Beneficial Effect of Oral Tigecycline Treatment on Clostridium difficile Infection in Gnotobiotic Piglets

Abstract

The efficacy of oral tigecycline treatment (2 mg/kg of body weight for 7 days) of Clostridium difficile infection (CDI) was evaluated in the gnotobiotic pig model, and its effect on human gut microflora transplanted into the gnotobiotic pig was determined. Tigecycline oral treatment improved survival, clinical signs, and lesion severity and markedly decreased concentrations of Firmicutes but did not promote CDI. Our data showed that oral tigecycline treatment has a potential beneficial effect on the treatment of CDI.

TEXT

Clostridium difficile is a major cause of antibiotic-associated diarrhea worldwide. Currently, standard treatment for C. difficile infection (CDI) includes one of several antibiotics such as metronidazole, vancomycin, or the newly developed fidaxomicin. These treatments are effective in the short term but are also associated with a recurrence rate of 15% to 35%, possibly due to disrupting the gut microbiome (1, 2). Therefore, new therapeutic agents are being investigated. One candidate is tigecycline, which is a broad-spectrum glycylcycline antibiotic with potent activity against C. difficile in vitro (3). Although tigecycline did not reduce C. difficile shedding in mice, it also did not predispose mice to CDI (4), and this reflects the clinical situation, where tigecycline does not induce CDI (3). In fact, tigecycline is a salvage therapy in patients with refractory CDI (5). Together, these data suggest that tigecycline may be a useful first-line treatment for CDI, but further testing is needed. We hypothesized that the oral tigecycline therapy would reduce C. difficile burden in our disease model (infection and then tigecycline administration) and would result in limited perturbation of gut microflora in our model of gut homeostasis (human gut microbiome transplantation first and then tigecycline administration followed by infection). The aims of this study were to (i) evaluate the effects of oral tigecycline treatment on CDI; (ii) determine how oral tigecycline alters transplanted human gut microflora; and (iii) identify whether tigecycline itself predisposes subjects to CDI. Using the gnotobiotic pig CDI model (infection and then tigecycline administration) (6), we compared the treatment efficacy of tigecycline to our previously published success with vancomycin (7). For the first time, we also determined how tigecycline alters transplanted human gut microflora in gnotobiotic piglets (human gut microbiome transplantation first and then tigecycline administration followed by infection).

To evaluate the effects of oral tigecycline treatment on CDI, gnotobiotic piglets were derived via cesarean section (C-section) at full-term gestation (114 days), housed in sterile isolators, and fed Similac milk replacer (Abbott) 3 times daily. A total of 12 5-day-old gnotobiotic piglets were orally inoculated with 1 × 10⁶ spores of C. difficile UK6. By 48 h postinoculation, all piglets developed the yellow, watery diarrhea typical of CDI. At 48 h postinoculation (7days of age), 8 were treated with 2 mg of tigecycline/kg of body weight for 7 consecutive days. Four control piglets received no treatment. Piglets were monitored daily for clinical signs of disease.

Within 7 days postinoculation (dpi), 2 of 4 control piglets developed signs of systemic disease, including lethargy, weakness, anorexia, and dyspnea (Table 1). By necropsy at 10 dpi, control piglets showed lesions characteristic of CDI, including dilation and edema of the gut and pseudomembrane formation in the rectum. In 7 of 8 tigecycline-treated piglets, the gastrointestinal system had no gross lesions, and only 1 piglet had mild signs of CDI (ascites and mesocolonic edema but no pseudomembranous colitis). Microscopic examination revealed a normal spiral colon in tigecycline-treated piglets, while untreated piglets showed pseudomembranes, neutrophil infiltration, and submucosal edema (Fig. 1). Final bacterial counts at 10 dpi obtained by fecal culture on TCCFA plates (8) showed that six of eight tigecycline-treated piglets had negative C. difficile cultures and that the other 2 piglets had lower bacterial counts than control piglets (Table 1).

TABLE 1.

Clinical outcomes of control and antibiotic-treated piglets

Treatment group (no. of piglets)	No. of piglets with indicated symptom severity		Posttreatment C. difficile count (CFU/g)	No. of piglets with fatal systemic disease/total no. of pigletsb	Source or reference
	Diarrhea posttreatment	GI lesionsa
Control (4)c	2 moderate, 2 severe	Moderate to severe	104–107	2/4	This study
Tigecycline at 2 mg/kg (8)	7 resolved,d 1 severe	7 None to mild, 1 severe	0–104	1/8	This study
MBX-500 at 400 mg/kg (6)	4 resolved, 2 mild	None to mild	0–105	0/6	7
Vancomycin at 20 mg/kg (4)	3 resolved, 1 mild	None to mild	0	0/4	7

^aGastrointestinal (GI) lesions were noted during necropsy at 10 dpi and on histopathologic examination of tissues. Lesion severity was based on the extent of dilatation, mesocolonic edema, hemorrhage, mucosal erosions and ulcerations, and pseudomembrane formation and on the degree of neutrophilic inflammation.

^bDetermination of severe systemic illness was based on the presence of respiratory distress, lethargy, and dehydration and the development of pleural or abdominal effusion.

^cNo antibiotic treatment.

^dPiglets began to recover from CDI diarrhea starting at 2 to 3 days post-tigecycline treatments, and the diarrhea was resolved within 7 days post-tigecycline treatments.

FIG 1.

Histopathologic lesions in control and antibiotic-treated piglets. (A and B) Tigecycline-treated piglet tissue (magnified 40 times [A] and magnified 200 times [B]). The large intestine is normal by microscopic examination, with an empty lumen, intact epithelium, and no submucosal edema. (C and D) Untreated piglet tissue (magnified 40 times [C] and magnified 200 times [D]). The asterisks and line delineate the region of submucosal edema in panel C. The arrow points to the ulcerated epithelium that was infiltrated and replaced by neutrophils, exiting to the lumen to the left of the photomicrograph, where the cells degenerated, forming a pseudomembrane. These changes are typical of CDI lesions. Hematoxylin and eosin staining was used.

We further evaluated cytokines in the gut. Interleukin-8 (IL-8) is a potent neutrophil chemoattractant and may contribute to tissue damage in CDI in humans (9–11). Tigecycline treatment significantly reduced IL-8 levels compared to control results (Fig. 2). Other cytokines, including IL-4, IL-6, IL-10, IL-12, tumor necrosis factor (TNF), transforming growth factor beta (TGF-β), and gamma interferon (IFN-γ), showed no significant differences, which we have previously reported (7). This shows that tigecycline treatment reduced the level of a biomarker (IL-8) of CDI. The lack of changes for the other cytokines may reflect the fact that gnotobiotic piglets have underdeveloped immune systems due to the lack of normal microbial stimuli (12). Overall, our results indicate that oral tigecycline is effective in the gnotobiotic piglet CDI model and is comparable to other drugs in this regard.

FIG 2.

Intestinal IL-8 concentrations in infected control and antibiotic-treated piglets. Cytokine concentrations in intestinal contents were measured via enzyme-linked immunosorbent assay. Kruskal-Wallis analysis demonstrated significantly greater mean IL-8 levels in control piglets than in those treated with antibiotics (P < 0.05). Values for MBX-500 and vancomycin were adopted from our data previously generated in our laboratory and published (7).

We next assessed how tigecycline altered human microflora in a transplanted model of homeostasis and determined whether the changes predisposed subjects to CDI. This is important, since the intestinal microflora contribute to health and protect against pathogenic bacteria. Furthermore, disruptions of human gut microflora are closely associated with CDI (13, 14), and some bacterial subpopulations may be particularly vulnerable (15). Despite recent progress in understanding complex microbial communities, we do not know exactly how antibiotics alter the gut microbiome or how these alterations promote CDI in susceptible individuals. In part, this is due to the lack of suitable animal models. We have recently established a gnotobiotic pig model that harbors only the adult human gut microbiome (16). In the current study, we utilized this model and next-generation sequencing to perform a comprehensive analysis of the human microbial population in the germfree pig gut in response to tigecycline therapy. Briefly, piglets were delivered by C-section, maintained in sterile isolators to prevent exposure to unwanted microbes, and fed milk replacer as described above. When the piglets were 2 to 3 days old, 3 ml of human fecal inoculum was administered to six piglets following our published methods (16). Starting 5 to 7 days after fecal transplantation to allow engraftment of human microflora, piglets received 2 mg/kg of oral tigecycline for 6 to 8 consecutive days. Two days after the course of tigecycline, piglets were inoculated with 1 × 10⁷ spores of C. difficile UK6. The C. difficile dose was increased to 10⁷ spores at 14 and 11 days of age for experiments 1 and 2, respectively, to maximize the chance of detecting whether tigecycline leads to clinical CDI.

16S rRNA gene amplicon sequencing targeting regions V1 and V2 was used to characterize the human bacterial communities in gnotobiotic pigs (16). Phylogenetic assessments were performed using the RDP classifier implemented in QIIME with a bootstrap cutoff of 80%, and diversity indices were calculated with an operational taxonomic unit definition at an identify cutoff of 97% (17, 18).

A total of 3,082,558 and 2,362,959 DNA sequence reads were generated from 2 independent experiments. Over 68% and 73% of the total numbers of sequence reads from the two experiments passed the quality control implemented in this study as previously described (19, 20), resulting in 2,124,151 and 1,733,814 sequences. The average numbers of sequence reads generated per pig using a HiSeq2000 sequencer were 126,992 for experiment 1 and 99,921 for experiment 2. At the phylum level, the bacterial communities from piglets that received human microflora were composed primarily of Bacteroidetes, Firmicutes, and Proteobacteria (Fig. 3). However, the composition of the bacterial microbiome experienced a clear change with tigecycline treatment. In gnotobiotic pigs colonized only with human intestinal microflora, tigecycline oral treatment markedly decreased concentrations of Firmicutes within days after antibiotic treatment (Fig. 3). Averages of 73.22% and 40.2% of the gut microflora of three piglets in experiment 1 (piglets 1 to 3) and in experiment 2 (piglets 4 to 6) belonged to phylum Firmicutes before the tigecycline treatment; however, the proportions decreased to 3.17% and 4.03% after the tigecycline treatment. In the previous studies where gnotobiotic piglets were transplanted with the human microflora, Firmicutes did not disappear when they were established in the pig gut (16). Also, the transplanted human gut bacterial communities with high proportions of Bacteroidetes, Firmicutes, and Proteobacteria were stable (21). The proportion of members of the phylum Proteobacteria increased concurrently with tigecycline treatment (Fig. 3). Proteobacteria constituted an average of 11.43% and 14.16% of the gut microflora of three piglets in experiment 1 and experiment 2 before the tigecycline treatment; however, the proportions of Proteobacteria increased to 38.40% and 30.84% after the tigecycline treatment. An approximate return to pretreatment conditions also occurred in this study within days after cessation of tigecycline treatment as described by others for experiments using different antibiotics (22, 23). The proportion of Firmicutes increased whereas that of Proteobacteria decreased days after the cessation of tigecycline treatment. Interestingly, the tigecycline treatment markedly decreased concentrations of Bacteroides and Bifidobacteria in a chemostat model of human intestinal microflora (24). At the class level, the same trends of human gut bacterial shifts as those shown at the phylum level were observed. The proportions of members of Firmicutes, including Bacilli and Clostridia, decreased with tigecycline treatment and then recovered their relative concentrations when the tigecycline treatment was over. The proportion of gammaproteobacteria, a member of Proteobacteria, increased concurrently with the tigecycline treatment and then decreased after the cessation of tigecycline treatment (see Fig. S1 in the supplemental material). Shannon-Weaver and Simpson diversity indices were used to calculate the diversity of microbial communities. As experienced with a taxon analysis, these values also decreased concurrently with the tigecycline treatment (see Fig. S2). The average Shannon-Weaver and Simpson index values showed high diversity and richness with mean values of 5.18 and 0.94 for experiment 1 and mean values of 4.57 and 0.89 for experiment 2 before the tigecycline treatment (19). The calculated mean values of the Shannon-Weaver and Simpson indices decreased to 2.40 and 0.70 for experiment 1 and to 2.33 and 0.69 for experiment 2 after tigecycline treatment. The diversity indices, however, increased to approximately the pretreatment values within days after the tigecycline treatment was over. The average Shannon-Weaver and Simpson index values increased to 4.07 and 0.90 for experiment 1 and to 4.65 and 0.93 for experiment 2 (see Fig. S1). The recovery of bacterial diversity happened together with taxon recovery of the pig gut bacteria. In this study, we compared the bacterial compositions of the human gut microbiome before and after tigecycline treatment. Even though we previously showed that the human gut bacterial communities were stable in gnotobiotic piglets, the inclusion of gnotobiotic piglets transplanted with the human microbiome without any treatment in the study will strengthen the results, which link CDI with perturbation of microbiome. No piglets colonized with human gut microflora and treated with tigecycline followed by C. difficile infection developed any signs of CDI, suggesting that the gut microbiome alteration caused by tigecycline did not promote CDI (3, 4). In contrast, one study (25) showed that the tigecycline treatment increased susceptibility to CDI, which is contradictory with results from this study and others (3, 4). This discrepancy might be the consequence of the use of different doses of tigecycline, durations of treatment, routes of drug administration, strains of C. difficile, animal species, and ages of animals used in each study (see Table S1 in the supplemental material). Effects of these variables should be investigated further to elucidate potential roles of tigecycline in the pathogenesis of CDI.

FIG 3.

Taxonomic classification of the sequences at the phylum level. Phylogenetic assessments were performed using the RDP classifier implemented in QIIME with a bootstrap cutoff of 80%. Ino, inoculation. (A) In experiment 1, piglets were treated with tigecycline (2 mg/kg) for 8 consecutive days when the piglets were between 10 and 17 days of age. (B) In experiment 2, piglets were treated with tigecycline (2 mg/kg) for 6 consecutive days when the piglets were between 7 and 12 days of age.

An analysis of the human gut microbiome showed that it had shifted after tigecycline treatment (Fig. 3; see also Fig. S1 in the supplemental material). Once the human microbiome was transplanted, the gut microbiome resembled that of the inoculum. The treatment with tigecycline almost depleted the Firmicutes population and increased the proportion of Proteobacteria in the pig gut. In contrast, Firmicutes did not disappear in our other experiments when they were established in the pig gut (16). When piglets were challenged with C. difficile UK6 in the absence of Firmicutes, they did not develop CDI. Therefore, data from this study suggest that the tigecycline treatment altered human gut bacterial populations but did not promote CDI.

In conclusion, our data showed that oral tigecycline is an effective treatment for experimental CDI in vivo and that oral tigecycline treatment alone does not promote CDI but does change the composition of the gut microflora. We expect that our data will aid clinicians in developing alternative strategies for treatment of CDI.