Helicobacter pylori perturbs iron trafficking in the epithelium to grow on the cell surface

Abstract

Helicobacter pylori (Hp) injects the CagA effector protein into host epithelial cells and induces growth factor-like signaling, perturbs cell-cell junctions, and alters host cell polarity. This enables Hp to grow as microcolonies adhered to the host cell surface even in conditions that do not support growth of free-swimming bacteria. We hypothesized that CagA alters host cell physiology to allow Hp to obtain specific nutrients from or across the epithelial barrier. Using a polarized epithelium model system, we find that isogenic ΔcagA mutants are defective in cell surface microcolony formation, but exogenous addition of iron to the apical medium partially rescues this defect, suggesting that one of CagA's effects on host cells is to facilitate iron acquisition from the host. Hp adhered to the apical epithelial surface increase basolateral uptake of transferrin and induce its transcytosis in a CagA-dependent manner. Both CagA and VacA contribute to the perturbation of transferrin recycling, since VacA is involved in apical mislocalization of the transferrin receptor to sites of bacterial attachment. To determine if the transferrin recycling pathway is involved in Hp colonization of the cell surface, we silenced transferrin receptor expression during infection. This resulted in a reduced ability of Hp to colonize the polarized epithelium. To test whether CagA is important in promoting iron acquisition in vivo, we compared colonization of Hp in iron-replete vs. iron-deficient Mongolian gerbils. While wild type Hp and ΔcagA mutants colonized iron-replete gerbils at similar levels, ΔcagA mutants are markedly impaired in colonizing iron-deficient gerbils. Our study indicates that CagA and VacA act in concert to usurp the polarized process of host cell iron uptake, allowing Hp to use the cell surface as a replicative niche.

Figure 1. Hp acquires iron from host cells during colonization of the polarized epithelium.

(A) Addition of iron allows ΔcagA growth on the apical cell surface. Polarized cells in the Transwell system were infected with WT or ΔcagA. Co-culture media (+) was added basally. Solid lines indicate conditions with DMEM apically. Dashed lines indicate conditions with 100 µM ferric chloride added to the apical DMEM. Samples were taken daily from the apical chamber and plated for CFU counts. (B) ΔcagA response to iron is concentration dependent. Polarized cells in the Transwell system were infected with WT or ΔcagA. Co-culture media was added basally, and different amounts of ferric chloride (FeCl₃) in DMEM added to the apical chamber. Graph shows Hp CFU counts from the apical chamber at day 5 post-infection. (C) Exogenous iron rescues microcolony growth of ΔcagA on the cell surface. 3D confocal images of WT or ΔcagA colonizing the polarized epithelium 5 days post-infection, in the absence or presence of 100 µM ferric chloride (Fe³⁺). Bacteria are visualized with anti-Hp antibodies (green) and cell junctions are stained red (anti-ZO-1). Scale bar 10 µm.

Figure 2. Holotransferrin promotes Hp microcolony growth on the cell surface.

(A) Partially saturated transferrin inhibits Hp colonization of the apical cell surface. Polarized cells in the Transwell system were infected with WT. Co-culture media was added basally. Solid line indicates conditions where DMEM was present apically. Dashed line represents conditions where 75 µg/ml partially saturated transferrin (Tf) was added to the apical DMEM. Samples were taken daily from the apical chamber and plated for CFU counts. (B) Apical Hp microcolonies are protected by the epithelium from partially saturated transferrin in the basal chamber. Polarized cells in the Transwell system were infected with WT or ΔcagA. Solid lines indicate conditions where DMEM was present apically and co-culture media basally. Dashed lines represent conditions where DMEM was present apically and 75 µg/ml partially saturated transferrin (Tf) added to the co-culture media basally. Samples were taken and plated as in (A). (C) Holotransferrin partially rescues ΔcagA growth on the apical cell surface. Polarized cells in the Transwell system were infected with WT or ΔcagA. Co-culture media (+) was added basally. Solid lines indicate conditions with DMEM apically. Dashed lines indicate conditions with 75 µg/ml holotransferrin (holo-Tf) added to the apical DMEM. Samples were taken and plated as in (A). (D) ΔcagA forms microcolonies on the apical cell surface in the presence of holotransferrin. 3D confocal images of WT or ΔcagA colonizing the polarized epithelium 5 days post-infection, in the absence or presence of 75 µg/ml holotransferrin (holo-Tf) in the apical media. Bacteria are visualized with anti-Hp antibodies (green) and cell junctions are stained red (anti-ZO-1). Scale bar 10 µm.

Figure 3. Hp colonization of the apical cell surface increases internalized transferrin.

(A) 3D confocal images of fluorescent transferrin signal in polarized MDCK cells stably expressing human transferrin receptor, uninfected or infected with WT or ΔcagA. Fluorescent transferrin was added to the basal chamber and incubated on ice for 30 minutes, unbound transferrin washed away, then immediately fixed (top panels, cross-sectional view, 0 minutes post-uptake), or further incubated for 30 minutes at 37°C to allow uptake of bound transferrin (middle and bottom panels, top and cross-sectional views, 30 minutes post-uptake). Fluorescent transferrin is shown in red, and bacteria are visualized with anti-Hp antibodies (blue). Scale bars 10 µm. (B) Hp colonization does not significantly affect host cell transferrin receptor expression. Polarized cells in the Transwell system were infected for 2 days with WT or ΔcagA. Whole-cell lysates from these infections were separated by SDS-PAGE, transferred to a nitrocellulose membrane, then immunoblotted with antibodies against transferrin receptor (top panel) and against GAPDH as a loading control (bottom panel). (C) Quantification of transferrin fluorescence 30 minutes post-uptake in polarized epithelial monolayers. Monolayers were infected for 2 days with the indicated Hp strains, fixed after 30 minutes of transferrin uptake, and total transferrin fluorescence measured from multiple 3D confocal images. EPISA is a mutant expressing a mutated CagA that cannot be phosphorylated. CagA* is the complemented ΔcagA mutant. p-values were obtained with a Mann-Whitney statistical test. N.S. indicates no statistical significance.

Figure 4. Transferrin receptor is mislocalized apically to sites of bacterial microcolonies.

(A) Polarized MDCK cells in the Transwell system were infected with WT, ΔcagA, or ΔvacA for 2 days. Apical staining with anti-transferrin receptor antibodies was carried out on non-permeabilized samples. Bacteria are visualized with anti-Hp antibodies (blue), transferrin receptor (tfR) is stained red, and phalloidin staining of f-actin is shown in yellow. 3D confocal images are shown, and cross-sectional view is also presented for WT (second row). Scale bars 5 µm. (B) Quantitative data of the transferrin receptor (tfR) fluorescence intensity associated with the bacterial microcolonies, determined by the fluorescence voxel volume of each microcolony stained with anti-Hp antibodies, and the fluorescence sum of the transferrin receptor signal associated with the microcolonies, measured from multiple 3D confocal images. VacA* is the complemented ΔvacA mutant. Each point on the graph represents a microcolony. p-values were obtained with a Mann-Whitney statistical test. N.S. indicates no statistical significance.

Figure 5. Hp induces mislocalization of basolateral proteins to the apical surface at sites of bacterial attachment.

(A) Cell surface proteins of polarized cells on Transwell filters were selectively labeled with biotin in either the basal (top panels) or apical (bottom panels) chambers. Streptavidin labels biotinylated proteins in green, phalloidin stains f-actin in red and DAPI stains cell nuclei in blue. Scale bar 5 µm. (B and C) Polarized cells infected apically with WT for 2 days were selectively biotinylated on ice at the basolateral surface, and either immediately fixed (top panels, B and C), or first incubated for 30 minutes at 37°C (bottom panels, B and C) before fixation and apical streptavidin staining. Top views (B) and cross-sectional views (C) are shown. Bacteria are visualized with anti-Hp antibodies (red) and biotin is marked in green. Phalloidin staining of f-actin is shown in blue. Scale bars 5 µm. (D) Quantification of basolateral proteins associated with Hp microcolonies. Open squares are data from monolayers fixed immediately after basolateral biotinylation. Closed circles are data from monolayers incubated at 37°C for 30 minutes after basolateral biotinylation. Each point represents the total fluorescence intensity of apically-exposed biotin associated with a microcolony, divided by the number of bacteria present in that microcolony. p-value was obtained with a Mann-Whitney statistical test.

Figure 6. VacA contributes to Hp colonization of the apical cell surface.

(A) VacA aids Hp colonization of the apical cell surface. Using the Transwell system, cells were infected with WT or ΔvacA, and co-culture media added only to the basal chamber (+). DMEM was added to the apical chamber (−). Samples were taken daily from the apical chamber and plated for CFU counts. (B) Exogenous addition of iron apically rescues ΔvacA growth on the polarized epithelium. Polarized cells were infected as in (A). Solid lines indicate conditions with DMEM apically. Dashed lines indicate conditions with 100 µM ferric chloride (Fe³⁺) added to the apical DMEM. Samples were taken and plated as in (A). (C) 3D confocal images of WT or ΔvacA colonizing the cell surface of polarized MDCK cells in the Transwell system. Cells were infected for 5 minutes and then unattached bacteria washed away and media replaced. At 1, 3 and 5 days post-infection, samples were fixed and processed for immunofluorescence. Bacteria are visualized with anti-Hp antibodies (green) and cell junctions are stained red (anti-ZO-1). Scale bar 10 µm. (D) Quantification of WT, ΔvacA and VacA* microcolony sizes over time (1, 3 and 5 days), determined by fluorescence volume measured from multiple 3D confocal images. VacA* is the complemented ΔvacA mutant. Each point on the graph represents a microcolony. p-value was obtained with a Mann-Whitney statistical test. (E) CagA and VacA work in concert to enable Hp colonization of the polarized epithelium. Polarized cells were infected as in (A) with the strains indicated. Samples were taken and plated as in (A). (F) CagA and VacA work together to aid Hp acquisition of iron from host cells. Polarized cells were infected as in (A). DMEM (solid lines) or DMEM + 100 µM ferric chloride (Fe³⁺, dashed lines) was added apically. Samples were taken and plated as in (A).

Figure 7. Down-regulation of host cell transferrin receptor decreases Hp microcolony growth on the cell surface.

(A) siRNA knockdown of transferrin receptor expression in MDCK cells. Cells were transfected with a combination of two siRNAs directed against transferrin receptor, or with siRNA directed against enhanced GFP (eGFP) as a control. 3 days post-transfection, the cells were collected and lysates separated by SDS-PAGE, transferred to a nitrocellulose membrane, then immunoblotted with antibodies against transferrin receptor (top panel) and against GAPDH as a loading control (bottom panel). (B) Residual transferrin receptor is mislocalized apically to sites of bacterial microcolonies. MDCK cells were transfected with siRNA against canine transferrin receptor (ctfR, top panels) or eGFP as a control (bottom panels). After polarization, the cells were infected with WT for 1 day. Bacteria are visualized with anti-Hp antibodies (blue), transferrin receptor (tfR) is stained red, and phalloidin staining of f-actin is shown in yellow. Scale bar 5 µm. (C) Hp form smaller microcolonies on the apical cell surface when transferrin receptor expression is knocked down. 3D confocal images of WT colonizing the polarized epithelium 2 days post-infection, on cells either transfected with siRNAs against canine transferrin receptor (ctfR) or eGFP as a control. Bacteria are visualized with anti-Hp antibodies (green) and transferrin receptor is stained red. Scale bar 10 µm. (D) Quantification of Hp microcolony sizes on cells transfected with siRNAs directed against transferrin receptor or eGFP as a control. Data from 0 and 2 days post-infection are shown. Microcolony sizes were determined by fluorescence volume measured from multiple 3D confocal images. Each point on the graph represents a microcolony. p-values were obtained with a Mann-Whitney statistical test. N.S. indicates no statistical significance.

Figure 8. Hp colonization of the cell surface leads to transferrin transcytosis.

30 µg of biotin-albumin and 75 µg of biotin-transferrin were added to the basal chamber of uninfected or 2 day-infected polarized monolayers. The apical supernatant was sampled after a 24 hour incubation, and 10 µl of these samples separated by SDS-PAGE, blotted onto nitrocellulose, and the biotinylated albumin and transferrin were visualized with fluorescent streptavidin. The lane labeled “standard” is a 1∶250 dilution of the basal media containing biotin-albumin + biotin-transferrin. CagA* is the complemented ΔcagA mutant. Each band was quantified with the LI-COR Odyssey Scanner. Biotin-albumin amounts were used to normalize for loading. The graph depicts the average result from 6 experiments. p-values were obtained with a Wilcoxon signed rank test, using a hypothetical median of 1. N.S. indicates no statistical significance.

Figure 9. Host iron depletion decreases ΔcagA fitness in vivo.

Mongolian gerbils maintained either on a regular, iron-replete diet, or on an iron-deficient diet were infected with Hp strain 7.13 WT or its isogenic ΔcagA mutant. 6–8 weeks post-inoculation, the animals were sacrificed and bacterial counts obtained from the stomach. Each point on the graph represents one animal. Animals from which no bacteria could be recovered are represented at 100 CFU/g, which is the limit of detection. p-values were obtained with a Mann-Whitney statistical test. N.S. indicates no statistical significance.