Alpha-defensin-dependent enhancement of enteric viral infection

Abstract

The small intestinal epithelium produces numerous antimicrobial peptides and proteins, including abundant enteric α-defensins. Although they most commonly function as potent antivirals in cell culture, enteric α-defensins have also been shown to enhance some viral infections in vitro. Efforts to determine the physiologic relevance of enhanced infection have been limited by the absence of a suitable cell culture system. To address this issue, here we use primary stem cell-derived small intestinal enteroids to examine the impact of naturally secreted α-defensins on infection by the enteric mouse pathogen, mouse adenovirus 2 (MAdV-2). MAdV-2 infection was increased when enteroids were inoculated across an α-defensin gradient in a manner that mimics oral infection but not when α-defensin levels were absent or bypassed through other routes of inoculation. This increased infection was a result of receptor-independent binding of virus to the cell surface. The enteroid experiments accurately predicted increased MAdV-2 shedding in the feces of wild type mice compared to mice lacking functional α-defensins. Thus, our studies have shown that viral infection enhanced by enteric α-defensins may reflect the evolution of some viruses to utilize these host proteins to promote their own infection.

Fig 1. MAdV-2 but not MAdV-1 productively infects mouse small intestinal enteroids.

Parallel cultures of (A) enteroids from C57BL/6 mice or (B) CMT-93 cells were infected with wild type MAdV-1 (closed squares, dotted line), MAdV-2 (closed circles, solid line), or MAdV-2.IXeGFP (open circle, solid line, CMT-93 cells only). Titers of progeny virus from samples harvested on the indicated days were determined on CMT-93 cells. For A, a single well was infected at each time point, and the integrated density of the immunofluorescence signal of the well is depicted. For B, the concentration of virus was calculated in fluorescence forming units (FFU) per mL from the TCID₅₀ for each sample. Data are the average of 2 (A) or 3 (B) independent experiments ±SD.

Fig 2. Naturally secreted α-defensins enhance MAdV-2 infection of small intestinal enteroids.

(A) Wild type and Mmp7^-/- enteroids were infected with MAdV-2.IXeGFP via microinjection (circles, black line), basolateral infection (squares, green line), or disruption and mixing (triangles, blue line). The anticipated location of α-defensins (α) and virus (blue) relative to the enteroid cell layer (solid lines) are depicted for each route of infection in the schematics on the right. Viral titers from samples harvested on the indicated days were determined on CMT-93 cells. Data is the relative titer of progeny virus from wild type enteroids compared to Mmp7^-/- enteroids from 3 independent experiments ± SD. (B) Expression of cryptdin 4 (Defcr4) relative to expression of ribosomal protein L5 (Rpl5) in wild type and Mmp7^-/- enteroids from small intestine (SI) and colon was measured by qPCR four days after enteroid passage. Data are the average of two replicate experiments ± SD. (C) Representative images and (D) quantification of total GFP positive cells in wild type (WT, black columns) and Mmp7^-/- (KO, white columns) small intestinal (SI) and colonic (C) enteroids at 24 h post-infection with MAdV-2.IXeGFP by microinjection. Virus was mixed with Texas Red-conjugated dextran (red in C) to mark injected enteroids. Data is the average number of GFP positive cells from ten microinjected enteroids from at least 9 independent experiments ± SEM. (E) Relative light units of wild type and Mmp7^-/- small intestinal enteroids at 24 h post-infection with MAdV-2.IX2AFFluc by microinjection. Bars are colored as in (D). Data is the average of triplicate samples from 3 independent experiments ± SEM. **P<0.001; *P<0.05; ns, not significant.

Fig 3. Fecal shedding of MAdV-2 is increased in mice expressing functional enteric α-defensins.

Data are viral genomes per fecal pellet at the indicated times post infection for each wild type (solid black lines) or Mmp7^-/- mouse (grey dashed lines) after oral infection with (A and B) 1x10⁷ infectious units/mouse or (C and D) 1x10⁶ infectious units/mouse of wild type MAdV-2. Dashed black line in A and C indicates the limit of detection. (B and D) Total virus shed per mouse was calculated by log-transforming the data and determining the area under the curve (AUC) for the time range indicated in A and C for each mouse. N = 7–10 mice per group. In scatter plots, lines are mean ± SD. *P<0.05.

Fig 4. Purified enteric defensins but not pro-defensins bind to and enhance infection by MAdV-2 in cell culture.

CMT-93 cells were infected with (A) MAdV-2 or (B) MAdV-2.IXeGFP that was pre-incubated with the indicated concentrations of the α-defensin cryptdin 2 (circles) or pro-cryptdin 2 (squares). Data is expressed relative to control cells infected in the absence of α-defensin (100%) and are the means of at least three independent experiments ± SD. *P<0.05, **P<0.01, ****P<0.0001 comparing cryptdin 2 to pro-cryptdin 2 at each concentration. (C) The z-average diameter of MAdV-2 (closed symbols) or MAdV-2.IXeGFP (open symbols) upon incubation with the indicated concentrations of cryptdin 2 (circles) or pro-cryptdin 2 (squares) was generated from cumulant analysis of dynamic light scattering. Results are the means of three independent experiments ± SD. ****P<0.0001 applies to both viruses when comparing cryptdin 2 to pro-cryptdin 2 at each concentration. (D) CMT-93 cells were infected with MAdV-2.IXeGFP that was pre-incubated with the indicated concentrations of the α-defensin cryptdin 3 (open triangles) or cryptdin 4 (closed triangles). Data is expressed relative to control cells infected in the absence of α-defensin (100%). Results are the means of at least three independent experiments ± SD. *P<0.05 relative to no defensin control.

Fig 5. α-defensins allow MAdV-2 entry in a receptor-independent manner.

(A) MAdV-2.IXeGFP (circles, solid line) or MAdV-1.IXeGFP (squares, dotted line) was added to CMT-93 cells pretreated with the indicated concentrations of MAdV-2 fiber knob, and infection was quantified 48 h post-infection. Results are the mean of two independent experiments ± SD. (B) MAdV-2.IXeGFP was pre-incubated with 5 μM cryptdin 2 (Crp2) or left untreated and then added to CMT-93 cells that had been pre-treated with 1.0 μM MAdV-2 fiber knob (FK) or left untreated. Infection was quantified 48 h post-infection. Data is expressed relative to control cells infected in the absence of α-defensin or FK (100%). Results are the mean of four independent experiments ± SD. Representative cell monolayers in 96 well plates were imaged at a resolution of 50 μm. Grayscale intensity correlates with eGFP expression. *P<0.05 relative to control.