ratio of 41, which is consistent with the 10–100-fold ratios reported in other studies [17, 19, 28–30]. In a significant proportion of the leukocyte samples that we analyzed, negative-strand HCV would have been detected with our assay, were such replication occurring under the conditions assumed. We also examined the possibility that HCV could replicate in peripheral B cells, but only under altered physiological conditions, such as immunosuppression or cellular activation. Four persons enrolled in this study (PB6–PB8 and PB18; table 1) were immunosuppressed posttransplant patients and had the highest serum HCV RNA levels, yet evidence of negative-strand HCV was not found in their peripheral leukocytes. To test whether cellular activation after in vitro stimulation might induce HCV replication, enriched B cells from 4 patients were either stimulated with CD40L and IL-4 or cultured without stimulation. In this experimental system, we did not detect evidence of replication. Of note, although B cells are clearly activated via CD40-CD40L interactions in the presence of IL-4 (figure 4), several other mechanisms of cellular activation that may induce molecular changes within the host B cell differ from those that occur under these stimulation conditions. HCV replication under these alternate stimulating conditions (e.g., antiIgM or LPS) might be examined in future studies. In summary, although the association of HCV with B cells is clear, the nature of that association is more obscure. Our data support the argument against a productive infection of B cells or any other leukocyte subsets analyzed in this study. Certainly, within the limits of our assays, data consistent with replication in B cells could not be obtained. Of importance, our assay was sufficiently sensitive to detect viral replication, with the assumptions stated above, in a significant proportion of the samples tested. Finally, we also show that in vitro stimulation of HCV-positive peripheral B cells via CD40L and IL-4 does not lead to HCV replication in culture.

5. 6.

7.

8.

9. 10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

Acknowledgments We thank C. Rice (Rockefeller University), for the generous gift of pTET/T7HCV(D)BglII/5corr plasmid, and M. Roederer (National Institutes of Health), J. J. Campbell (Harvard University), A. Warford, M. Parry, Y. Lu, H. Monge, N. Feng, and S. Levy (Stanford University), for advice and technical assistance.

20.

21.

22. References 1. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States 1988 through 1994. N Engl J Med 1999; 341:556–62. 2. Wasley A, Alter MJ. Epidemiology of hepatitis C: geographic differences and temporal trends. Semin Liver Dis 2000; 20:1–16. 3. Agnello V, Chung RT, Kaplan LM. A role for hepatitis C virus infection in type II cryoglobulinemia. N Engl J Med 1992; 327:1490–5. 4. Agnello V. The etiology and pathophysiology of mixed cryoglobulinemia

23. 24.

25.

26.

JID 2001;184 (1 October)

secondary to hepatitis C virus infection. Springer Semin Immunopathol 1997; 19:111–29. Pascual M, Perrin L, Giostra E, Schifferli JA. Hepatitis C virus in patients with cryoglobulinemia type II. J Infect Dis 1990; 162:569–7. Pozzato G, Mazzaro C, Crovatto M, et al. Low-grade malignant lymphoma, hepatitis C virus infection, and mixed cryoglobulinemia. Blood 1994; 84: 3047–53. Zuckerman E, Zuckerman T, Levine AM, et al. Hepatitis C virus infection in patients with B-cell non-Hodgkin lymphoma. Ann Intern Med 1997; 127:423–8. Qian C, Camps J, Maluenda MD, Civeira MP, Prieto J. Replication of hepatitis C virus in peripheral blood mononuclear cells: effect of alpha-interferon therapy. J Hepatol 1992; 16:380–3. Zignego AL, Macchia D, Monti M, et al. Infection of peripheral mononuclear blood cells by hepatitis C virus. J Hepatol 1992; 15:382–6. Bouffard P, Hayashi PH, Acevedo R, Levy N, Zeldis JB. Hepatitis C virus is detected in a monocyte/macrophage subpopulation of peripheral blood mononuclear cells of infected patients. J Infect Dis 1992; 166:1276–80. Wang JT, Sheu JC, Lin JT, Wang TH, Chen DS. Detection of replicative form of hepatitis C virus RNA in peripheral blood mononuclear cells. J Infect Dis 1992; 166:1167–9. Muller HM, Pfaff E, Goeser T, Kallinowski B, Solbach C, Theilmann L. Peripheral blood leukocytes serve as a possible extrahepatic site for hepatitis C virus replication. J Gen Virol 1993; 74:669–76. Schmidt WN, Klinzman D, LaBrecque DR, Macfarlane DE, Stapleton JT. Direct detection of hepatitis C virus (HCV) RNA from whole blood, and comparison with HCV RNA in plasma and peripheral blood mononuclear cells. J Med Virol 1995; 47:153–60. McGuinness PH, Bishop GA, McCaughan GW, Trowbridge R, Gowans EJ. False detection of negative-strand hepatitis C virus RNA. Lancet 1994; 343:551–2. Lanford RE, Chavez D, Chisari FV, Sureau C. Lack of detection of negativestrand hepatitis C virus RNA in peripheral blood mononuclear cells and other extrahepatic tissues by the highly strand-specific rTth reverse transcriptase PCR. J Virol 1995; 69:8079–83. Willems M, Moshage H, Yap SH. PCR and detection of negative HCV RNA strands. Hepatology 1993; 17:526. Radkowski M, Wang LF, Vargas HE, Rakela J, Laskus T. Detection of hepatitis C virus replication in peripheral blood mononuclear cells after orthotopic liver transplantation. Transplantation 1998; 66:664–6. Laskus T, Radkowski M, Wang LF, Vargas H, Rakela J. The presence of active hepatitis C virus replication in lymphoid tissue in patients coinfected with human immunodeficiency virus type 1. J Infect Dis 1998;178:1189–92. Laskus T, Radkowski M, Wang LF, Jang SJ, Vargas H, Rakela J. Hepatitis C virus quasispecies in patients infected with HIV-1: correlation with extrahepatic viral replication. Virology 1998; 248:164–71. Lerat H, Rumin S, Habersetzer F, et al. In vivo tropism of hepatitis C virus genomic sequences in hematopoietic cells: influence of viral load, viral genotype, and cell phenotype. Blood 1998; 91:3841–9. Mellor J, Haydon G, Blair C, Livingstone W, Simmonds P. Low level or absent in vivo replication of hepatitis C virus and hepatitis G virus/GB virus C in peripheral blood mononuclear cells. J Gen Virol 1998;79:705–14. Luka J. Activation of Epstein-Barr virus in latently infected cell lines. J Virol Methods 1988; 21:223–7. Kwok S, Higuchi R. Avoiding false positives with PCR. Nature 1989; 339: 237–8. Umlauft F, Wong DT, Oefner PJ, et al. Hepatitis C virus detection by singleround PCR specific for the terminal 3 noncoding region. J Clin Microbiol 1996; 34:2552–8. Shindo M, Di Bisceglie AM, Silver J, Limjoco T, Hoofnagle JH, Feinstone SM. Detection and quantitation of hepatitis C virus RNA in serum using the polymerase chain reaction and a colorimetric enzymatic detection system. J Virol Methods 1994; 48:65–72. Novak JE, Kirkegaard K. Improved method for detecting poliovirus negative

JID 2001;184 (1 October)

27. 28.

29.

30.

Analysis of HCV in Peripheral Leukocytes

strands used to demonstrate specificity of positive-strand encapsidation and the ratio of positive to negative strands in infected cells. J Virol 1991;65: 3384–7. Andino R, Rieckhof GE, Baltimore D. A functional ribonucleoprotein complex forms around the 5 end of poliovirus RNA. Cell 1990; 63:369–80. Laskus T, Radkowski M, Wang LF, Cianciara J, Vargas H, Rakela J. Hepatitis C virus negative strand RNA is not detected in peripheral blood mononuclear cells and viral sequences are identical to those in serum: a case against extrahepatic replication. J Gen Virol 1997; 78:2747–50. Laskus T, Radkowski M, Wang LF, Vargas H, Rakela J. Lack of evidence for hepatitis G virus replication in the livers of patients coinfected with hepatitis C and G viruses. J Virol 1997; 71:7804–6. Laskus T, Radkowski M, Wang LF, Vargas H, Rakela J. Detection of hepa-

31. 32. 33. 34. 35.

835

titis G virus replication sites by using highly strand-specific Tth-based reverse transcriptase PCR. J Virol 1998; 72:3072–5. Pileri P, Uematsu Y, Campagnoli S, et al. Binding of hepatitis C virus to CD81. Science 1998; 282:938–41. Herzenberg LA, Stall AM, Lalor PA, Sidman C, Moore WA, Parks DR. The Ly-1 B cell lineage. Immunol Rev 1986; 93:81–102. Hardy RR, Hayakawa K, Shimizu M, Yamasaki K, Kishimoto T. Rheumatoid factor secretion from human Leu-1⫹ B cells. Science 1987;236:81–3. Hayakawa K, Hardy RR. Development and function of B-1 cells. Curr Opin Immunol 2000; 12:346–53. Curry MP, Golden-Mason L, Nolan N, Parfrey NA, Hegarty JE, O’Farrelly C. Expansion of peripheral blood CD5⫹ B cells is associated with mild disease in chronic hepatitis C virus infection. J Hepatol 2000; 32:121–5.