Introduction

Waterborne spread of schistosomiasis has prevailed in the Nile delta in Egypt, mediated by small shellfish that play the natural host. Cross-sectional studies conducted in sera from blood banks as well as community-based studies (Darwish et al. 1993; Kamel et al. 1992; Hibbs et al. 1993; Nafeh et al. 2000) suggested an association between self-reported parenteral treatment for schistosomiasis and hepatitis C virus (HCV). To cope with this endemic, more than 2 million intravenous injections with tartar emetic (potassium antimony sulfate) were given to 250,000 patients per year during 1964 through 1982 (Frank et al. 2000). Association is suggested between parenteral treatment and HCV for the highest endemicity of HCV in the world ever, involving >10% of the national population in Egypt. The prevalence of antibody to HCV (anti-HCV) is extremely high in patients with parenteral treatment for schistosomiasis (Frank et al. 2000; Hibbs et al. 1993; Kamel et al. 1992). These observations were further corroborated by Rao et al. (2002), who reported clustering of HCV infections and parenteral treatment for schistosomiasis within the same households, which provides further evidence for an association between the schistosomiasis treatment campaigns and high rates of HCV seroprevalence currently observed in the Nile delta of Egypt.

In Egypt, a high homology of HCV subtypes, 90% of which is accounted for by genotype 4a, strongly suggests a local epidemic of HCV throughout the general population (Dusheiko et al. 1994; McOmish et al. 1994; Ray et al. 2000). To elucidate the population dynamic history of HCV, such a unique setting is very suitable because it allows for a viral population study based on the neutral theory (Stumpf and Pybus 2002). By means of this theory, the history of changes in effective population size of a target virus, which is proportional to the diversity among a limited number of isolates, can be examined; the population size is estimated from a phylogenetic tree reconstructed by nucleotide sequences of the virus genome in the context of an accurate molecular clock (Pybus et al. 2000, 2001; Pybus and Rambaut 2002; Tanaka et al. 2002). Recently, Pybus et al. (2003) adopted such an indirect way of addressing the sampled sequences of HCV E1 genes in Egypt. In this study, we also adopted this framework to investigate the demographic history of HCV in Egypt on the basis of sequences of independent HCV NS5B genes and found the growth of HCV-4a to have increased exponentially between the 1940s and 1980 in Egypt.

Materials and Methods

Sample Collection

In Egypt during 1999, 317 serum samples positive for anti-HCV were screened of 3608 blood donors at 13 governorates in four districts [Central (Gharbiya, Kalyobiya, Monofiya, Beheira, Alexandria, Kafr El-Sheikh), North (Dakahliya, Dumietta), East (Sharkiya, Ismailiya, Port Said, Sinia) and South (Assuit)]. Serum samples were tested by a second-generation immunoassay (HCV EIA 2.0 enzyme; Abott Diagnostika, Wieabaden, Germany). The study protocol conformed to the 1975 Declaration of Helsinki and was approved by ethic committees of the institutions. All blood donors gave informed consents for the virological research of HCV from their sera.

Genotyping and Sequencing

Nucleic acids were extracted with use of a SepaGean RV-R Nucleic acid extracting kit (Sanko Junyaku Co., Ltd., Tokyo) in accordance with the manufacturer’s protocol. They were reverse-transcribed to cDNA using Superscript II Rnase H– Reverse Transcriptase (Invirogen Corp., USA) and random hexamer primer (TaKaRa Shuzo Co. Ltd., Tokyo) by the method described previously (Ohno et al. 1997).

The cDNA obtained was amplified by polymerase chain reaction (PCR) with primers deduced from the 5-noncoding region according to the reported protocol (Takeuchi et al. 1999). Genotypes were determined by PCR with the mixture of primers deduced from the core gene and specific for one or more of 1a, 1b, 2a, 2b, 3a, 3b, 4, 5a, and 6a (Ohno et al. 1997).

A sequence spanning 339 nucleotides (nt) in the NS5B region was amplified by PCR with previously described primers (Tanaka et al. 2002). PCR products were directly sequenced with Prism Big Dye (Applied Biosystems, USA) in the ABI 3100 DNA automated sequencer. To reduce the number of artificial substitutions arising during PCR, PLATINUM Pfx DNA Polymerase (Invirogen Corp.) with a very high fidelity was used. The sequences determined were used to confirm HCV genotypes and construct phylogenetic trees.

Demographic Modeling

A reconstructed tree was built on the NS5b sequence of 339 nt by a heuristic maximum-likelihood topology search with the stepwise addition and nearest-neighbor interchange algorithms. Tree likelihood scores were calculated using HKY85 with the molecular clock enforced by PAUP version 4.0b8. As estimates of demographic history, a nonparametric function N(t), called the skyline plot, was obtained by transforming the coalescent intervals of an observed genealogy into a piecewise plot that represents an effective population size through time. A parametric maximum likelihood was estimated by several models in the computer software Genie version 3.5 to set up a statistical framework for estimating the demographic history of a population on the basis of phylogenies reconstructed on sampled DNA sequences (Pybus and Rambaut 2002). The coalescent theory examines the history of changes in virus effective population size, proportional to the diversity among a population, based on a phylogenetic tree reconstructed by nucleotide sequences of the virus genome in the context of an accurate molecular clock.

Results

During 1999, 3608 blood donors with a mean age of 34.4 ± 9.5 years including 2699 (74.8%) males, from 13 governorates in four districts in or surrounding the Nile valley, were tested for anti-HCV, and 317 of them (8.8%) with a mean age of 38.8 ± 9.0 years gave positive results by a second-generation immunoassay. HCV RNA was detected in 225 of the 317 (71.0%) donors with anti-HCV (Table 1). The prevalence of anti-HCV in the Nile delta (Central and North regions) was significantly higher than that in the other regions (12.4 vs. 5.2%; p < 0.01 [χ2 test]). Genotype 4 was detected in 205 (91.1%) donors with HCV RNA, 1b in 15 (6.7%), and 1b plus 4 in 3 (1.3%); HCV RNA in the remaining two donors was untypable by PCR with type-specific primers (Ohno et al. 1997).

Table 1 Markers of HCV infection in Egyptians living in different districts and HCV-4a isolates recruited for estimating the population size
Full size table

When 50 samples from the 205 carriers infected with HCV genotype 4, which were selected to represent the four districts, were subjected to amplification of the NS5B sequence of 339 nt, 47, including the 13 reference sequences deposited in the database (AB103424–AB103457 and AF271800–AF271812), formed the specific cluster of HCV genotype 4a (Fig. 1).

Figure 1
figure 1

A phylogenetic tree on the basis of HCV sequences in NS5B region as reconstructed by a heuristic maximum-likelihood topology search. Forty-seven, including 13 reference, sequences (AF271800–AF271812) in Egypt retrieved from the database formed the specific cluster of HCV genotype 4a. These isolates were obtained from 13 governorates in four districts (Central, North, East, South).

Full size image

The phylogenetic tree constructed on NS5B-region sequences of the 47 HCV-4a isolates recovered in the year 1999 indicates that the most recent common ancestor of HCV-4a in Egypt would have existed approximately 100 years ago. Figure 2 shows the skyline plots and population growth based on a specific demographic model in Genie version 3.5 with four parameters (called constant + exponential + constant), which had the highest log likelihood in this study. As recently reported (Pybus et al. 2003), the specific demographic model, which has a constant size in the past, changes to exponential growth, and changes back to a constant size until the present, would be suitable for investigating the Egyptian endemic. Based on the demographic parameter estimates, HCV infection in Egypt shifted from a low level to rapid growth around 1940 (spread time), when the spread time of HCV is estimated as 10% of the present population size of HCV infections. The exponential growth continued during the 1940s through 1980 and has led to a more than 10-fold increase in the effective number of infections (Fig. 2).

Figure 2
figure 2

The maximum-likelihood estimate of N(t) on the effective population size of infection with HCV genotype 4a in Egypt. The parametric model is indicated by the blackline and stepwise plots by the gray line, which represents corresponding nonparametric estimates of N(t) (number as a function of time). Genetic distances were transformed into the time scale of years using estimates of the molecular clock in the NS5B region.

Full size image

Discussion

The molecular clock has been successfully applied to long-term serial serum samples containing HCV from the United States and Japan and estimated the spread time of HCV in the 1930s in Japan as 30 years earlier than that in the United States in the 1960s (Tanaka et al. 2002). It was extended in this study to reconstruct the national spread of HCV in Egypt, which has infected >10% of the general population.

The HCV RNA sequences studied are ideal for a coalescent analysis for the following reasons: (a) they represent a homogeneous viral population (HCV genotype 4a) that is outstandingly prevalent in Egypt and indigenous there; (b) they represent various geographic regions in Egypt (Table 1); (c) information about the sampling time is available (the year 1999); and (d) the molecular clock (a constant nucleotide substitution rate) in the NS5B region can be estimated on the basis of our previous molecular evolutionary studies (Tanaka et al. 2002).

The analysis (population study) of 47 HCV-4a samples by the molecular clock indicates that the current high prevalence of HCV in Egypt is the result of a rapid exponential spread during the past. The fast growth rate reflects an extremely efficient route of infection that must have operated during four decades in the twentieth century in Egypt. The national campaign with intravenous antimony for treatment of schistosomiasis (Frank et al. 2000; Hibbs et al. 1993; Kamel et al. 1992) is the major candidate for a viable route. The exponential expansion of Egyptian HCV-4a strains since the 1940s was corroborated by an independent analysis of E1-region sequences (Pybus et al. 2003).

Note that the spread time of HCV-4a strains is consistent with previous cohort studies in Egypt (Frank et al. 2000), and the period of exponential growth coincides with mass-control campaigns with intravenous antimony conducted from 1961 until 1986, when oral medications became available there (WHO 1993; Frank et al. 2000). This intensive transmission through inadequately sterilized needles and syringes in the past would have established a large reservoir of chronic HCV infection and is responsible for the high prevalence of HCV infection at present.

The molecular clock has been a very powerful tool in looking back at epidemic spreads of HCV infection in the United States and Japan (Tanaka et al. 2002), as well as Egypt in the present study. It would be useful, also, in predicting the population size and extent of HCV infection in various epidemiological settings. Studies for foreseeing future spreads would be required to cope with and prevent health care problems where de novo HCV infections are increasing. The virtue of the molecular clock method, capable of accurately estimating the dynamics of HCV based on a limited number of isolates (Pybus et al. 2001), will extend its application anywhere in the world where clinical sequelae of persistent HCV infection pose ever increasing burdens on the public health care of nations.

In summary, the molecular clock based on the neutral theory has estimated the exponential spread of HCV-4a in Egypt during 1940 through 1980, consistent with the duration of intravenous treatment campaigns for schistosomiasis.