INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Quantitation of hepatitis C virus (HCV) in the blood of infected patients is now widely used to monitor therapy. An RNA extraction step from blood or blood fractions is necessary for all the major methods used to quantify HCV because they involve amplification of HCV RNA. Whether to use serum, plasma, or whole blood to obtain the greatest sensitivity for determining viral clearance, i.e., undetectable virus in the blood, has been controversial. This question is of clinical importance because the determination of viral clearance is an important parameter used to gauge the duration of therapy. Most commercial HCV RNA assays use serum samples. Equivalent quantitative results are obtained from serum and plasma samples (4, 10; L. Cook, A. M. Ross, G. B. Knight, and V. Agnello, unpublished data). However, it has been postulated that the use of whole blood provides greater sensitivity in determining clearance of HCV from the circulation because the presence of HCV in the white blood cells (WBC) is not detected when serum or plasma is used (15, 16, 20).

The amount of virus present in WBC of the blood has been controversial. Since 1992, more than 60 articles have reported levels of HCV in mononuclear cells from more than 700 patients with HCV infection treated and untreated with interferon. In the early studies of untreated patients, the HCV RNA positive (virion) strand was found in the mononuclear cells from a majority of patients, and the negative (replicative) strand was found in the mononuclear cells from a minority of patients (3, 13, 21). Viral replication in cultured lymphocyte cell lines (17-19) was also reported. Data from the large number of follow-up studies have, in general, confirmed the presence of the HCV RNA positive-strand virus in mononuclear cells of infected patients and demonstrated a decrease in the percentage of positive results during and after interferon therapy. On the other hand, recent studies have demonstrated that the methods for the detection of the negative-strand virus are technically difficult and frequently generate false-positive results. A recent study (11) conducted with a methodology that is less susceptible to technical artifacts found that none of the 27 patients infected with HCV had detectable negative-strand virus.

The majority of published studies of HCV in mononuclear cells have used nonquantitative screening methods for the detection of the virus. Of the patients whose serum or plasma tested positive for HCV RNA, 20 to 60% tested negative for HCV RNA in their mononuclear cells. In one study (7) in which the HCV in the serum and mononuclear cells was quantitatively determined by the branched-DNA assay, the level of HCV virus was found to be 100 to 5,000 times less in the cells. These studies support the idea that HCV is not present in large quantities in the WBC. A few studies (9, 12, 14) have separated the mononuclear cells into T cells, B cells, monocytes, and neutrophils (polymorphonuclear leukocytes) and demonstrated that HCV can be found in all of the cell fractions. The amount of HCV in each of the cell fractions has not been determined.

In contrast, studies (15, 16, 20) from one laboratory using a new extraction agent, Catrimox-14, have suggested that HCV RNA was present in significant concentrations in blood cells. These studies showed 100- to 1,000-times-higher levels of HCV RNA in Catrimox-treated whole blood samples than in those from traditional methods of extraction from serum or plasma samples.

To determine whether extraction from whole blood is a more sensitive and useful method than extraction from serum or plasma for HCV RNA measurements for the clinical laboratory, we used a variety of different extraction methods on whole blood, serum samples, and mononuclear cells and measured the amount of HCV RNA present by competitive reverse transcription (RT)-PCR.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study samples were obtained from patients with HCV infection who were being routinely tested for HCV RNA concentration in the Department of Gastroenterology at the Lahey Clinic Medical Center. The majority of samples were collected from individuals undergoing standard therapy with interferon. Some additional samples were drawn from individuals who were participating in an interferon-ribavarin therapy study protocol. After consent was obtained from each patient, two tubes were drawn for the testing: one 10-ml non-SST and one 10-ml acid-citrate-dextrose-A tube (Vacutainer tubes; Becton Dickinson, Brea, Calif.). Samples were collected from July 1998 to March 1999. For some of the samples, a routine complete blood count (with differential) was performed as part of routine patient care. Hemoglobin and hematocrit levels were determined for these patients with a Coulter StakS Plus (Beckman-Coulter Corp., Irvine, Calif.).

Serum RNA extraction methods. The 10-ml non-SST tube was processed by centrifugation within 2 h of draw, and the serum was separated and frozen at 70°C for 1 to 6 days. The samples were thawed, and the RNA was extracted. Extraction from 100 µl of serum was performed with the EZNA total RNA kit (Omega Biotech, Norcross, Ga.) or with Trizol (Life Technologies, Rockville, Md.) (5), an improvement of the original Chomczynski and Sacchi method (6). Extraction from 100 µl of serum was performed with Trizol according to the manufacturer's instructions, with the exception that rRNA carrier was added to facilitate quantitative precipitation of the isolated RNA.

Whole blood RNA extractions. The acid-citrate-dextrose tube was processed in multiple ways. The samples were processed within 0 to 10 h after venipuncture. Three types of extractions were performed: Trizol extraction of RNA present in the WBC, solubilization of the whole blood with Catrimox-14 (Iowa Biotechnology, Ames, Iowa) followed by extraction of the RNA with Trizol according to previously published methods (16, 20), and extraction of RNA from 100 µl of whole blood with Trizol.

RNA extraction from WBC. First, 100 µl of blood was added to a 50-ml sterile centrifuge tube, and 9.0 ml of sterile 0.09% saline solution was added to hypotonically lyse the red blood cells. The tube was mixed vigorously for 20 s, and 1.0 ml of 9.0% saline solution was added to return the cells to a normal salt concentration. The tube was filled with an isotonic saline solution (Isoflow; Beckman-Coulter Corp.) and centrifuged at room temperature for 7 min at 1,000 rpm (2,500 × g) to pellet the WBC. A second hypotonic lysis was performed in an identical manner, and the pellet was resuspended in 500 µl of Isoflow and transferred to a 1.5-ml sterile microcentrifuge tube. An additional 500 µl of Isoflow was then used to wash and transfer the remaining cells. The tube was then centrifuged at 14,000 × g for 5 min to pellet the cells completely. The supernatant was removed, and 1.0 ml of Trizol was added and mixed thoroughly until all particulate matter was dissolved. The tube was immediately frozen at 70°C and held for up to 1 month.

RNA extraction from whole blood. For the Catrimox-Trizol method, 1.0 ml of Catrimox was put in a 1.5-ml sterile microcentrifuge tube, and 100 µl of whole blood was added. The sample was mixed and incubated at room temperature for 10 min and centrifuged at 14,000 × g for 5 min. The supernatant was removed, and the pellet was resuspended with 1.0 ml of DEPC water. The tube was vortexed to mix the contents and centrifuged again for an additional 5 min. The pellet was washed once more with an additional 1.0 ml of DEPC water, and the final pellet was resuspended in 1.0 ml of Trizol and mixed thoroughly to dissolve the contents completely. The sample was frozen immediately at 70°C and held for up to 1 month. For a limited number of samples prepared by the Catrimox-Trizol method, the extraction was performed according to the method described above, except the two washes of the pellet with DEPC water were omitted to ensure that no RNA was being lost during the washing steps. Quantitative results for these alternately processed samples were essentially the same as for the routine processing procedure (data not given). For the Trizol method, 1.0 ml of Trizol was placed in a 1.5-ml sterile microcentrifuge tube, and 100 µl of blood was added. The sample was mixed thoroughly until all particulate matter was dissolved. When all material was completely in solution, the tube was frozen immediately at 70°C and held for up to 1 month.

DNA sequencing. The PCR products were separated on agarose gels, bands were excised, and the DNA was released into 50 µl of DEPC water by a freeze-thaw method. The DNAs were subjected to direct cycle sequencing for each strand using an ABI Prism 377 DNA sequencer with each of the HCV primers and Big Dye terminators.

Calculations. The amount of HCV present in serum was adjusted to take into account the red cell volume using the hematocrit (HCT) for the correction factor as follows: [1  (HCT · 0.01)] × HCV copies/ml of serum = adjusted HCV copies/ml of serum. This calculation is appropriate if the assumption is made that essentially all virus is found in the fluid phase of the blood and that no significant amount of virus is present in the cells.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Thirty-nine samples were analyzed by a variety of RNA extraction methods for serum and whole blood samples. Of the 39 samples, 25 were positive, with a range of quantitative results for all extraction methods (Table 1). For 21 of the samples, the results of all of the different extraction methods were within 0.5 log unit of each other. The whole blood RNA extraction methods gave results for some of the samples that were slightly lower but within 0.5 log unit of the serum RNA extraction results. Four samples (patients 22 to 25) gave results that were more than 0.5 log unit different by the different extraction methods. Two of the samples, from patients 23 and 24, were >0.5 log unit lower by the Catrimox-Trizol method than by the other three methods. The sample from patient 22 was about 0.5 log unit higher by the EZNA method than by the other three methods.

For the samples with available hematocrit values, the adjusted levels of HCV for both the EZNA-treated serum and the Trizol-treated serum (Table 2) correlated slightly better with HCV levels obtained with the whole-blood RNA extraction methods (Table 1) than did the unadjusted HCV serum levels (Table 1). No significant differences in the quantitative results were seen with the different specimen types based on the red cell volume for the patient.

The final discrepant sample from patient 25 (Table 1) was weakly positive for 5,000 RNA copies/ml by the EZNA extraction method but was negative by all the other extraction methods, including the Catrimox-Trizol extraction from whole blood. This discrepant result from a patient who was weakly positive was particularly informative. This sample was taken from an individual in the initial stages of interferon therapy who had a good response, and the virus was rapidly cleared from the blood (Fig. 1). The sample was taken from the patient at a time when the viral titer was very low and just before the time when the viral titer in the serum became undetectable. At that time, the EZNA extraction was weakly positive at 5,000 copies/ml, whereas the virus was undetectable in the extractions obtained by the other three methods.

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Serial studies of HCV RNA quantitations during interferon-ribavarin therapy for one individual over the course of 6 months. The sample indicated by the arrow was studied by all of the extraction methods. The dates at the bottom indicate month/day. gE, genome equivalent.

All 13 of the remaining samples studied were determined to be negative by all of the extraction methods. The EZNA and Trizol serum RNA extraction methods were clearly negative with all 13 of these samples. Most of the 13 samples treated by the whole blood RNA extraction methods had significant amounts of contaminating PCR products that, when visualized on the agarose gel, gave bands that were both higher and lower than the expected molecular weight of the PCR product (Fig. 2). These artifact bands were not seen in any of the samples with positive serum results. Representative bands were excised from several of the lanes, and the PCR products were sequenced to determine their possible source. None of the sequences determined were consistent with either primer artifacts or HCV gene sequences. Significant sequence homologies were found with a variety of human cDNAs, including rRNA sequences. A human genome Basic Local Alignment Search Tool (BLAST; National Center for Biotechnology Information, National Institutes of Health, Bethesda, Md.) search of the primers used in this study and those of Schmidt et al. (15, 16) revealed 14- to 16-nucleotide matches for all primers with human sequences, but none of the matches included the 3'-terminal ends of the primers. Without the 3'-terminal end annealing, PCR amplification cannot proceed. From this BLAST analysis, there appear to be no potential differences among the abilities of any of the primers to anneal with sequences from human DNA contamination.

View larger version (75K): [in this window] [in a new window]   FIG. 2.   Agarose gel containing HCV RNA amplicon products. Four patient samples (lanes 1 to 3, patient 33; lanes 4 to 6, patient 35; lanes 7 to 9, patient 37; and lanes 10 to 11, patient 39) of either whole blood (lanes 1, 2, 4, 5, 7, 8, 10, and 11) or cells (lanes 3, 6, and 9) were treated by the Catrimox protocol (lanes 1, 4, 7, and 10) or with Trizol (lanes 2, 3, 5, 6, 8, 9, and 11). STD, lanes containing the 100-bp ladder. The sizes of the wild-type (WT) and mutant (21) HCV amplicon products are indicated.

Two strategies were used for a further determination of the presence or absence of HCV sequence in these negative samples. In an effort to eliminate the artifact bands, higher and lower annealing temperatures and variation of the magnesium concentration in the reaction buffers were tried (data not shown). Attempts to adjust the PCR conditions to increase the stringency without significantly decreasing the detection sensitivity minimized but did not completely eliminate these artificial bands. No conditions were found that eliminated the artifact bands and maintained the assay sensitivity. In additional studies with about half of the samples with prominent artifact bands, a small quantity of HCV-related sequence, the 21 standard, was added to the PCR mixture. The 21 standard was added to a copy number of 100, equivalent to an original serum concentration of 10,000 copies/ml. All samples with prominent artifact bands doped with 21 standard produced only the 21 amplicon product (Fig. 3). The results from the doping experiment demonstrated that, in the absence of any HCV product, a low level of priming and amplification was taking place with other human RNAs present in the WBC.

View larger version (67K): [in this window] [in a new window]   FIG. 3.   Agarose gel containing HCV RNA amplicon products amplified from whole blood specimens from 10 patients which were treated with Catrimox (lanes 1 to 11), an equal mix of wild-type and mutant HCV RNAs (lane 12), wild-type HCV RNA (lane 13), mutant HCV RNA (lane 14), and a water extraction blank control (lane 15). STD, lanes containing a 100-bp ladder. Lanes 1 through 10 also contained 100 copies of 21 RNA. The sizes of the wild-type (WT) and mutant (21) HCV amplicon products are indicated.

For a further comparison of the Catrimox-Trizol whole-blood RNA extraction method with the serum RNA extraction methods, a sample was obtained from a patient undergoing interferon therapy just as the HCV RNA level was becoming negative in the routine assay used in our clinical laboratory. Serial HCV quantitations conducted with the patient's serum performed with the routine serum RNA extraction method showed a response to the interferon therapy with a conversion from positive to negative at about 6 months after the initiation of the interferon therapy (Fig. 1). At low levels of viremia, frequent blood samples were studied. The whole blood sample (21 May 1999) treated with Catrimox-Trizol at the time HCV RNA was no longer detected in the serum by the routine assay was also negative.

All 14 patients with negative HCV RNA results were receiving therapy at the time of the study or had previously been treated with either interferon or interferon-ribivarin. About half of them had at least one other negative serum HCV RNA assay with a different draw date. For these patients, the time from initiation of therapy ranged from 4 weeks to 8 months, and the time since therapy was stopped ranged from 2 weeks to 2.5 years.

The results for the extraction of RNA from the isolated blood cells from the patients with positive serum and whole-blood results are shown in Table 3. Eleven of the samples with positive serum results had negative WBC RNA extractions (<2,500 copies/ml), and the remaining 14 showed levels of virus at least 2 to 3 log units lower than that found in the serum or blood. There was no correlation between the concentrations of HCV RNA in the serum and the WBC (r = 0.139). There was also no correlation between the concentration of HCV RNA in WBC and the total WBC count (r = 0.559) or absolute counts of neutrophils (r = 0.148), lymphocytes (r = 0.179), or monocytes (r = 0.101). All 14 samples that were negative in the serum and whole blood RNA extractions were also negative in the WBC RNA extractions.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

These studies demonstrate that the use of whole blood specimens for routine quantitation of HCV RNA in the blood did not provide any greater sensitivity than the use of serum specimens. In every specimen tested from patients with HCV infection in which HCV RNA became undetectable in the serum after therapy, the whole blood specimen was also negative. This was further validated by the careful longitudinal study of a patient during interferon therapy. At the point in treatment at which the HCV RNA became undetectable in the serum, it was also undetectable in the whole blood specimen.

Our findings from the studies of isolated WBC from infected sera were consistent with those of whole blood; insignificant amounts of HCV RNA were present in the cells. Moreover, there was no correlation of the amount of HCV RNA in the cellular fraction and the amount of HCV RNA in the serum, nor was there a correlation with the absolute number of neutrophils, lymphocytes, or monocytes.

The results obtained in our studies using the Catrimox extraction method differed greatly from previous results (15, 16, 20). The reason for this discrepancy is not readily apparent; however, in the previous studies, only a semiquantitative HCV RNA assay was employed rather than the quantitative assay used in our study. Because a precisely quantitative assay was not employed, the conclusion that whole blood specimens provided greater sensitivity than plasma specimens was based on the demonstration of HCV RNA in whole blood specimens in which the plasma specimens were negative rather than a systemic comparison of the two types of specimens for positive sera. The whole blood-positive, plasma-negative specimens in the study were from HCV antibody-negative patients, and studies of liver biopsy specimens for HCV RNA were not performed (16, 20). The implication that HCV was present in cells but not in the sera of seronegative patients is untenable without the demonstration of HCV RNA in the livers of these patients. In the only study (8) that has demonstrated HCV RNA in the liver for patients in whom HCV RNA was not detectable in the serum, all the patients were HCV antibody positive.

There is no definitive evidence that HCV can replicate in peripheral blood mononuclear cells (PBMC) in vivo, and the presence of high concentrations of HCV in the blood is prima facie evidence that HCV is not phagocytosed in any significant amounts by neutrophils. Hence, the notion that increased HCV concentrations in whole blood compared with those in plasma is the result of virus replicating in PBMC and/or phagocytosis by neutrophils is speculation. The findings that only a portion of serum HCV RNA-positive specimens in this study were HCV RNA positive in the cellular fraction and that the amounts of HCV RNA in the cellular fraction did not correlate with either serum HCV concentrations or the absolute number of any WBC fraction may be explained by the recent findings that HCV is endocytosed by means of the low-density lipoprotein (LDL) receptor (1). In that study, only very small amounts of HCV were endocytosed by PBMC, and the amount of HCV endocytosed could be increased by increasing LDL receptor expression on the cell membrane. Hence, the seemingly random presence of HCV in the cellular fraction in our study may only be apparent; the activation of PBMC that is known to up-regulate LDL receptors may be responsible.

Another notable observation in our study was the artifactual RT-PCR bands in the HCV-negative whole blood specimens. These bands were shown by DNA sequencing to be normal cellular RNA sequences, indicating that, in the presence of very high concentrations of cellular RNA and the absence of HCV sequences, the HCV-specific primers amplify non-HCV sequences. This phenomenon of promiscuous annealing of primers and subsequent amplification cannot be circumvented entirely by increasing the stringency of the RT-PCR. The poor quality of amplification resulting in many nonspecific bands, shown in Fig. 2 of this paper and in Fig. 3 and 5 of Schmidt et al. (15), exemplifies a further disadvantage of the whole blood methods compared with serum RNA extraction methods. Differences in primer selection and amplification strategies (nested versus direct) probably have little impact on the outcome because all of the primers had approximately equal sequence identity with human sequences in GenBank and neither strategy eliminated nonspecific bands. Sequencing these bands for routine confirmation would be an unacceptable practice because of considerations of time and expense. Hence, caution must be exercised in the use of whole blood in RT-PCR assays for HCV. In most cases, serum-based RT-PCR assays give similar and clearer data by using simpler extraction methods.