Dissecting the Contribution of IgG Subclasses in Restricting Airway Infection with Legionella pneumophila

Although cell-mediated adaptive immunity is generally known to be pivotal to fight intracellular pathogens, recent studies demonstrated an unexpected efficacy of Abs in mediating protection against such microorganisms (1, 2). Using the infectious agent of Legionnaires’ disease, Legionella pneumophila, as a model intracellular bacterial pathogen, we could show previously that a powerful L. pneumophila–specific Ab response is raised upon a primary infection in mice that provides local protection upon a secondary airway infection (3). Abs opsonize the bacteria and confer an FcR-mediated signal in L. pneumophila target cells that vetoes the bacterial evasion strategies from intracellular degradation by efficiently targeting the bacteria to lysosomal degradation (4).

Recently, we were able to identify protein B cell Ags of L. pneumophila that we successfully used to vaccinate mice (5). During these vaccination studies, we found that the vaccination regimen and adjuvants employed highly contributed to the efficacy of the induced L. pneumophila–specific Abs in restricting infection. Hence, not all animals having high L. pneumophila–specific total IgG titers were protected to the same extent from a secondary infection.

In mice there are four different IgG subclasses known, namely IgG1, IgG2a/c, IgG2b, and IgG3, which are determined by their respective C regions γ1, γ2a/c, γ2b, and γ3. The mutually exclusive occurrence of either IgG2a or IgG2c is determined by the genetic background of the mouse strain; however, both IgG2a and IgG2c are thought to have similar characteristics. The C region confers effector properties to the Ab subclass such as half-life, complement activation, and interactions with FcRs (6). For example, IgG1 is the only murine IgG subclass that is not able to activate complement. Additionally, the different subclasses have varying affinity for the activating FcRs FcγRI, FcγRIII, and FcγRIV as well as for the inhibitory FcγRIIB. The seroprevalence of the different IgG subclasses differs with IgG2a/c present at comparably high concentrations (∼1100 μg/ml) and IgG1 and IgG2b at intermediate levels (∼100 μg/ml), followed by IgG3 (∼20 μg/ml). Class switching of B cells from producing Ag-specific IgM to producing the IgG subclasses commonly depends on productive interactions with Th cells. The selection of a particular C region for class switch recombination is regulated by the cytokine milieu present during the cell-to-cell interactions. In the presence of inflammatory Th1 cytokines such as IFN-γ or other proinflammatory agents such as LPS, B cells tend to switch to the production of IgG2a/c and IgG3, whereas in the presence of the Th2 hallmark cytokine IL-4 and TGF-β, the production of IgG1 and IgG2b is induced, respectively (7, 8). Thus, characteristics of the infectious agent or Ag and the type of Th cell response elicited will skew the IgG subclass Ab profile. Protein Ags preferentially elicit an IgG1 response in mice, whereas viral Ag and bacterial polysaccharide Ags favor the class switch to IgG2a/c and IgG3, respectively (9-11).

In this study, we specifically aimed to analyze and compare the relative amounts of the different IgG subclasses present in animals exhibiting different ability to limit the magnitude of pulmonary L. pneumophila infection, and to assess their relative protective potential in the same model of airway infection.

To test the individual contribution of the different IgG subclasses in the pulmonary L. pneumophila infection model we employed two complementary approaches to achieve pure IgG subclass L. pneumophila–specific Ab samples. First, affinity purification of different IgG subclasses from polyclonal immune serum resulted in a mixture of IgG Abs of one subclass, encompassing different specificities and affinities for L. pneumophila, but also for other Ags. Second, the generation of monoclonal L. pneumophila–specific Abs allowed us to engineer L. pneumophila–specific Abs with identical specificity but differing IgG sublasses and hence effector function.

We found that although the Th1 hallmark subclasses IgG2c and IgG3 were prevalent in immunized mice, which efficiently restricted L. pneumophila infection, all IgG subclasses individually had the capacity to reduce bacterial burdens in a model of passive immunization when administered prophylactically at high doses. At lower concentrations or when normalized to their prevalence after L. pneumophila infection, IgG2c proved to be slightly more effective compared with the other IgG subclasses. Moreover, we show that i.v. application of purified monoclonal IgG (mIgG) is not only able to function in a prophylactic setting but also in a therapeutic one, as mAb transfer was able to reduce bacterial titers also when initiated after established L. pneumophila infection. Therefore, we propose that Ag-specific IgG might be a potent therapeutic agent against L. pneumophila infection in situations where antibiotics cannot be applied due to resistance or intolerance.

The L. pneumophila strains used in this study were the laboratory wild-type strain JR32 (Philadelphia-1, serogroup [sg] 1) (12) and an aflagellated mutant of JR32 (ΔflaA) (5). Additionally, we used isolates of sg4, sg6, and sg10 (Swiss National Reference Center for Legionella, Bellinzona, Switzerland) (13) as well as wild-type Legionella strains Corby (sg1) (14) and AA100 (sg1) (15) and Escherichia coli and Salmonella enterica serovar Typhimurium to determine Ab specificities. L. pneumophila was grown for 3 d at 37°C on charcoal yeast extract agar plates before use.

C57BL/6 mice (Janvier Labs, Le Genest Saint Isle, France) were used at 7–20 wk of age (sex and age matched within the experiments). All animal experiments were performed according to institutional policies and have been reviewed by the cantonal veterinary office.

Mice were immunized and boosted twice i.v. with 5 × 10⁶ CFU L. pneumophila JR32 to harvest immune serum 4 wk after the last boost. Similarly, mice were vaccinated three times s.c. with a mixture of four recombinantly expressed B cell protein Ags (Lpg1362, Lpg2271, Lpg0688, and Lpg2025, 10 μg each, emulsified in IFA supplemented with 5 μg CpG) or with PBS (emulsified in IFA with CpG s.c.) (5). For passive immunizations, mice were injected i.v. with 25 μl immune serum or with 5 μg purified Ab samples unless stated otherwise.

For intranasal (i.n.) infections mice were anesthetized with an i.p. injection of 5 μg xylazine/100 μg ketamine per gram body weight, and 5 × 10³ CFU L. pneumophila (ΔflaA) resuspended in 20 μl PBS were directly applied to one nostril using a micropipette (16). Where indicated, bacteria were opsonized prior to immunization by incubation for 1 h at 37°C with PBS, naive or immune murine serum, with serum harvested from s.c. vaccinated mice, or with purified Ab preparations. To determine bacterial titers in bronchoalveolar lavage (BAL), mice were euthanized and perfused on day 2 postinfection unless stated otherwise. BAL was extracted in 1 ml PBS and plated in serial dilutions on charcoal yeast extract plates. Colonies were enumerated after 3–4 d of incubation at 37°C. Mice from which no CFU could be recovered were assumed not to have been appropriately infected and were omitted from subsequent analysis.

Polyclonal Ab subclasses (polyclonal IgGs) were affinity purified from immune mouse serum by chromatography using murine subclass-specific Abs (SouthernBiotech, Birmingham, AL) coupled to AminoLink coupling resin (Thermo Scientific, Rockford, IL). Ab fractions were eluted using 0.1 M glycine buffer (pH 2.7). Eluates were immediately neutralized using 0.04 volumes 1 M Tris (pH 9). The presence and purity of L. pneumophila–specific Ab subclasses in samples were assessed by ELISA.

Splenocytes of a C57BL/6 mouse infected with 5 × 10⁶ JR32 and boosted on day 13 with the same dose were fused 3 d after the boost with the mouse myeloma cell line X63AG8.653 (17). This is a nonsecreting clone derived from P3X63AG8 (a BALB/c myeloma) and is recommended as a fusion partner for production of hybridomas (18). Fused cells were diluted and dispensed as single-cell suspensions in round-bottom 96-well plates. Growing clones were cultivated and screened for L. pneumophila specificity by ELISA, and positive clones were subjected to an additional round of single-cell dilutions to harvest monoclonal cell cultures. One positive clone (4E12.14) was used for mRNA extraction, reverse transcription, and amplification of rearranged variable light and heavy Ig gene segments. Variable L and H chain regions were cloned in-frame into plasmids containing the constant κ L chain region or different constant H chain regions γ1 (19), γ2c, γ2b, or γ3 (this study), respectively. The combined variable and constant gene segment of the κ L chain was combined with each of the four combined variable and constant γ gene segments separately to produce four novel plasmids, each encoding both L and H chains of Abs of one specific subclass, and with identical specificity for L. pneumophila. The four plasmids (pSW205, pSW206, pSW207, and pSW208 encoding anti–L. pneumophila IgG1, IgG2c, IgG2b, and IgG3, respectively) were transfected into HEK 293T cells (originally referred to as 293tsA1609neo) (20) using the Amaxa cell line Nucleofector kit V according to the manufacturer’s recommendations. Stably transformed cells were selected by their resistance to puromycin and were subjected to two rounds of single-cell dilution cloning to harvest monoclonal cell lines. Abs produced by the cells were collected and purified by protein G chromatography and elution under acidic conditions (0.1 M glycine, pH 2.7). Eluted fractions were immediately neutralized using 0.04 volumes 1 M Tris (pH 9) (mIgG1, mIgG2c, and mIgG2b) or with 10 volumes PBS (mIgG3). Purity and specificity were assessed by ELISA, SDS-PAGE, and Western blot. As a control Ab we used a similarly purified mIgG2c against lymphocytic choriomeningitis virus harvested from cell culture supernatant of a hybridoma cell line (clone 8B7-A100).

For Western blot analysis, Abs (1 μg/lane) or whole bacteria (1 × 10⁷ CFU/lane) were separated by SDS-PAGE and blotted onto nitrocellulose membranes. Membranes were blocked in TBS with 0.1% Tween 20 supplemented with 4% milk powder and incubated for 1 h at room temperature with purified Ab samples and with secondary or tertiary Abs. Abs used included HRP-conjugated goat anti-mouse IgG (Sigma-Aldrich Chemie, Buchs, Switzerland), goat anti-mouse IgG1, IgG2c, IgG2b, and IgG3 (SouthernBiotech), and HRP-conjugated donkey anti-goat IgG (Abcam, Cambridge, U.K.). Ultimately, stained and washed membranes were incubated with ECL reagent (Amersham Biosciences, Otelfingen, Switzerland) according to the manufacturer’s recommendations, and exposed medical x-ray films (Fuji, Tokyo, Japan) were analyzed with a Fujifilm developer (FPM 800A).

ELISA plates (Nunc MaxiSorp immunoplates, Thermo Scientific, Waltham, MA) were coated with 4 × 10⁶ CFU L. pneumophila per well in coating buffer (0.1 M NaHCO₃, pH 9.6), and ELISA experiments were performed according to standard protocols using serum from mice (naive, immunized by i.v. infection, or vaccinated with purified B cell Ags or L. pneumophila s.c.) or purified Ab samples as primary Ab and HRP-conjugated secondary Abs against mouse IgG (Sigma-Aldrich Chemie) (16). Alternatively, subclass-specific Ab determination was performed by using subclass-specific secondary Abs (goat anti-mouse IgG1, IgG2c, IgG2b, and IgG3, SouthernBiotech) and an HRP-conjugated tertiary Ab against goat IgG (Abcam) (16). Plates were developed with 100 μl/well of a solution containing 0.2 mg/ml ABTS, 0.1 M NaH₂PO₄, and 0.04% H₂O₂ (pH 4) and the plates were read at 405 nm in a SpectraMax Plus spectrophotometer (Molecular Devices, Sunnyvale, CA). Endpoint titer values were defined as the highest dilution giving a specific absorbance value OD greater than the threshold (e.g., 0.5) multiplied with the respective dilution factor.

Statistical analyses included verification of normal population distribution among experimental groups with the Shapiro–Wilk test and calculation of p values comparing two individual groups using a t test with Welch’s correction for unequal variances. A p value < 0.05 was considered significant.

Recently, we were able to identify protein B cell Ags of L. pneumophila and evaluated their potency in inducing a protective humoral immunity upon i.n. infection (5). To this end, we immunized mice s.c. with a pool of four recombinant B cell Ags (Lpg1362, Lpg2271, Lpg0688, and Lpg2025, emulsified in IFA supplemented with CpG) or infected mice with live L. pneumophila (i.v.) or injected PBS in IFA with CpG (s.c.) as positive and negative controls, respectively. After two boosts with the same regimen we challenged the animals i.n. with a dose of 5000 L. pneumophila and analyzed bacterial counts in BAL 2 d later (Fig. 1A). Mice that were immunized with live L. pneumophila were able to clear the bacteria very efficiently compared with mice that were treated with PBS. Mice that were vaccinated with the purified proteins also showed reduced bacterial titers 2 d after challenge; nevertheless, the live vaccine seemed to be more effective in controlling the infection. When we analyzed the total IgG titers in these animals, we found that despite the difference in efficacy of controlling the infection, both groups of vaccinated mice showed comparable amounts of L. pneumophila–specific Ab in serum (data not shown). Because the total IgG in mice is composed of four individual IgG subclasses, we next addressed the question of whether these subclasses were also comparably present in the vaccinated mice. By determining the titers of the different IgG subclasses, we found that in mice infected with live L. pneumophila, the subclasses IgG2c and IgG3 were more prevalent compared with mice that received the protein vaccine, which showed higher titers of IgG1 and IgG2b (Fig. 1B). Indeed, a formal correlation analysis of bacterial burden in the BAL with IgG titers in the serum showed that the presence of IgG2c and IgG3 correlated with better bacterial control during i.n. challenge with L. pneumophila (Fig. 1C). To exclude a potential contribution of adaptive cell–mediated immunity in our challenge model, we harvested the sera of the immunized and boosted animals and used these to opsonize L. pneumophila in vitro prior to inoculation into naive mice (Supplemental Fig. 1A). This procedure per se does not impair bacterial viability, as was confirmed by plating (data not shown). In line with the previous experiments, we recovered fewer bacteria from mice that were infected with bacteria that had been opsonized with serum from live L. pneumophila–infected mice in comparison to mice infected with bacteria treated with serum from protein-vaccinated animals (Fig. 1D). Hence, we conclude that the active agent in our experiments is a serum component that binds specifically to L. pneumophila, that is, Abs.

To evaluate the protective potential of individual IgG subclasses, we next aimed at purifying the different IgG subclasses from immune serum. To this end, we infected and boosted mice with live L. pneumophila i.v. and harvested the immune serum 4 wk after the last injection. This serum was applied sequentially to four IgG subclass-specific affinity columns whose resins consisted of subclass-specific Abs coupled to agarose beads (Supplemental Fig. 2A). This procedure allowed us to purify polyclonal murine Abs of the different subclasses independent of specificity. Because we used immune serum, the purified preparations contained L. pneumophila–specific Abs as well as Abs of various other specificities. The assessment of the purity of the preparations showed only minor contaminations with Abs of other subclasses (Supplemental Fig. 2B).

The IgG subclass purifications from immune serum were used to opsonize bacteria in vitro prior to i.n. infection of naive mice (Supplemental Fig. 1A). In a first attempt, we opsonized the bacteria with an excess of purified IgG of the individual subclasses. We found that at this high concentration of opsonizing agent all of the different IgG subclasses had the potential to reduce bacterial counts in the BAL within 2 d of infection, whereas significantly more bacteria were recovered from mice that had been infected with PBS-treated L. pneumophila (Fig. 2A). These high amounts of the different IgG subclass Abs used in this experiment do not correspond to the quantity of the IgG subclasses present in immune serum, where IgG2c and IgG3 are predominant (Supplemental Fig. 2). Thus, we decided to reduce the opsonizing agent for a subsequent experiment, now reflecting the ratio of the different IgG subclasses present in immune serum. Two days after the infection of naive mice we recovered strongly reduced bacterial counts from mice that were infected with immune serum–treated bacteria as well as from mice that received IgG2c-opsonized bacteria (Fig. 2B). Opsonization with the other subclasses also reduced bacterial counts in the BAL at day 2 postinfection, but to a lower extent. This implies that IgG2c plays a dominant role in restricting secondary L. pneumophila infection. The method of opsonizing L. pneumophila in vitro prior to infection has the limitation that only the inoculated population of bacteria is exposed to the L. pneumophila-specific Abs, whereas subsequent generations emerging after in vivo intracellular replication are no longer subject to opsonizing Ab. To expose extracellular L. pneumophila continuously to L. pneumophila–specific Abs in vivo, we decided to transfer the IgG subclass purified Abs i.v. prior to i.n. infection with L. pneumophila (Supplemental Fig. 1B). Ab transudes from the blood circulation into the alveolar space (21), where it first encounters the bacteria inhaled during i.n. infection and might still be present at sizeable amounts during later stages of infection. To this end we i.v. injected mice with equal quantities (∼0.2 μg) (Fig. 2C) of purified IgG subclasses or immune serum, followed by i.n. L. pneumophila infection 5 h later. Bacterial titers in BAL were determined 2 d later. We found that systemically applied polyclonal IgG1, IgG2c, and IgG3 (and less pronounced IgG2b) reduced bacterial titers in the BAL 2 d postinfection (Fig. 2D). Interestingly, we could not recover L. pneumophila–specific Ab from the BAL of infected animals, whereas transferred Ab was readily detectable in serum (Fig. 2E, 2F).

Because Ab could not be recovered and detected in BAL samples, we reasoned that transferring larger quantities of purified IgG subclasses i.v. prior to infection would increase their protective capacity. However, purifying large quantities of specific Abs from immune serum is not easily feasible. Therefore, we decided to engineer mAbs against L. pneumophila to achieve pure subclass samples of L. pneumophila–specific Abs. This approach enabled us not only to produce large quantities of clean Ab subclasses, but also to engineer Abs of different subclasses with identical Ag specificity, thereby also normalizing potential differences in L. pneumophila avidities that might be present in polyclonal IgG subclass preparations.

To this end we fused splenocytes of an L. pneumophila–immunized mouse with a mouse myeloma cell line to generate hybridoma cells (18). Viable clones were screened by ELISA for anti–L. pneumophila Ab activity. cDNA of a positive clone (4E12.14) was used to amplify the rearranged V regions of the Ig H and L chains, which were subsequently cloned into four Ab expression plasmids of the individual IgG subclasses (Supplemental Fig. 3A). These were separately transfected into HEK 293T cells to allow the expression and purification of mAbs of individual IgG subclasses but identical Ag specificity. Purity of the preparations was assessed by ELISA as well as by SDS-PAGE and Western blot (Supplemental Fig. 3B, 3C, and 3D, respectively). As a negative control we employed an mAb of unrelated specificity (control IgG2c). The specificity of the mAbs against L. pneumophila LPS of sg1 was assessed and confirmed by L. pneumophila sg–specific ELISA and by Western blot (Supplemental Fig. 3E, 3F). Detection of LPS in Western blot results in very strong characteristic signals ranging from high to low molecular mass (5). Our mAb is specific for L. pneumophila LPS of sg1 and does not cross-react with L. pneumophila of other sgs or with bacteria of other genera such as Escherichia or Salmonella.

We next tested whether the individual L. pneumophila–specific IgG subclasses were able to reduce bacterial counts 2 d after i.n. infection. To this end, we transferred a high quantity (50 μg per mouse) of the purified Ab preparations i.v. prior to i.n. infection with L. pneumophila. We found that all IgG subclasses were equally able to reduce the bacterial burden after i.n. infection with L. pneumophila (data not shown). The level of protection was comparable to i.v. transfer of 25 μl immune serum. Furthermore, the respective transferred IgG subclasses could be recovered from serum as well as from BAL at day 2 postinfection (data not shown). To more accurately assess potential differences in bacteria restricting efficacy, we decided to reduce the amount of transferred mAb 10-fold down to 5 μg per mouse (Fig. 3A). Decreased systemic concentrations of all L. pneumophila–specific mAb subclasses were still able to reduce bacterial counts in the BAL 2 d postinfection (Fig. 3B). Of note, subclasses IgG2c and IgG1 seemed to have a slightly but significantly increased capacity in reducing bacterial burdens in the BAL compared with IgG3 and IgG2b. Remarkably, as also found for serum transfer, the transferred mAbs could only be recovered from serum, but not from BAL, at this time point, indicating that levels of mAb in the BAL below the detection limit have an L. pneumophila restricting effect (Fig. 3C, 3D). Further decreasing the amount of transferred mAb to 1 μg per mouse failed to restrict bacterial burden (data not shown).

So far our data demonstrate a restricting effect when mAbs were transferred prior to infection. We therefore wondered whether mAbs might also have a therapeutic effect in L. pneumophila infection. To test this, we infected animals i.n. with L. pneumophila and treated them either 5 h before or 1 or 2 d after the infection with 5 μg monoclonal anti–L. pneumophila IgG2c i.v. (Supplemental Fig. 4A). The animals were sacrificed 2 d postinfection and bacterial titers in BAL were determined. As shown previously, application of the mAbs before the infection resulted in 10-fold reduced bacterial count in BAL (Fig. 4A). Interestingly, mice that received the IgG injection 1 d after the infection showed slightly reduced bacterial titers compared with untreated controls. In contrast, mice that received the mAb injection shortly before bacterial load analysis in the BAL were not able to reduce L. pneumophila counts, probably due to the short time given. Because the bacterial load in the BAL peaks on day 2 postinfection, we still wondered whether mice were able to clear the infection more efficiently when injected with L. pneumophila–specific mAbs on day 2, but given more time to resolve the infection. To this end, we infected mice i.n. with L. pneumophila and transferred purified anti–L. pneumophila IgG2c i.v. 2 d later and analyzed the BAL for bacterial titers on the following days (Supplemental Fig. 4B). Although already 1 d after injection a trend toward more efficient clearance of the bacteria is apparent, reduction of bacterial titers becomes significant after 2 d in the presence of pathogen-specific Ab. (Fig. 4B). This indicates that mAbs are able to act therapeutically against an established lung infection characterized by the presence of very high bacterial titers.

We could previously identify Abs as pivotal players reducing bacterial counts in secondary L. pneumophila infection (4). Opsonizing Abs are able to activate phagocytic host cells via their FcγRs, thereby targeting the bacteria for lysosomal degradation. During the development of a safe protein-based vaccine against L. pneumophila and its evaluation in the model of pulmonary infection in mice (5), we found that despite the presence of comparably high L. pneumophila–specific total IgG titers in serum, L. pneumophila immune mice were more efficient in restricting secondary L. pneumophila infection compared with protein-vaccinated mice. By comparing the different IgG subclasses in infected versus immunized animals, we found that the presence of IgG2c and IgG3 (as opposed to IgG1 and IgG2b) correlated with good restriction of a secondary L. pneumophila infection.

We employed two complementary approaches to analyze the individual contribution of the different IgG subclasses in limiting the magnitude of L. pneumophila titers upon infection. First, we purified polyclonal IgG from immune serum that harbors L. pneumophila–specific Abs of all subclasses. These preparations contained L. pneumophila–specific Abs of various specificities, thus being able to efficiently opsonize the bacteria by recognizing various B cell Ags on the bacterial surface. Second, we engineered mAbs of all IgG subclasses. These all shared the identical variable Ab regions and were specific for L. pneumophila LPS. The advantage of these is that despite their identical specificity for L. pneumophila LPS they are recombinantly expressed and affinity purified, thus being devoid of contaminating IgG subclasses and other serum-derived factors.

Using these different Ab preparations, we could show that all IgG subclasses have a certain ability to restrict pulmonary L. pneumophila infection in mice. Nevertheless, the high prevalence of IgG2c in immune serum and the slightly increased protective potential suggest that this IgG subclass has an important role in restricting bacterial growth in vivo. Interestingly, we could not recover any detectable Ab from BAL samples of animals that had received i.v. transfer of small but effective quantities of polyclonal Abs or mAbs. This could either be due to only very limited, but obviously sufficient, amounts of IgG transuding from the circulation into the respiratory tract, or owing to the adsorption of all free IgG by the infecting bacteria.

We showed previously that Ab-mediated protection from L. pneumophila infection depends on activating FcRs (4). Because there is a controversy on whether IgG3 is able to bind any known FcγR (22–25), the manner by which this subclass mediates bacterial control is unclear. Our results show that IgG3 can indeed limit L. pneumophila infection, and hence we think that this subclass may actually interact with a receptor to mediate its restricting effect. Because we know from previous studies that an activating FcR signal is necessary for Ab-mediated protection from pulmonary L. pneumophila infection (4), we assume that IgG3 can initiate receptor signaling through the common γ-chain.

In this study, we did not address the role of IgA in pulmonary L. pneumophila infection. IgA is the Ab isotype known to be transported through mucosal epithelia, thus playing an important role in agglutinating foreign intruders in mucosal tissues such as the intestinal or the upper respiratory tract. Nevertheless, in the lower respiratory tract and in the alveoli, where alveolar macrophages first interact with the invading bacteria, IgG transuding from the blood circulation is known to be physiologically more important (21). Thus, although we were previously able to show a protective capacity of IgA in the pulmonary L. pneumophila infection model (4), we think that IgG is also very important in vivo to combat lower respiratory L. pneumophila infections. Indeed, in our model of i.v. Ab transfer prior to as well as after infection, we confirm that IgG transuding from the blood circulation has the potential to reduce bacterial counts. Moreover, we show that s.c. or i.v. applied vaccination regimens raise Abs (IgGs, and not IgA; data not shown) that limit bacterial growth in the lung, indicating that it is possible to vaccinate against this respiratory infection without the use of a pulmonary route of vaccination that would give rise to IgA-producing B cells.

It has long been thought that affinity and specificity of Abs to Ag are solely determined by the variable Ab regions whereas the C regions are involved in effector functions. Nevertheless, recent studies implicated that the C region, in particular the CH1 domain, can mediate allosteric changes in the V region, thus influencing specificity and affinity of Abs (26–28). With our experiments we were not able to analyze increased or reduced affinities of our mAbs to L. pneumophila LPS, but because all Abs were able to reduce bacterial titers, we assume that this effect is negligible in the context of our experiments.

Because we use recombinant mAbs and transfer them i.v. into the circulation of subsequently infected animals, to our knowledge we show in this study for the first time that Abs themselves are able to enhance the recovery from pulmonary L. pneumophila infection. With this model, we are able to exclude the contribution of other serum components that might be transferred by opsonizing bacteria in vitro prior to infection.

In addition to enhanced control of L. pneumophila infection by prophylactic administration of Abs, a therapeutic application of pathogen-specific Abs i.v. also resulted in reduced bacterial burden in BAL of infected mice. Owing to emerging antibiotic resistances of a variety of bacterial pathogens, our study implicates that Ab therapy may be an interesting alternative for the treatment of patients in which antibiotic treatment is no longer effective or is impossible due to drug intolerance. The relevance of specific IgG subclasses in mediating enhanced control of airway L. pneumophila infection and increasing knowledge about Ab–FcR interactions may translate into refined active and passive immunization strategies not only against this pathogen, but possibly also against other human intracellular bacterial pathogens such as Mycobacterium tuberculosis or Salmonella enterica.

We thank Franziska Wagen, Nathalie Oetiker, Stefanie Joller, Iris Scherwitzl, and Kirsten Richter for experimental help and excellent technical assistance; Roger Beerli, Patrik Maurer, and Martin Bachmann (Cytos Biotechnology, Schlieren-Zurich, Switzerland) for the provision of cloning constructs; and members of the Oxenius group for fruitful discussions.

The authors have no financial conflicts of interest.