Paneth cell ablation in the presence of Klebsiella pneumoniae induces necrotizing enterocolitis (NEC)-like injury in the small intestine of immature mice

Abstract

Necrotizing enterocolitis (NEC) is a leading cause of morbidity and mortality in premature infants. During NEC pathogenesis, bacteria are able to penetrate innate immune defenses and invade the intestinal epithelial layer, causing subsequent inflammation and tissue necrosis. Normally, Paneth cells appear in the intestinal crypts during the first trimester of human pregnancy. Paneth cells constitute a major component of the innate immune system by producing multiple antimicrobial peptides and proinflammatory mediators. To better understand the possible role of Paneth cell disruption in NEC, we quantified the number of Paneth cells present in infants with NEC and found that they were significantly decreased compared with age-matched controls. We were able to model this loss in the intestine of postnatal day (P)14-P16 (immature) mice by treating them with the zinc chelator dithizone. Intestines from dithizone-treated animals retained approximately half the number of Paneth cells compared with controls. Furthermore, by combining dithizone treatment with exposure to Klebsiella pneumoniae, we were able to induce intestinal injury and inflammatory induction that resembles human NEC. Additionally, this novel Paneth cell ablation model produces NEC-like pathology that is consistent with other currently used animal models, but this technique is simpler to use, can be used in older animals that have been dam fed, and represents a novel line of investigation to study NEC pathogenesis and treatment.

Fig. 1.

Infants with NEC have a loss of Paneth cell granules compared with age-matched controls with SIP. Ileal samples were obtained from human infants between 25 and 28 weeks of gestation who developed either SIP or surgical NEC (Bell stage 3) (n=10 for each group). Intestinal sections were stained with the Paneth-cell-specific markers phloxine/tartrazine (Phlox), lysozyme (Lys) and Alcian Blue/periodic acid Schiff (AB/PAS). Tissue sections from NEC samples showed a loss of Paneth cell granules (arrowheads) in all staining methods. To quantify this loss, Paneth cells were counted per 100 total epithelial cells by a single blinded investigator. Paneth cell numbers were significantly decreased compared with age-matched controls (graph; *P<0.001). Scale bars: 0.1 mm.

Fig. 2.

Mice treated with dithizone have selective Paneth cell ablation. P14-P16 CD-1 mice were randomly divided into two groups and given an i.p. injection of dithizone (75 mg/kg) or an equivalent volume i.p. injection of Li₂CO₃ buffer as control and monitored for 16 hours before sacrifice (n=4 for each group). Ileal sections were harvested and stained with either H&E, AB/PAS (to mark Paneth cells) or Alcian Blue alone (to mark goblet cells). Paneth cells were counted per crypt and goblet cells were counted per 100 epithelial cells. Dithizone caused a significant loss of Paneth cells (*P=0.008) but no change in goblet cell number. Arrowheads indicate positive stained cells. Scale bars: 0.1 mm.

Fig. 3.

Selective ablation of Paneth cells in combination with Klebsiella induces an NEC-like injury in 14- to 16-day-old mice. P14-P16 CD-1 mice were divided into four groups (control n=31, dithizone only n=42, Klebsiella only n=13 and Dith/Kleb n=30). Mice were given either an i.p. injection of dithizone (75 mg/kg) or an equivalent volume of Li₂CO₃ buffer. At 6 hours after injection, mice were given either a gavage feeding of 11011 CFU Klebsiella/kg body weight suspended in nutrient broth, or nutrient broth only. Mice were monitored for 10 hours following gavage and then euthanized for small intestine tissue harvest. Using a blinded scoring system evaluating villus integrity, mice treated with Dith/Kleb were significantly different from other groups and their average score was >2 (red line), which is considered to be significant for NEC-like damage (*P<0.001). In the graph, each mouse is represented by a circle. Horizontal lines for each group show the mean. Scale bars: 0.1 mm.

Fig. 4.

Mice without mature Paneth cells are unaffected by Dith/Kleb treatment. P5 CD-1 mice were divided into three groups. Control animals were given an i.p. injection of Li₂CO₃ buffer and 6 hours later were given sterile nutrient broth by gastric gavage (n=3). Klebsiella mice were given an i.p. injection of Li₂CO₃ and 6 hours later were enterally infected via gastric gavage with 1×10¹¹ CFU Klebsiella/kg body weight suspended in nutrient broth (n=3). Dith/Kleb mice were given an i.p. injection of 75 mg/kg dithizone followed 6 hours later with a Klebsiella gavage (n=7). Using a blinded scoring system evaluating villus integrity, mice treated with Dith/Kleb were not significantly different from other groups and all scores were <2, which is considered as having no NEC-like damage. Scale bars: 0.1 mm.

Fig. 5.

NEC induced by Paneth cell ablation causes an increase in inflammation in 14- to 16-day-old mice. P14-P16 CD-1 mice were given either an i.p. injection of dithizone (75 mg/kg) followed 6 hours later by a gavage of Klebsiella (1×10¹¹ CFU/kg body weight) or equivalent volumes of Li₂CO₃ and media as controls (n=8 for each group). At 10 hours after gavage, animals were sacrificed and their ileal tissue was homogenized and harvested for mRNA and protein evaluation of inflammatory cytokines. Reverse transcriptase (RT)-PCR evaluation of tissues revealed a 40% decrease in IL-10 and MMP9, a 40–50% increase in IL-1β and IL-12, and a threefold increase in TNFα. Tissue homogenates were also examined for protein concentration of IL-1β, IL-1, IL-6, IL-10, INF-γ, IL-12 and TNFα using a Meso-Scale Discovery 7-plex pro-inflammatory ultra-sensitive plate and compared with controls. All cytokines were significantly elevated except for IL-12 (P=0.5) and TNFα (P=0.1).

Fig. 6.

The Paneth cell ablation model of NEC is equivalent to other models of NEC. Three models of murine NEC were compared. In the formula + stress (hypoxia-hypothermia) group, P4 mice were surgically given a gastrostomy tube or a sham surgery (controls). Control animals were placed back with their dams (n=5). Control + stress animals were dam fed but given BID (twice daily) hypoxia and hypothermia (n=8). Formula + stress animals were fed a prepared formula through a gastrostomy tube and given BID hypoxia and hypothermia treatments (n=8). In the LPS+PAF group, P10 mice were treated with PAF followed by LPS (n=14) or given an equivalent volume of saline for controls (n=18). In the Dith/Kleb group, P14-P16 mice were divided into four groups. Dithizone mice were given an i.p. injection of dithizone (75 mg/kg) and were monitored for 16 hours before sacrifice (n=42). Control animals were given an equivalent volume i.p. injection of Li₂CO₃ buffer (n=31). Klebsiella mice were given an i.p. injection of Li₂CO₃ and 6 hours later were given a gavage feeding of 1×1011 CFU Klebsiella/kg body weight suspended in nutrient broth (n=13). Dith/Kleb mice were given an i.p. injection of dithizone followed 6 hours later with a Klebsiella gavage (n=30). (A) The formula + stress animals in the formula hypoxia-hypothermia group had significantly more NEC (*P<0.02) than their control animals and had mean scores >2. In the LPS+PAF group, LPS+PAF mice had significantly higher scores than their control animals (*P<0.001) but did not reach a mean score >2. In the Dith/Kleb group, animals treated with Dith/Kleb had significantly higher NEC scores than their controls (*P<0.001) and reached a mean score >2. (B,C) Representative slides from the formula hypoxia-hypothermia group (B) and the LPS+PAF group (C). Scale bars: 0.1 mm.

Fig. 7.

Treatment with Zn can prevent dithizone-induced damage but not Dith/Kleb-induced damage. Ileal segments (2×0.5 cm) were obtained from P14-P16 CD-1 mice and placed on 500 μm mesh Netwell inserts in DMEM with 0.5% FBS at the air-media interface (A). Explants were maintained in 5% CO₂ at 37°C for 30 minutes prior to treatment and during experiments. All experiments were 2 hours in duration. (B–E) A 15-minute treatment with 0.5% dithizone causes significant NEC-like injury (E) compared with lower doses of dithizone (D) or controls (B,C). Adding Klebsiella causes similar injury to dithizone alone (F). Addition of 15 μM zinc protects against dithizone-induced injury (G), but not the injury induced by dithizone and Klebsiella combined (H). Scale bars: 0.1 mm.

Fig. 8.

Surgical placement of gastrostomy tubes. Representative photograph of a 4-day old mouse with gastrostomy tube placed (A). Representative pup connected to syringe pump in ‘pup in cup’ set-up (B).