Potential Mechanisms of Mucin-Enhanced Acinetobacter baumannii Virulence in the Mouse Model of Intraperitoneal Infection

Porcine mucin has been commonly used to enhance the infectivity of bacterial pathogens, including Acinetobacter baumannii, in animal models, but the mechanisms for enhancement by mucin remain relatively unknown. In this study, using the mouse model of intraperitoneal (i.p.) mucin-enhanced A. baumannii infection, we characterized the kinetics of bacterial replication and dissemination and the host innate immune responses, as well as their potential contribution to mucin-enhanced bacterial virulence.

ABSTRACT

Porcine mucin has been commonly used to enhance the infectivity of bacterial pathogens, including Acinetobacter baumannii, in animal models, but the mechanisms for enhancement by mucin remain relatively unknown. In this study, using the mouse model of intraperitoneal (i.p.) mucin-enhanced A. baumannii infection, we characterized the kinetics of bacterial replication and dissemination and the host innate immune responses, as well as their potential contribution to mucin-enhanced bacterial virulence. We found that mucin, either admixed with or separately injected with the challenge bacterial inoculum, was able to enhance the tissue and blood burdens of A. baumannii strains of different virulence. Intraperitoneal injection of A. baumannii-mucin or mucin alone induced a significant but comparable reduction of peritoneal macrophages and lymphocytes, accompanied by a significant neutrophil recruitment and early interleukin-10 (IL-10) responses, suggesting that the resulting inflammatory cellular and cytokine responses were largely induced by the mucin. Depletion of peritoneal macrophages or neutralization of endogenous IL-10 activities showed no effect on the mucin-enhanced infectivity. However, pretreatment of mucin with iron chelator DIBI, but not deferoxamine, partially abolished its virulence enhancement ability, and replacement of mucin with iron significantly enhanced the bacterial burdens in the peritoneal cavity and lung. Taken together, our results favor the hypothesis that iron at least partially contributes to the mucin-enhanced infectivity of A. baumannii in this model.

INTRODUCTION

Acinetobacter baumannii is a major cause of nosocomial and community-acquired pneumonia, skin and urinary tract infections, and bacteremia (1–3). Moreover, A. baumannii infections are becoming increasingly difficult to treat due to the bacterium’s rapid development of resistance to multiple antibiotics, often leading to significant morbidity and mortality (4). The World Health Organization has recently listed A. baumannii as a top-priority bacterial pathogen that requires urgent research and development for new anti-infective products (https://www.who.int/news-room/detail/27-02-2017-who-publishes-list-of-bacteria-for-which-new-antibiotics-are-urgently-needed).

Development of novel therapeutics against A. baumannii requires the availability of well-characterized animal models for pathogenesis studies and for efficacy evaluation of candidate products. Most A. baumannii clinical isolates and type strains have low virulence in immunocompetent mice and generally require a large inoculum (50% lethal dose [LD₅₀] of >10⁸ viable bacterial cells) to establish a reproducible and measurable infection (1, 5, 6). To circumvent this challenge, many researchers have used virulence enhancers (such as porcine mucin) or immunosuppressed mice (such as neutropenic and diabetic mice) to enhance A. baumannii virulence and infection in mice (1, 7–10). In this regard, the mouse model of admixture of A. baumannii with porcine mucin (A. baumannii-mucin) has been used as a valuable, first-line tool in screening the in vivo efficacies of anti-Acinetobacter vaccines and therapeutics because of its relative simplicity, low cost, and short learning curve to establish the model (1, 11–14).

Enhancement of in vivo bacterial virulence and infection by porcine mucin is not a phenomenon unique to A. baumannii. Indeed, mucin enhances the infection or virulence of a wide range of Gram-positive and -negative bacteria, although the magnitude of the effect varies from bacterium to bacterium and even from isolate to isolate (15–19). Apart from intraperitoneal inoculation, admixture of mucin with bacterial inoculum has also been applied to enhance other routes (intranasal [i.n.], intratracheal [i.t.], and intravenous [i.v.]) of infection (12–14, 19–21). Although the first description of the mucin-enhanced bacterial infection can be dated back as early as the 1930s (15), the mechanisms and potential contributing factors of mucin remain largely unknown and appear to be varied among pathogens (16–18, 22, 23). In this study, we characterized the local (peritoneal cavity) and systemic (serum) innate cellular and cytokine responses to i.p. inoculation with A. baumannii-mucin in an attempt to better identify the potential mechanisms in this model.

RESULTS AND DISCUSSION

Mucin exacerbates i.p. infection with A. baumannii strains of different virulence.

Several laboratories have used porcine mucin to enhance the virulence of various A. baumannii clinical and type strains in mice (13, 21, 24–27). In this study, we assessed the relative virulence of 3 A. baumannii strains (ATCC 19606^T, ATCC 17961, and LAC-4) by the inclusion of mucin in the inocula. These three strains have been previously shown to have low, medium, or high virulence, respectively, in mice when infected through i.n, i.v., or i.p. routes (9, 10, 28, 29). As shown in Fig. 1, admixture of 5% porcine mucin with A. baumannii significantly increased their relative virulence in that all susceptible mice succumbed to the infection within 24 h. The estimated LD₁₀₀ for ATCC 19606^T or ATCC 17961 was at least 100-fold less when combined with mucin than for the bacteria alone (Fig. 1A and B). Moreover, admixture of mucin with LAC-4, a hypervirulent strain in mice, also increased its virulence at least 50-fold (Fig. 1C). A similar virulence enhancement was also observed in C57BL/6 mice when it was tested with ATCC 17961 (Fig. 1D), suggesting that the mucin-enhanced virulence in A. baumannii is not restricted to BALB/c mice. In this regard, other mouse strains, such as ICR-Swiss, Swiss White, and CF-1, previously have been used in similar models (16, 18, 19, 26), although actual magnitudes of the enhancement in some of those studies had not been reported in detail.

FIG 1.

Effect of porcine mucin on the survival rates of mice following i.p. inoculation with A. baumannii. Groups of 5 BALB/c mice were i.p. inoculated with various numbers of ATCC 19606^T (A), ATCC 17961 (B), or LAC-4 (C) in the presence or absence of mucin, as indicated, and their clinical outcome was monitored daily for 5 days. (D) Groups of 5 BALB/c and C57BL/6 mice were i.p. inoculated with 10⁵ CFU of ATCC 17961 in the presence of mucin.

Our results, together with the previous reports of others (13, 21, 25–27, 30, 31), indicate that mucin is able to enhance the virulence of a number of A. baumannii type and clinical strains, regardless of their natural in vivo virulence. Importantly, the magnitude of the virulence enhancement appears to be somewhat related to the known natural virulence of a given strain. These data therefore suggest the potential for broad application of this model in A. baumannii research and anti-infective development, since the model is unlikely to be restricted to the use of certain strains or isolates of clinical relevance.

Kinetics of local bacterial replication and systemic dissemination.

To understand the potential mechanisms of the mucin-enhanced A. baumannii infection, we examined the kinetics of bacterial replication at the site of the inoculation and dissemination to blood and lungs in the mice following i.p. inoculation with 7 × 10⁴ CFU of ATCC 17961-mucin. As shown in Fig. 2, rapid local replication of the bacteria was noted by 2 h postinfection (hpi) and the bacterial burden was highest at the initial site of infection, although substantial numbers of bacteria were found in the blood and, to a lesser extent, the lung even at this early time point. The bacteria replicated continuously in the next 6 h, reaching approximately 10⁸ CFU in all compartments. As anticipated, mice inoculated with 7 × 10⁴ CFU A. baumannii-saline showed only ∼100 CFU/ml in peritoneal lavage fluid at 2 h (see Fig. S1A in the supplemental material) and similar loads in the lungs at 2 and 4 h postinoculation. The bacterial burdens in the blood and at other time points in the peritoneal lavage fluid and lungs were below the detection limit. These results suggested that mucin decreases the host’s capacity to control the local replication and systemic dissemination of A. baumannii, turning an otherwise self-limited infection into a lethal one.

FIG 2.

Bacterial burdens in the peritoneal cavity, blood, and lungs of BALB/c mice following i.p. inoculation with A. baumannii-mucin. Groups of BALB/c mice (n = 5) were i.p. inoculated with 7 × 10⁴ CFU of A. baumannii ATCC 17961 admixed with 5% porcine mucin. Bacterial burdens in the peritoneal lavage fluid, blood, and lungs at various times postinoculation were determined by quantitative bacteriology. The data are presented as means ± SD and represent one of at least two experiments with similar results. The detection limits for bacterial burdens at each respective tissue are indicated as black dotted lines. ***, P < 0.001; ****, P < 0.0001.

Mucin fails to enhance the in vitro growth of A. baumannii.

To examine the possibility of a direct nutrient effect of mucin on bacterial replication, we examined the effect of mucin on the in vitro growth of A. baumannii. As shown in Fig. 3A, the comparative initial growth rates of A. baumannii between mucin-supplemented and standard tryptic soy broth (TSB) were similar for the first 2 h, but growth then became limited in TSB-mucin, leveling off to a lower Y_max (maximum population density in a culture) by 5 h of culture. The exact reasons for this are unknown, but it is possible that the high viscosity of TSB–5% mucin leads to less local availability of nutrients or oxygen than TSB alone, even under a shaking culture condition. Nevertheless, our observations excluded the possibility that mucin enhances the growth of A. baumannii directly by nutrient supply over what is available in TSB. Moreover, we also found that A. baumannii cultured in mucin-supplemented TSB (10 mg/liter) was no more virulent in mice than that cultured in TSB alone, in that no mice showed any clinical signs after being i.p. inoculated with 1 × 10⁵ CFU ATCC 17961, and all mice survived this dose of infection, as anticipated (Fig. 3B). Those results indicate that mucin per se does not enhance the replication rate or virulence of A. baumannii in vitro when the standard TSB bacterial culture medium is used. This raised the possibility that mucin enhances the virulence of A. baumannii primarily through interaction with the host innate immune systems. Indeed, we found that although mucin has to be given in vivo to enhance the virulence of A. baumannii, it is not essential to physically admix mucin with the bacteria or to administer the mucin and bacteria at the same site of inoculation, as simultaneous injection of mucin and A. baumannii at separate sites of the peritoneal cavity induced the same clinical outcome as preadmixture of mucin and A. baumannii (Fig. 3C). In this regard, DeWitt showed that separate routes of injection of mucin (i.v.) and bacteria (i.p.) also enhanced E. coli infection (19). Those observations suggest that mucin is not simply physically shielding and protecting bacteria from host defense mechanisms, i.e., by entrapping the bacteria in the mucin matrix and hiding bacteria from phagocytes.

FIG 3.

(A) Effect of mucin on in vitro growth of A. baumannii. Cells (10⁷ CFU) of freshly grown A. baumannii ATCC 17961 were inoculated into mucin-supplemented TSB (10 mg/liter, TSB/mucin) or standard TSB (TSB), and their growth rates were monitored hourly for 5 h. Data are presented as means ± SD. (B) Effect of A. baumannii cultured in the presence or absence of mucin on infectivity in mice. Groups of 5 BALB/c mice were i.p. inoculated with ∼1 × 10⁵ CFU of ATCC 17961 grown in the mucin-supplemented TSB (TSB/mucin) or standard TSB (TSB), and their clinical outcome was monitored daily for 5 days. (C) Effect of mucin administration on the infectivity of A. baumannii in mice. Groups of 5 BALB/c mice were either i.p. inoculated with ∼1 × 10⁵ CFU of ATCC 17961 admixed with 5% porcine mucin (admixed) or inoculated with the bacteria and mucin at different injection sites of the peritoneal cavity (separate), and their clinical outcome was monitored daily for 5 days.

Host inflammatory cell responses to i.p. inoculation with A. baumannii-mucin.

To dissect and better understand the host and bacterial interactions, we next characterized the local (peritoneal cavity) inflammatory and innate immune cell responses to the i.p. injection of mucin alone and to A. baumannii-mucin infection. Mucin alone and A. baumannii-mucin infection both resulted in moderate increases in the total number of peritoneal cells with the predominant recruitment of neutrophils (Fig. 4). Neutrophil abundance increased from 0% to 40% at 2 hpi and then reached nearly 80% by 8 hpi, whereas macrophage abundance was relatively decreased from an initial ∼30% of total cells to ∼5% by 2 h and then remained at this low level throughout the experiment (8 hpi) (P < 0.05). The changes in peritoneal lymphocyte abundance were also moderate compared to those for neutrophils, decreasing from 70% at 0 h to 50% at 2 h and then to just 20% by 8 hpi. In addition, there was a small but significant increase in the percentage of eosinophils in the mice infected with A. baumannii-mucin by 4 and 8 hpi, but this was not observed in mice treated with mucin only. A similar effect was also observed in the total numbers of eosinophils (Fig. S2). The total numbers and percentages of mast cells were overall low and showed little change following i.p. mucin or A. baumannii-mucin administration (data not shown). Thus, increases in total peritoneal cells for both mucin-treated and A. baumannii-mucin-infected mice were predominantly due to the increased neutrophil recruitment. On the other hand, the reduction in the percentage of peritoneal macrophages and the increase in the peritoneal neutrophil recruitment in mice inoculated with A. baumannii-saline were overall much milder than those for the mice with mucin or A. baumannii-mucin administration (Fig. S1B).

FIG 4.

Peritoneal inflammatory cell responses in mice i.p. inoculated with mucin alone or A. baumannii-mucin. Groups of BALB/c mice (n = 5) were inoculated i.p. with 2.5 × 10⁵ CFU of ATCC 17961 admixed with 5% mucin or mucin only at 0 h. The mice were sacrificed at 0 (before inoculation), 2, 4, and 8 h postinoculation. The peritoneal cavity was lavaged, and the total and differential cell counts were determined. Data are presented as means ± SD and are representative of at least two independent experiments with similar results. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; each versus results for the 0-h group.

It is somewhat surprising but interesting that the changes in the numbers of peritoneal cells induced by i.p. inoculation of mice with A. baumannii-mucin were comparable to those induced by mucin treatment on its own (Fig. 4). This suggests that most of the observed peritoneal inflammatory cell recruitment in A. baumannii-mucin-inoculated mice resulted from the mucin administration itself. In this regard, ATCC 17961 is of relatively low virulence in mice (Fig. 1B) (32). At this i.p. challenge dose without mucin, bacteria are rapidly eliminated and there is only a minimal and transient local inflammatory response observed (32). In addition, the relevance of observed peritoneal eosinophil recruitment is currently unknown, but the observation was somewhat consistent with the increased serum and peritoneal interleukin-4 (IL-4) levels (Fig. 5A and B). We postulate that the changes in eosinophils and IL-4 are related to the nature of intraperitoneal infection rather than to mucin administration, because similar observations were also seen when mice were intraperitoneally infected with lethal or sublethal doses of LAC-4 (32).

FIG 5.

Cytokine and chemokine levels in the peritoneal lavage fluid (A) and sera (B) of mice following i.p. inoculation with A. baumannii-mucin or mucin only. Groups of BALB/c mice (n = 5) were i.p. inoculated with 2.5 × 10⁵ CFU of ATCC 17961 admixed with 5% mucin (ATCC/mucin) or mucin only and sacrificed at the indicated hours postinoculation. Cytokine and chemokine levels were determined in the sera and peritoneal lavage fluid using 14-plex Milliplex MAP mouse cytokine/chemokine kits on a Luminex MAGPIX system. The detection limit is indicated by the dotted line or was <10 pg/ml for other cytokines/chemokines. All values are means ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; each versus results at 0 h.

The results presented in Fig. 4 and Fig. S2 clearly showed that i.p. injection of mucin alone induces a significant reduction in the relative abundance and absolute number of peritoneal macrophages and increase in the neutrophil recruitment compared to that of nontreated, control mice. Previous studies by us and others have shown that macrophages play important roles in the early host defense to A. baumannii infection (29, 33, 34). Therefore, our results raised the possibility that mucin enhances A. baumannii virulence by depleting peritoneal macrophages. To examine this possibility, groups of BALB/c mice were treated i.p. with clodronate-liposomes to deplete peritoneal macrophages or with phosphate-buffered saline (PBS)-liposomes as a control 3 days before i.p. challenge with 10⁵ CFU of ATCC 17961 admixed with 5% mucin. We found no significant differences in the bacterial burdens in the peritoneal cavity, blood, or lung between the mice treated with clodronate-liposomes and the mice treated with PBS-liposomes (Fig. S3), suggesting peritoneal macrophages play little or no role in the host defense against this dose of mucin-enhanced ATCC 17961 challenge.

The results described above were puzzling, since mucin strongly recruited neutrophils but nonetheless enhanced the infection. Neutrophils are crucial to host defense against A. baumannii infection, including i.p. infection, in immunocompetent mice (28, 35). Depletion of neutrophils in mice enhances host susceptibility to A. baumannii infection significantly and converts an otherwise self-limited infection to a lethal infection (28, 35). Moreover, we have previously shown that enhancement of neutrophil recruitment by MIP-2 or c-di-GMP significantly increased the host resistance to sublethal infection with both clinical and ATCC isolates (36, 37). In the mucin model, however, despite the recruitment of large numbers of neutrophils to the peritoneal cavity, A. baumannii replicated and disseminated rapidly with a lethal infection ensuing (Fig. 1 and 2), suggesting that those neutrophils play little role in host defense against A. baumannii infection in this model. Indeed, we found that depletion of neutrophils had no overt effect on the magnitude or course of the infection in mice inoculated with A. baumannii-mucin (Fig. S4A). Taken together, these data further supported the notion that it is the competency and quality and not just the quantity of the neutrophils that contributes to the host innate defense against A. baumannii. Indeed, we previously demonstrated that reactive oxygen species play a key role in neutrophil-mediated host defense against A. baumannii infection with the observation that gp91^−/− mice, with defective reactive oxygen species production, showed a normal neutrophil recruitment and phagocytosis response to A. baumannii infection but failed to kill bacteria effectively (38). In this regard, we also attempted to determine the effect of depletion of neutrophils or macrophages on the enhancement of i.p. A. baumannii-saline infection. We found that the bacterial burdens in the neutrophil-depleted mice were about 100 (in lungs)- to 1,000 (in peritoneal lavage fluid and blood)-fold lower than those in mucin-enhanced infection (Fig. 2 versus Fig. S4B). As anticipated, depletion of macrophages showed no effect on the bacterial burdens in the mice inoculated with this dose of A. baumannii-saline (data not shown), because it is generally recognized that the role of macrophages in mouse models of A. baumannii infection is more subtle than that of neutrophils (32, 33). This suggests that neutrophil function at most only marginally contributes to the mucin-enhanced A. baumannii virulence. The relative competence of the mucin-recruited neutrophils was not studied, but a future investigation of this may provide additional information as to the effects of mucin on neutrophil activity.

Local and systemic proinflammatory cytokine and chemokine responses to i.p. challenge with A. baumannii-mucin.

To further characterize the host innate immune responses to i.p. mucin administration and A. baumannii-mucin infection, we profiled and compared the local (peritoneal) and systemic (serum) cytokine/chemokine responses. As reported in other mouse models of A. baumannii infection (9, 28, 35, 39–42), i.p. administration of mucin alone or A. baumannii-mucin induced significant increases of keratinocyte-derived chemokine (KC), IL-6, IL-10, and, to lesser magnitudes, macrophage inflammatory protein 1α (MIP-1α) and tumor necrosis factor alpha (TNF-α) levels in the peritoneal cavity. Interestingly, we found the cytokine/chemokine responses were more transient in mice treated with mucin alone than for mice infected with A. baumannii-mucin. Cytokine/chemokine levels in mucin-treated mice returned relatively rapidly to baseline levels by 8 hpi. However, the levels of these same cytokines/chemokines remained elevated in A. baumannii-mucin-infected mice (Fig. 5A). These results suggested that the early cytokine responses at 2 and 4 hpi were driven by the mucin and largely transient, while these responses were sustained to at least 8 hpi by the bacterial infection. In addition, i.p. A. baumannii-mucin infection induced significant increases of granulocyte-macrophage colony-stimulating factor (GM-CSF), monokine induced by gamma interferon (MIG), monocyte chemoattractant protein 1 (MCP-1), and IL-1β in the peritoneal cavity. On the other hand, the levels of gamma interferon (IFN-γ), IL-2, IL-4, IL-12p40, and IL-13 were near or below the assay detection limit in all groups of mice (Fig. 5A), suggesting that the infection induced primary proinflammatory and chemotactic cytokines with very minimal effects on the cytokines that are involved in T cell differentiation and function.

Intraperitoneal injection with mucin alone or A. baumannii-mucin infection also induced significant increases of serum KC, IL-6, MIP-1α, MCP-1, MIG, and TNF-α, although the levels for most of these cytokines/chemokines were significantly higher in A. baumannii-mucin-infected mice than in mice treated with mucin alone (Fig. 5B). This suggested that disseminated A. baumannii infection is the major driving force in inducing circulatory cytokines. Similar to the cytokine responses in the peritoneal cavity, the mucin-alone-induced systemic cytokine responses were more transient and largely limited at 2 and 4 hpi, whereas the cytokines induced by A. baumannii-mucin generally remained significantly elevated at 8 hpi (Fig. 5B). In addition, GM-CSF, IFN-γ, and IL-12p40 were significantly increased only until 8 hpi. Those results further support the possibility that the systemic cytokine responses are largely pathogen driven, whereas the local peritoneal responses were initiated primarily by the mucin.

As anticipated, the proinflammatory cytokine and chemokine responses to A. baumannii-saline administration in the peritoneal lavage fluid and serum were mild and transient. In the peritoneal lavage fluid (Fig. S1C), only MIP-1α was significantly increased at 2, 4, and 8 hpi. The levels of IL-1β, IL-6, KC, MCP-1, and TNF-α were significantly increased at 2 hpi, whereas IL-10 and IL-12p40 were significantly elevated at both 2 and 4 hpi. On the other hand, serum proinflammatory cytokine and chemokine responses were highly variable between individual mice within groups, and only IL-6, KC, and MCP-1 were significantly increased at 2 hpi (Fig. S1D), corroborating well with the tissue and blood bacterial burden data (Fig. S1A).

IL-10 showed a substantial increase by 2 hpi in peritoneal lavage fluid of mice treated with mucin alone (Fig. 5A). IL-10 is a pleotropic cytokine which is known to play important roles in the anti-inflammatory response and immunosuppression (43). It is therefore conceivable that mucin enhances the in vivo bacterial virulence through upregulating IL-10 production, which in turn suppresses or inhibits the antibacterial function of innate immune cells (such as macrophages and neutrophils). To assess this possibility, IL-10^−/− mice or anti-IL-10 monoclonal antibody (MAb)-treated BALB/c mice were inoculated i.p. with ATCC 17961-mucin. In addition, we challenged the recombinant murine IL-10-treated BALB/c mice with A. baumannii ATCC 17961 without mucin. However, we found no difference in the bacterial burdens between those treatment groups, indicating that IL-10 plays no clear role in the mucin-enhanced A. baumannii infection (Fig. S5).

Potential role of iron in the mucin enhancement of A. baumannii virulence in mice.

The studies described above did not support a potential direct and significant contribution of several innate immune defense components, including peritoneal macrophages, neutrophils, or IL-10, to the mucin-enhanced A. baumannii infection of mice, leaving the question of just how mucin enhances the in vivo virulence of A. baumannii. It has previously been suggested that mucin promotes bacterial virulence by inhibiting the recruitment and function (phagocytosis and killing) of peritoneal phagocytes through its anticomplement and anticoagulation functions or by increasing the capillary permeability (16–19, 44, 45). Others have postulated that mucin enhances the bacterial virulence simply by its coating effect on bacteria (26) or through its growth-promoting abilities or replenishing the nutrition for bacteria (22). Indeed, recent transcriptomic studies showed that the mucin acts as a nutrient source and a signal to trigger an upregulation of cellular processes and the virulence pathway in vitro (22). None of our results provides support for any of these previously postulated mechanisms.

Iron supply to bacteria is well recognized to be critical to the virulence of bacterial pathogens, including A. baumannii (46–51), and porcine mucin contains a substantial amount of iron and other metals (48). Indeed, earlier studies have shown that the mucin-enhanced virulence of Neisseria meningitidis is primarily mediated by the iron in the mucin (16). Analysis of porcine mucin used in this study by iCAP Q inductively coupled plasma mass spectrometry (ICP-MS) shows the presence of substantial amounts (277.28 μg/g [dry weight]) of iron and other metal elements (Table S1). We therefore investigated iron as a potential contributor to the mucin-enhanced A. baumannii virulence. We found that pretreatment of mucin with a novel iron chelator, DIBI (Fig. 6B), but not with deferoxamine (Fig. 6A) significantly diminished its capability to enhance A. baumannii virulence and reduced bacterial burdens in the peritoneal lavage fluid, blood, and lungs (P < 0.005), supporting the notion that iron is an important contributor to the mucin-enhanced A. baumannii virulence. In this regard, it has been previously demonstrated that the accessibility of deferoxamine-chelated iron by A. baumannii and other bacterial pathogens (16) varies in that deferoxamine is totally noninhibitory to A. baumannii (MIC of >1,280 μg/ml), as the bacterium can utilize deferoxamine-Fe as an iron source, but it is highly sensitive to DIBI inhibition (MIC of <5 μg/ml) because the bacterium is not able to access Fe-DIBI for its iron requirement (52). DIBI has been chemically characterized in other studies (52, 53) and shown to be highly specific for iron and not to affect other trace essential metals. In addition, we showed that coadministration of A. baumannii with DIBI failed to exacerbate the infection (Fig. 6C). This excluded the possibility that DIBI functioned solely by starving A. baumannii of needed iron, i.e., either its endogenous iron as acquired from the culture medium prior to inoculation or from host iron sources during infection.

FIG 6.

Effect of iron on the mucin-enhanced infectivity of A. baumannii in mice. (A and B) Groups of BALB/c mice (n = 5) were i.p. inoculated with 7 × 10⁴ CFU of ATCC 17961 admixed with 5% untreated mucin (ATCC/mucin) or mucin pretreated with the iron chelator deferoxamine (ATCC/mucin/deferoxamine) or DIBI (ATCC/mucin/DIBI). (C) Groups of BALB/c mice (n = 5) were i.p. inoculated with 7 × 10⁴ CFU of ATCC 17961 admixed with 0 (ATCC/saline) or 160 mg/ml DIBI (ATCC/saline/DIBI). (D) Groups of BALB/c mice (n = 5) were i.p. administered 90 mg/kg FeSO₄ 18 and 0 h before being i.p. inoculated with 7 × 10⁴ CFU of ATCC 17961. Additional groups of mice were i.p. inoculated with ATCC 17961 only (ATCC/saline) and served as controls. (E) Groups of BALB/c mice (n = 5) were i.p. inoculated with 7 × 10⁴ CFU of ATCC 17961 admixed with 5% mucin with (ATCC/mucin + FeSO₄) or without (ATCC/mucin) the coadministration of 90 mg/kg FeSO₄. Additional groups of mice were i.p. inoculated with ATCC 17961 only (ATCC/saline) and served as controls. Four hours after the infection, mice were euthanized and the bacterial burdens in the peritoneal lavage fluid, blood, and lung were determined by quantitative bacteriology. Data are presented as means ± SD (n = 5). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Moreover, we found that the function of mucin in this model can be partially replaced with iron, in that i.p. injection of FeSO₄ before and at the time of A. baumannii inoculation significantly increased bacterial burden in the lung and, to a lesser magnitude, in the peritoneal cavity, the site of initial inoculation, but surprisingly not in the blood (Fig. 6D). This was a curious result, as mucin did not enhance the in vitro growth of the bacterium in TSB. It seems likely that TSB is sufficiently rich in iron such that the iron from mucin would not be essential as an additional in vitro source. However, iron availability is known to be very low in the host, representing an additional innate immune response (54); thus, the additional iron provided from mucin in vivo could serve to support infection. Indeed, we found that administration of FeSO₄ at the time of A. baumannii-mucin challenge had no effect on the infection (Fig. 6E), suggesting that the iron in the mucin is sufficient to enhance A. baumannii virulence. Taken together, our results strongly suggest iron as an important contributor to the mucin-enhanced A. baumannii virulence. This finding emphasizes the notion that mucin exerts its ability to enhance bacterial virulence through diverse mechanisms. In this regard, it has recently been shown that mucin enhances the virulence of enteropathogenic Bacillus cereus by upregulating the production of enterotoxin and other putative virulence factors (23).

MATERIALS AND METHODS

Mice.

Eight- to 12-week-old female, specific-pathogen-free BALB/c and C57BL/6 mice were purchased from Charles River Laboratories (St. Constant, Quebec, Canada), and 7- to 10-week-old female, specific-pathogen-free IL-10 knockout mice (B6.129P2-Il10^tm1Cgn/J) and wild-type control mice (C57BL/6J) were purchased from the Jackson Laboratory (Bar Harbor, ME). The animals were maintained and used in accordance with the recommendations of the Canadian Council on Animal Care Guide to the Care and Use of Experimental Animals (55), and experimental procedures were approved by the institutional animal care committee.

A. baumannii isolates and bacterial culture.

A. baumannii ATCC 19606^T, ATCC 17961, and LAC-4 strains were used in this study. LAC-4 is a hypervirulent clinical isolate previously described by us (9, 56). In some in vitro culture experiments, 1 × 10⁷ CFU of freshly grown A. baumannii were inoculated into tryptic soy broth (TSB) with or without the addition of porcine mucin (10 mg/liter; type II; MilliporeSigma, Oakville, Ontario, Canada), and the number of bacteria was determined over a period of 5 h by plating 10-fold serial dilutions on brain heart infusion (BHI) agar plates.

i.p. A. baumannii inoculation and sample collection.

For i.p. inoculation of mice, fresh inocula were prepared for each experiment from frozen stocks of A. baumannii strains. Stocks were thawed to streak on a chocolate-enriched agar plate and cultured at 37°C for 18 h, followed by culture in TSB for 2 to 3 h at 37°C with shaking (200 rpm). In some experiments, 10 mg/liter mucin was added to the TSB prior to culture. The cultures were harvested, centrifuged at 5,000 × g for 10 min, and resuspended in saline, and cells were counted using a bacterial counting chamber. Bacteria were then diluted in either saline or 5% mucin (11) to the appropriate inoculation concentration, and nonanesthetized mice were inoculated i.p. with 0.2 ml of the appropriate numbers of various A. baumannii strains. In some experiments, bacterial cells and mucin were administered separately at different sites of the peritoneal cavity. In other experiments, mucin was mixed for 18 h at room temperature with the iron chelator DIBI (160 mg/ml) (Chelation Partners, Inc., Halifax, Nova Scotia, Canada) or deferoxamine (40 mg/ml) (MilliporeSigma). Chelator-treated mucin samples without further processing were compared to untreated mucin for their infection enhancement activities. In still other experiments, mice were treated with 0.5 ml of 24 mM FeSO₄ (90 mg/kg, i.p.) 24 h before and at the time of A. baumannii challenge or at the time of A. baumannii-mucin inoculation.

Actual inocula in each experiment were determined by plating 10-fold serial dilutions on BHI agar plates. The clinical appearance of the mice was monitored and scored as described previously (9). Groups of five infected mice were sacrificed at various times postinoculation as indicated. Additional groups of untreated mice were also sacrificed and served as controls (referred to as 0 h), when applicable. The peritoneal cavity was lavaged as described below, and blood and lungs were aseptically removed and used for quantitative bacteriology. In some cases, serum and peritoneal lavage supernatant were collected for cytokine/chemokine analysis.

Peritoneal lavage.

The peritoneal cavity was lavaged with 10 ml phosphate-buffered saline (PBS) supplemented with 3 mM ethylenediaminetetraacetic acid and 1% fetal bovine serum as previously described (10). The total number of peritoneal cells in peritoneal lavage samples was determined with a hemacytometer, and differential cell counts were determined by examining 200 cells on cytospin slides (Cytospin 3; Shandon, Pittsburgh, PA) stained with Hema-3 (Fisher Scientific, Ottawa, Ontario, Canada). The lavage fluid was then centrifuged at 14,000 × g for 7 min, and aliquots of supernatant were collected and stored at –20°C.

Quantitative bacteriology.

The lungs were homogenized in sterile saline using aerosol-proof homogenizers. Blood was diluted 1:10 in sterile water immediately upon sample collection. Aliquots (100 μl) of 10-fold serial dilutions of the tissue homogenates, peritoneal lavage fluid, and blood were cultured on BHI agar plates to quantify the number of viable A. baumannii organisms in the respective samples (9). Bacteriology data are presented as total CFU per organ or compartment, assuming a blood volume of 1.5 ml/mouse and taking into account the volume of peritoneal lavage fluid used (10 ml).

Luminex cytokine and chemokine analysis.

The levels of cytokines and chemokines in the sera and peritoneal lavage fluid were measured using an 11- or 14-plex Milliplex MAP mouse cytokine/chemokine kit (MilliporeSigma) on a Luminex MAGPIX system (Luminex, Austin, TX). Samples were assayed in duplicate, and cytokine/chemokine concentrations were calculated against the standards using Milliplex Analyst software (ver 3.5; MilliporeSigma) (9).

Quantitation of iron in mucin.

Mucin samples were dried for 24 h at 90°C, and samples of known weight were microwave digested in 2% (wt/wt) nitric acid using a Discover SP-D microwave digester (CEM Corporation, Matthews, NC). Digestates were analyzed for metals (V, Cr, Mn, Fe, Co, Ni, Cu, and Zn) using an iCAP Q ICP-MS (Thermo Fisher Scientific, Waltham, MA) paired with an electrospray ionization SC-4DXS autosampler (Elemental Scientific, Omaha, NE). A six-point external calibration curve was used to determine unknown concentrations of analytes in the samples. Metal contents were reported as micrograms per gram of oven-dried dry mucin (see Table S1 in the supplemental material).

In vivo depletion of peritoneal macrophages.

Peritoneal macrophages were depleted by i.p. administration of liposome-encapsulated dichloromethylene diphosphonate (clodronate-liposome) (Clodronate Liposomes, Haarlem, Netherlands). Groups of five BALB/c mice were administered i.p. with 0.2 ml of clodronate-liposomes or PBS-liposomes, as recommended by the supplier and published previously (33). The depletion was confirmed to be >95% by differential counts of peritoneal lavage cells as detailed above. Seventy-two hours later, the mice were inoculated i.p. with ∼10⁵ CFU A. baumannii ATCC 17961, admixed with or without 5% mucin as described above. Infected mice were sacrificed at 4 hpi for sample collection as previously described (35).

MAb treatment.

For in vivo depletion of neutrophils or neutralization of endogenous IL-10 activities, groups of five BALB/c mice were treated i.p. with rat anti-mouse monoclonal antibody (MAb) RB6-8C5 (100 μg in 0.2 ml sterile saline; Bio X Cell, West Lebanon, NH), which recognizes the neutrophil surface marker Gr-1, or rat anti-mouse MAb JES5-2A5 (450 μg in 0.2 ml sterile saline; Bio X Cell), which neutralizes the mouse IL-10 activities in vivo or an equivalent amount of purified rat IgG (MilliporeSigma) as described previously (35). The treatments were administered 18 h before (RB6-8C5) and at the time (RB6-8C5 and JES5-2A5) of A. baumannii challenge. The depletion of neutrophils was confirmed to be >95% by differential counts of peritoneal lavage cells as detailed above.

Recombinant IL-10 treatment.

Groups of five BALB/c mice were treated i.p. with recombinant IL-10 (1.0 μg in 0.2 ml sterile saline; Peprotech, Rocky Hill NJ), or an equivalent amount of bovine serum albumin (MilliporeSigma), 45 min prior to A. baumannii challenge.

Statistical analysis.

Data are presented as means ± standard deviations (SD) for each group, unless otherwise specified. Differences in quantitative measurements were assessed by Student's t test or one-way or two-way analysis of variance, followed by Dunnett’s post hoc multiple-comparison tests, when appropriate. Differences were considered significant at a P value of <0.05.