|Home | About | Journals | Submit | Contact Us | Français|
PARV4 is a human parvovirus first detected and cloned from an individual with a HIV seroconversion-like illness and which subsequently persists in lymphoid tissue and bone marrow. In contrast to B19V, PARV4 infections are most frequently detected in injecting drug users (IDUs), particularly those co-infected with HIV-1. To investigate its transmission routes and whether infections are acquired through plasma-derived blood products, we developed a novel anti-PARV4 ELISA to determine seroprevalence in parenterally exposed and non-exposed subjects.
PARV4 VP2 was expressed and used as antigen in an indirect ELISA to detect anti-PARV4 IgG.
All 50 non-parenterally exposed adult controls were anti-PARV4 negative, in contrast to 67% and 33% antibody frequencies in HIV-positive and –negative IDUs respectively. Predominantly parenteral transmission was confirmed by the finding of similar infection frequencies (11/20 and 4/15) among HIV-coinfected and –uninfected haemophiliacs treated with non-virally inactivated FVIII/IX, whereas all but one of their 35 non-haemophiliac siblings were seronegative (despite close household contact).
This study provides convincing evidence that PARV4 is primarily transmitted parenterally. Evidence for widespread infection of haemophiliacs treated with non-virally inactivated clotting factor creates fresh safety concerns for plasma-derived blood products, particularly as parvoviruses are relatively resistant to virus inactivation.
PARV4 is a recently discovered and as yet poorly characterised member of the virus family Parvoviridae . The virus was originally cloned from an individual at risk for infection with human immunodeficiency virus (HIV), who displayed an “acute infection syndrome” resembling that of primary HIV infection. Since this original report, no further cases of acute infection by primary PARV4 infections have been characterised, although viral DNA sequences have been detected by polymerase chain reaction (PCR) at low frequencies in blood donors and plasma pools used for blood product manufacturing and in individuals at risk for parenterally transmitted viruses such as injecting drug users (IDUs) [2–4].
Infections with parvoviruses are generally acute with rapid resolution of both symptoms and clearance of viral DNA from blood and the respiratory and gastrointestinal tracts. However, several parvoviruses, such as the human erythrovirus B19, establish lifelong persistence with restricted replication and absence or rarity of detectable long term viraemia [5–7]. Persistence has similarly been shown to occur in PARV4 infections  and detection of viral DNA sequences in autopsy or biopsy samples by PCR provides a (rather laborious and necessarily indirect) method to determine frequencies of past infections with PARV4 in different risk groups [8,9]. Among UK study subjects, high frequencies of past exposure were found among IDUs, with higher frequencies in those co-infected with human immunodeficiency virus type 1 (HIV-1). In contrast, no evidence of past exposure was found in male homosexuals whether infected with HIV-1 or not, and in non-parenterally exposed, “low-risk” age matched controls. These findings, along with previous data showing higher frequencies of viraemia in IDUs suggest the unusual possibility that PARV4 may be a blood-borne virus [4,9], a mode of transmission quite unlike those of other parvoviruses where respiratory and gastrointestinal routes are extensively described.
To further explore this possibility and to overcome the sampling difficulties associated with autopsy/biopsy sample screening and viraemia detection, we developed a serological assay for IgG reactivity to the recombinant structural PARV4 protein, VP2. This allowed a more extensive investigation of risk group associations of PARV4 and, in particular, a detailed analysis of PARV4 transmission through therapy with non-virally inactivated factor VIII and IX concentrates.
Samples of plasma from IDUs were obtained from previously untreated HIV-infected subjects attending the Regional Infectious Diseases Unit (RIDU), Western General Hospital, Edinburgh. Plasma samples from HIV-uninfected HCV-infected IDUs were obtained from the Hepatitis Clinic, John Radcliffe Hospital, Oxford. Samples from HIV-1 infected and non-infected male homosexuals (MSMs), and from study subjects infected through heterosexual contact were obtained from RIDU and the Department of Genitourinary Medicine, Wycombe General Hospital.
Samples from 35 haemophiliacs and their 35 non-haemophiliac siblings were obtained from the Haemophilia Growth and Development Study (HGDS) cohort . Members of the group with haemophilia were born between 1972 and 1982, were between 7 to 16 years of age at study entry, and 10 to 21 years at the time the sample was taken. Their siblings ranged in age at entry from 7 to 20, and 9 to 22 years at the time the sample was taken. Samples for PARV4 analysis were drawn within six months in approximately 64% of subject-sibling pairs. All HGDS study subjects with haemophilia used non-virally inactivated clotting factor concentrate in the two years prior to enrolment. Nine or more infusions over that period, or 100+ U/kg of body weight of factor per year over the two years were required for eligibility. Twenty of the subjects with haemophilia were HIV-infected, 15 were HIV-negative. All were anti-HCV positive, whereas all non-haemophilic siblings were both HIV and HCV uninfected. Further low risk controls samples were obtained from anonymised archived surplus blood samples collected from adult attendees of orthopaedic outpatient clinics in Edinburgh, and present a cross-section of the adult population without compounding risk factors for virus infections.
Research was approved by the relevant institutional review boards or ethics committees and that all human participants or guardians gave written informed consent where required.
Nested PCR primers for the VP1 open reading frame (ORF) were designed based on published PARV4 sequences from genotype 1 and 2 viruses with appropriate restriction sites for sub cloning added to the inner primers (VP1_OS GGT ATG GAG CTG GGG ACA TTG AG [5′ base at position 2060, numbering here and elsewhere in the manuscript based on the NC_007018 reference sequence]; VP1_IS+Sal GGG TCG ACA TGT CTG CTG CTG AYG CTT AYC GT [8th base from 5′ end at position 2377]; VP1_OAS GCG TAT TTC CGC TTC CGG TCC [position 5268]; VP1_IAS+Not GGG CGG CCG CTT AHA RCA AAT GHG ART AAT TDC GCG C [11th base from 5′ end at position 5122]). PCRs were performed using GoTaq (Promega) according to manufacturers’ instructions using the following reaction conditions for both rounds: 30 cycles of [18 seconds at 94°C, 21 seconds at 50°C and 180 seconds at 72°C] and a final extension of 6 minutes at 72°C. Amplicons were cloned into the pCR-Blunt II-TOPO vector (Invitrogen) after blunting with mung bean endonuclease (NEB). Clones were screened by PCR using M13F and M13R primers and appropriately sized amplicons were sequenced using a combination of M13 and PARV4 primers.
The VP2 coding sequence was amplified using a single round PCR as described above with VP1_IAS+Not and VP2-specific sense (VP2_IS+Sal GGG TCG ACA TGT CCG TGG AAC CAG CTG G [9th base from 5′ end at position 3464]) primers and the cloned VP1 as a template.
Recombinant Autographa californica multiple nuclear polyhedrosis viruses (AcMNPV) were produced using the Bac-to-Bac Baculovirus Expression System (Invitrogen) following manufacturers’ instructions. Briefly, the cloned VP1 and VP2 sequence was subcloned into the pFastBac transfer vector. Appropriate insertions were confirmed by PCR and sequencing. Transfer vectors were then used to transform DH10Bac competent E. coli which contain the AcMNPV parent bacmid. Recombinant clones were selected by antibiotic selection and confirmed by PCR and sequencing. Recombinant bacmids were isolated using the alkaline lysis method recommended by the manufacturer.
Monolayers of Spodoptera frugiperda 9 (Sf9) cells were transfected with recombinant bacmid DNA using the cellfectin reagent according to manufacturers’ instructions. The resulting baculoviruses were collected in the supernatants of transfected cells and passaged to naïve monolayers to increase viral titres. High titre stocks were used to infect SF9 cells for recombinant protein expression. Infected cells were collected 5 days post infection in phosphate-buffered saline (PBS) containing Complete Protease Inhibitor Cocktail (Roche). Cells were frozen and thawed rapidly three times and further homogenised by repeated passage through a small gauge needle. Cell debris was removed by low-speed centrifugation (1,000 × g) for 10 minutes in a benchtop microcentrifuge. Supernatants were layered on a 20% (wt/wt) sucrose cushion and centrifuged for 2 hours at 100,000 × g (SW28 rotor, Beckmann L8–70M ultracentrifuge). Pellets enriched in high density proteins consistent with virus-like particles (VLPs) were resuspended in 5% glycerol/PBS (v/v) containing complete protease inhibitor cocktail (Roche Molecular Biochemicals).
Protein samples were placed on formvar/carbon coated copper slot grids for 10 minutes. Excess solution was then removed and samples were stained with a 1% aqueous uranyl acetate solution 30 seconds. Following removal of excess staining solution, grids were allowed to air dry and viewed in a Philips CM120 transmission electron microscope. Images were captured using a Gatan Orius 1000 Digital Camera.
Whole-cell lysate and enriched protein samples were analysed on sdium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis (PAGE) gels. Gels were either stained with Coomassie-Blue (Sigma) or blotted onto Protran BA 85 nitrocellulose membranes (Whatman) using a semi-dry electroblotter. Membranes were blocked overnight in 2% bovine serum albumin (BSA)/0.1% Tween-20/PBS (wt/v/v). Membranes were then incubated for 1 hour at room temperature with the serum from an IDU suspected of being seropositive for PARV4 diluted 1:100 in 2% BSA/PBS (wt/v). After 6 washes in 0.1% Tween-20/PBS (v/v) for a total of 2 hours, the membrane was incubated with horseradish peroxidase (HRP)-conjugated goat anti-human IgG antibody (Serotec) diluted 1:1000 in 2% BSA/PBS (w/v). Following 4 washes in 0.1% Tween-20/PBS (v/v) for a total of 1 hour, bound antibody was visualised with 3,3′-Diaminobenzidine (DAB) Enhanced Liquid Substrate System according to the manufacturer’s protocol (Sigma).
High bind 96-well ELISA plates (Greiner Bio-One) were coated overnight with VLP enriched protein fractions (0.5ug recombinant protein in 100ul carbonate buffer) or an equivalent volume of protein isolated from cells infected with non-recombinant baculovirus collected and processed in parallel with the VLP sample. Plates were washed with 250ul of 1% Tween-20/PBS (v/v) and coated wells were blocked with 150ul of 1% BSA/PBS (wt/v) for 2 hours at room temperature. After one round of washing, test sera diluted 1:100 in 100ul of 1% BSA/PBS (wt/v) were added to the wells and incubated for 1 hour. The wells were washed 6 times with 250ul of 1% Tween-20/PBS (v/v), each time incubated for 15 minutes, and then incubated for 30 minutes with 100ul of HRP-conjugated goat anti-human IgG antibody (Serotec) diluted 1:1000 in 1% BSA/PBS (wt/v). Following four rounds of washing, plates were developed by adding 70μl of the HRP substrate (2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) Liquid Substrate System) to each well according to the manufacturer’s protocol (Sigma). Plates were allowed to develop for 20–25 minutes and read at 405nm. The immunoreactivity of sera to control wells was subtracted from VLP immunoreactivity prior to analysis.
The putative PARV4 VP1 coding sequence (positions 2378–5122) was amplified and cloned from DNA extracted from an archived clotting factor VIII sample as template. Several individual clones of PARV4 genotype 1 were recovered and the sequence showing the highest amino acid identity to published genotype 1 sequences was selected for recombinant baculovirus production. To minimise variation between VP1 and VP2 sequences, the selected VP1 clone was used as the template for VP2 (nucleotides 3464–5122) amplification. The VP1 and VP2 sequences were cloned into the Autographa californica multiple nuclear polyhedrosis virus (AcMNPV) using the Bac-to-Bac kit (Invitrogen). In addition to the two recombinant viruses (VP1-Bac and VP2-Bac) a non-recombinant control virus (Con-Bac) was produced in parallel.
Sf9 monolayers were collected 5 days post infection with high titre baculovirus stocks. Clear expression of both VP1 protein with a molecular weight of approximately 110 KDa (Figure 1A) and VP2 protein with a molecular weight of approximately 62 KDa (Figure 1B) were observed in infected cells at this time following SDS-PAGE analysis of whole-cell lysates. The identity of these two proteins as PARV4 antigens is supported by their reactivity to the serum antibodies in plasma samples from two HIV co-infected IDU by Western blot (data not shown).
PARV4 capsid proteins are predicted to self assemble to form virus-like particles (VLPs) in the absence of virus genomic DNA as has been reported for other parvoviruses [11–15]. To determine the ability of the expressed proteins to form VLPs, Sf9 cells were infected with either VP2-Bac alone (the predicted major capsid protein) or both VP1- and VP2-Bac (the predicted major and minor capsid proteins). Infected cells were collected 5 days post infection and homogenised. Soluble, high density protein aggregates were enriched from cleared cell lysates by ultracentrifugation through a sucrose cushion. Proteins pelleted by this procedure were analysed by SDS-PAGE (Figure 1C), Western blot (Figure 1D) and transmission electron microscopy (Figure 2). Particles with an icosahedral appearance and predicted parvovirus capsid diameter of around 20nm can be seen in pelleted protein preparations from VP2-Bac infected cells (Figure 2). This suggests that, like other parvoviruses, VP2 alone is sufficient for capsid assembly. Enrichment of a protein consistent with VP2 can be clearly seen in the preparation from VP2-Bac infected cells (Figure 1C, lane 2). This protein is similarly enriched in the preparation from VP1- and VP2-Bac infected cells (Figure 1C, lane 1). No clear VP1 is seen in this sample, even after Western blot staining (Figure 1D, lane 1), suggesting that little if any is incorporated into VLPs. As Western blot staining of VLP preparations show a strong signal over control preparations, these samples were used as antigen for the ELISA without further purification.
Following the observation of marked differences in PARV4 detection in autopsy samples between different risk groups for HIV-1 and HCV infection and low risk controls [8,9], we assayed plasma samples from these different study subject categories by ELISA using recombinant VP2 as antigen (Table 1; Fig. 3). Serological reactivity varied substantially between groups with high net OD’s recorded from over half of the samples from HIV-positive IDUs, and infrequent or absent reactivity among 84 samples assayed from low-risk controls. The latter group (comprising siblings of haemophiliacs [excluding the single immunoreactive sample] and orthopaedic outpatients; see Methods) showed a tight distribution of low OD values, with a mean of −0.002 and a standard deviation (SD) of 0.039. The cut-off for the assay was therefore set conservatively as the mean + 3 SD’s (0.113) and used to assign samples as anti-PARV4 positive and negative.
High frequencies of PARV4 sero-reactivity were observed among IDUs (co-infected with HCV and HIV-1; 67%) and haemophiliacs exposed to non-virally inactivated clotting factor concentrates (55% among those co-infected with HIV-1, 27% in those who were HIV-negative; Table 1). These frequencies of inferred past exposure to PARV4 were significantly higher than found in control groups assembled for the study (1 from 84; see above), and from (anti-HCV negative) individuals infected with HIV-1 through heterosexual contact in the UK (0/14; Table 1). However, four from 17 HIV-positive MSMs showed low level serological reactivity in the PARV4 ELISA, although their anti-HCV negative status demonstrates likely non-parenteral routes of transmission of PARV4 in this risk group.
This study describes the development, evaluation and application in seroepidemiological studies of a novel serological assay for detection of antibodies to the newly discovered parvovirus, PARV4. Its design was based on previously developed assays for B19 and human bocavirus (HBoV), employing recombinant antigen expressed from the full length VP2 gene. The protein expressed in baculovirus-infected Sf9 cells assembled in virus-like particles indicating that it adopted an approximately native configuration and thus likely presenting most linear and conformational epitopes expressed on native virion particles. The recombinant protein was indeed highly antigenic, with sera from risk groups previously shown to be at risk for PARV4 infection (HIV-positive IDUs) showing frequent reactivity in ELISA, in contrast to the uniform negativity of samples from low risk adult control and all but one of the haemophiliac siblings. As this is the first description of a serological assay for PARV4 antibodies, we lacked previously defined anti-PARV4 positive and negative sera. However, from the distribution of ODs from low-risk study subjects in the ELISA, a mean value plus 3 standard deviations allowed provisional assignment of serological anti-PARV4 positive and negative status and minimised “indeterminate” reactivity (Fig. 3).
Testing samples from risk groups previously analysed in autopsy based studies [8,9] strongly supported our previous hypothesis for a predominantly parenteral route of transmission. Frequencies of seropositivity of 67% among HIV-infected IDUs and 33% in HIV-uninfected IDUs indeed closely matched frequencies of exposure inferred through autopsy screening of study subjects from the same geographical area and similar age distribution [8,9]. In contrast, screening of an extended number of samples from HIV-infected MSMs along with those infected with HIV-1 through heterosexual contact revealed lower (4/17) or zero frequencies (0/14) of seroreactivity respectively, the former consistent with the previously reported absence of detectable PARV4 sequences in autopsy samples from MSMs (0/13; ).
However, what most firmly links PARV4 infection to parenteral exposure is the finding of high frequencies of anti-PARV4 antibody among those treated for haemophilia with non-virally inactivated clotting factor concentrates in the late 1970s – early 1980s (Table 1). The parallel finding of almost complete seronegativity among sibling controls demonstrates an absence of PARV4 infection among normal children and adolescents that contrasts strongly with rapidly increasing seroprevalence with age of the other human parvoviruses, B19 and HBoV during childhood [14–20]. It similarly demonstrates that even close domestic contact with individuals during acute infection does not efficiently transmit PARV4, in marked contrast to the highly infectious nature and extensive virus shedding in respiratory secretions during primary B19 and HBoV infections. In the case of the one seropositive sibling control, it is quite conceivable that inadvertent parenteral exposure may occasionally occur in a household setting. For example, skin trauma followed contract with infectious body fluids from a sibling during the period of acute, viraemic stage of PARV4 infection may transmit the virus.
The likelihood of PARV4 transmission through virally inactivated clotting factor concentrates is currently difficult to assess. PARV4 DNA was detected in 16%–33% of non-virally inactivated coagulation factors (FVII, FVIII and FIX), but also in 9% of more recently manufactured solvent- or heat-inactivated concentrates [21,22]. Parvoviruses are non-enveloped, and B19 along with model viruses such as canine parvovirus have been previously shown to be resistant to solvent-detergent inactivation methods and moderately resistant to heat-inactivation [23,24]. In the specific case of B19, PCR-based screening of plasma used for blood product manufacture to remove highly viraemic plasma donations has proven to be a highly effective additional step that reduces transmission risk of B19 to haemophiliacs and other users of plasma-derived products . However, inactivation of B19 is undoubtedly assisted by the high seroprevalence of B19 antibodies among normal donors that likely contributes to the inactivation of any residual virus in the plasma pools prior to manufacture . Furthermore, high frequencies of past infection protect the majority of recipients from infection. The low seroprevalence of antibodies to PARV4 in normal blood or plasma donors predicted from our study along with an absence of immunity in recipients would conversely increase the risk of PARV4 transmission though blood products, as well by labile components such as red cell concentrates and platelets. Further studies investigating the seroprevalence of PARV4 in recipients of solvent- and heat-inactivated plasma derived products and of recombinant FVIII and FIX are underway.
Although this study provides unequivocal evidence for primarily parenteral transmission routes of PARV4 in Western countries, the finding of frequent infection among study subjects infected with HIV-1 in Africa without HCV co-infection  and the low frequency of antibody detection in HIV-positive MSMs in the current study suggests other routes may be involved in different settings and likely reduced efficiency. Among heterosexually-infected African men, genetic characterisation of PARV4 amplified from autopsy tissue from HIV-infected individuals from the Congo and Nigeria revealed the presence of a PARV4 genotype genetically distinct from those circulating among IDUs in Western countries , providing further evidence for its spread in different transmission networks. Further work is required to understand the natural history of PARV4 infections, the time-scale for its emergence in IDUs and other at-risk populations and its pathogenicity during acute and persistent infections. The availability of a reliable serological assay for PARV4 antibodies will assist considerably in these future studies.
The authors would like to thank the Sheila Morris and other staff of the Regional Infectious Diseases Unit, Western General Hospital, Edinburgh for their invaluable help in collection of blood samples and provision of clinical data and to staff for assistance in preparation and data collection of samples. We would additionally like to thank Mary McNally (Science Applications International) National Cancer Institute-Frederick for managing and shipping clinical samples and the clinical information from haemophiliacs enrolled in the Haemophilia Growth and Development Study Cohort, and staff at the Royal Infirmary of Edinburgh for additional anti-HCV testing. Transmission electron microscopy expertise was provided by Stephen Mitchell at the Institute of Molecular Plant Sciences Electron Microscope facility at the University of Edinburgh.
Funding: This study was supported by an unrestricted investigator-initiated grant from Baxter Healthcare and by the National Institutes of Health, National Institute of Child Health and Human Development (R01 HD41224). Studies in Oxford were supported by the James Martin School for 21st Century, the Wellcome Trust, NIHR Biomedical Research Centre programme and the MRC (UK).
Conflict of Interest: Eric Delwart and his institution (Blood Systems Research Institute and Department of Laboratory Medicine, University of California - San Francisco) have filed a patent describing the discovery of PARV4 and claim intellectual property rights. There are no actual or potential conflicts of interests for other authors.