A novel, helminth-derived immunostimulant enhances human recall responses to hepatitis C virus and tetanus toxoid and is dependent on CD56⁺ cells for its action

Abstract

We have described previously an immunostimulant derived from Onchocerca volvulus, the helminth parasite that causes onchocerciasis. Recombinant O. volvulus activation-associated secreted protein-1 (rOv-ASP-1) was a potent adjuvant for antibody and cellular responses to protein, polypeptide and small peptide antigens. Our aims were to determine whether rOv-ASP-1 is immunostimulatory for human peripheral blood mononuclear cells (PBMC) and, if so, whether it could augment cellular responses against human pathogen antigens in vitro. Cytokines from rOv-ASP-1-stimulated human PBMC were measured by a fluorescence activated cell sorter-based multiplex assay. Recall responses of normal healthy donor (NHD) and chronic hepatitis C virus (c-HCV)-infected patient PBMC to tetanus toxoid (TT) or HCV core (HCVco) antigen, respectively, were measured by interferon-γ enzyme-linked immunospot assays. Interferon-γ was the predominant cytokine induced by rOv-ASP-1. 77·3% of NHD anti-TT and 88·9% of c-HCV anti-HCVco responses were enhanced by rOv-ASP-1. The immunostimulant effect was dependent upon contact between CD56⁺ and CD56⁻ fractions of PBMC. We have described a helminth-derived protein that can act as an immunostimulant for human recall responses in vitro to TT and, perhaps more importantly, HCV antigens in patients with chronic HCV infection. Our longer-term goal would be to boost anti-viral responses in chronic infections such as HCV.

Introduction

With the increasing prevalence of chronic viral infections such as human immunodeficiency virus (HIV) and hepatitis C virus (HCV), there is a pressing need for safe and effective immunotherapies that will drive the host's immune effector mechanisms towards viral clearance. Chronic hepatitis C infection affects 170 million people worldwide and is characterized by ineffective cellular immune responses to HCV that are thought to contribute to the inability to eliminate infection and disease chronicity. Approximately 60% of infected patients can clear HCV in response to treatment with pegylated interferon (IFN)-α and ribavirin. An effective immune stimulant might prove to be an effective treatment either alone or in conjunction with existing or new anti-virals.

We have described previously a novel immunostimulant derived from Onchocerca volvulus, the helminth parasite that causes onchocerciasis, known more commonly as river blindness. Highly purified, lipopolysaccharide (LPS)-free recombinant O. volvulus activation-associated secreted protein-1 (rOv-ASP-1) was a potent adjuvant for inducing T helper 2 (Th2) and Th1-associated antibodies to protein (ovalbumin), polypeptide (HIV-1 gp120-CD4 chimera) and small peptide (37-mer from S protein of SARS-CoV) antigens. Unusually for a helminth product, rOv-ASP-1 as an adjuvant with ovalbumin induced an exclusively Th1 (IFN-γ) response to ovalbumin in vaccinated mice [1].

Activation-associated secreted proteins (ASPs) of parasitic nematodes are highly immunogenic and have been studied as potential vaccine components, particularly hookworm ASPs [2–5]. O. volvulus-ASP-1 (Ov-ASP-1) is typical of nematode ASPs in that it is intrinsically immunogenic [6], highly charged (rOv-ASP-1 is basic with a pI of 9·6) and contains a highly conserved cysteine-rich domain with structural homology to pathogenesis-related proteins in plants and also to vespid venom antigen 5 [7].

Having shown that rOv-ASP-1 is a potent vaccine adjuvant in mice, the aims of the present study were to determine whether the protein has immunostimulatory properties for human leucocytes and, if so, whether it could augment cellular responses against human pathogen antigens in vitro. To address this, we studied normal healthy donor (NHD) peripheral blood mononuclear cell (PBMC) responses to a commonly used recall antigen, tetanus toxoid (TT) and, in patients infected chronically with HCV, PBMC responses to HCV core (HCVco) antigen. If rOv-ASP-1 was able to enhance in vitro responses to these antigens, our longer-term goal would be to boost antigen-specific responses in chronic infections such as HCV.

Methods

Human subjects

Ethical approval was obtained from Southampton and South-west Hampshire Joint Research Ethics Committee, and all patients gave informed consent in writing prior to participating in the study. A total of 18 patients with chronic HCV infection (c-HCV, 15 male and three female, median age 48 years, range 31–61 years) were recruited from the hepatology clinics run by Southampton University Hospitals National Health Service Trust. All patients had detectable HCV RNA (11 genotype 1a/b, seven non-genotype 1). Patients were excluded if they had received treatment for HCV infections 6 months or less prior to the study or tested positive with hepatitis B virus or HIV. Twenty-two uninfected NHD (16 male and six female, median age 46 years, range 26–77 years), vaccinated previously against tetanus and with no known risk factors for blood-borne virus infection, consented to give blood for this study.

Preparation of recombinant Ov-ASP-1

The rOv-ASP-1 was expressed as a histidine-tagged protein in Escherichia coli and purified as described previously [1]. It should be noted that our purified rOv-ASP-1 protein tested negative in a Limulus amoebocyte lysate assay. In addition, quantitative LPS testing by Cambrex Bio Science (Baltimore, MD, USA) showed that it contained less than 0·25 endotoxin units per milligram of protein (≈ 25 pg endotoxin/mg). For functional confirmation, in three experiments using NHD PBMCs, polymyxin B inhibited LPS (0·1, 1·0, 10·0 ng/mL)-induced IFN-γ and interleukin (IL)-10 production but had no effect on cytokine secretion stimulated by 5 μg/mL rOv-ASP-1 (data not shown). The same, essentially LPS-free, stock was used in all experiments. As a control for rOv-ASP-1, in preliminary experiments we used E. coli-recombinant chloramphenicol acetyl transferase (rCAT) and other O. volvulus recombinant proteins that were cloned and identified as putative vaccine targets for River blindness.

Preparation and stimulation of PBMC and measurement of cytokines

Fifty ml of freshly drawn blood was obtained from NHD subjects (n = 22) or c-HCV patients (n = 18). Blood was collected into K₃-ethylenediamine tetraacetic acid and separated immediately by centrifugation over Lymphoprep (Robbins Scientific, Solihull, UK). PBMCs were recovered according to the manufacturer's recommended protocol. PBMCs, 2 × 10⁵ per duplicate well, were cultured in U-bottomed 96-well culture plates, in complete RPMI medium [RPMI-1640 with phenol red (Invitrogen, Paisley, UK), with 2 mmol/l l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin (Sigma, Poole, UK) and 10% heat-inactivated human antibody serum (Sigma)] with or without rOv-ASP-1.

Initial experiments with rOv-ASP-1 and concanavalin A stimulation established the optimal time-points for PBMC supernatant collection as 24 h for tumour necrosis factor (TNF)-α, IL-2 and IL-4 and 72 h for IFN-γ, IL-10, IL-5 and granulocyte–macrophage colony-stimulating factor (GM-CSF). Similarly, a concentration of 5 μg/ml was determined to result in maximal cytokine secretion from PBMCs.

Interferon-γ, TNF-α, IL-2, IL-4, IL-5, IL-10 were measured in cell culture supernatants using a human Th1/Th2 cytometric bead array (BD Biosciences, San Jose, CA, USA). GM-CSF was measured by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN, USA) according to the manufacturers' protocols.

In IFN-γ enzyme-linked immunospot (ELISPOT) assays, antigens [1 μg/ml TT, Merck Chemicals (Calbiochem), Beeston, UK] or 5 μg/ml recombinant HCVco protein (Mikrogen, Neuried, Germany), with or without rOv-ASP-1, were added to cells in triplicate or quadruplicate wells (2 × 10⁵/well) and incubated for 18 h in U-bottomed 96-well culture plates (Nunc, Fisher UK, Loughborough, UK) before transferring the cells to coated and blocked IFN-γ ELISPOT plates for a further 24 h. We used human IFN-γ ELISPOT reagents (Mabtech, Nacka Strand, Sweden) following the manufacturer's instructions and Multiscreen HTS IP plates (Millipore, Watford, UK). ELISPOT plates were counted using an AID ELR02 automated counter (Autoimmun Diagnostika GmbH, Strassberg, Germany).

Preparation of CD56⁺ cells and PBMC depleted of CD56⁺ cells

CD56⁺ cells were purified from PBMC using magnetic human CD56 MicroBeads and an AutoMACS^TM Separator using the depletes program (Miltenyi Biotec, Bisley, UK). By labelling with anti-CD56^– antigen-presenting cells (APC) antibody (Miltenyi Biotec), isolated CD56 cells were typically ≥ 98% pure on fluorescence activated cell sorter analysis and CD56-depleted fractions contained virtually no CD56⁺ cells (< 0·05%).

In some experiments, Transwell^TM 0·4 μm culture inserts (Corning Life Sciences, the Netherlands) were added to 24-well culture plates (Nunc) to separate CD56⁺ from CD56⁻ cells.

Data analysis

Paired Student's t-tests to compare means and, where appropriate, Mann–Whitney tests to compare medians (data in Fig. 4) or repeated-measures analysis of variance (anova) with Dunnett's multiple comparison test (data in Figs 2a and 3a) were used. Analyses were performed using prism version 4·0 for Windows (GraphPad Software, San Diego, CA, USA).

Fig. 4.

Effect of recombinant Onchocerca volvulus activation-associated secreted protein-1 (rOv-ASP-1) on interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) responses of CD56⁺ cells from (a) normal healthy donors (NHD, n = 7) or (b) patients with chronic hepatitis C virus infection (c-HCV, n = 7). CD56⁺ cells were incubated in culture medium alone (open circles) or medium + 5 μg/ml rOv-ASP-1 for 18 h (solid circles). Bold solid lines indicate median values.

Fig. 2.

(a) Interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) responses of peripheral blood mononuclear cells (PBMC) from normal healthy donors (n = 22) in response to recombinant Onchocerca volvulus activation-associated secreted protein-1 (rOv-ASP-1) (ASP-1), tetanus toxoid (TT), TT + ASP-1 and net response to TT in the presence of ASP-1 [(TT + ASP-1) − ASP-1]. Horizontal bars indicate mean values. *P < 0·01 compared with untreated control wells (medium). (b) Tetanus toxoid (TT, open circles)-specific IFN-γ ELISPOT responses in normal healthy donors (n = 22) are enhanced by rOv-ASP-1 (ASP-1). (TT + ASP-1) − ASP-1 = net anti-TT response after subtracting the response because of ASP-1 alone (solid circles). Mean values are indicated by bold solid lines. (c) Representative IFN-γ ELISPOT assay showing responses of untreated PBMC (column 1), PBMC + TT (column 2), PBMC + rOv-ASP-1 (column 3), PBMC + TT + rOv-ASP-1 (column 4). PBMC were seeded into triplicate wells (2 × 10⁵/well) and numbers of spots are indicated for each well.

Fig. 3.

(a) Interferon (IFN)-γ enzyme-linked immunospot (ELISPOT) responses of peripheral blood mononuclear cells (PBMC) from chronic hepatitis C virus (c-HCV)-infected patient (n = 18) in response to recombinant Onchocerca volvulus activation-associated secreted protein-1 (rOv-ASP-1) (ASP-1), hepatitis C virus (HCV) core (HCVco), HCVco + ASP-1 and net response to HCVco in the presence of ASP-1 ([HCVco + ASP-1] − ASP-1). Horizontal bars indicate mean values. *P < 0·01 compared with untreated control wells (medium). (b) Anti-HCV core (HCVco) PBMC IFN-γ ELISPOT responses in chronic HCV patients (n = 18) are enhanced by rOv-ASP-1. PBMC responses to HCVco are indicated by open circles (HCVco + ASP-1) − ASP-1 = net anti-HCVco response after subtracting the response because of ASP-1 alone (solid circles). Mean values are indicated by bold solid lines. (c) Representative IFN-γ ELISPOT assay showing responses of untreated PBMC (column 1), PBMC + HCVco (column 2), PBMC + rOv-ASP-1 (column 3), PBMC + HCVco + rOv-ASP-1 (column 4). PBMC were seeded into triplicate wells (2 × 10⁵/well) and numbers of spots are indicated for each well.

Results

Onchocerca volvulus-ASP-1 induces proinflammatory (IFN-γ, GM-CSF, TNF-α) and immunoregulatory (IL-10) cytokine secretion from normal human PBMCs

Recombinant, essentially LPS-free rOv-ASP-1 at an optimal concentration of 5 μg/ml (Fig. 1) stimulated significant (P < 0·05 versus unstimulated PBMCs) production of nanogram quantities of IFN-γ and IL-10 (although approximately fivefold more of the former than the latter cytokine on a molar basis) and picogram amounts of GM-CSF and TNF-α from PBMCs isolated from NHD (n = 14). The donors were never exposed to the helminth parasite O. volvulus. There was no significant induction of IL-4, IL-5 or IL-2 by the protein. In preliminary experiments, 5 μg/ml Ov-ASP-1 was established as the optimal dose for cytokine stimulation and the relative amounts of cytokines did not vary with the dose of Ov-ASP-1 used (data not shown). The control rCAT and recombinant O. volvulus proteins that we tested did not have any biological activity.

Fig. 1.

Mean amounts of cytokine secretion (pg/10⁶ cells ± standard deviation) induced by recombinant O. volvulus activation-associated secreted protein-1 (rOv-ASP-1) (5 μg/ml) from normal healthy donor peripheral blood mononuclear cells (PBMC) (n = 14). Supernatants were collected at 24 h for tumour necrosis factor (TNF)-α, interleukin (IL)-2 and IL-4 and 72 h for interferon (IFN)-γ, IL-10, IL-5 and granulocyte–macrophage colony-stimulating factor (GM-CSF). *P < 0·05 compared with untreated control wells.

Recall responses to TT are enhanced by rOv-ASP-1

To test the immunostimulatory capacity of rOv-ASP-1 for human antigen-specific responses we initially tested the effect of the protein on recall responses to TT by assaying IFN-γ secretion by PBMC in human NHD. We chose to focus upon IFN-γ secretion as our readout for Ov-ASP-1 bioactivity because: (i) IFN-γ was, by far, the most abundant cytokine induced by the parasite protein (Fig. 1); and (ii) we were particularly interested in a response relevant to the clearance of HCV [8].

The responses of NHD (n = 22) PBMC in IFN-γ ELISPOT assays to rOv-ASP-1 alone, TT alone and TT + rOv-ASP-1 are shown in Fig. 2a. Although 18 of 22 (82%) of the subjects responded to TT, when analysed on a population level there was no significant increase in the frequency of IFN-γ-secreting PBMC in the presence of the antigen. Because the parasite protein itself induced IFN-γ secretion significantly in the vast majority of subjects (P < 0·01 versus controls in medium alone), the data for responses in the presence of antigen + rOv-ASP-1 are expressed as the response to TT + rOv-ASP-1 minus that because of rOv-ASP-1 alone for each donor, i.e. the net antigen-specific response (Fig. 2a, extreme right-hand column). Both TT + ASP-1 and, most importantly, net anti-TT responses (after subtracting the response to rOv-ASP-1 alone) were increased significantly (P < 0·01) compared with unstimulated controls (medium).

From the data shown in Fig. 2a, we then dissected out the key component, i.e. the net response to TT with or without added rOv-ASP-1 to investigate the effect of the parasite protein on the antigen-specific response; these data are presented in Fig. 2b. Co-culturing PBMCs with rOv-ASP-1 and TT significantly (P = 0·018, n = 22) enhanced NHD PBMC IFN-γ ELISPOT responses to the TT recall antigen (Fig. 2b). The mean response to TT alone was 268·3 IFN-γ spots/million PBMCs and 591·8 (net value) in the presence of rOv-ASP-1. Overall, 17 of 22 donors' (77·3%) net anti-TT responses increased in the presence of rOv-ASP-1. A typical ELISPOT assay result is shown (Fig. 2c).

Recombinant Onchocerca volvulus-ASP-1 enhances anti-HCVco responses in chronic HCV patients

We then tested the immunostimulatory capacity of the protein in a disease setting, i.e. the effect on HCV antigen-specific responses in the context of c-HCV infection. As donors were infected with different genotypes of HCV, we used HCVco as a stimulatory antigen that is well conserved across HCV genotypes. The data were analysed and presented in the same way as the NHD PBMC responses to TT. On a population level, c-HCV PBMC responses to rOv-ASP-1 were not significantly enhanced compared with unstimulated control wells (Fig. 3a), despite 94% of patients responding to the protein by showing an increased frequency of IFN-γ-secreting cells. Similarly, 17 of 18 (94%) of the cohort showed a positive (although not statistically significant on a population level) response to HCVco. Like the NHD anti-TT results, the net response to anti-HCVCo became significant (P < 0·01) in the presence of rOv-ASP-1. Similar to anti-TT responses in NHDs, rOv-ASP-1 significantly (P = 0·018, n = 18) improved net anti-HCVco IFN-γ responses (antigen alone mean = 159·4; net antigen response in the presence of rOv-ASP-1 = 624·8) with 16 of 18 patients (88·9%) having increased net anti-HCVco responses in the presence of rOv-ASP-1 (Fig. 3b). A representative ELISPOT assay result is shown in Fig. 3c.

CD56⁺ cells secrete IFN-γ in response to rOv-ASP-1 and are essential for enhancement of recall anti-TT and anti-HCVco responses

In preliminary depletion experiments designed to identify cell populations that may respond to rOv-ASP-1 by secreting IFN-γ in the ELISPOT assay, it became apparent that CD56⁺ cells were the principal source. CD4⁺ T cells made a minimal contribution in the absence of antigen. In addition, direct stimulation of purified CD8⁺ cells with rOv-ASP-1 resulted in no increase in IFN-γ spots (data not shown). In contrast, rOv-ASP-1 induced IFN-γ secretion from purified CD56⁺ cells (Fig. 4).

In seven experiments where isolated CD56⁺ cells (typical purity = 98–99%) were treated with rOv-ASP-1, it was clear that rOv-ASP-1 increased the numbers of IFN-γ-positive cells, significantly so in the case of cells isolated from NHD (Fig. 4a, P = 0·021, n = 7). Interestingly, the frequency of constitutively IFN-γ-secreting CD56⁺ cells (i.e. cultured in medium alone) was significantly higher (P = 0·011, n = 7) in c-HCV-infected versus NHD subjects. In contrast to the NHD subjects, there was no significant increase on a population level in the frequency of IFN-γ-secreting CD56⁺ cells isolated from c-HCV donors following treatment with rOv-ASP-1 (Fig. 4b).

Having determined that CD56⁺ cells were the main cells responding directly to rOv-ASP-1 stimulation by secreting IFN-γ in our system, we next investigated if they may also contribute to the rOv-ASP-1-mediated enhancement of net anti-TT and anti-HCVco responses described in Figs 2 and 3. We tested whether rOv-ASP-1 enhancement of the recall antigen responses could occur in the absence of CD56⁺ cells.

Net anti-TT IFN-γ responses were augmented significantly by rOv-ASP-1 in intact PBMC (Fig. 5a, mean anti-TT IFN-γ response without rOv-ASP-1 = 256·3 IFN-γ spots/million cells versus 601·6 net anti-TT response in the presence of rOv-ASP-1; P = 0·018, n = 9), but not when the CD56⁺ cells were removed (mean anti-TT IFN-γ response without rOv-ASP-1 = 214·3 IFN-γ spots/million cells versus 315·3 net anti-TT response in the presence of rOv-ASP-1; not significant, n = 9) (CD56-depleted PBMC were ≥ 98% free of CD56⁺ cells and were used at the same final concentration as the unfractionated PBMC preparations in both NHD and c-HCV experiments).

Fig. 5.

Comparison of the effect of recombinant Onchocerca volvulus activation-associated secreted protein-1 (rOv-ASP-1) (5 μg/ml) on whole peripheral blood mononuclear cells (PBMC) (P) or PBMC depleted of CD56⁺ cells (P-CD56) on net [(antigen + ASP-1) − ASP-1] interferon (IFN)-γ enzyme-linked immunospot responses to: (a) tetanus toxoid (TT) in normal healthy donors (NHD, n = 9); (b) HCV core (HCVco) in patients with chronic HCV infection (c-HCV, n = 16). Responses to antigen in the absence of rOv-ASP-1 are indicated by open circles and by solid circles in the presence of rOv-ASP-1. Bold solid lines indicate mean values.

Enhancement of c-HCV anti-HCVco responses by rOv-ASP-1 was similarly dependent upon the presence of the CD56⁺ fraction (Fig. 5b, mean anti-HCVco IFN-γ response without rOv-ASP-1 = 155·3 IFN-γ spots/million cells versus 486·2 net anti-HCVco response in the presence of rOv-ASP-1; P = 0·007, n = 16. Mean anti-HCVco IFN-γ response of CD56⁺-depleted PBMC without rOv-ASP-1 = 176·1 IFN-γ spots/million cells versus 126·5 net anti-HCVco response in the presence of rOv-ASP-1; no significant enhancement of net anti-HCVco response, n = 16). Thus we established a critical role of CD56⁺ cells in the immunostimulatory effect of rOv-ASP-1 on recall responses to both TT (in NHDs) and HCVco (in c-HCV patients).

In both sets of experiments using TT and HCVco antigens, it is important to stress that the dependence of the immunostimulatory effect of rOv-ASP-1 on the presence of CD56⁺ cells was not due simply to the removal of IFN-γ secreting rOv-ASP-1-reactive CD56⁺ cells from the PBMCs added to the ELISPOT assays. This is because we subtracted the response because of rOv-ASP-1 alone from that induced by antigen + rOv-ASP-1 to give the net response which is the antigen-specific component.

Contact is required between CD56⁺ and CD56⁻ cells for Ov-ASP-1-mediated enhancement of anti-TT responses to occur

To dissect further the role of CD56⁺ cells in rOv-ASP-1-mediated immunostimulation, we tested whether a soluble product was involved or if cell–cell contact was required. For logistic and ethical reasons, it was not possible to recruit c-HCV donors again to donate repeat blood samples within the time-scale of the study. Therefore, using four TT-reactive volunteer donors who had shown enhancement previously by rOv-ASP-1, we set up experiments where CD56⁻ and CD56⁺ fractions were either mixed (in the same proportions as the original unfractionated PBMC) or separated by a Transwell^TM membrane and co-exposed to TT antigen with or without rOv-ASP-1 (Fig. 6). Net anti-TT responses in reconstituted CD56⁺ and CD56⁻ mixed populations were enhanced significantly by rOv-ASP-1 (Fig. 6a, P = 0·042, n = 4). However, when the two cell populations were separated by a permeable Transwell^TM membrane, enhancement was completely abrogated (Fig. 6b, not significant, n = 4). This confirmed further the absolute requirement for CD56⁺ cells to be present for rOv-ASP-1-mediated antigen-specific enhancement of the IFN-γ response to take place and also demonstrated that cell–cell contact between CD56⁺ and CD56⁻ is critical.

Fig. 6.

Enhancement of anti-tetanus toxoid (TT) interferon (IFN)-γ enzyme-linked immunospot responses by recombinant Onchocerca volvulus activation-associated secreted protein-1 (rOv-ASP-1) (5 μg/ml) requires contact between CD56⁺ and CD56⁻ cells. The left-hand panel (a) shows normal healthy donor (n = 4) anti-TT (open circles) and net anti-TT responses in the presence of rOv-ASP-1 [(TT + ASP-1) − ASP-1, solid circles] of mixed, fractionated CD56⁻ and CD56⁺ cells reconstituted in the same proportions as the original unfractionated peripheral blood mononuclear cells (PBMCs). The right-hand panel (b) shows responses of CD56⁻ cells reconstituted in the same proportions as the original unfractionated PBMCs, but with the CD56⁺ cells separated by a Transwell^TM membrane. Mean values are indicated by bold solid lines.

Discussion

We show here that a parasite protein, rOv-ASP-1, has immunostimulatory properties for human IFN-γ cellular recall responses in vitro to HCVco (in c-HCV donors) and TT (in NHD subjects previously vaccinated against tetanus). This finding extends the immunostimulatory activity of rOv-ASP-1 beyond our previous report of the adjuvanticity of the parasite protein for primary immune responses in vaccinated mice [1] to enhancing human responses to recall pathogen antigens in vitro. With the aim of elucidating the potential mechanism by which rOv-ASP-1 enhances human PBMC responses to HCVco and TT, we have also identified CD56⁺ cells as being critical to the immunostimulatory effect of the protein.

All NHD subjects tested reacted to rOv-ASP-1 by secreting IFN-γ, albeit with a wide range of absolute responses. Similarly, in our earlier vaccine studies, the recombinant protein induced IFN-γ secretion by mouse spleen cells from five Balb/c and five C57/BL6 animals with less variation in response than in humans. Denaturing the protein by boiling for 10 min removed its bioactivity, suggesting that the conformation of the protein or a subunit is essential for binding to a receptor (A. J. MacDonald, unpublished observations). The nature of the receptor and which cells bind Ov-ASP-1 are subjects of ongoing studies.

In 77·3% of NHD (Fig. 2b) and 88·9% of c-HCV subjects (Fig. 3b), rOv-ASP-1 increased the frequency of antigen-specific IFN-γ-secreting PBMCs in response to TT and HCVco antigens respectively. When we looked for possible determinants of enhancement of recall responses by rOv-ASP-1, regression analysis revealed no significant relationship between the amount of enhancement and either the magnitude of the donor's initial response to antigen (i.e. without addition of rOv-ASP-1) or the response to rOv-ASP-1 alone in both NHD and c-HCV subjects. In both groups, there were incidences of previously undetectable responses to antigen being induced to detectable levels by the addition of rOv-ASP-1 (two of 22 anti-TT in NHD and three of 18 anti-HCVco in c-HCV patients). This phenomenon could be of particular therapeutic benefit in restoring anti-HCV responses in c-HCV, where such responses tend to decline with years of infection [9,10]. Interestingly, the infecting genotype of the c-HCV donors (see Materials and methods) in the present study neither determined the responsiveness to HCVco (implying that the antigen has conserved T cell epitopes across different HCV genotypes) nor the ability of rOv-ASP-1 to enhance the anti-HCVco IFN-γ response. Patients with genotypes 2/3 responded to HCVco and were enhanced by Ov-ASP-1 as well as those infected with genotype 1 from which the HCVco antigen was derived. HCVco has many T cell epitopes that are conserved across different genotypes [11,12].

Based on the results shown in Fig. 1, rOv-ASP-1-induced IL-10 may also have been present in ELISPOT assays. Despite this, antigen-specific IFN-γ responses were enhanced in the great majority of NHD and, especially c-HCV donors. This is analogous to the adjuvant effect of Ov-ASP-1 in vaccinated mice where, despite the theoretical presence of Ov-ASP-1-induced IL-10 in vivo (based on stimulation of normal mouse splenocytes in vitro), the protein has potent Th1-inducing immunostimulatory activity [1]. It is possible that in natural infection with O. volvulus larvae, co-induction of IL-10 helps to limit damaging effects of too potent a Th1 response. Our sequence and structural homology analyses of rOv-ASP-1 with the crystallized closely related hookworm secreted recombinant protein Na-ASP-2 [13] suggested that rOv-ASP-1 has two or three distinct domains, each of which could contain the bioactive sites (S. Lustigman, unpublished observations). If the IL-10 and IFN-γ-inducing activities are associated with different domains, then it may be possible to express a subunit that stimulates IFN-γ without concurrent IL-10 secretion.

From the results of early cell fractionation experiments, it became clear that the CD56⁺ fraction was the primary population responding to Ov-ASP-1 by secreting IFN-γ. By comparing whole PBMCs with CD56⁺-depleted PBMCs, it was apparent that enhancement of TT and HCVco-specific responses was dependent on the presence of CD56⁺ cells (Fig. 5) and that cell contact was required between these cells and other PBMCs (Fig. 6). The CD56⁺ fraction comprises natural killer (NK) T and NK cells and both cell types have been shown to secrete IFN-γ in response to helminth infection. Hepatic invariant NKT cells produce IFN-γ and IL-4 when exposed to schistosome eggs [14]. Of more relevance to rOv-ASP-1 was the observation of a secreted protein from the human hookworm, Necator americanus, that binds selectively to human NK cells and induces IFN-γ production [15]. The authors did not isolate the protein; however, it is known that hookworm excretory/secretory products contain ASPs homologous to the native Ov-ASP-1 protein and that are immunostimulatory in rodents [3,7,16].

CD56⁺ NK T cells can secrete IL-4 as well as IFN-γ[17] and a helminth closely related to O. volvulus is known to stimulate IL-4 production from mouse NK T cells [18]. However, given that we have failed to detect IL-4 (Fig. 1) in either human or mouse cells (A. J. MacDonald, unpublished observations) treated with rOv-ASP-1, it is likely that NK cells rather than NK T cells are being activated to secrete IFN-γ in our system.

In whole PBMC, there was no significant difference between responsiveness of c-HCV versus NHD cells to rOv-ASP-1 and no difference in baseline (unstimulated) IFN-γ secretion. However, after purification, unstimulated CD56⁺ cells from c-HCV donors secreted IFN-γ in higher frequencies than those from NHDs. This could be due to removal of an inhibitory cell type by the purification process or simply that CD56⁺ cells were more ‘activated’ in the c-HCV donors. However, it should be stressed that c-HCV CD56⁺ cells in six of seven donors still increased their frequency of IFN-γ secretion further (albeit not statistically significantly in the small group tested) after rOv-ASP-1 treatment (Fig. 4b) and, enhancement of anti-HCVco responses still required the presence of CD56⁺ cells (Fig. 5b). NK cells can show impaired functioning and altered subset distribution in c-HCV versus NHD donors [19]. Our study shows that even if there may be functional differences in CD56⁺ cells from c-HCV compared with NHD subjects, cells from both populations respond to rOv-ASP-1 and are equally amenable to its immunostimulatory effect on antigen-specific responses in vitro. It is important to stress that we have measured only one readout of CD56⁺ cell activation (i.e. the frequency of IFN-γ-secreting cells) and the most important index of activation by rOv-ASP-1, probably a cell-surface molecule (soluble factors seem not to be essential as indicated by the data presented in Fig. 6), remains unknown at present.

Once activated by Ov-ASP-1, NK cells may facilitate IFN-γ production by antigen-specific T cells via cell–cell interaction with either T cells (bystander activation) or rOv-ASP-1-activated APC which would have enhanced antigen uptake and processing as well as co-stimulatory activity leading to a greater frequency of antigen-specific IFN-γ T cells.

From the perspective of a host–parasite relationship, the biological advantage of the helminth in stimulating a strong IFN-γ (Th1-type) response by the host could lie in the resulting down-regulation of protective Th2 responses, as suggested by Hsieh et al. [15], with the hookworm NK-cell activating secreted protein. An alternative, somewhat unconventional possible explanation comes from a study where host NK cells were required for the growth and development of the filarial helminth, Brugia malayi (closely related to O. volvulus). In NK-deficient mice, the parasite failed to develop. Activation of remaining NK cells in another mouse strain with diminished NK cell activity made the animals permissive to the infection [20]. Thus, a product(s) of activated NK cells may be necessary for the successful development of O. volvulus larvae.

In summary, we have described a helminth-derived protein that can act as an immunostimulant for human recall responses to TT and, perhaps more importantly, HCV antigens in patients with chronic HCV infection. As well as being an adjuvant for acquired immune responses, rOv-ASP-1 has possible applications for the direct stimulation of NK cells against viral pathogens and tumours.