Proc. Nati. Acad. Sci. USAVol. 89, pp. 187-191, January 1992Medical Sciences

Importance of primer selection for the detection of hepatitis Cvirus RNA with the polymerase chain reaction assay

(non-A, non-B hepatltis/5' noncoding sequence/genetic heterogeneity)

JENS BUKH, ROBERT H. PURCELL*, AND ROGER H. MILLERHepatitis Viruses Section, Laboratory of Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda,MD 20892

Contributed by Robert H. Purcell, October 2, 1991

ABSTRACT We compared four primer sets from con-served regions of the hepatitis C virus (HCV) genome for theirability to detect HCV RNA in a "nested" cDNA polymerasechain reaction assay on sera from 114 anti-HCV antibody-positive individuals from around the world. The differentprimer sets had equivalent sensitivity, detecting <1 chimpan-zee IDSO (dose that infects 50%) when tested against referencestrain H of HCV. We tested equal amounts of RNA extractedfrom the serum of each individual with the four primer sets.The set derived from two highly conserved domains within the5' noncoding (NC) region of the HCV genome, which also sharesignificant similarity with Pestivirus 5' NC sequences, was themost effective at detecting HCV RNA. AU samples positive forHCV RNA with any other primer set were also positive with theprimer set from the 5' NC region, and the latter was at least 3times more likely to detect HCV infection than a primer setfrom within the nonstructural protein 3-like gene region (P <0.001). We had no false positive results in >500 negativecontrols interspersed among the test samples. The 5' NC regionprimer set detected HCV-specific RNA, verified by high-stringency Southern blot hybridization and DNA sequencing,in 100% of 15 acute and 33 chronic non-A, non-B hepatitispatients from the United States, Europe, and Asia, and 10hepatocellular carcinoma patients from Africa and Asia thattested negative for the hepatitis B virus-encoded surface anti-gen. In conclusion, use of an appropriate primer set is crucialfor detecting HCV RNA in the serum of infected individuals.

Hepatitis C virus (HCV) is a positive-stranded RNA virusthat appears to be the etiological agent of most posttransfu-sion non-A, non-B (NANB) hepatitis cases (1-3). Amplifi-cation ofHCV RNA sequences by reverse transcription andcDNA polymerase chain reaction (cDNA PCR) is the onlypractical method currently available to demonstrate viremiain patients with HCV infection (3-14). However, geneticheterogeneity among different HCV strains (15-27) results infalse-negative results in cDNA PCR assays because ofprimerand template mismatch. Therefore, to develop a reliable PCRassay for the diagnosis of HCV infection, we designedprimers from conserved regions of the HCV genome (15,28-30) and tested their ability to detect HCV RNA in a nestedcDNA PCR assay in serum from 114 anti-HCV antibody-positive individuals from around the world. The resultsemphasize the importance of primer selection for the detec-tion of HCV RNA.

MATERIALS AND METHODSWe tested serum samples from 114 individuals from 12countries (Denmark, Dominican Republic, Germany, HongKong, India, Italy, Peru, South Africa, Sweden, Taiwan,

United States, and Zaire) who were anti-HCV antibody-positive by a first-generation anti-HCV test (2). The Germanand South African samples were selected for high titers ofanti-HCV from larger anti-HCV-positive groups.There were 33 patients with chronic NANB hepatitis

(Italy-12, Denmark-8, Taiwan-6, Hong Kong-5, and Sweden-2), 15 patients with acute NANB hepatitis (U.S.-12, Den-mark-2; and Sweden-1), 2 patients with chronic hepatitis whowere positive for hepatitis B virus-encoded surface antigen(HBsAg) (Denmark, Hong Kong), 3 HBsAg-negative pa-tients with liver cirrhosis (Denmark-2, Taiwan-i), and 17patients with hepatocellular carcinoma (HCC; South Africa-13, Taiwan-3, Hong Kong-1). The remaining 44 anti-HCV-positive individuals were found by screening high-risk pop-ulations and included 10 Peruvian health care workers, 14individuals with a history of sexual promiscuity (Zaire-9,Dominican Republic-5), 2 dialysis patients (Hong Kong), 2HBsAg carriers (Hong Kong, India), 1 patient with hemo-philia (Hong Kong), 1 Swedish blood donor who transmittedNANB hepatitis to a recipient, 6 commercial blood donors aswell as 2 blood transfusion recipients from India, and 6German recipients of anti-Rh immunoglobulin.To reduce the risk of contamination in PCR analysis, a

number of precautions were taken (31). Solutions and ali-quots were prepared in a laboratory that was never used forHCV research. Extraction of RNA and transfer of samplesfrom first to second PCR reaction tubes were performedunder sterile conditions with sterile reagents. These proce-dures as well as gel electrophoresis of the final PCR productwere performed in separate, self-contained areas that hadnever been exposed to recombinant HCV DNA. All solutionswere pipeted with positive displacement pipets or with dis-posable pipets or pipet tips containing an aerosol barrier. Inaddition, gloves were changed after each contact. A negativecontrol was included for every test sample to monitor forcontamination as a source of false-positive results.

All sera were stored at -800C after receipt in our labora-tory. RNA was extracted by a modification of the hot phenolmethod (32). Specifically, 100 1.l of serum or plasma wasmixed with 400 p.1 of solution A containing 4.2 M guanidiniumisothiocyanate, 25 mM Tris-HCI (pH 8.0), 0.5% sarkosyl,0.7% (vol/vol) 2-mercaptoethanol, and 50 ,ul of solution Bcontaining 1.0 M Tris HCI (pH 8.0), 0.1 M EDTA, and 10%(wt/vol) NaDodSO4, followed by extraction with an equalvolume of water-saturated phenol/chloroform, 1:1 (vol/vol)at 65°C for 30 min. The aqueous phase was removed andsaved. The organic phase (with the interphase) was extractedwith 300 j. of solution A at 65°C for 5 min. The two aqueous

Abbreviations: HCV, hepatitis C virus; NANB hepatitis, non-A,non-B hepatitis; PCR, polymerase chain reaction; NC, noncoding;NS3, nonstructural protein 3; SBH, Southern blot hybridization;HBsAg, hepatitis B virus-encoded surface antigen; HCC, hepato-cellular carcinoma; CID50, chimpanzee ID50 (dose that infects 50% ofchimpanzees).*To whom reprint requests should be addressed.

187

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

188 Medical Sciences: Bukh et al.

phases were pooled and extracted at room temperature withan equal volume of water-saturated phenol and chloroform(1:1) and then with chloroform alone. RNA was precipitatedby the addition of 400 p.l of 7.5 M ammonium acetate and 2volumes of isopropanol and incubation at -20TC for at least16 hr. After centrifugation, the RNA pellets were washedonce with 70%6 ethanol, dried at 37TC, and resuspended in 1001.l of water containing 10 mM dithiothreitol and 200 units ofhuman placental ribonuclease inhibitor (RNasin, Promega).This RNA solution was immediately divided into 10-pul ali-quots and stored at -80TC. In other studies, the RNA pelletswere resuspended in 10 p.l of the above solution, and theentire RNA sample was used for synthesis of cDNA.

Synthetic oligonucleotides, based on published HCV se-quences (15, 33), were synthesized on an Applied Biosystemsmodel 391 DNA synthesizer. Primer sets a-d were designedfrom conserved regions of the HCV genome (Table 1). Priorto cDNA synthesis, a 10-pl aliquot of the RNA solution wasthawed at room temperature, heated to 65TC for 2 min, chilledon ice, and used immediately. Synthesis of cDNA wasperformed in a 20-Al' reaction mixture containing 10 mMTris HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl2, 1 mM of eachof the four dNTPs (Pharmacia), 40 units of RNasin (Pro-mega), 2.5 gM of the reverse external primer, and 8 units ofavian myeloblastosis virus reverse transcriptase (Promega)with incubation at 43TC for 60 min. cDNA was denatured byadding 61.5 1.l of water and heating the mixture to 95TC for 5min. The first round ofPCR amplification (34) was performedin a reaction volume of 100 A.l containing 10mM Tris HCl (pH8.4), 50 mM KCl, 2.5 mM MgCl2, 0.2 mM of each of the fourdNTPs, 0.5 p.M of the external sense and antisense oligonu-cleotide primers, and 2.5 units of Taq DNA polymerase(Amplitaq, Perkin-Elmer/Cetus), which was overlaid with0.1 ml of molecular biology grade mineral oil (Sigma). Am-

plification was performed for 35 cycles in an automaticthermocycler (Perkin-Elmer/Cetus) with denaturation at940C for 1 min, primer annealing at 450C for 2 min, andamplification at 720C for 3 min. For the second round ofPCRamplification, a 10-pl aliquot of the first PCR reaction wasamplified as described above except for the substitution ofinternal sense and antisense primers. Finally, 10 p1 of thePCR product from the second round of amplification was

analyzed by electrophoresis on a 2% agarose gel, stainingwith ethidium bromide, and visualization under ultravioletlight.

Specificity of selected DNA bands was confirmed byhigh-stringency Southern blot hybridization (SBH) and bypartial sequence analysis. SBH ofDNA bands amplified withprimer set a was performed as described (35) except that thehybridization probe was a 5'-end-labeled oligonucleotide(Table 1) labeled with [_y-32P]dATP (Amersham; 3000 Ci/mmole, 1 Ci = 37 GBq) by T4 polynucleotide kinase(GIBCO/BRL), and the low-salt wash was performed atroom temperature. DNA for sequencing was purified byfractionation on agarose gels. DNA bands were excised andplaced in 10 mM Tris (pH 8)/1 mM EDTA to facilitatediffusion. Agarose was removed by filtration and ethidiumbromide was removed by 1-butanol extraction. The DNA wasextracted with phenol (2 times) and chloroform (1 time)followed by precipitation with ethanol. Approximately 100 ngof DNA was used for sequencing by the dideoxynucleotidechain-termination method (36) with phage T7 DNA polymer-ase (Sequenase, United States Biochemical).

Statistical analysis was performed to determine the signif-icance of the differences in the ability of the different primersets to detect HCV RNA. The binomial and McNemar testswere used (37).

Table 1. Synthetic oligonucleotide sequences

Primer set Nucleotide position (5'-3')* Oligonucleotide sequences (5'-3')ta

ExternalSense -285 to -256 ACTGTCTTCACGCAGAAAGCGTCTAGCCATAntisense -14 to -43 CGAGACCTCCCGGGGCACTCGCAAGCACCC

InternalSense -276 to -247 ACGCAGAAAGCGTCTAGCCATGGCGTTAGTAntisense -21 to -50 TCCCGGGGCACTCGCAAGCACCCTATCAGG

bExternalSense -323 to -294 GCGACACTCCACCATAGATCACTCCCCTGTAntisense 75 to 46 CGGGAACTTGACGTCCTGTGGGCGACGGTT

InternalSense -312 to -283 CCATGGATCCCTCCCCTGTGAGGAACTACTAntisense 69to 40 CTTGGGATCCTGTGGGCGACGGTTGGTGTT

c

ExternalSense -74 to -45 GTCGCGAAAGGCCTTGTGGTACTGCCTGATAntisense 497 to 468 GTTGCATAGTTCACGCCGTCTTCCAGAACC

InternalSense -66 to -37 AGGAATCCTGGTACTGCCTGATAGGGTGCTAntisense 442 to 413 CAGCGAATCCAAGAGGGGCGCCGACGAGCG

dExternal

Sense 3969 to 3998 CACATCCATCTTGGGCATCGGCACTGTCCTAntisense 4560 to 4531 GCACTCACAGAGGACGGACGAGTCGAACAT

InternalSense 3985 to 4014 ATCGQAATTCTCCTTGACCAAGCAGAGACTAntisense 4546 to 4517 CGGAGAATTCGAACATGCCGGAGGGGCGCT

Hybridizationprobe: -95 to -56 CTGCTAGCCGAGTAGTGTTGGGTCGCGAAAGGCCTTGTGG

*Map position 1 is the start of the genome polyprotein (15, 33).tNucleotides in boldface are restriction enzyme sites engineered into the primers to facilitate cloning.

Proc. Natl. Acad. Sci. USA 89 (1992)

Proc. Natl. Acad. Sci. USA 89 (1992) 189

RESULTSThe purpose of this study was to compare different primersets to determine whether a particular set was significantlybetter for detecting HCV RNA by PCR in sera from anti-HCV antibody-positive individuals from around the world.First, we analyzed the available HCV genomic sequences(15-28, 30) and found well-conserved domains in the virus 5'NC and protein-encoding gene regions. The 5' NC region, themost conserved region of the virus genome, possessed twodomains that shared statistically significant similarity toPestivirus 5' NC sequences (28, 30). The structural andnonstructural regions were found to have various degrees ofsequence heterogeneity, with the putative core and nonstruc-tural protein number 3 (NS3)-like region being the mostconserved. The protein-encoding gene regions did not pos-sess statistically significant similarity to Pestivirus se-quences. We selected four well-conserved regions of theHCV genome and designed four primer sets (i.e., a set is twopairs ofnested primers) for comparison in acDNA PCR assay(Table 1). Primer set a was from the two highly conserveddomains within the 5' NC region of the HCV genome. Primerset b spanned the 5' end of the 5' NC region to the 5' end ofthe putative core gene sequence. Primer set c was derivedfrom the 3' end of the 5' NC region to the 3' end of the coregene region, while primer set d was derived from within aconserved domain of the NS3-like gene of HCV. To assessthe relative sensitivity of the four primer sets, we analyzed10-fold serial dilutions of HCV RNA purified from HCVreference strain H (14, 30), which has an infectivity titer of106 5 CID50 doses/ml (a CID50 is the dose that infects 50%o ofchimpanzees). We found that all four primer sets could detectHCV RNA at a dilution of 10-7 but not at 10-8 in our assay.Thus, the sensitivity of all primer sets was <1 CID50. Next,we tested the ability of the four primer sets to detectPestivirus sequences. Only primer set a was capable ofdetecting bovine viral diarrhea virus RNA (strain NADL,American Type Culture Collection) in our cDNA PCR assay.This was an expected finding since primer set a was designedto correspond to the domain of the HCV 5' NC region thatshared sequence similarity with animal Pestiviruses. Todistinguish genuine HCV sequences from putative humanPestivirus sequences (38) that might be amplified by primerset a, we hybridized PCR products with an oligonucleotideprobe specific for HCV but not Pestivirus sequences by SBH.We tested sera from a total of 114 anti-HCV-positive

individuals from 12 countries for the presence ofHCV RNAusing primer sets a-d in a cDNA PCR assay (Tables 2 and 3).No false-positive results were found in >500 controls in theexperiment. Primer set a was the most effective for detectingHCV RNA. A total of 84 (74%) of the 114 samples werepositive with this primer set (Table 2). The DNA bands fromall 84 samples hybridized to our HCV-specific oligonucleo-tide probe under stringent conditions, thereby confirming thepresence ofgenuine HCV RNA sequences in the test sera. Ofthe 84 positive samples, 80 were positive with at least oneother primer set (Table 3). The DNA bands from the foursamples that had HCV RNA detectable only with primer seta were sequenced and each was found to represent a uniqueHCV sequence. Primer sets b and c were less effective at

Table 2. Detection of HCV RNA in 114 anti-HCV antibody-positive individuals from around the worldPrimer HCV RNA-positive,

set Primer location no. (%)a Within 5' NC 84 (74)b 5' end of 5' NC to 5' end of core 77 (68)c 3' end of 5' NC to 3' end of core 59 (52)d Within NS 3 27 (24)

detecting HCV RNA, with 68% and 52% of the sera testingpositive, respectively. Primer set d was the least effectiveprimer set tested and detected HCV RNA in only 24% of thesera. Differences could not be explained by loss of RNA,since under the storage conditions used, no degradation ofRNA occurred during the course of the experiments (unpub-lished data). Thus, primer set a was at least 3 times morelikely to detect HCV RNA than primer set d in our study.

Analysis of the results obtained from patients with liverdisease revealed the sensitivity of our PCR test, both in thestandard assay (i.e., testing RNA extracted from the equiv-alent of 10 AI ofserum) and in a more sensitive assay that usedthe most effective primer set (a) to detect HCV RNAextracted from 100 A.l of sera (Table 4). In our cDNA PCRassay using primer set a and RNA extracted from 10-100 Ilof serum, we were able to detect HCV RNA in 100% ofanti-HCV-positive patients with acute or chronic NANBhepatitis from the United States, Europe, and Asia, as wellas HBsAg-negative HCC patients from Africa and Asia.Interestingly a patient with unspecified liver cirrhosis andanother patient with biopsy-proven alcoholic liver cirrhosiswere both HCV RNA-positive. In a third patient, withinactive liver cirrhosis, HCV RNA could not be detected.This initially anti-HCV-positive patient was later found to benegative with other commercially available anti-HCV assaysand most likely had had autoimmune liver disease.

DISCUSSIONThere are several important factors to consider when com-paring primer sets in a cDNA PCR assay for their ability todetect HCV RNA in sera from different regions of the world.First, the genomic sequences used to design the primers mustbe well conserved to ensure optimal sensitivity. We designedour primer sets to correspond to regions of the HCV genomethat are conserved in known HCV sequences. Second, it wasnecessary to demonstrate that all of the primer sets can detectan equivalent amount of HCV RNA under standard condi-tions. We used the well-characterized pool of HCV strain H(14, 30) to show that our four primer sets could detect HCVRNA at a level of <1 CID50. A third important factor was theelimination of false positive results caused by contaminationof samples with exogenous HCV RNA. We paired each of the114 sera with a negative control in the RNA extractionprocedure and then performed each of the four PCR assaysin parallel with a negative control. We did not have a singlefalse-positive result in over 500 control samples tested. Also,SBH and DNA sequence analysis were used to verify thatselected DNA bands, identified by agarose gel electropho-resis, represented unique HCV sequences. The final factorconsidered was that each of the four PCR reactions for agiven serum sample should be performed on an equivalentamount of extracted RNA. With these conditions met, webelieve that the results of the comparison accurately reflectthe ability of the four primer sets to detect HCV RNA in seracollected worldwide.

Several groups have examined the effectiveness of differ-ent primer sets in detecting HCV RNA by PCR (4, 6, 13, 14,34, 39-41). The most comparable study, by Okamoto andcoworkers (4), examined 10 sets of primers and found thatthose derived from the 5' NC and the putative core generegion were equivalent in their ability to detect HCV RNA in10 anti-HCV-positive sera from Japanese individuals. Otherstudies have demonstrated enhanced detection ofHCV RNAwith primers from the 5' NC region compared with thosespecific for other regions (6, 39, 40). Although these studiesdid not specifically address the factors discussed above, theresults suggest that primers from the 5' end of the HCVgenome are effective at detecting HCV RNA by PCR anal-ysis.

Medical Sciences: Bukh et al.

190 Medical Sciences: Bukh et al.

Table 3. Primer set comparison by country

No. with individual pattern of response to primers

abcd abcd abcd abcd abcd abcd abcd abcdCountry No. tested ++++ +++- ++-- +--- ++-+ +-++ +-+-

Denmark 13 4 4 1 1 - 3Sweden 4 3 1 -

Germany 6 3 2 1Italy 12 3 4 4 1United States 12 4 1 1 1 1 4Dominican Republic 5 4 - 1Peru 10 1 1 - 8South Africa 13 9 1 1 - 2Zaire 9 1 4 1 3Hong Kong 11 4 3 2 2Taiwan 10 1 5 2 1 1India 9 1 2 - 6Total 114 23 33 18 4 3 1 2 30

The goal of our investigation was to identify a primer setthat had the lowest false-negative rate possible for thedetection of HCV RNA worldwide in a cDNA PCR assay.Such a primer set would have to be specific for a conservedregion of the virus genome. Therefore, we examined the 5'NC region, already known to be the most conserved regionof the HCV genome, for domains that were completelyconserved among available HCV sequences but also wereconserved among available Pestivirus 5' NC sequences. Ourrationale was that such sequences are likely to representbiologically important domains conserved in virus evolutionand should be identical, or nearly identical, in all HCVgenomes. The primer set derived from these conserveddomains, primer set a, was the most effective set in detectingHCV RNA. HCV RNA was found in 84 (74%) of the seratested from 114 anti-HCV-positive individuals. It was im-pressive that all samples that scored positive for HCV RNAwith any of the other three primer sets also scored positivewith primer set a. SBH and DNA sequence analysis dem-onstrated that all 84 samples possessed authentic HCVsequences. This indicates that putative human Pestivirussequences were not present. Primer sets b and c, derivedfrom different domains within both the 5' NC and theconserved core gene region (Tables 1 and 2) were able todetect HCV RNA in 92% and 70%, respectively, of the serathat were positive with primer set a. However, primer set d,derived from the NS3-like region, detected HCV RNA inonly 32% ofthose sera positive for HCV RNA with primer seta. The difference between detection of HCV RNA usingprimer set a and the other three primer sets is statisticallysignificant: with primer set b, the difference is significant atthe P = 0.016 level, whereas with primer sets c and d, thedifference is significant at the P < 0.001 level. Our resultsconfirm the importance of using primer sets from the mostconserved region of the HCV genome to avoid false-negativeresults caused by mismatch between primer and template

resulting from the genetic heterogeneity (nucleotide substi-tutions, deletions, or insertions) among HCV strains.

Titration studies in chimpanzees have shown that theinfectivity titer of HCV in patients with acute or chronicNANB hepatitis can be low, and the level ofHCV RNA canbe at or below the detection limit (<30 CID50/ml) of anycDNA PCR assay using 10 jl of serum. Therefore, a secondaspect of this study was to perform PCR analysis with ourmost effective primer set on RNA extracted from 100 Al ofselected sera. We found that there were five additionalpatient sera that possessed detectable HCV RNA: onechronic and four acute NANB hepatitis patients. However,there were also five anti-HCV-positive sera that did notcontain detectable HCV RNA even after analysis of RNAfrom 100 ul of serum using primer set a (see Results). Thelack of detection of virus RNA in these cases could resultfrom: (i) a level of virus below the detection limit of ourassay, (ii) the fact that the presence of anti-HCV was from aprevious infection, (iii) nonspecificity of the antibody test(42, 43), or (iv) loss of virus RNA due to poor handling ofsera. We conclude that, for optimal detection of serum HCVRNA, primer set a should be used in a cDNA PCR assay onRNA extracted from 100 ,ul or more of serum.Beyond the comparison of primer sets, we have made

several noteworthy observations that have clinical signifi-cance. We found that 100% of patients with acute or chronicNANB hepatitis from the United States, Europe, and Asia,as well as HBsAg-negative HCC patients from Africa andAsia, were HCV RNA-positive in our cDNA PCR assay withprimer set a. First, we were able to detect HCV RNA in all33 chronic NANB hepatitis patients examined from Denmark(8 patients), Sweden (2 patients), Italy (12 patients), HongKong (5 patients), and Taiwan (6 patients). One other study(4) has reported a similar high frequency ofdetection ofHCVRNA in a large number of patients from a single country (i.e.,Japan). Taken together, these data clearly suggest that HCV

Table 4. Detection of HCV RNA in anti-HCV antibody-positive individuals with or without liver disease

Assay of RNA from 10 Retest of RNA from 100 HCVAl of serum 1.l of negative samples RNA-positive,

Diagnosis No. tested Positive Negative Positive Negative total no. (%)

Acute NANB hepatitis 15 11 4 4 0 15 (100)Chronic NANB hepatitis 33 32 1 1 0 33 (100)HCC; HBsAg- 10 10 0 10 (100)Liver cirrhosis; HBsAg- 3 2 1 0 1 2 (67)Chronic hepatitis; HBsAg' 2 1 1 0 1 1 (50)HCC; HBsAg' 7 4 3 0 3 4 (57)No recognized liver disease 44 24 20 NT NT

NT, not tested.

Proc. Natl. Acad. Sci. USA 89 (1992)

Proc. Natl. Acad. Sci. USA 89 (1992) 191

is the causative agent in all, or nearly all, of the cases ofchronic NANB hepatitis in many parts of the world. Otherstudies (3, 5, 7, 22, 24, 41) of patients with chronic NANBhepatitis reported lower frequencies of the presence of HCVRNA. It is likely that there was a high rate of false negativeresults in the latter studies because primers were used thatwere derived from more heterogeneous regions of the HCVgenome. Second, we detected HCV RNA in the serum of all15 acuteNANB hepatitis patients examined from Denmark (2patients), Sweden (1 patient), and the United States (12patients), suggesting that HCV is also the major causativeagent in this group ofpatients in the countries studied. Similardata have been reported by others (8, 14). Third, 14 of 17HCC patients (13 from South Africa, 1 from Hong Kong, and3 from Taiwan) were positive for HCV RNA in our PCRassay. In fact, all 10 HBsAg-negative patients (8 from SouthAfrica and 2 from Taiwan) were HCV RNA-positive. Thesedata confirm that HCV may play a major role in the devel-opment of HCC (7, 22).

In the diagnosis ofHCV infection by PCR, both sensitivityand specificity of the assay are of crucial importance (44, 45).We find that careful attention to the selection of primer setsis essential for optimal detection of serum HCV RNA.

We thank Ms. T. Tsarev for synthesis of oligonucleotides, Ms. D.Wong for serological analysis, Dr. D. Ailing for assistance with thestatistical analysis, Drs. P. Farci and N. Ogata for helpful discussionson HCV RNA extraction and PCR methodology, and Dr. M. Kew forcomments on the manuscript. We gratefully acknowledge the gener-osity of the following investigators for providing sera for this study:Drs. H. Alter, V. Arankalle, D.-S. Chen, P. Farci, M. Kew, K.Krogsgaard, A. Lok, T. Quinn, F. Renger, M. Sjogren, and A. Widell.

1. Choo, Q.-L., Kuo, G., Weiner, A. J., Overby, L. R., Bradley,D. W. & Houghton, M. (1989) Science 244, 359-362.

2. Kuo, G., Choo, Q.-L., Alter, H. J., Gitnick, G. L., Redeker, A. G.,Purcell, R. H., Miyamura, T., Dienstag, J. L., Alter, M. J., Ste-vens, C. E., Tegtmeier, G. E., Bonino, F., Colombo, M., Lee,W.-S., Kuo, C., Berger, K., Shuster, J. R., Overby, L. R., Bradley,D. W. & Houghton, M. (1989) Science 244, 362-364.

3. Weiner, A. J., Kuo, G., Bradley, D. W., Bonino, F., Saracco, G.,Lee, C., Rosenblatt, J., Choo, Q.-L. & Houghton, M. (1990) Lancet335, 1-3.

4. Okamoto, H., Okada, S., Sugiyama, Y., Tanaka, T., Sugai, Y.,Akahane, Y., Machida, A., Mishiro, S., Yoshizawa, H., Miyakawa,Y. & Mayumi, M. (1990) Jpn. J. Exp. Med. 60, 215-222.

5. Kato, N., Yokosuka, O., Omata, M., Hosoda, K. & Ohto, M. (1990)J. Clin. Invest. 86, 1764-1767.

6. Shibata, M., Morishima, T., Kudo, T., Maki, T., Maki, S. & Nagai,Y. (1991) Blood 77, 1157-1160.

7. Ohkoshi, S., Kato, N., Kinoshita, T., Hijikata, M., Ohtsuyama, Y.,Okazaki, N., Ohkura, H., Hirohashi, S., Honma, A., Ozaki, T.,Yoshikawa, A., Kojima, H., Asakura, H. & Shimotohno, K. (1990)Jpn. J. Cancer Res. 81, 862-865.

8. Zonaro, A., Puoti, M., Fiordalisi, G., Mantero, G., Castelnuovo,F., Primi, D. & Cariani, E. (1991) J. Infect. Dis. 163, 923-924.

9. Garson, J. A., Tedder, R. S., Briggs, M., Tuke, P., Glazebrook,J. A., Trute, A., Parker, D., Barbara, J. A. J., Contreras, M. &Aloysius, S. (1990) Lancet 335, 1419-1422.

10. Garson, J. A., Tuke, P. W., Makris, M., Briggs, M., Machin, S. J.,Preston, F. E. & Tedder, R. S. (1990) Lancet 336, 1022-1025.

11. Simmonds, P., Zhang, L. Q., Watson, H. G., Rebus, S., Ferguson,E. D., Balfe, P., Leadbetter, G. H., Yap, P. L., Peutherer, J. F. &Ludlam, C. A. (1990) Lancet 336, 1469-1472.

12. van der Poel, C. L., Cuypers, H. T. M., Reesink, H. W., Weiner,A. J., Quan, S., di Nello, R., van Boven, J. J. P., Winkel, I.,

Mulder-Folkerts, D., Exel-Oehlers, P. J., Schaasberg, W., Leen-tvaar-Kuypers, A., Polito, A., Houghton, M. & Lelie, P. N. (1991)Lancet 337, 317-319.

13. Ulrich, P. P., Romeo, J. M., Lane, P. K., Kelly, I., Daniel, L. J. &Vyas, 0. N. (1990) J. Clin. Invest. 86, 1609-1614.

14. Farci, P., Alter, H. J., Wong, D., Miller, R. H., Shih, J. W., Jett,B. & Purcell, R. H. (1991) N. Engl. J. Med. 325, 98-104.

15. Okamoto, H., Okada, S., Sigiyama, Y., Yotsumoto, S., Tanaka, T.,Yoshizawa, H., Tsuda, F., Miyakawa, Y. & Mayumi, M. (1990)Jpn. J. Exp. Med. 60, 167-177.

16. Takamizawa, A., Mori, C., Fuke, I., Manabe, S., Murakami, S.,Fujita, J., Onishi, E., Andoh, T., Yoshida, I. & Okayama, H. (1991)J. Virol. 65, 1105-1113.

17. Kato, N., Hijikata, M., Ootsuyama, Y., Nakagawa, M., Ohkoshi,S., Sugimura, T. & Shimotohno, K. (1990) Proc. Nati. Acad. Sci.USA 87, 9524-9528.

18. Choo, Q.-L., Richman, K. H., Han, J. H., Berger, K., Lee, C.,Dong, C., Gallegos, C., Coit, D., Medina-Selby, A., Barr, P. J.,Weiner, A. J., Bradley, D. W., Kuo, G. & Houghton, M. (1991)Proc. Natd. Acad. Sci. USA 88, 2451-2455.

19. Takeuchi, K., Kubo, Y., Boonmar, S., Watanabe, Y., Katayama,T., Choo, Q.-L., Kuo, G., Houghton, M., Saito, I. & Miyamura, T.(1990) J. Gen. Virol. 71, 3027-3033.

20. Hijikata, M., Kato, N., Mori, S., Ootsuyama, Y., Nakagawa, M.,Sugimura, T., Ohkoshi, S., Kojima, H., Meguro, T., Taki, M.,Takayama, S., Yamada, K. & Shimotohno, K. (1990) Jpn. J. CancerRes. 81, 1195-1197.

21. Takeuchi, K., Boonmar, S., Kubo, Y., Katayama, T., Harada, H.,Ohbayashi, A., Choo, Q.-L., Kuo, G., Houghton, M., Saito, I. &Miyamura, T. (1990) Gene 91, 287-291.

22. Kato, N., Hijikata, M., Ootsuyama, Y., Nakagawa, M., Ohkoshi,S. & Shimotohno, K. (1990) Mol. Biol. Med. 7, 495-501.

23. Kato, N., Ohkoshi, S. & Shimotohno, K. (1989) Proc. Jpn. Acad.65, 219-223.

24. Enomoto, N., Takada, A., Nakao, T. & Date, T. (1990) Biochem.Biophys. Res. Commun. 170, 1021-1025.

25. Hoikata, M., Kato, N., Ootsuyama, Y., Nakagawa, M., Ohkoshi,S. & Shimotohno, K. (1991) Biochem. Biophys. Res. Commun. 175,220-228.

26. Weiner, A. J., Brauer, M. J., Rosenblatt, J., Richman, K. H.,Tung, J., Crawford, K., Bonino, F., Saracco, G., Choo, Q.-L.,Houghton, M. & Han, J. H. (1991) Virology 180, 842-848.

27. Kubo, Y., Takeuchi, K., Boonmar, S., Katayama, T., Choo, Q.-L.,Kuo, G., Weiner, A. J., Bradley, D. W., Houghton, M., Saito, I. &Miyamura, T. (1989) Nucleic Acids Res. 17, 10367-10372.

28. Han, J. H., Shyamala, V., Richman, K. H., Brauer, M. J., Irvine,B., Urdea, M. S., Tekamp-Olson, P., Kuo, G., Choo, Q.-L. &Houghton, M. (1991) Proc. Natl. Acad. Sci. USA 88, 1711-1715.

29. Miller, R. H. & Purcell, R. H. (1990) Proc. Natl. Acad. Sci. USA87, 2057-2061.

30. Ogata, N., Alter, H. J., Miller, R. H. & Purcell, R. H. (1991) Proc.Natd. Acad. Sci. USA 88, 3392-3396.

31. Kwok, S. & Higuchi, R. (1989) Nature (London) 339, 237-238.32. Maniatis, T., Fritsch, E. F., Sambrook, J., eds. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY) pp. 194-195.

33. Hougton, M., Choo, Q.-L. & Kuo, G. (1988) Eur. Patent Appl.88,310,922.5 and Publ. 0318,216.

34. Kaneko, S., Unoura, M., Kobayashi, K., Kuno, K., Murakami, S.& Hattori, N. (1990) Lancet 335, 976.

35. Miller, R. H., Lee, S.-C., Liaw, Y.-F. & Robinson, W. S. (1985) J.Infect. Dis. 151, 1081-1092.

36. Bachmann, B., Luke, W. & Hunsmann, G. (1990) Nucleic AcidsRes. 18, 1309.

37. Siegel, S., ed. (1956) in Non-parametric Statistics (McGraw-Hill,New York), pp. 35-94.

38. Yolken, R., Dubovi, E., Leister, F., Reid, R., Almeido-Hill, J. &Santosham, M. (1989) Lancet i, 517-520.

39. Garson, J. A., Ring, C., Tuke, P. & Tedder, R. S. (1990) Lancet336, 878-879.

40. Tedder, R. S., Gilson, R. J. C., Briggs, M., Loveday, C., Cameron,C. H., Garson, J. A., Kelly, G. E. & Weller, I. V. D. (1991) Br.Med. J. 302, 1299-1302.

41. Cristiano, K., DiBisceglie, A. M., Hoofnagle, J. H. & Feinstone,S. M. (1991) Hepatology 14, 51-55.

42. Wong, D. C., Diwan, A. R., Rosen, L., Gerin, J. L., Johnson,R. G., Polito, A. & Purcell, R. H. (1990) Lancet 336, 750-751.

43. McFarlane, I. G., Smith, H. M., Johnson, P. J., Bray, G. P., Ver-gani, D. & Williams, R. (1990) Lancet 335, 754-757.

44. Alberti, A. (1991) J. Hepatol. 12, 279-282.45. Houghton, M., Weiner, A., Han, J., Kuo, G. & Choo, Q.-L. (1991)

Hepatology 14, 381-388.

Medical Sciences: Bukh et al.