The Host Antimicrobial Protein Calgranulin C Participates in the Control of Campylobacter jejuni Growth via Zinc Sequestration

ABSTRACT

Campylobacter jejuni is a leading cause of bacterially derived gastroenteritis worldwide. Campylobacter is most commonly acquired through the consumption of undercooked poultry meat or through drinking contaminated water. Following ingestion, Campylobacter adheres to the intestinal epithelium and mucus layer, causing toxin-mediated inflammation and inhibition of fluid reabsorption. Currently, the human response to infection is relatively unknown, and animal hosts that model these responses are rare. As such, we examined patient fecal samples for the accumulation of the neutrophil protein calgranulin C during infection with Campylobacter jejuni. In response to infection, calgranulin C was significantly increased in the feces of humans. To determine whether calgranulin C accumulation occurs in an animal model, we examined disease in ferrets. Ferrets were effectively infected by C. jejuni, with peak fecal loads observed at day 3 postinfection and full resolution by day 12. Serum levels of interleukin-10 (IL-10) and tumor necrosis factor alpha (TNF-α) significantly increased in response to infection, which resulted in leukocyte trafficking to the colon. As a result, calgranulin C increased in the feces of ferrets at the time when C. jejuni loads decreased. Further, the addition of purified calgranulin C to C. jejuni cultures was found to inhibit growth in a zinc-dependent manner. These results suggest that upon infection with C. jejuni, leukocytes trafficked to the intestine release calgranulin C as a mechanism for inhibiting C. jejuni growth.

INTRODUCTION

Campylobacter jejuni significantly impacts human health as it is the leading cause of bacterially derived gastroenteritis in the world (1). This pathogen is fairly ubiquitous as it resides asymptomatically in the gastrointestinal tracts of cattle, pigs, poultry, and other livestock (2). The most common route of infection in the developed world is through consumption of undercooked poultry, whereas in the developing world, infection most commonly occurs via drinking contaminated water (1). Following ingestion, Campylobacter adheres to the intestinal epithelium and mucus layer, causing toxin-mediated inflammation and inhibition of fluid reabsorption (3). This leads to the development of mild to severe diarrhea that may be accompanied by abdominal cramps, fever, and frank blood in the stool. While Campylobacter infection is often self-limiting, campylobacteriosis can lead to serious postinfectious diseases such as Guillain-Barré syndrome, reactive arthritis, and inflammatory bowel disease (4). In low-resource settings, C. jejuni infection can also lead to persistence and stunting when acquired at a young age (5).

In addition to problems associated with the frequency and severity of these infections, Campylobacter is becoming increasingly resistant to clinically important antibiotics such as ciprofloxacin and azithromycin, leading the CDC and the WHO to classify it as a serious threat to public health (2). Due to the increasing antibiotic resistance of C. jejuni, it is necessary that the field identify and develop new technologies to reduce Campylobacter infection in humans. Most of these efforts have focused on reducing colonization of poultry in an effort to prevent contamination of food and transmission to humans, while relatively less has been done to identify factors that contribute to colonization of humans and leveraging those observations to develop direct treatments (6, 7).

S100 proteins are calcium-binding regulators of several important cell processes in vertebrates, including energy metabolism, growth, motility, cell cycle regulation, and apoptosis (8). Many of these proteins are commonly used as markers of gastrointestinal inflammation in humans. Calprotectin (also called S100A8/S100A9 or MRP 8/14) and calgranulin C (also called S100A12 or EN-RAGE) are quantifiable in feces and are elevated during intestinal inflammation. These proteins indicate the presence of neutrophils as they make up approximately 40% of all neutrophil cytosolic proteins (9). Extracellularly, calprotectin has been shown to be chemotactic for neutrophils (10), while calgranulin C is chemotactic for monocytes and mast cells (11), thus promoting a proinflammatory response. Both calprotectin and calgranulin C have antimicrobial properties due to their abilities to sequester copper, manganese, iron, and zinc (8, 12). These proteins have been shown to inhibit the growth of fungi, certain bacterial species such as Staphylococcus aureus, and some parasites (13–15). The molecular details of the mechanism behind this antimicrobial action remains unknown, but it is speculated to be due to the inhibition of essential metal-dependent functions such as the reduction of superoxide (16).

Weaning-age ferrets have been used in previous studies to determine the effects of Campylobacter infection in an animal model because they are one of the few animals that display clinical signs similar to human campylobacteriosis although ferrets aged 6 to 7 weeks tend to display more severe disease than ferrets aged 11 weeks and older (17, 18). These signs include mild to moderate diarrhea and mucus and/or blood in the stool. Most of the previous studies simply looked at whether various C. jejuni strains could colonize and cause diarrhea and did not examine the effects of infection on the animal. A more in-depth analysis of the host factors that are impacted by C. jejuni infection could contribute to understanding disease in humans through, for example, noninvasive testing of the levels of inflammatory markers in fecal samples.

Due to the high prevalence, severity, and increasing antibiotic resistance of C. jejuni, it is necessary that we increase our knowledge of how Campylobacter causes disease, how the human host combats the infection, and whether nonantibiotic methods can be developed to treat this pathogen. In response to these needs, this work found that C. jejuni infection leads to accumulation of the major antimicrobial protein calgranulin C in humans. Additionally, ferrets were successfully infected with Campylobacter and exhibited a productive immune response that was characterized by increased levels of systemic tumor necrosis factor alpha (TNF-α) and increased leukocyte trafficking to the colon. This trafficking was found to contribute to increased gastrointestinal calgranulin C levels, which were subsequently found to effectively inhibit growth of C. jejuni in a zinc-dependent manner. Calgranulin C treatment led to a bacterial response that was characterized by increased transcript abundance of genes involved in the production of energy and decreased transcript abundance of genes associated with translation. This work represents the first observation of calgranulin C accumulation during Campylobacter infection and examination of the role of host-derived zinc sequestration as a mechanism of inhibiting Campylobacter growth.

RESULTS

Calgranulin C increases in the gastrointestinal tract of patients infected with C. jejuni.

To initially determine whether calgranulin C accumulation occurs during human infection, we examined the presence of the protein in clinical stool samples. After samples were confirmed to be either negative for a panel of gastrointestinal pathogens, including Campylobacter, or positive for Campylobacter alone, they were compared for calgranulin C levels by enzyme-linked immunosorbent assay (ELISA). Human fecal samples that came from patients infected only with Campylobacter exhibited a statistically significant (approximately 2-fold; P < 0.05) increase in calgranulin C levels compared to levels in samples from uninfected patients (Fig. 1).

FIG 1.

Calgranulin C concentrations in human fecal samples. Clinical residual stool samples were analyzed for calgranulin C (S100A12) concentrations by ELISA. The mean and standard deviation are displayed. Data were compared using an unpaired two-tailed Student's t test with Welch's correction at an α value of 0.05.

Campylobacter effectively colonizes both male and female ferrets, which develop clinical signs of infection.

To determine whether C. jejuni infection impacts male or female ferrets differently, we infected cohorts of each gender and noninvasively monitored them for disease. Following oral challenge with C. jejuni, Campylobacter was detected in the feces of infected ferrets at various levels. In general, we observed three trends: (i) some ferrets became infected to high levels early in infection but rapidly cleared the bacteria; (ii) some ferrets became infected, peaked at day 3, and cleared the infection by day 12; and (iii) one male was relatively resistant to infection, yielding countable colonies only at day 3 postinfection. Also, we did not detect drastic differences in colonization levels between male and female ferrets (Fig. 2). For the cohort, bacterial loads peaked at day 3 postinfection and steadily declined until C. jejuni was no longer detected in any ferret by day 12 postinfection. C. jejuni was not detected in the feces of mock-infected ferrets at any time point throughout the study (data not shown).

FIG 2.

C. jejuni loads in the feces of infected ferrets. Campylobacter CFU were enumerated in feces of infected ferrets using selective medium. Loads in individual males (blue) and females (orange) are displayed as the number of CFU per gram of feces, along with the mean and standard error of the mean for the cohort at each time point.

Frank blood was not observed in the stool of infected ferrets at any time point, but occult blood was assayed for as previously described (19). Occult blood was detected frequently in all animals (infected and uninfected) throughout the experiment, possibly due to stress or lack of dietary restriction (see Fig. S1 in the supplemental material). In contrast, only infected animals exhibited watery stool and lethargy during early infection, with no clinical signs being observed in any animal by day 12. To determine whether Campylobacter infection impacted development, we recorded the length and weight of each animal every week for 6 weeks postinfection. These measurements did not significantly differ between infected and uninfected ferrets at any time point, suggesting that acute infection does not adversely affect development (Fig. S2A and B). There was a slight effect on the weight of males during early infection, which was likely due to decreased appetite and did not persist past clearance. As expected, males were generally larger and heavier than females, with lengths and weights increasing steadily for both genders throughout the experiment.

Interleukin-10 (IL-10) and TNF-α are significantly increased during C. jejuni infection in male and female ferrets.

To confirm that an immunogenic infection was established, we examined cytokine production throughout infection. Initially, we analyzed serum samples from days 6 and 7 postinfection for seven cytokines (gamma interferon [IFN-γ], IL-1β, IL-6, IL-10, IL-12p40, IL-12p70, and TNF-α) using a Luminex murine Cytokine 1 panel (data not shown). These cytokines were chosen based on previous work that analyzed cytokine production during C. jejuni infection of human dendritic cells (20). From this initial screen, we determined that IL-1β, IL-10, IL-12p40, and TNF-α could be detected using the murine panel. Accordingly, the levels of these four cytokines were determined for serum samples from infected and uninfected animals at days 1 to 12 postinfection. When the levels of IL-1β, IL-10, IL-12p40, and TNF-α were compared, we observed a statistically significant (P < 0.05) increase in the levels of IL-10 (up to 3-fold) and TNF-α (up to 2-fold) in infected ferrets (Fig. 3B and D). There was a trend toward increased levels of IL-1β and IL-12p40 in infected versus uninfected ferrets, but these increases were not statistically significant (Fig. 3A and C). Again, these results did not appear to be affected by the gender of the animal.

FIG 3.

Cytokine levels in ferret serum. Cytokine levels for IL-1β, IL-10, IL-12p40, and TNF-α, as indicated, were measured using a Luminex mouse Cytokine 1 panel. Levels in infected individual males (closed blue) and females (closed orange) correspond with the same symbols in Fig. 2. The mean and standard error of the mean are displayed. This infected group was compared to the uninfected group (open symbols) by a Mann-Whitney U test (*, P < 0.05). LoD, limit of detection.

Calgranulin C accumulated in the feces of infected ferrets as C. jejuni numbers decreased.

To examine whether, similar to what we observed in human samples, calgranulin C levels increased in the ferret, we evaluated fecal calgranulin C levels by ELISA at each time point. At day 7 postinfection, infected ferret fecal samples exhibited a 2-fold increase (376.6 ± 175.6 pg/ml in uninfected ferrets versus 754.8 ± 110.8 pg/ml in infected ferrets) in the abundance of calgranulin C compared to levels in uninfected ferret fecal samples (Fig. 4). This result was similar to what was observed in humans and was found to be significant using an unpaired two-tailed Student's t test with Welch's correction (P < 0.05). At all other time points, there were no differences detected between infected and uninfected ferrets, indicating that calgranulin C diffusion from the tissue occurs at discrete time points late in infection (data not shown).

FIG 4.

Calgranulin C concentrations in ferret fecal samples. Levels of calgranulin C (S100A12) were determined for infected and uninfected ferrets at day 7 postinfection by ELISA. The mean and standard deviation are displayed. Infected and uninfected groups were compared using an unpaired two-tailed Student's t test with Welch's correction at an α value of 0.05.

Campylobacter infection results in leukocyte accumulation in blood and intestinal tissue.

Following observations that Campylobacter infection results in increased calgranulin C levels in the feces, a more invasive study was performed on male ferrets to examine intestinal pathology and leukocyte migration during infection. As no differences in colonization were observed between male and female ferrets, as noted in Fig. 2, naive male ferrets were infected, and Campylobacter loads were determined at days 1, 3, 6, and 7 postinfection. As expected, these animals were efficiently colonized and exhibited reduced numbers of C. jejuni bacteria until full elimination by day 6 postinfection, representing a slight increase in the rate of clearance compared to that of the animals described above (data not shown).

Extracted colonic tissue from these animals was imaged following hematoxylin and eosin (H&E) staining, which indicated an increase in macrophages and neutrophils in infected animals at days 1 and 3 postinfection (Fig. 5). By day 7, the infected tissue became indistinguishable from that of the uninfected control tissue. In order to quantify the infiltration of the leukocytes into tissue, flow cytometry was performed on both blood and colonic tissue. This analysis found that both probable macrophages (Cd11b⁺ mononuclear cells) and neutrophils (Cd11b⁺/SSC^high granulocytes, where SSC^high indicates a high side scatter light-scatter phenotype) increase in the blood during early infection (days 1 and 3) and return to levels observed in uninfected controls by day 7 postinfection (Fig. 6A and B). Restoration of levels to those of uninfected controls was presumably due to the chemotaxis of leukocytes toward the site of infection in the colon. To examine leukocyte infiltration into the intestinal tissue, we similarly quantified both probable macrophage and neutrophil populations in ferret colons. Similar to what was observed in the blood, both populations of leukocytes increased in intestinal tissue early in infection, peaked at day 3, and remained slightly elevated following clearance (Fig. 6C and D).

FIG 5.

Histology of ferret colonic tissue. Representative H&E-stained images from infected and uninfected ferret colons from days 1, 3, and 7 postinfection are presented. Apparent neutrophils (orange arrows) and macrophages (green arrows) are indicated. Magnification, ×40.

FIG 6.

Quantification of leukocyte populations in ferret tissues. Mononuclear cell and granulocyte abundances were determined for both blood and colonic tissue in infected and uninfected ferrets at each time point. Counts for each cell type are presented as relative abundance compared to the level in the uninfected control for that time point. The mean and standard error of the mean are presented for each infected pair at each time point.

Calgranulin C is able to inhibit in vitro growth of C. jejuni in a zinc-dependent manner.

To examine whether calgranulin C is capable of reducing C. jejuni growth in vitro, the bacterium was grown in the presence of purified calgranulin C. Dose-response analysis of calgranulin C-treated C. jejuni cultures determined that the protein was able to inhibit growth in a concentration-dependent manner with a 50% inhibitory concentration (IC₅₀) of 494.3 μg/ml (Fig. 7A). To determine a possible mechanism for this growth restriction, C. jejuni was grown in the presence of 500 μg/ml of calgranulin C supplemented with calcium, copper, or zinc (each at 100 μM) to examine whether growth could be restored (Fig. 7B). Calgranulin C-treated C. jejuni cultures grew significantly less than C. jejuni treated with vehicle alone. Neither calcium nor copper could significantly restore growth in the presence of calgranulin C, but cultures treated with added zinc grew significantly better than those treated with vehicle (phosphate-buffered saline [PBS]) alone (P < 0.01).

FIG 7.

Calgranulin C inhibition of C. jejuni growth in vitro. (A) Dose-response analysis using purified calgranulin C (S100A12) was determined at 48 h at 37°C under microaerobic conditions. (B) Supplementation of medium for growth restoration during calgranulin C treatment. Medium with 500 μg/ml calgranulin C was supplemented with 100 μM CuCl₂, CaCl₂, or ZnCl₂, and cultures were grown for 48 h at 37°C under microaerobic conditions. Groups were compared using an unpaired two-tailed Student's t test (***, P < 0.01). (C) Comparison of znuA mutant susceptibility to calgranulin C growth inhibition. Wild-type C. jejuni with empty vector (EV), the znuA mutant with empty vector, and the mutant complemented with pBWznuA were grown for 48 h at 37°C under microaerobic conditions in medium treated with 62.5 μg/ml calgranulin C. Groups were compared using an unpaired two-tailed Student's t test (***, P < 0.0001).

The ZnuABC zinc transport system was previously shown to be involved in the uptake of zinc from the environment. To determine whether this system is involved in combating zinc sequestration by calgranulin C, we treated wild-type C. jejuni, a znuA mutant, and a complemented mutant with subinhibitory concentrations (62.5 μg/ml) of calgranulin C. As expected, a significant (approximately 2.5-fold) reduction in growth occurred in znuA mutant cultures (Fig. 7C) (P < 0.0001). When znuA was reintroduced on a plasmid, growth was restored in calgranulin C-treated medium to levels indistinguishable from the wild-type C. jejuni level (P = 0.1).

Transcriptome analysis of calgranulin C-treated C. jejuni indicates that calgranulin C may affect energy metabolism.

Using transcriptome sequencing (RNA-seq), calgranulin C-treated and untreated C. jejuni cultures were analyzed for differential gene expression. Surprisingly, increased transcript abundance of zinc transport systems was not observed when calgranulin C-treated cultures were compared to untreated cultures. RNA-seq analysis identified 15 clusters of orthologous groups (COGs) that exhibited significant increases in transcript abundance (P < 0.05) in calgranulin C-treated cultures, with energy production and conversion representing the most increased genes (Fig. 8A). Four different COGs were found to exhibit significantly decreased transcript abundance (P < 0.05) in calgranulin C-treated cultures, with translation and ribosomal structure and biogenesis accounting for over half of all genes showing decreased expression (Fig. 8B).

FIG 8.

Differences in transcript abundance of calgranulin C-treated and untreated C. jejuni cultures. Levels of transcript abundance were compared between treated and untreated C. jejuni cultures. Transcripts that exhibited a significant change (P < 0.05) in abundance are presented. (A) Functions of upregulated genes (n = 32). (B) Functions of downregulated genes (n = 7).

DISCUSSION

Campylobacter is the leading cause of bacterially derived gastroenteritis in the world. Following ingestion, the bacterium adheres to and transits through the intestinal mucus, where it adheres to the intestinal epithelium and can invade the tissue, causing inflammation (3). Additionally, the bacterium expresses the cytolethal-distending toxin (CDT), which causes cell cycle arrest at the G₁/S or G₂/M transition. C. jejuni strains encoding the CDT have increased invasion in host cells compared to CDT-negative strains. Infection with virulent strains of C. jejuni causes mild to severe diarrhea that may contain blood (21).

The inflammatory response to this process is not well understood in humans, especially compared to what is known about other gastrointestinal pathogens. It has been previously shown that C. jejuni infection stimulates the production of multiple proinflammatory cytokines, including IL-1β, IL-8, and TNF-α, likely through NF-κB activation (20). Based on the role of these cytokines in neutrophil recruitment, we hypothesized that the presence of the neutrophil-derived protein calgranulin C would increase during acute infection of patients with C. jejuni infection. As predicted, we observed that in fecal samples from patients infected with only C. jejuni, calgranulin C was significantly increased compared to the levels in samples of uninfected patients. This supports our hypothesis that proinflammatory cytokine production during infection leads to neutrophil recruitment, subsequent degranulation of the neutrophils in the gastrointestinal tissue, and accumulation of calgranulin C, which may inhibit growth of Campylobacter during infection.

Research into the host responses that occur during C. jejuni infection has been impeded by the absence of a good animal model that exhibits some of the same responses. One model that has been used sparingly is the ferret, which has previously been shown to develop clinical signs that resemble human symptoms. Additionally, during early infection, epithelial cells are lost, and neutrophil influx occurs in intestinal tissue, with disease resolution occurring at around day 6 (17, 18). Based on these studies, we hypothesized that in ferrets, similar to the response in humans, proinflammatory cytokines would increase during early infection and the subsequent recruitment of neutrophils would lead to calgranulin C accumulation in the lumen. Such an observation would lend further support for the use of the ferret as an adequate model for studying Campylobacter-induced disease in humans.

In our experiments during infection with Campylobacter, bacterial loads in ferret feces increased and peaked at day 3 postinfection, followed by C. jejuni numbers steadily decreasing in all individuals, with full clearance of all animals by day 12 postinfection. This peak and duration of infection are consistent with previous studies that used ferrets as a disease model for Campylobacter (22). Based on a previous study that examined cytokine production by human dendritic cells in response to C. jejuni infection, we initially quantified levels of the pro- and anti-inflammatory cytokines IFN-γ, IL-1β, IL-6, IL-10, IL-12p40, IL-12p70, and TNF-α in ferret serum during infection. While we were unable to detect a few of the cytokines using the murine Luminex panel (IFN-γ, IL-6, and IL-12p70), we were able to detect appreciable amounts of IL-1β, IL-10, IL-12p40, and TNF-α.

Ferrets exhibited a significant increase in serum levels of IL-10 early in infection, which was accompanied by an increase in TNF-α production that continued beyond that of IL-10 into days 6 and 7 postinfection. Following the TNF-α increase, IL-10 returned at days 9 and 12 postinfection. This early increase in TNF-α, followed by a later increase in IL-10, was also observed in the previous study using human dendritic cells (20). As was concluded using human cells, our observations support the potential role of TNF-α in combating Campylobacter infection in ferrets, likely through promoting intestinal inflammation and neutrophil recruitment to the site of infection (23). Interestingly, dysregulation of TNF-α is observed in chronic inflammatory disorders such as inflammatory bowel disease and rheumatoid arthritis—two increasingly recognized chronic consequences of Campylobacter infection (24).

The accumulation of IL-10, an anti-inflammatory cytokine with a relatively unknown mechanism of action, was initially elevated, likely in an attempt to control the effects of increased TNF-α, but declined during peak infection to allow for inflammation-mediated clearance of Campylobacter. Eventually, as bacterial loads decreased, IL-10 likely returned to restore the immune response back to baseline. The anti-inflammatory role of IL-10 during C. jejuni infection is well known in mice since mice deficient for IL-10 are one of the only animal models that exhibit intestinal inflammation in response to inoculation with Campylobacter. As a result, they are occasionally used to study Campylobacter-induced disease (25). Such an observation begs the question as to how IL-10 in mice seems to be more resilient in its anti-inflammatory role during C. jejuni infection, often inhibiting inflammation despite colonization.

Interestingly, there was one male ferret that was particularly resistant to colonization by C. jejuni, producing only a single fecal sample that contained a low number of viable bacteria. After examining the cytokine profile in all animals, we found that this ferret exhibited an exaggerated immune response relative to the responses of the other individuals. This may further support the role of these cytokines in combating Campylobacter infection since this animal's robust response may have been the driving factor behind the observed colonization resistance.

As mentioned above, concentrations of calgranulin C were significantly increased in fecal samples from patients that were confirmed to be infected with only C. jejuni. Similarly, in response to proinflammatory cytokine production in the ferret, calgranulin C levels were significantly elevated during late infection. While we did examined calgranulin levels at other time points, only day 7 samples exhibited significantly elevated levels compared to those of uninfected ferrets. This timing is consistent with an earlier study and our study, both of which found that neutrophil recruitment occurs during peak Campylobacter infection, which may result in diffusion of calgranulin C from the site of infection into the intestinal lumen as Campylobacter loads decrease. It is these relatively dilute concentrations that we are detected during surveillance. Since ferret calgranulin C levels reflect the same 2-fold increase in infected individuals that is observed in humans, we can conclude that ferrets can be used to model the effects of calgranulin C—and possibly other neutrophil-derived proteins—during C. jejuni infection of humans (26).

Using dose-response analysis, we found that calgranulin C inhibited C. jejuni growth in vitro at concentrations lower than 500 μg/ml. This concentration is physiologically relevant as it is has been shown previously that calgranulin C levels are comparable to calprotectin levels in a neutrophil, and neutrophilic concentrations of calprotectin can exceed 1 mg/ml. Furthermore, the concentration of calgranulin C that has been employed in our study has also been shown to result in antimicrobial activity against other pathogens (13, 14). Since calgranulin C is known to bind essential metals, this antimicrobial effect may be due to the sequestration of calcium, copper, or zinc. If so, such an effect on Campylobacter growth would indicate that the bacterium may require one or more of these metals for essential processes, supporting the possible targeting of metal import as an anti-infective strategy for the treatment of Campylobacter (16).

To determine whether the sequestration of metals was responsible for restricted growth, metals previously shown to be bound by calgranulin C were supplemented back to treated cultures. When C. jejuni was treated with calgranulin C, growth could not be recovered with calcium or copper supplementation. In contrast, when calgranulin C-treated C. jejuni was supplemented with zinc, growth was restored to levels that exceeded those of untreated bacteria. These results suggest that calcium or copper is not necessary for C. jejuni growth but that calgranulin C's ability to restrict zinc has profound effects on the growth potential of C. jejuni. This likely means that the host restricts zinc availability in gastrointestinal tissue by releasing calgranulin C from neutrophils in an attempt to limit C. jejuni growth during infection. Previous work identified the C. jejuni zinc transport system, ZnuABC, and found that it was necessary for acquiring zinc from the environment (27). As such, we predicted that a znuA mutant would be particularly sensitive to growth inhibition by calgranulin C zinc sequestration. As predicted, since the system is required for efficient zinc uptake by Campylobacter, this mutant exhibited increased susceptibility to calgranulin C. This result, along with results of the initial ZnuABC study, indicates that Campylobacter possesses systems that aid in zinc acquisition in the host and that these systems may be important for combating host-initiated metal sequestration during Campylobacter infection.

Using RNA-seq, cultures treated with calgranulin C were analyzed for differential gene expression. While we did not see upregulation of zinc transport systems, we did observe 15 COGs that were upregulated in calgranulin C-treated cultures, with the category energy production and conversion being the most represented. Additionally, four COGs were downregulated in calgranulin C-treated cultures, with the translation, ribosomal structure, and biogenesis genes accounting for over half of all downregulated genes. Interestingly, this system also appears to be upregulated in treated cultures. A possible explanation for this was recently demonstrated by Shin and Helmann, who found that Bacillus subtilis scavenges for zinc from intracellular sources before undertaking energy-dependent zinc uptake using systems like ZnuABC (28). Similar to the above study, we may be observing the downregulation of enzymes that have high zinc requirements, while Campylobacter begins to harvest zinc from existing enzymes (27). As the scavenged zinc is incorporated into other processes, the bacterium can then produce new enzymes that do not contain zinc. In the future, it would be interesting to determine whether, like B. subtilis, longer treatments with calgranulin C lead to induction of znuABC expression as intracellular zinc pools become depleted.

In summary, calgranulin C accumulates in the gastrointestinal tract during C. jejuni infection of humans, which is reflected in the ferret model of campylobacteriosis, likely through neutrophil recruitment during infection. Calgranulin C was found to restrict the growth of C. jejuni in a zinc-dependent manner in vitro, indicating that nutritional immunity may be an important aspect of host-initiated growth restriction during infection. This is a novel insight into the host-microbe interactions that occur during human C. jejuni infection.

MATERIALS AND METHODS

Human fecal assay.

Residual human fecal samples from standard-of-care testing were obtained from the University of Nebraska Medical Center's Department of Pathology and Microbiology (approval UTK IRB-17-03795-XM). Fecal specimens were evaluated for the presence of gastrointestinal pathogens using a BioFire FilmArray gastrointestinal panel. Briefly, protease inhibitor tablets were dissolved in PBS (one tablet per 15 ml of PBS). Approximately 0.5 g of each fecal sample was weighed out in a Falcon tube, and 5 ml of the protease inhibitor mixture was added to each sample. Samples were then vortexed vigorously, and 50 μl of each sample was used in a human calgranulin C (S100A12) ELISA kit (Life span Biosciences, Inc.). Absorbances at a wavelength of 450 nm (and 620 nm as a reference) were determined using a plate reader immediately following the final step of the manufacturer's protocol. A statistical analysis was performed using a t test with Welch's correction.

Animals.

Twelve weaning-aged ferrets (5.5 to 6 weeks of age), six males and six females, were obtained from Marshall BioResources (North Rose, NY). Ferrets were individually housed at the Walters Life Sciences (WLS) Animal Facility at the University of Tennessee Knoxville. They were acclimated in the WLS animal facility for 7 days to allow the animals to recover following transportation. Ferrets were fed Envigo Teklad global ferret diet and given water ad libitum throughout this study, except for a restriction 2 to 3 h prior to inoculation with Campylobacter jejuni. All animal procedures were conducted under University of Tennessee IACUC protocol 2519.

Bacterial cultures and inoculation.

The principal C. jejuni strain used in this study was 81-176. This strain was grown on Campylobacter-specific medium and Gram strained to ensure purity of the culture. The pure culture was then streaked again on Campylobacter-specific medium and incubated at 37°C under microaerophilic conditions for 24 h. Six suspensions of C. jejuni 81-176 were made in 5 ml of phosphate-buffered saline (PBS) and diluted to an optical density at 600 nm (OD₆₀₀) of 2. Ferrets were anesthetized with isoflurane, and inoculum was administered via orogastric tubing and 12-ml syringe. Six ferrets were infected with 5 ml of C. jejuni 81-176 at an OD₆₀₀ of 2 (an infectious dose of 1.5 × 10⁹ CFU) in PBS combined with 5 ml of 5% sodium bicarbonate buffer. Six ferrets were mock infected with 5 ml of PBS combined with 5 ml of 5% sodium bicarbonate buffer. Morphine was administered to each ferret approximately 1 h following inoculation.

Sample collection and Campylobacter enumeration.

Ferrets were weighed and measured once per week postinfection. Fecal and serum samples were collected on days 1, 3, 6, 7, 9, 12, and 14 postinfection, followed by sample collection once per week. Three fecal samples were collected from each cage per time point in the morning. Approximately 200 mg of feces from one sample from each animal was immediately weighed out and diluted 1:100 in PBS. The remaining samples were immediately frozen at −80°C. Diluted samples were further serially diluted in PBS to 10⁻⁸, and 100 μl of each dilution was plated on Campylobacter-specific medium. The plates were then incubated at 37°C under microaerophilic conditions for 48 h before Campylobacter loads were determined.

Ferrets were individually anesthetized with isoflurane prior to blood collection. Blood was drawn from the vena cava and was limited to 0.2% of the animal's body weight for time points during days 1 to 12 postinfection, followed by 1.5% of the animal's body weight for blood draws once per week. After each collection, blood was allowed to clot in an upright position for at least 20 min. Samples were then centrifuged at 3,500 rpm for 15 min. Serum was then pipetted into clean 1.5-ml Eppendorf tubes and frozen at −80°C. The remainder of the clotted sample was disposed of.

Serum assay.

Serum samples collected on days 1 through 14 were analyzed at Vanderbilt University Medical Center (VUMC) using a Luminex mouse Cytokine 1 panel. For this assay, a 96-well plate was washed prior to the addition of standards and samples. The appropriate matrix was then added to the wells, followed by antibody-immobilized beads. The plate was then incubated overnight (16 to 18 h) on a shaker at 4°C. After this incubation, plate contents were removed, and the plate was washed three times with buffer. Detection antibody was then added to each well, and the plate was shaken for 1 h at room temperature. Following incubation, streptavidin-phycoerythrin was added to each well, and the plate was incubated for 30 min at room temperature with shaking. The contents of the plate were removed, and the plate was washed three times with buffer. After the plate was washed, drive fluid was added to each well, and the plate was shaken at room temperature for 5 min. Finally, the plate was read on a Magpix using Exponent software, and results were analyzed using Milliplex Analyst software. Using this method, levels of IL-1β, IL-10, IL-12p40, and TNF-α were compared in uninfected and infected ferret samples.

Fecal assays.

Several assays were performed on fecal samples from the first seven time points (days 1 through 12 postinfection). An occult blood test was performed using a Beckman Coulter Hemoccult II test kit. This procedure involved smearing a thin layer of feces over the specified box, incubating the sample for 3 to 5 min, applying two drops of Hemoccult developer solution, and reading the result within 60 s. Results were subjectively scored using 0 as a negative, 0.5 as a weak positive, and 1 as a strong or regular positive, which is shown as presence/absence of a blue appearance.

Levels of calgranulin C were analyzed in ferret fecal samples similar to the method described above. Briefly, protease inhibitor tablets were dissolved in PBS (one tablet per 15 ml of PBS). Approximately 0.5g of each fecal sample was weighed out in a Falcon tube, and 5 ml of the protease inhibitor mixture was added to each sample. Samples were then vortexed vigorously, and 50 μl of each sample was used in a human calgranulin C (S100A12) ELISA kit (Life span Biosciences, Inc.). This same procedure was performed to measure calgranulin C levels in Campylobacter-infected and uninfected human fecal samples as well. Absorbance of the samples was determined using a plate reader immediately following the final step of the ELISA. ELISA plates were read at a wavelength of 450 nm, with 620 nm as the reference wavelength. To measure alpha-1 protease inhibitor levels in ferret fecal samples, a canine (Texas A&M) immunoassay and a human (Biovendor) ELISA kit were used.

Preparation of H&E-stained colonic tissue.

The terminal 1 cm of the ferret colon was removed following sacrifice, and the lumen was washed five times with 1 ml of sterile, cold PBS. Tissue was placed in 10% buffered formalin and fixed for 4 h at room temperature. Fixed tissue was embedded in formalin, and 4-μm sections were stained with hematoxylin and eosin. Stained slides were visualized, and representative images are presented.

Purification and enumeration of leukocytes from whole blood and colon.

Whole blood collected by vena cava puncture was stored in lithium heparin tubes. Peripheral blood leukocytes (PBLs) were obtained by ammonium-chloride-potassium (ACK) lysis.

Cells were isolated from colon as previously described for murine colons (29). Briefly, the entire colon was removed from ferrets immediately following euthanasia. The lumen was washed with 5 ml of cold PBS and then separated into 1-cm sections. These sections were opened, exposing the lumen. These opened sections were placed in a petri dish and washed three times with cold PBS. Sections were minced into 1-mm sections and collagenase digested. Digested products were passed through a 40-μm-pore-size cell strainer using a 5-ml plunger to obtain single-cell suspensions. These single-cell suspensions were washed with cold PBS, counted, and then resuspended to the appropriate density in fluorescence-activated cell sorting (FACS) buffer.

Flow cytometry assays were conducted as previously described (30). Briefly, single-cell suspensions were Fc receptor blocked using 1% goat serum in FACS buffer. Cells were then washed and stained with an antibody recognizing ferret CD11b (clone M1/70-APC; Biolegend). Data were acquired using an LSR II instrument (BD Bioscience, San Jose, CA) and analyzed using FlowJo software (Tree Star, Ashland, OR).

C. jejuni growth in the presence of calgranulin C in vitro.

Dose-response assays were performed on C. jejuni wild-type strain DRH212 using calgranulin C as a growth inhibitor. Calgranulin C was expressed recombinantly and purified from Escherichia coli as previously described (31). The concentration of calgranulin C started at 2,000 μg/ml and was serially diluted (1:1 in Helicobacter pylori medium) to 15.625 μg/ml to determine the minimal concentration necessary to inhibit C. jejuni growth. These assays were carried out in triplicate using Helicobacter pylori medium (brucella broth with fetal bovine serum and calprotectin buffer) as the growth medium. To each calgranulin C concentration, an equal volume (100 μl) of C. jejuni culture at an OD₆₀₀ of 0.05 was added. These plates were read at 600 nm following 48 h of incubation in a microaerobic environment at 37°C.

Role of zinc in calgranulin C-dependent growth inhibition.

Once the MIC of calgranulin C (500 μg/ml) was determined, assays were carried out with wild-type C. jejuni and elements known to be sequestered by calgranulin C. Copper or zinc was individually added to wells of a 96-well plate containing an inhibitory concentration of calgranulin C (100 μl of 100 μM CaCl₂, CuCl₂, and ZnCl₂ was added to a 100-μl mixture of C. jejuni at an OD₆₀₀ of 0.05 with 500 μg of calgranulin C). Plates were incubated in a microaerobic environment at 37°C and read at 600 nm following 48 h of incubation to determine if any of these elements could restore C. jejuni growth in vitro following calgranulin C treatment. Statistical analysis was performed using Student's t test.

To determine whether the C. jejuni znuA mutant exhibits increased susceptibility to calgranulin C-dependent growth inhibition, wild-type C. jejuni carrying the empty vector pBW210, the ΔznuA mutant with empty pBW210, and the ΔznuA pBWznuA (pLD36) complemented strain were grown overnight under microaerobic conditions at 37°C on Mueller-Hinton (MH) agar with kanamycin (100 μg/ml) (27). Cells were harvested and normalized to an OD₆₀₀ of 0.05 in H. pylori medium and then added to an equal volume (100 μl:100 μl) of H. pylori medium containing 125 μg/ml calgranulin C (62.5 μg/ml final concentration) in the wells of a 96-well plate. Cultures were grown for 48 h at 37°C under microaerobic conditions and then read at 600 nm. Statistical analysis was performed using Student's t test.

Transcriptome analysis of C. jejuni cultures.

C. jejuni (strain DRH212) was incubated overnight at 37°C under microaerobic conditions on Mueller-Hinton agar plates containing 10 μg/ml trimethoprim. Cells were harvested and normalized to an OD₆₀₀ of 1 in 160 ml of H. pylori medium. This suspension was separated into four replicate tubes of 40 ml, and two were treated with 500 μg/ml purified calgranulin C. All replicates were grown at 37°C under microaerobic conditions for approximately 6 h. Samples were subsequently centrifuged at 4,000 rpm for 10 min at 4°C, the supernatant was discarded, and the pellet was resuspended in 1 ml of TRIzol. Samples were frozen at −80°C until RNA extraction.

To extract RNA, the two untreated and two calgranulin C-treated TRIzol samples were thawed on ice, vortexed, and incubated at room temperature for 5 min. Subsequently, 100 μl of chloroform was added to each sample and vortexed. Samples were incubated at room temperature for 5 min, followed by centrifugation at 15,000 rpm for 5 min at 4°C. The aqueous phase was collected and added to 700 μl of 100% isopropanol in a fresh tube. The tubes were vortexed and centrifuged at 15,000 rpm for 10 min at 4°C. The supernatant was discarded, and the nucleic acid pellets were washed with 70% isopropanol. These pellets were centrifuged at 15,000 rpm for 5 min at 4°C. The isopropanol was removed, and the pellet was air dried for approximately 15 min at room temperature. Finally, pellets were resuspended in 50 μl of UltraPure nuclease-free water and frozen at −20°C.

Samples were DNase treated to ensure that only RNA was present. For DNase treatment, 20 μl of DNase I 10× reaction buffer was added to 30 μg of RNA. DNase (8 μl) was added to each sample and brought to a volume of 200 μl using UltraPure nuclease-free water. Samples were incubated at 37°C for 30 min, and an additional 8 μl of DNase was added and incubated for another 30 min. Samples were cleaned using a Zymo RNA Clean and Concentrate kit, and the absence of DNA was confirmed by standard PCR.

The four DNA-free samples were treated with an Illumina Ribo-Zero Magnetic kit to remove bacterial rRNA. DNA libraries were then made from each RNA sample using a MiSeq Reagent kit, version 3, according to the manufacturer's protocol, and labeled with individual indexes. Libraries were quality checked using an Agilent bioanalyzer. These libraries were sequenced using an Illumina MiSeq instrument, and reads were analyzed for differential gene expression using Geneious, version 10.1.3.