Human IgG Subclasses: In Vitro Neutralization of and In Vivo Protection against West Nile Virus

Abstract

West Nile virus (WNV)-neutralizing intravenous immune globulins (IVIG) were fractionated into IgG subclasses, and the contribution of each subclass to in vitro neutralization of and in vivo protection against WNV was evaluated. The results indicate that IgG1 (i) is the main subclass induced following WNV infection of humans, (ii) contained nearly all the in vitro WNV neutralization capacity, and (iii) mediates effector functions in vivo that render it superior to other subclasses in protection against WNV. The importance of human IgG1 indicates that a candidate WNV vaccine should induce an immune response that includes WNV-specific IgG1.

West Nile virus (WNV), a flavivirus which has become endemic in North America (1) and has now also been recognized as a major public health concern in Europe (6), causes mostly asymptomatic infections (in ∼80% of cases), and less than 1% of WNV-infected persons develop neuroinvasive disease (7). There are currently no WNV-specific antiviral drugs available, and treatment of patients is limited to supportive care. To optimize efficacy of a potential vaccine, an understanding of the factors involved in successful WNV clearance after infection of humans is important. Some aspects of the immune response to WNV infection have been studied in detail, especially the affinities and specificities of WNV-neutralizing antibodies (Abs), properties which are crucial for effective virus neutralization (5, 8, 16, 26). During these investigations, it became evident that the murine immune response to WNV differs to some degree from that of humans; e.g., the strongly neutralizing Abs that recognize an epitope on the lateral ridge of domain III of the WNV envelope protein in mice are generated far less frequently in the human immune response (16, 27). Reports of protective immune responses following the induction of WNV-specific Abs without virus-neutralizing properties highlighted the involvement of Ab-mediated effector responses such as Ab-dependent cellular cytotoxicity, complement activation, and Ab interaction with Fc-γ receptors (FcγRs) in protection and immunity (3, 14, 15). The apparent importance of effector functions beyond virus neutralization in host defense renders an understanding of the profile of human IgG subclasses induced following WNV infection and the respective contributions of the subclasses to in vivo protection particularly interesting. The four human IgG subclasses, IgG1 to IgG4, differ in their affinities for FcγRs and therefore in their abilities to induce effector responses (2). In addition, IgG subclass-dependent effects on in vitro virus neutralization were shown recently for another flavivirus, dengue virus (20). Although accumulating evidence highlights the differential involvement of IgG subclasses in the humoral response against flavivirus infection, no information on the contributions of the different human IgG subclasses to WNV neutralization in vitro and, more importantly, their contributions to in vivo protection is currently available.

In this work, we used intravenous immune globulin (IVIG) lots manufactured from plasma samples collected in the United States, which have been shown to contain increasing amounts of WNV-neutralizing Abs since 2003 (18), reflecting the increasing cumulative exposure of the U.S. population to WNV. As IVIG is manufactured from plasma samples from thousands of healthy donors, the WNV-neutralizing Abs in IVIG are the result of successful immune responses against WNV infection. Two 10% IVIG lots (Gammagard Liquid; Baxter Healthcare Corp) (19, 25) known to contain WNV-neutralizing Abs (mean 50% virus neutralization titer [NT₅₀] ± standard deviation [SD] for lot 1, 27.6 ± 4.9 [no. of replicates, n = 7], and for lot 2, 10.1 ± 2.9 [no. of replicates, n = 6]) (Fig. 1 A) were fractionated into the respective IgG subclasses (Table 1) by using pH gradient elution from a protein A-conjugated affinity column in a procedure similar to one described previously (21). Fractions were diluted to identical protein concentrations of 30 mg/ml and tested for in vitro WNV neutralization using essentially a method described previously (18). A significantly higher in vitro WNV neutralization capacity was present within the IgG1 subclass fraction (NT₅₀ ± SD for lot 1, 14.1 ± 2.5 [n = 5]) than within the IgG2 subclass (NT₅₀ ± SD for lot 1, 1.1 ± 0.4 [n = 5]) or the IgG3 subclass (NT₅₀ ± SD for lot 2, 2.1 ± 0.4 [n = 6]) or even the parental IVIG preparations (NT₅₀ ± SD for lot 1, 8.2 ± 1.6 [n = 5]; NT₅₀ ± SD for lot 2, 2.6 ± 0.3 [n = 2]) (Fig. 1A). The contribution of each IgG subclass to WNV neutralization was then calculated relative to the overall WNV neutralization capacity of each IVIG lot and the physiological concentration of each IgG subclass in IVIG. (i) The contribution of each IgG subclass to WNV neutralization was calculated from the mean NT₅₀ determined for the individual IgG subclass, which was divided by the mean NT₅₀ determined for the whole IVIG lot from which the subclass was fractionated (lot 1 for IgG1 and IgG2 and lot 2 for IgG3) (Table 1) and multiplied by the factor 100; e.g., the contribution of IgG1 was calculated as follows: the IgG1 mean NT₅₀ of 14.1 was divided by 8.2 (the mean NT₅₀ of IVIG lot 1) and multiplied by the factor 100, yielding a WNV NT 172% of that of parental IVIG (mean percent WNV NT of parental IVIG ± SD: IVIG, 100% ± 17% [n = 7]; IgG1, 172% ± 31% [n = 5]; IgG2, 13% ± 5% [n = 5]; and IgG3, 81% ± 14% [n = 6]) (Fig. 1B). (ii) The value was then adjusted for the actual subclass content in 10% IVIG, i.e., 60 mg/ml for IgG1, 30 mg/ml for IgG2, and 5 mg/ml for IgG3; e.g., the value for IgG1 was adjusted as follows: 172% × 0.6 = 103% (Fig. 1B; Table 1). The obtained results indicate that nearly all WNV neutralization activity in the tested IVIG lots was due to IgG1 (calculated value, 103%), whereas IgG2 and IgG3 had only minor WNV-neutralizing activities (calculated values, 4% each) (Fig. 1B). Previous work already showed the importance of WNV-neutralizing Abs in viral clearance, with details on immunodominant WNV epitopes and neutralizing Ab affinity, neutralization kinetics, and mechanisms of neutralization described in the earlier studies (8, 14, 17, 26, 28). This is the first investigation, however, of the involvement of the different human IgG subclasses, as induced in humans after WNV infection, in WNV neutralization, and the data show that WNV-neutralizing Abs of humans who have successfully overcome a WNV infection are mainly of the IgG1 subclass.

FIG. 1.

In vitro WNV (isolate 385-99) neutralization capacities of whole human polyclonal IgG or human polyclonal IgG subclass fractions. (A) Two IVIG lots were tested undiluted or diluted to a protein content of 30 mg/ml or were fractionated into IgG subclasses and tested at a protein content of 30 mg/ml. Neutralization assays were done at least five times, with the exception of the assay of lot 2 at 30 mg/ml, which was performed twice. (B) The contribution of each IgG subclass to WNV neutralization (expressed as a percentage) was calculated from the mean NT₅₀ determined for the individual IgG subclass, which was divided by the mean NT₅₀ determined for the whole IVIG lot from which the subclass fraction was fractionated and multiplied by the factor 100 (results are represented by black diamonds). The value was then adjusted for the actual subclass content in 10% IVIG (results are represented by gray bars), i.e., 60 mg/ml for IgG1, 30 mg/ml for IgG2, and 5 mg/ml for IgG3 (Table 1). Error bars represent standard deviations determined by Student's t test (unpaired, two-tailed); statistical analysis was performed using GraphPad Prism v5.0 software (San Diego, CA). ***, P < 0.0005 versus the parental IVIG lot. NS, not significant.

TABLE 1.

Biochemical parameters of the IVIG lots and the respective fractionated IgG subclasses^a

IVIG lot and subclass fraction	Amtd (g)	Protein			ELISAb result (mg/ml) for:				HP-SEC result (% peak area)c
		Concn (%)	Content (g)	Yielde (%)	IgG1	IgG2	IgG3	IgG4	Polymers (>450 kDa)	Dimers (∼310 kDa)	Monomers (∼150 kDa)	Fragmented (<100 kDa)
Lot 1		10	98.8	100	60.5	27.6	5.0	2.2	0.2	11.0	88.2	0.7
IgG1	614	7.7	47.4	48/69f	56.6	6.8	0.4	2.1	0.5	11.9	87.4	0.2
IgG2	222	7.5	16.7	17/24f	<3.1	69.8	1.9	0.8	0.3	3.5	96.1	0.2
IgG3g	67	2.4	1.6	2/3f	0.2	<0.04	28.4	0.02			4.1	95.8
Lot 2		10	98.8	100	71.8	33.9	5.4	2.3	0.2	9.5	89.8	0.5
IgG3	59	2.4	1.5	1.5/3f	0.1	>0.0	37.6	>0.0	3.1	8.8	74.9	13.2

^aTwo lots of IVIG were fractionated into the respective IgG subclasses, and the yield, purity, and structural integrity of each fraction were determined by measurement of protein content, ELISA, and high-performance size exclusion chromatography (HP-SEC), respectively.

^bA quantitative IgG subclass-specific ELISA, which was based on the single incubation multilayer immune technique (11) and had results calibrated using international certified reference material 470 (4) for which the IgG subclass distribution is known (22), was used (details are available upon request).

^cIgG solutions were applied to a 7.5- by 600-mm TSK-GEL column (Tosoh Corp.), and protein was detected using UV detector 1790 (Bio-Rad).

^dAmount of pooled and concentrated fraction.

^eYield (%) after ultra-/diafiltration = volume (g)/98.8 × protein concentration (%).

^fFraction yield after affinity chromatography, prior to ultra-/diafiltration; the difference indicates unspecific IgG loss at the concentration step.

^gThis IgG3 fraction was not stable during the course of the work, and only the IgG3 fraction of lot 2 was used in all experiments.

In vivo protection against an otherwise mostly lethal WNV challenge was evaluated using IgG subclass aliquots diluted to identical in vitro WNV neutralization titers, and significantly superior in vivo protection was mediated by the human IgG1 subclass (Fig. 2 B). Using WNV-susceptible inbred BALB/c mice (1), we determined the WNV neutralization titer required to protect 50% of the animals from an otherwise mostly lethal WNV infection (Fig. 2A). The animals were challenged subcutaneously (s.c.) with 10⁵ 50% tissue culture infective doses (TCID₅₀), which resulted in only 27% survival in control mice (mean survival rate ± standard error of the mean [SEM], 27% ± 11% [n =15]) (Fig. 2B). When mice were given IVIG or IgG subclass fractions that were adjusted to an NT₅₀ of 2.0, i.e., a concentration that conveyed near-complete protection (Fig. 2A), 2 h prior to WNV challenge, a mean survival rate ± SEM of 95% ± 5% (n = 20) was achieved by treatments with IVIG and ≥80% survival was seen with all the IgG subclass preparations (data not shown). However, when mice were treated with IVIG adjusted to the suboptimal neutralization capacity of an NT₅₀ of 0.2, only 50% (mean survival rate ± SEM, 50% ± 9% [n = 30]) survived the WNV challenge (Fig. 2). When mice were given IgG subclass fractions equally adjusted to an NT₅₀ of 0.2, similar survival rates for IgG2- and IgG3-treated mice were seen (P = 0.71; mean survival rate ± SEM, 33% ± 12% [n = 15] for both treatments), and these rates did not differ significantly from that for the untreated control group (mean survival rate ± SEM, 27% ± 11% [n = 15]; IgG2-treated group versus control group, P = 0.48; IgG3-treated group versus control group, P = 0.29) (Fig. 2B). Animals treated with IgG3 did, however, survive longer after the WNV challenge than IgG2-treated or untreated animals, where maximum lethality was reached on day 20 (d20) for IgG3-protected animals but already on d14 or d12 for IgG2-protected or unprotected animals, respectively (median survival times, 13d, 10d, and 9d for IgG3-protected, IgG2-protected, and unprotected animals; IgG2-treated group versus control group, P = 0.48; IgG3-treated group versus control group, P = 0.13) (Fig. 2B). There was a significant difference in the survival of IgG1-treated animals versus the control group, as 67% of animals that received IgG1 at the suboptimal neutralization capacity of an NT₅₀ of 0.2 survived the challenge (mean survival rate ± SEM, 67% ± 12% [n = 15]; P = 0.003), compared to 27% of animals in the untreated control group (Fig. 2B). In addition, treatment with IgG1 resulted in more animals surviving longer than in the IVIG-treated group (Fig. 2B). As all groups received IgG treatments with identical in vitro WNV neutralization capacities, the increased survival in the IgG1 group indicates the involvement of additional factors that are affected by IgG1, but not IgG2 or IgG3, during in vivo protection. IgG1, IgG2, and IgG4 have similar biologic half-lives (∼21d), whereas IgG3 is short-lived (half-life, ∼7d) (10), and the low level of IgG3-mediated in vivo protection could therefore be due to the more rapid catabolism of this IgG subclass.

FIG. 2.

In vivo protection against WNV (isolate 385-99) infection by IVIG and human IgG subclasses. (A) Suboptimal protection against WNV infection by IVIG preparations adjusted to different neutralization capacities. Six- to 8-week-old female BALB/c mice (Charles River Laboratories) were treated via s.c. injection on the contralateral side of the back with phosphate-buffered saline (PBS; 0) or IVIG that had been adjusted to a mean NT₅₀ of 0.2, 2.0, 2.6, or 4.4 2 h prior to s.c. challenge with 10⁵ TCID₅₀/200 μl WNV. Survival was monitored for 28d, and the NT₅₀ that conferred 50% survival was determined through nonlinear regression analysis. The results of 8 independent experiments that included at least 10 mice per treatment are shown; data are presented as mean and SEM values. (B) Protection against WNV infection by IVIG and the IgG subclasses fractionated therefrom, when adjusted to the suboptimal neutralization capacity of an NT₅₀ of 0.2. Mice were treated with 200 μl IVIG or IgG subclass-enriched fractions 2 h prior to challenge with 10⁵ TCID₅₀ WNV, and survival was monitored for 28d. Results are given as mean survival rates (expressed as percentages) from three independent experiments, with five mice per group used in each series. Survival curves were created by the Kaplan-Meier method using the product limit method and compared by the log-rank (Mantel-Cox) test. **, P = 0.003 versus the WNV-challenged control group. All animal experiments were done in compliance with the Austrian Animal Experiments Act (no. 501/1989).

Previous work already showed more effective adaptor functions of IgG1 than of other IgG subclasses with other parts of the immune system (13), mediated through FcγR specificity (12) and variations in the composition of the Fc fragment sugar moieties (13, 14). The IgG1 Fc portion was shown to be important in protection: e.g., the F(ab′)₂ portion of a nonneutralizing monoclonal antibody (MAb) against the NS1 protein of the flavivirus yellow fever virus was protective when joined to a mouse IgG2a (homologous to human IgG1) Fc portion but not when joined to mouse IgG1 or IgG2b (a homologue of human IgG4 or IgG3, respectively) (23). Similar results were later obtained with a virus-neutralizing MAb against the yellow fever virus E protein (24). Further work will be needed to elucidate the exact mechanisms of the IgG1-mediated in vivo protection against WNV.

In conclusion, the results obtained in this study demonstrate the importance of the IgG1 subclass in a protective immune response against WNV. As subclass restriction to IgG1 and IgG3 is mediated by cytokines of the Th1 subset of T helper cells, a candidate WNV vaccine should induce a balanced Th1 response, which can be mediated by, e.g., a whole virus particle rather than a split-virus formulation, as was shown recently for an H1N1 vaccine (9).