Envelope-Specific IgG3 and IgG1 Responses Are Associated with Clearance of Acute Hepatitis C Virus Infection

Hepatitis C virus (HCV) can be cleared naturally in a subset of individuals. However, the asymptomatic nature of acute HCV infection makes the study of the early immune response and defining the correlates of protection challenging. Despite this, there is now strong evidence implicating the humoral immune response, specifically neutralising antibodies, in determining the clearance or chronicity outcomes of primary HCV infection. In general, immunoglobulin G (IgG) plays the major role in viral neutralisation. However, there are limited investigations of anti-HCV envelope protein 2 (E2) isotypes (IgM, IgG, IgA) and IgG subclasses (IgG1–4) in early HCV infection. In this study, using a rare cohort of 14 very recently HCV-infected individuals (4–45 days) with varying disease outcome (n = 7 clearers), the timing and potency of anti-HCV E2 isotypes and IgG subclasses were examined longitudinally, in relation to neutralising antibody activity. Clearance was associated with anti-E2 IgG, specifically IgG1 and IgG3, and appeared essential to prevent the emergence of new HCV variants and the chronic infection outcome. Interestingly, these IgG responses were accompanied by IgM antibodies and were associated with neutralising antibody activity in the subjects who cleared infection. These findings provide novel insights into the early humoral immune response characteristics associated with HCV disease outcome.


Introduction
Hepatitis C virus (HCV) is a major cause of chronic liver disease globally [1,2]. It is predominantly transmitted via blood-to-blood contact associated with injecting drug use (IDU), via sharing of injecting equipment [3,4]. Following acute infection, approximately 75% of people fail to clear the virus, resulting in chronic hepatitis [5][6][7]. While direct-acting antiviral (DAAs) treatments are decreasing the number of people living with HCV, control remains challenging globally due to the limitations in health infrastructure and high drug cost for treatment of the marginalised population affected, in addition to high rates of re-infection [8][9][10][11][12][13]. As HCV can be cleared naturally in a subset of individuals (~25%), and as subsequent re-infections are characterised by higher clearance rates (up to~80%), it is reasonable to hypothesise that a vaccine could be designed to elicit effective immune responses that confer protection [14,15].
were collected every six months to screen for seroconversion and for HCV RNA positivity; upon incident infection, subjects were sampled frequently for 24 weeks until infection outcome was resolved and antiviral treatment offered, and then three to six monthly thereafter. Early incident cases (n = 14), designated when a time point was available that was antibody-negative and HCV RNA-positive, were selected for this analysis. The date of infection was estimated by subtracting the average pre-seroconversion window period, which was estimated at 51 days [52][53][54][55], from the midpoint between the last seronegative and first seropositive time points. Details of the study protocol are reported elsewhere [12,[49][50][51].

Ethics Statement
Ethical approvals were obtained from New South Wales Department of Corrective Services (reference number 05/0884), the University of New South Wales (reference numbers 05094, 08081), and Human Research Ethics Committees of Justice Health, (reference numbers GEN 31/05 and G304/11), 6 December 2011, all located in Sydney, Australia. Written informed consent was obtained from the participants. All methods were performed in accordance with the ethical guidelines of the 1975 Declaration of Helsinki.

Virologic Assessments
All sera were tested for HCV antibodies using the Abbott ARCHITECT anti-HCV chemiluminescent microparticle immunoassay (Abbott Diagnostics, Chicago, IL, USA). HCV RNA detection was performed using either the VERSANT HCV RNA Qualitative Transcription Mediated Amplification (TMA) assay (Bayer Diagnostics, Emeryville, CA, USA; lower limit of detection: 3200 copies/mL) for samples collected prior to July 2008, or the COBAS AmpliPrep/COBAS TaqMan HCV assay (Roche, Basel, Switzerland; lower limit of detection 223 genome copies/mL) for samples from August 2008 onward [56].
Transient transfection of expression plasmids was performed using 293-F cells (FreeStyle 293 HEKs, Thermo Fisher Scientific, Waltham, MA, USA) using DNA-293fectin reagent (Life technologies, Carlsbad, CA, USA) as per the manufacturer's instructions. Media was harvested at approximately 96 h post transfection for purification. This supernatant was then passed through a 1-mL HiTrap Chelating HP chromatography column (Amersham, Buckinghamshire, UK) charged with 0.1 M NiSO 4 ·6H 2 O and washed with 10 mL of binding buffer. Proteins were eluted from the column by passing through 50 mM imidazole in binding buffer. Elution fractions were pooled, and native PAGE, SDS-PAGE, and Western blotting were performed with the HCV-specific monoclonal antibody, CBH4G, as well as the AviTag antibody. Additionally, an ELISA was performed with nine previously characterised HCV monoclonal antibodies (mAbs) to confirm the conformation of the recombinant E2 (rE2) protein.

Enzyme-Linked Immunoassays
Microtitre 96-well plates 96-well (Nunc Maxisorb, Thermo Fisher Scientific) were prepared with rE2 (H77 for genotype 1 and 2 infected subjects or UNK3a.13.6 for genotype 3 infected subjects) and incubated for one hour . Plates were washed three times with Tris-buffered saline with Tween (TBS-T) and then blocked for one hour with blocking buffer (5% non-fat dry milk in TBS-T). The bound rE2 was then incubated with plasma (heat inactivated by incubation at 56 • C for 30 min) at a final dilution of 1:10 for 1.5 h. The bound plasma was then incubated for one hour with anti-human IgG conjugated with horseradish peroxidase (HRP) (Jackson Immunoresearch, West Grove, PA, USA, 1:6000 for genotype 1a and 3a), anti-human IgA (α-chain-specific)-HRP (Sigma-Aldrich, St. Louis, MO, USA, 1:3000 for genotype 1a and 1:1500 for genotype 3a), anti-human IgM (µ-chain-specific)-HRP (Sigma-Aldrich, 1:3000 for genotype 1a and 1:1500 for genotype 3a), anti-human IgG1 Fc-HRP (SouthernBiotech, Birmingham, AL, USA, 1:6000 for genotype 1a and 1:1500 for genotype 3a), anti-human IgG2 Fc-HRP (SouthernBiotech, 1:6000 for genotype 1a and 1:1500 for genotype 3a), anti-human IgG3 hinge-HRP (SouthernBiotech, 1:6000 for genotype 1a and 3a), and anti-human IgG4 Fc-HRP (SouthernBiotech, 1:6000 for genotype 1a and 1:1500 for genotype 3a). The 3,3 ,5,5 -tetramethylbenzidine (TMB) liquid substrate for ELISA (Sigma-Aldrich) was added to the plate and incubated for 15 min after absorbance was read at 450 nm. In order to define a true positive result, a cut-off value was calculated for each assay as the mean + 3 SD of signal/noise (S/N) values in plasma from 10 healthy donors; these were as follows: genotype 3a; 1.87 S/N for IgA, 1.39 S/N for IgG and 2.89 S/N for IgM, 1.54 S/N for IgG1, 1.54 S/N for IgG2, 1.33 S/N for IgG3, and 2.9 S/N for IgG4 and genotype 1a; 1.67 S/N for IgA, 1.68 S/N for IgG and 3.91 S/N for IgM, 1.19 S/N for IgG1, 1.17 S/N for IgG2, 1.74 S/N for IgG3, and 2.55 S/N for IgG4. Results were normalised and converted to S/N for analysis. The mAbs CBH-5 and CBH-7 were used as a positive control for IgG and IgG1, respectively. For the remaining subclasses and isotypes, serum samples that previously tested positive in comparable ELISAs were kindly supplied as positive controls (Dr Kim Wilson, National Reference Laboratory, Melbourne, Australia). Unless specified, all subclass and isotype assays were performed in intra-assay duplicates.

Statistical Analysis
All data analysis was performed and all graphs were created using GraphPad Prism Software (version 7.0, La Jolla, CA, USA) for Macintosh. Wilcoxon rank-sum tests, Kolmogorov-Smirnov tests and Mann-Whitney tests were used (where appropriate) to evaluate statistically significant differences between groups. Statistical significance was defined as a p-value less than 0.05.
For the analysis of temporal trends in relation to viraemia, the period during primary infection in which the HCV viral load consistently declined was designated here as at least two consecutive time points, immediately after the peak HCV viral load time point (occurring at <90 days post infection (DPI), as described previously [61,62]). To be included in this analysis, consecutive time points were required to decline by >0.076 log 10 IU/mL when compared to the HCV viral load of the previous time point (corresponding with the intra-assay variability previously reported for the COBAS Taqman Assay [63,64]). All sequential time points that declined were included. The endpoint for inclusion in the analysis was an HCV RNA increase by >0.076 log 10 .

Patient Description
Longitudinally collected samples from 14 newly viraemic, seronegative subjects were selected for this study. These 14 subjects were described extensively elsewhere [19]. In brief, the 14 subjects had a primary HCV infection with the most prevalent Australian strains, including genotypes 1a (n = 5), 1b (n = 3), 3a (n = 4), and 2b (n = 2) [13]. The estimated days post infection (DPI) of the initial infection time point ranged from four to 45 days (median 30). Of the 14 subjects, seven naturally cleared the primary infection, and seven developed chronic infection (Table 1).

The Timing and Magnitude of Anti-E2 IgA and IgM Are Not Associated with Clearance, Whereas an Early Anti-E2 IgG Response Is
Anti-E2 IgM, IgG, and IgA responses were examined in the longitudinally collected sera to determine which isotype responses were being generated and whether these responses were associated with clearance or chronic infection outcomes.
In this study, five of seven chronic progressors had detectable anti-E2 IgM reactivity, whereas all seven clearers had detectable anti-E2 IgM, at any one time during infection ( Figure 1). In 10 of the 12 subjects with anti-E2 IgM, responses persisted and were relatively stable throughout infection. In the two remaining subjects, IgM responses were transient. The timing and magnitude of anti-E2 IgM responses were assessed in the clearers and chronic progressors with a reactive response to determine whether anti-E2 IgM was associated with disease outcome ( Three of seven clearer subjects were found to have anti-E2 IgA responses, whereas only one of the seven chronic progressors was found to have anti-E2 IgA responses. In two of the subjects with anti-E2 IgA, responses were stable and persisted. In contrast, in the remaining two subjects, responses were transient ( Figure 1). As only one chronic progressor had detectable anti-E2 IgA, the timing and magnitude of these responses were not compared statistically ( Figure 2).
All seven clearers and all seven chronic progressors were found to have anti-E2 IgG responses. For the clearers, the anti-E2 IgG responses increased in magnitude up until~100 DPI. Following this increase, levels were sustained for n = 3 subjects (277_Cl, 360_Cl, and 4032_Cl), whereas levels started to decrease concurrent with the loss of detectable virus in the plasma for the remaining n = 4 subjects (168_Cl, 306_Cl, 4087_Cl, and 686_Cl). This contrasted with the pattern observed in chronic progressors, where anti-E2 IgG levels increased over the course of infection, concordant with ongoing viraemia ( Figure 1).
Clearers had an earlier anti-E2 IgG response (median of 44 DPI (range 5-74)) when compared to chronic progressors (median 80 DPI (range 58-365), Mann-Whitney; p = 0.0029). These dynamics were similar to those observed with nAb activity as recently published [19]. Additionally, clearers had an earlier peak anti-E2 IgG response (median 85 DPI (range 44-245)), when compared to chronic progressors (median of 310 DPI (range 96-704), Mann-Whitney; p = 0.0122) ( Figure 2). However, no difference was observed in the magnitude of the response between clearers (median of 26.73 S/N (range 1.72-30.42)) and chronic progressors (median of 24.14 S/N (range 10.39-31.38), Mann-Whitney; p = 0.3176).  The timing of anti-E2 IgG responses was compared to the appearance of non-envelope antibody responses (i.e., anti-core, NS3 and NS5) measured using the Abbott ARCHITECT anti-HCV CIA assay. In clearers, anti-E2 IgG responses were detected before or in close proximity to development of responses against core and NS proteins (median of −15 days (range −54-0 days), Wilcoxon ranksum test; p = 0.0625). In chronic progressors, anti-E2 IgG responses were often delayed until after the response to core and NS proteins (19 median days, range 0-307 days, Wilcoxon rank-sum test; p = 0.0625, Figure 3A). This time difference (seroconversion to core/NS (DPI)-appearance of anti-E2 IgG (DPI)) between clearers and chronic progressors was significant (p = 0.0035, Figure 3B). The timing of anti-E2 IgG responses was compared to the appearance of non-envelope antibody responses (i.e., anti-core, NS3 and NS5) measured using the Abbott ARCHITECT anti-HCV CIA assay. In clearers, anti-E2 IgG responses were detected before or in close proximity to development of responses against core and NS proteins (median of −15 days (range −54-0 days), Wilcoxon rank-sum test; p = 0.0625). In chronic progressors, anti-E2 IgG responses were often delayed until after the response to core and NS proteins (19 median days, range 0-307 days, Wilcoxon rank-sum test; p = 0.0625, Figure 3A). This time difference (seroconversion to core/NS (DPI)-appearance of anti-E2 IgG (DPI)) between clearers and chronic progressors was significant (p = 0.0035, Figure 3B). Viruses 2020, 12, x FOR PEER REVIEW 10 of 21 Figure 3. Comparison of E2-specific IgG binding and non-E2 specific binding in clearer and chronic progressor subjects. The timing (days post infection) of anti-E2 specific IgG as measured by ELISA was compared to the timing of core/non-structural protein (NS) seroconversion as measured by the commercial Abbott assay (A) for clearers (blue) and chronic progressors (red). The difference in days between the appearance of anti-E2 IgG and anti-core/NS was compared between clearers (blue) and chronic progressors (red) (B).

An Early Anti-E2 IgG1 Response Is Associated with Clearance
The IgG subclass utilisation of the anti-E2 responses was examined to understand whether specific subclasses are utilised preferentially, as reported for HIV and other acute viral infections [21,65,66].  Comparison of E2-specific IgG binding and non-E2 specific binding in clearer and chronic progressor subjects. The timing (days post infection) of anti-E2 specific IgG as measured by ELISA was compared to the timing of core/non-structural protein (NS) seroconversion as measured by the commercial Abbott assay (A) for clearers (blue) and chronic progressors (red). The difference in days between the appearance of anti-E2 IgG and anti-core/NS was compared between clearers (blue) and chronic progressors (red) (B).

An Early Anti-E2 IgG1 Response Is Associated with Clearance
The IgG subclass utilisation of the anti-E2 responses was examined to understand whether specific subclasses are utilised preferentially, as reported for HIV and other acute viral infections [21,65,66].    (range 4.27-51.57)). Limited anti-E2 IgG2 was detected throughout infection, and the responses that were detected were transient and low in magnitude-only marginally above cut off (Figure 4). Similarly, anti-E2 IgG4 responses were not detected in any of the seven clearer subjects and in only three of the seven chronic progressors (Figure 4) in whom the responses were transient and low in magnitude.

The Initial Decline in HCV RNA Is Associated with Anti-E2 IgG1 and IgG3 Responses in Subjects Who Clear
At HCV RNA peak time points, all chronic progressors and the majority of clearer subjects were negative for IgG1 and IgG3. Therefore, no significant difference in IgG1 (Mann-Whitney; p > 0.9999) or IgG3 titres (Mann-Whitney; p = 0.4615) ( Figure 6) was observed. The period referred to here as the HCV RNA decline (see Section 2 for definition) was evident in all seven clearers and seven chronic progressors between 16 and 130 DPI. At the HCV RNA decline (Table S1, Supplementary Materials) a potent IgG1 and IgG3 response was detected in all seven clearers. In contrast, the IgG1 and IgG3 Limited anti-E2 IgG2 was detected throughout infection, and the responses that were detected were transient and low in magnitude-only marginally above cut off (Figure 4). Similarly, anti-E2 IgG4 responses were not detected in any of the seven clearer subjects and in only three of the seven chronic progressors (Figure 4) in whom the responses were transient and low in magnitude.

The Initial Decline in HCV RNA Is Associated with Anti-E2 IgG1 and IgG3 Responses in Subjects Who Clear
At HCV RNA peak time points, all chronic progressors and the majority of clearer subjects were negative for IgG1 and IgG3. Therefore, no significant difference in IgG1 (Mann-Whitney; p > 0.9999) or IgG3 titres (Mann-Whitney; p = 0.4615) (Figure 6) was observed. The period referred to here as the HCV RNA decline (see Section 2 for definition) was evident in all seven clearers and seven chronic progressors between 16 and 130 DPI. At the HCV RNA decline (Table S1, Supplementary Materials) a potent IgG1 and IgG3 response was detected in all seven clearers. In contrast, the IgG1 and IgG3 responses in chronic progressors during this period were low or not detected (Mann-Whitney; p = 0.0008 and p = 0.0111 for anti-E2 IgG1 and IgG3, respectively) ( Figure 6). responses in chronic progressors during this period were low or not detected (Mann-Whitney; p = 0.0008 and p = 0.0111 for anti-E2 IgG1 and IgG3, respectively) ( Figure 6).

Co-Occurrence of IgM/IgG Occurs in Samples with nAb Activity in Subjects Who Clear
To understand isotype(s) which may correlate with nAb activity, the longitudinal nAb data generated on the same cohort and time points used in this study were analysed in the context of anti-E2 isotypes (Table S1, Supplementary Materials) [19]. In samples where nAb responses (50% inhibition of HCVpp at 1/40 dilution of serum) were present, IgG reactivity accompanied by either an anti-E2 IgM or an anti-E2 IgA response was observed. Therefore, the ratio of anti-E2 IgA/IgG and anti-E2 IgM/IgG was calculated for time points where nAb responses were detected, and for time points where nAb responses were not detected. Clearers were found to have significantly higher anti-E2 IgM/IgG (Kolmogorov-Smirnov; p = 0.0240) (Figure 7) at time points with nAb when compared to chronic progressors. No significance was observed between clearers and chronic progressors for anti-E2 IgM/IgG in time points where nAb responses were not detected (Kolmogorov-Smirnov; p = 0.1178). Furthermore, no significance was observed for anti-E2 IgA/IgG between clearers and chronic progressors in time points both with (Kolmogorov-Smirnov; p = 0.0928), and without nAbs detected (Kolmogorov-Smirnov; p = 0.9840). Figure 6. Anti-E2 IgG1 and IgG3 responses were examined at the peak HCV RNA time point and at the initial decline in viral load for both clearer and chronic progressor subjects. Anti-E2 IgG subclass responses at the peak HCV RNA time points were compared for anti-E2 IgG1 (A) or anti-E2 IgG3 (B). Additionally, time points where initial HCV RNA decline occurred were compared for clearers and chronic progressors for anti-E2 IgG1 (C) and anti-E2 IgG3 (D) responses.

Co-Occurrence of IgM/IgG Occurs in Samples with nAb Activity in Subjects Who Clear
To understand isotype(s) which may correlate with nAb activity, the longitudinal nAb data generated on the same cohort and time points used in this study were analysed in the context of anti-E2 isotypes (Table S1, Supplementary Materials) [19]. In samples where nAb responses (50% inhibition of HCVpp at 1/40 dilution of serum) were present, IgG reactivity accompanied by either an anti-E2 IgM or an anti-E2 IgA response was observed. Therefore, the ratio of anti-E2 IgA/IgG and anti-E2 IgM/IgG was calculated for time points where nAb responses were detected, and for time points where nAb responses were not detected. Clearers were found to have significantly higher anti-E2 IgM/IgG (Kolmogorov-Smirnov; p = 0.0240) (Figure 7) at time points with nAb when compared to chronic progressors. No significant difference was observed between clearers and chronic progressors for anti-E2 IgM/IgG in time points where nAb responses were not detected (Kolmogorov-Smirnov; p = 0.1178). Furthermore, no significant difference was observed for anti-E2 IgA/IgG between clearers and chronic progressors in time points both with (Kolmogorov-Smirnov; p = 0.0928), and without nAbs detected (Kolmogorov-Smirnov; p = 0.9840). Anti-E2 isotype ratios were compared between clearers and chronic progressors at samples with neutralising antibody (nAb) activity and without nAb activity. Anti-E2 IgM/IgG and anti-E2 IgA/IgG was compared between clearers and chronic progressors in samples with nAb activity (nAbpositive, panels (A,B)) and without nAb activity (nAb-negative, panels (C,D)).

Discussion
This study is the first to examine longitudinal anti-E2 isotype and subclass responses in early primary HCV infection in subjects with clearance or chronic progression outcomes. Clearance was associated with earlier and more potent anti-E2 IgG, specifically IgG1 and IgG3, responses. Interestingly, these IgG responses were accompanied by IgM antibodies and correlated with nAb activity in the subjects who cleared infection. These results provide novel insights into the humoral immune response characteristics associated with disease outcome in early HCV infection.
It is generally believed in primary HCV infection that an anti-E2 IgG response is associated with nAb activity, which in turn is associated with viral clearance [20]. The results presented here show that chronic progressors develop anti-E2 IgG much later in infection than those that clear the virus. This delayed kinetics was most evident for IgG1 subclass responses, but anti-E2 IgG3 responses were also found in all clearer subjects but only a subset of chronic progressors. There were limited significant anti-E2 IgG2 or anti-E2 IgG4 responses found in either outcome group, consistent with previous reports [25,[67][68][69].
The initial viral load decline was correlated with both the anti-E2 IgG1 and anti-E2 IgG3 in subjects that cleared infection and is consistent with our recent finding in these same subjects that HCV RNA decline was associated with nAb activity in those who cleared infection [19]. IgG1 and IgG3 were implicated as the major mediators of nAb across various viral infections including West Nile virus, dengue, and HIV [42,[70][71][72]. These findings are also consistent with those of Chung et al., who observed in primary HIV infection that IgG3 antibodies likely co-operate with IgG1 to drive Fc- Figure 7. Anti-E2 isotype ratios were compared between clearers and chronic progressors at samples with neutralising antibody (nAb) activity and without nAb activity. Anti-E2 IgM/IgG and anti-E2 IgA/IgG was compared between clearers and chronic progressors in samples with nAb activity (nAb-positive, panels A and B) and without nAb activity (nAb-negative, panels C and D).

Discussion
This study is the first to examine longitudinal anti-E2 isotype and subclass responses in very early primary HCV infection in subjects with clearance or chronic progression outcomes. Clearance was associated with earlier and more potent anti-E2 IgG, specifically IgG1 and IgG3, responses. Interestingly, these IgG responses were accompanied by IgM antibodies and correlated with nAb activity in the subjects who cleared infection. These results provide novel insights into the humoral immune response characteristics associated with disease outcome in early HCV infection.
It is generally believed in primary HCV infection that an anti-E2 IgG response is associated with nAb activity, which in turn is associated with viral clearance [20]. The results presented here show that chronic progressors develop anti-E2 IgG much later in infection than those that clear the virus. This delayed kinetics was most evident for IgG1 subclass responses, but anti-E2 IgG3 responses were also found in all clearer subjects but only a subset of chronic progressors. There were limited significant anti-E2 IgG2 or anti-E2 IgG4 responses found in either outcome group, consistent with previous reports [25,[67][68][69].
The initial viral load decline was correlated with both the anti-E2 IgG1 and anti-E2 IgG3 in subjects that cleared infection and is consistent with our recent finding in these same subjects that HCV RNA decline was associated with nAb activity in those who cleared infection [19]. IgG1 and IgG3 were implicated as the major mediators of nAb across various viral infections including West Nile virus, dengue, and HIV [42,[70][71][72]. These findings are also consistent with those of Chung et al., who observed in primary HIV infection that IgG3 antibodies likely co-operate with IgG1 to drive Fc-dependent functions, such as antibody-dependent cell-mediated cytotoxicity (ADCC) [66]. Thus, future vaccine efforts in HIV and HCV, as well as other viral infections, may need to induce both anti-E2 IgG1 and anti-E2 IgG3.
The delayed anti-E2 IgG responses observed in chronic progressors in this study highlights a key unresolved question in the immunopathogenesis of primary HCV infection-why only some subjects develop an early and strong anti-E2 IgG1 response. The potential contributors to these delayed kinetics include the immunogenicity of the E1E2 of the particular transmitted-founder (T/F) variants [73], the regulation of humoral response, and the characteristics of the B-cell repertoire [74]. The delayed induction of IgG1 in chronic progressors could indicate delayed class switching and somatic hypermutation of the antibody due to inadequate help by cluster of differentiation 4-positive (CD4 + ) follicular helper T cells, a concept that is supported by a previous study examining HCV-specific CD4 + T cell responses which showed that, in a subset of subjects from the same cohort, chronic progressors had higher CD4 + regulatory-to-effector ratios. This, along with two recent studies investigating follicular helper T cells and HCV disease progression [75,76], suggests that HCV disease outcome may be pre-determined by the host [77]. It would be interesting to examine the IgG1 responses in chronic progressors to determine if they require extensive maturation to appear, thus indicating that the pre-existing B-cell repertoire is a natural determinant of outcome, as observed in HIV [78]. This is the first study to show that early co-occurrence of IgM is associated with the neutralising activity in the plasma of subjects with a clearance outcome from primary infection. An early IgM response may be protective because of its pentameric structure which results in high-affinity antibodies which can neutralise viruses before the development of an IgG response [38]. IgM also plays an important role in the maturation of B-cell responses regulating B-cell tolerance as well as class switching to IgG and IgA [40,79]. Results of this study are in keeping with those of vaccination studies in HIV, which showed that subjects with a higher anti-envelope IgA:IgG ratio after vaccination were more likely to become infected, potentially indicating reduced protective immunity [43]. It may be noteworthy that the present study examined these virus-specific antibody responses in association with clearance, whereas the HIV vaccine study was investigating protection-that is, reduced potential for establishment of infection [43]. It is well recognised that IgG and IgA antibodies can have neutralisationas well as Fc-dependent functions including ADCC [20,80].
Although the results reported here are both promising and informative for HCV pathogenesis and vaccine research, there are limitations in the current study. A key limitation was that the ELISA-based assays used here did not measure virion-bound antibody (that is, antibodies bound in immune complexes), only free antibodies in the plasma [81]. It was shown in HIV that the first detectable antibody responses are found in the form of immune complexes [82]. This may also be the case in primary HCV infection. Engagement in immune complex formation could explain why transient responses were observed in some subjects for both anti-E2 IgM and anti-E2 IgA. It should be noted, however, that several time points were examined including those which had very low levels of viraemia (i.e., available antigen) and transient responses were still observed. Another key area of further investigation to strengthen these data is the characterisation of class switching in HCV-specific B cells.
In conclusion, subjects able to mount an early anti-E2 IgG3 and anti-E2 IgG1 response, along with IgM responses, were more likely to achieve natural clearance of HCV. Thus, future vaccine design efforts may seek to promote such anti-E2 IgG1 and anti-E2 IgG3 responses, potentially along with IgM to foster nAb and potentially non-nAb activities.