Main

Although HCV infection is rare in apparently healthy children(1), it is the major cause of posttransfusion hepatitis(2, 3). Our previous data demonstrated that the seroconversion rate of HCV in children who underwent open heart surgery was 4-5%, and there was no longer identified HCV infection after the screening of anti-HCV in the blood bank was begun(4). The clinical course of HCV infection varied. More than half of the children ran a chronic course, and their liver function was usually normal(5, 6). Thalassemic children must receive blood transfusions frequently, and they are at high risk for HCV infection. Since the advent of bone marrow transplantation, thalassemia is now a curable disease, and chronic HCV infection in thalassemic children has become an important issue(7). Previous studies on the occurrence of HCV infection in polytransfused thalassemic children are few and were performed by serologic tests only. From these limited studies, it was estimated that an anti-HCV prevalence rate in polytransfused thalassemic patients was 30-60% by a first generation of ELISA kit(8–10) and 90% by the more sensitive second generation kit(11). To better understand the prevalence of HCV infection in this area and the natural course of HCV infection in polytransfused thalassemic children, we longitudinally investigated the molecular virology of HCV in a group of thalassemic children. We aimed to elucidate the clinical significance of genotypes, viral titer, and genome evolution of HCV in the pediatric patient. A group of patients who were not HCV infected were also followed and acted as a control group. The information gained may provide important implications for patient management.

METHODS

Patients. A total of 61 thalassemic children were enrolled in the study with parental consent from June 1990 to February 1994; the total follow-up period was 44 mo. These children were regularly followed up in the thalassemia special clinic of this hospital and received regular transfusions(10 mL of packed red blood cells/kg), mostly every 2 or 3 wk. Blood samples were collected at the time of transfusion and stored at -70°C. Patients were divided into two groups: HCV-infected and HCV-noninfected. ALT was determined every 3 mo, unless abnormal data were shown, and then more frequent measurements were performed.

Biochemical and serologic tests and pathologic examinations. Liver function profiles, including ALT activity, were determined by an autoanalyzer (Hitachi 736, Tokyo). Anti-HCV status was assayed by a commercially available second-generation ELISA (hepatitis C II; Abbott Laboratories, North Chicago, IL). Histologic examinations were done with parental consents in those who had abnormal liver function and/or those who would undergo bone marrow transplantation. Liver biopsy was done by a modified Menghini biopsy set (Surecut, TSK Laboratory, Japan). The specimens were fixed in Bouin's solution and then processed for hematoxylin-eosin staining.

RT and nested PCR. All of the serum samples positive for anti-HCV were tested for HCV RNA by an RT-PCR procedure as described previously(6). The outer and inner pairs of primers were deduced from the 5′-noncoding region of the HCV genome. Their sequences and PCR conditions were described previously(4, 6). The nested PCR product was expected to be 158 bp in length. Strict procedures were followed to avoid contamination, as recommended by Kwok and Higuchi(12).

Genotyping. Serial serum samples with positive HCV RNA were subjected to genotyping by the method of Okamoto et al.(13). Briefly, the core region of the HCV genome was amplified by the first PCR. The product of the first PCR was then subjected to the second PCR with a 5′ common primer and a mixture of four 3′ primers for HCV genotypes I to IV. The HCV genome was typed by the different sizes of the second PCR products. Different sera of the patients were genotyped to determine whether there was any change.

Quantitation of HCV RNA by competitive PCR. The method was described in a previous report(14). Briefly, we first constructed a plasmid by inserting PCR fragments of the 5′-noncoding region (nucleotide position -341 to -1 of HCV) and the core region (nucleotide position 1 to 356 of HCV) into EcoRI and HindIII sites of pGEM-4 (Promega, Madison, WI), respectively, by blunt end ligation. Afterin vitro transcription, the final competitor, consisting of the 5′-noncoding region, followed by an internal 56-bp fragment derived from the pGEM-4 polylinker and the core sequence, was obtained. Various concentrations of competitor RNAs were mixed with the isolated seral RNA to run nested PCR. The sequence of primers and PCR conditions were described previously(14). The expected sizes of the nested PCR products for the wild type HCV and for the competitor RNA were 331 and 387 bp, respectively. After electrophoresis into a 2.5% agarose gel and by ethidium bromide staining, the amount of wild type cDNA was determined to be equivalent to the concentration of competitor cDNA. All tests were done in duplicate.

Cloning of HCV HVR. To examine the quasispecies nature of HCV in evolution, we cloned the HVR fragment of the HCV genome. RNA extraction from serum and RT-PCR procedures were the same as we stated above, but the primers used in PCR were different. The outer primers are: 5′-ATGGC, TTGGG, ACATG, ATGAT, GAACT, GGT-3′ (nucleotide position 952-979), 5′-GTAGT, GCCAG, CAATA, AGGCC-3′ (nucleotide position 1467-1448); the inner primers are: 5′-TTAGT, CGACT, GGGGA, GTCTG, GCGGGC-3′ (nucleotide position 1055-1074), the underlined sequence is the SalI site, and 5′-TTGCA, TGCCA, GCTGC, CATTG, GTGTT-3′ (nucleotide position 1263-1244), the underlined sequence is the SphI site. The expected size of the nested PCR product is 209 bp. The PCR product was purified by phenol/chloroform treatment and precipitated by ethanol; it was then subjected to SalI andSph I restriction enzyme digestion to expose the sticky end and ligated with pGEM-3Z with T4 DNA ligase (Boehringer Mannheim, Germany) at 16°C for more than 4 h. The colonies were selected and kept in one master plate for sequencing use.

Sequencing. The HVR fragment of HCV cDNA was sequenced by a cycle sequence system kit (Life Technologies, Inc., Gaithersburg, MD) based on the Sanger method. The sequencing primer was the aforementioned sense inner primer labeled with [γ-32P]ATP by T4 polynucleotide kinase. The sequencing reaction samples then were run on a 6% polyacrylamide gel. After electrophoresis, the film was dried and autoradiographed at-70°C.

Statistics. The Mann-Whitney-Wilcoxon test and χ2 exact test were used.

RESULTS

These 61 thalassemic children were divided into two groups: HCV-infected(n = 26, male:female = 13:13), mean age at enrollment = 16.4± 4.9 y; and HCV-noninfected groups (n = 35, male:female = 21:14), mean age at enrollment = 9.4 ± 5.2 y. To eliminate the age factor, we compared the ALT and number of transfusions between these two groups by age stratification (Table 1). The ALT and number of transfusions are significantly different (both p < 0.001) between these two groups at different ages.

Table 1 Age-stratified comparisons of ALT and number of transfusions
Full size table

In HCV-infected group, five contracted HCV infection during the time of follow-up as evidenced by newly appearing anti-HCV and HCV RNA. All five were infected before the initiation of anti-HCV screening of donated blood in July 1992 in Taiwan. Two of these patients were free of both anti-HCV and HCV viremia within 6 mo. The other three had persistent HCV viremia and high titer of anti-HCV. Their peak ALT levels were the highest among all patients. The other 21 patients had HCV infection from the time of their enrollment into this study. All except two patients had HCV viremia, a relatively high titer of anti-HCV, and an abnormal liver function. Eleven of the 21 patients had fluctuating ALT with at least two episodes of ALT surge, whereas eight patients had only one episode of abnormal liver enzymes, and two patients had normal levels during the follow-up period. Six patients in this group received a histologic examination, and five of them were found to have portal fibrosis in addition to hemochromatosis. Among these five cases, chronic active hepatitis and chronic persistent hepatitis were diagnosed in one case each according to the international classification with slight modification(15). None of these six patients was a HBsAg carrier.

In HCV-noninfected group, 10 of the 35 patients had abnormal ALT with the highest up to 163 IU/L, most were below three times the normal range. Three of these 10 were HBsAg positive, and the others had no serologic evidences of other hepatotropic virus infection. All of these cases had no HCV viremia, and their anti-HCV was negative. In this group, liver biopsies were performed in five patients for pre-bone marrow transplantation evaluation and in three for abnormal liver function. All eight biopsied patients were not HBsAg carriers, and two of the three with abnormal liver function had portal fibrosis. However, hemosiderin deposits in hepatocytes was always found.

Genotyping was performed in 19 patients who had at least two serial positive HCV viremia serum samples. Fifteen patients belonged to Okamoto type II/1b, two were type III/2a, and two were presumptive type IV/2b(16) (Fig. 1). None of the 19 patients showed a change in their genotype during the follow-up period. No mixed genotypes were found in our cases. Patients with different genotypes did not have statistically different peak ALT levels.

Figure 1
figure 1

Genotyping of HCV. The core region of HCV is amplified by the first PCR. The product of the first PCR was then subjected to nested PCR with a commmon primer of a 5′ end and a mixture of four 3′ primers specific for HCV Okamoto type I-IV. Electrophoresis of nested PCR products can distinguish genotypes by size. Lanes I-IV are positive controls for Okamoto type I-IV; N, negative control. Lanes 1-7 are samples; lane 2 belongs to type IV and lane 3 belongs to type III, whereas the others are type II. M, the marker is the DNA fragment of pBR328 digested byBgl I and pBR328 by HinfI.

Full size image

Quantitation of HCV RNA was performed in 10 patients because their HCV RNA had been positive for at least 2 y, thus we could study the relation between the quantity of HCV RNA and liver enzymes. The HCV RNA titer ranged from 1× 106 to 5 × 108 copies/mL as determined by comparing the intensity of the bands in gel electrophoresis(Fig. 2). The time course of HCV RNA versus ALT is listed in Table 2. We could demonstrate neither any significant fluctuation of HCV RNA titer during follow-up, nor any association between the HCV RNA titer and ALT level.

Figure 2
figure 2

Quantitation of HCV RNA by competitive RT-PCR. HCV RNA were extracted from the serum of four patients (1-4, labeled in the lower row) and mixed with four concentrations of competitive HCV RNA (5× 105 to 108 copies/mL, labeled in the upper row) in four different tubes, respectively, then subjected to RT-nested PCR. The concentration of serum HCV RNA is determined to be equal to competitive HCV RNA of similar band intensity. For example, in patient 1, the concentration of serum HCV RNA was determined to be 1 × 108 copies/mL because the intensity of the 331-bp band which is the wild type is less than the 5 × 108 copies/mL of competitor HCV RNA (387 bp) but greater than the 5× 107 copies/mL of competitor HCV RNA. M, size marker, the same as in Figure 1.

Full size image
Table 2 The time course of quantitation of HCV RNA vs ALT in 10 thalassemic children
Full size table

We randomly selected a patient to sequence the HVR of HCV cDNA. Eight clones (A1-A8) were obtained from her first serum samples, and another eight clones (B1-B8) were obtained from the serum collected 2 y later. ALT declined from 68 to 16 IU/L during this period (Fig. 3). The sequence of 209 bp in the HVR was read and compared with that of the Taiwan strain of HCV(17) (Fig. 4). The nucleotide sequences of A1-A8 are in 82-86% homology with the Taiwan strain(30-38 changes/209 bases), whereas those of B1-B8 are in 77-83% homology(37-47 changes/209 bases). Four clones (A4-A7) had identical sequences, whereas the others differed. The intrapatient variation of nucleotide sequences was 0-8%, namely, 10-13 nucleotide variations exist in A1-A8 clones. Among B1-B8 clones, the intrapatient interclonal variation was 1.5-11.9%, whereas B4-B8 were the major sequences. If we compare B4-B8 versus A4-A7, both are the dominant sequences in A and B clones, their nucleotide sequence differs in 5-7 bases, and the rate of variations would be 5-7 base differences/209 base of HVR/2 y = 1.2 × 10-2 to 1.7 × 10-2/nucleotide/y.

Figure 3
figure 3

ALT vs time course of the patient whose HCV RNA were sequenced. The A and B arrowheads indicate the points at which her sera were taken for cloning and sequencing.

Full size image
Figure 4
figure 4

Nucleotide sequences of the hypervariable region of HCV. The sequencing was done from nucleotide position 1055-1263, whereas only 1080-1240 is listed because there was no change in nucleotide positions 1055-1080 and 1241-1263. The Taiwan strain was used as a standard for comparison(17). Only the changed bases are listed; identical ones are replaced by a dash. A1-8 and B1-8 clones are explained in the text.

Full size image

DISCUSSION

It is important to evaluate the clinical features of HCV infection in the pediatric population, especially in terms of clarifying the natural course of chronic infection. Thalassemic children are good subjects for such a study because they require repeated blood transfusions. The older the patients, the more blood is transfused. The higher number of transfusions, the greater is the risk of HCV infection. During our 4-y follow-up period, most patients(24/26) in the HCV-infected group had elevated ALT levels, and more than half(14/24) had at least two episodes of ALT surges. In contrast, less than one-third of HCV-noninfected group had elevated ALT levels. Of course, as the thalassemia disease itself progressed, its complications of iron load including heart failure, diabetes mellitus, and hemochromatosis may all have contributed to the abnormal liver function. However, the age-stratified statistical method may help to eliminate those factors, and the results suggest that HCV infection is an independent important factor for liver impairment in thalassemic children. Histologic examinations also showed that liver damage is severer in the HCV-infected group than HCV-noninfected group, because more portal fibrosis was found in the HCV-infected group. Wonkeet al.(18) reported similar results in both pathology and ALT levels without stratifying the age factor in their studies. All five newly HCV-infected patients of the HCV-infected group contracted HCV before anti-HCV screening, supporting the effectiveness of anti-HCV screening(4). Three of the five (60%) ran a chronic HCV infection course, whereas two eliminated their viremia and anti-HCV. The chronicity rate is comparable with that reported in adults, which is from 50 to 86%(19, 20). Genotyping of HCV is important in tracing the infection sources and in establishing epidemiologic data, and it should be taken into account in vaccine development. It was also proposed that different genotypes varied in response to interferon therapy(21). In this country and other East Asia areas, Okamoto type II/1b is the most prevailing type, whereas type III/2a ranks second, type IV/2b is even less common, and type I/1a is almost absent(22, 23). Because nearly all of the transfusions were from domestic donors, we may expect that the recipients would have the same genotype distribution. Our study in these thalassemic children confirmed that type II/1b is the main HCV subtype in this area. However, we have two cases of type III/2a and presumptive IV/2b each. We cannot make assumptions based on this limited number of cases, but types other than II/1b and III/2a should not be neglected in this country. Different types may exhibit different biologic behavior, but we could not demonstrate a different peak ALT level among types II/1b, III/2a, and IV/2b.

Studies dealing with interferon-α treatment in HCV-infected children are limited in the literature(24). Quantitation of RNA by competitive PCR was initially demonstrated by Gilliland et al.(25). They created a restriction enzyme site by site-directed mutagenesis with RT-PCR. We have reported our method to produce the competitive HCV RNA by inserting a fragment of polylinker site of cDNA to avoid the complexity of site-directed mutagenesis(14). It has been proposed that HCV RNA titer may increase as the disease progresses(26). It has also been proposed that a high titer of HCV RNA is critical in determining mother-to-infant transmission(14, 27). Nevertheless, some authors question the importance of HCV RNA titer in the response to interferon-α and in the correlation of histologic grading(28). Our data failed to show a correlation between ALT change and quantities of HCV RNA. We found that HCV RNA did not fluctuate during the nearly 4-y follow-up period. The titer of HCV RNA of the thalassemic children is similar to that of adults, which is about 106 to 108 copies/mL. This may imply that HCV remains constant in serum for a period irrespective of the age of its host. However, even with such a titer of HCV RNA, the ALT level fluctuated mostly<100 IU/L as stated in Table 2. Although we did not perform a liver biopsy in these patients because there was no ALT surge, we speculate that no active hepatocellular damage is present, because the ALT is not high. Our previous reports also suggest that HCV-related liver damage may be relatively mild in children(4, 6)

The patient's serum samples (A and B) which we cloned were separated by 2 y. We found that a dominant sequence existed in each serum sample, A4-A7 and B4-B8. The sequence homology of the E2/NS1 region of HCV is 76-89% in the same type as reported by Cha et al.(29).

Our clones had 77-86% homology with the Taiwan strain. The diverse sequences of the clones in our study conform to the quasispecies nature of HCV(30). Kurosaki et al.(31) reported the variation rate of HVR in the flare-up stage (ALT 90-273 IU/L) of chronic HCV infection is 1.54 to 2.24 × 10-1/nucleotide/y, and that in the quiescent stage it is 0.13-1.21× 10-1/nucleotide/y. Our cases demonstrated 1.2-1.7 × 10-2/nucleotide/y. ALT was 68 IU/L then and decreased to 16 IU/L without flare-up during the 2 y. This means that, in the ALT quiescent stage of HCV infection, the variation rate is about this level. That four of the eight A clones have the same sequences, and five of the eight B clones have the almost same sequence, implies that a dominant sequence exists in the quasispecies. Whether the genetic evolution of HCV genome is a random event or a clinically relevant issue is an important but still controversial problem(32). In summary, HCV infection occurred in nearly half of the polytransfused thalassemic children. The older the age and the more blood was transfused, the higher chance the patients had to contract HCV infection. The HCV-infected patients had higher peak ALT levels than those without infection, even under the age-stratified statistics. Okamoto type II/1b is the mainstream in this area, and type III/2a and IV/2b are occasionally seen, whereas type I/1a is not detected in our cases. A higher titer of HCV RNA did not correlate with a higher ALT level. Their HCV RNA titer varied less than 100-fold in 4 y. The quasispecies nature of HCV is demonstrated, and the variation rate of HVR from a high ALT declining to a normal level would be 1.2-1.7 × 10-2/nucleotide/y, close to the previous report(28). All data obtained from thalassemic children delineate the basic molecular profiles of pediatric HCV infection and thus may contribute to the management of HCV.