Epithelial-Specific TLR4 Knockout Challenges Current Evidence of TLR4 Homeostatic Control of Gut Permeability

Abstract

Introduction

Toll-like receptor 4 (TLR4) is a highly conserved immunosurveillance protein of innate immunity, displaying well-established roles in homeostasis and intestinal inflammation. Current evidence shows complex relationships between TLR4 activation, maintenance of health, and disease progression; however, it commonly overlooks the importance of site-specific TLR4 expression. This omission has the potential to influence translation of results as previous evidence shows the differing and distinct roles that TLR4 exhibits are dependent on its spatiotemporal expression.

Methods

An intestinal epithelial TLR4 conditional knockout (KO) mouse line (Tlr4^ΔIEC, n = 6–8) was utilized to dissect the contribution of epithelial TLR4 expression to intestinal homeostasis with comparisons to wild-type (WT) (n = 5–7) counterparts. Functions of the intestinal barrier in the ileum and colon were assessed with tissue resistance in Ussing chambers. Molecular and structural comparisons in the ileum and colon were assessed via histological staining, expression of tight junction proteins (occludin and zonular occludin 1 [ZO-1]), and presence of CD11b-positive immune cells.

Results

There was no impact of the intestinal epithelial TLR4 KO, with no differences in (1) tissue resistance–ileum (mean ± standard error of mean [SEM]): WT 22 ± 7.2 versus Tlr4^ΔIEC 20 ± 5.6 (Ω × cm²) p = 0.831, colon WT 30.8 ± 3.6 versus Tlr4^ΔIEC 45.1 ± 9.5 p = 0.191; (2) histological staining (overall tissue structure); and (3) tight junction protein expression (% area stain, mean ± SEM)–ZO-1: ileum–WT 1.49 ± 0.155 versus Tlr4^ΔIEC 1.17 ± 0.07, p = 0.09; colon–WT 1.36 ± 0.26 versus Tlr4^ΔIEC 1.12 ± 0.18 p = 0.47; occludin: ileum–WT 1.07 ± 0.12 versus Tlr4^ΔIEC 0.95 ± 0.13, p = 0.53; colon–WT 1.26 ± 0.26 versus Tlr4^ΔIEC 1.02 ± 0.16 p = 0.45. CD11b-positive immune cells (% area stain, mean ± SEM) in the ileum were mildly decreased in WT mice: WT 0.14 ± 0.02 versus Tlr4^ΔIEC 0.09 ± 0.01 p = 0.04. However, in the colon, there was no difference in CD11b-positive immune cells between strains: WT 0.53 ± 0.08 versus Tlr4^ΔIEC 0.49 ± 0.08 p = 0.73.

Conclusions

These data have 2 important implications. First, these data refute the assumption that epithelial TLR4 exerts physiological control of intestinal physiology and immunity in health. Second, and most importantly, these data support the use of the Tlr4^ΔIEC line in future models interrogating health and disease, confirming no confounding effects of genetic manipulation.

Introduction

Polarized epithelial cells covering the intestinal tract form a highly selective barrier between the bacteria-filled gut lumen and the comparatively sterile subepithelial tissue [1]. This barrier maintains homeostasis within the gastrointestinal tract, allowing for nutrient absorption and regulation of water exchange [2]. Crucially, the intestinal epithelial lining is also a first-line of defense from pathogens, whereby innate immune pattern recognition receptors recognize harmful bacteria and promote protective inflammatory cascades [1]. Toll-like receptor 4 (TLR4) is a type of pattern recognition receptor expressed on a variety of cell types including immune [3] and epithelial cells [4]. TLR4 and its accessory proteins MD2 and CD14 are widely researched due to dual roles in homeostatic control and suspected involvement in multiple conditions, including inflammatory bowel diseases and chemotherapy-induced gastrointestinal toxicity [5, 6].

Based on its consistent implication with diseased states that are characterized by intestinal dysfunction, TLR4 has been regularly reported to be a key regulator of mucosal barrier function and thus intestinal permeability under physiological conditions [7, 8]. Intestinal permeability via the paracellular route is dictated via tight junction proteins located on the apical-lateral cell surface [9]. The multiple intercellular and bridging proteins of the tight junction, including occludin, zonular occludin 1 (ZO-1), and claudins, allow for the movement of solutes across their electro-osmotic gradient to maintain intestinal homeostasis [10]. A considerable body of evidence anecdotally supports TLR4-mediated barrier control, with TLR4 expression strongly correlating with functional assessments of intestinal permeability and molecular characteristics of tight junction proteins [6]. For example, a 2018 study by Bein et al. [11] found that in a necrotizing enterocolitis model, a decrease in TLR4 expression was significantly associated with a decrease in occludin, ZO-1, and claudin-4 and resulted in increased permeability. While this suggests a connection between functional TLR4 and the preservation of the tight junction complex, these findings are only secondary to original aims and do not fully explain the role of TLR4 in homeostasis. Previous research using global TLR4 knockout (KO) mice also shows that a lack of TLR4 expression does not impact tight junction protein development and barrier function; however, this study only analyzed tight junction expression post-chemotherapy challenges [6]. While these studies implicate TLR4 in the pathobiological control of the mucosal barrier, the majority of these data have been generated in models of disease, and as such, conclusions regarding its physiological control cannot be made.

Another major oversight in the literature regarding TLR4's regulatory control of the intestinal barrier is the lack of site-specific interrogation. TLR4 is not only expressed on epithelial cells of the intestinal mucosa but also immune cells of the submucosa. In fact, immune expression of TLR4 is considerably higher than that of epithelial expression [3], and as such, its impact on mucosal homeostasis and disease is arguably higher. A failure to address site-specific TLR4 mechanisms hampers our ability to dissect causative mechanisms and thus impairs translation of fundamental findings. This has the potential to misguide new interventions targeting TLR4 that may not be delivered in a manner that optimally targets TLR4. This paradox is particularly important in cancer research, where TLR4-dependent mechanisms are central to both the efficacy and toxicity of therapy; yet, a lack of site-specific interrogation has resulted in highly variable and contradictory findings in studies attempting to augment its activity.

There is a clear need to study TLR4-dependent control of the mucosal barrier in a manner that dissects epithelial versus immune mechanisms. As such, we have utilized a conditional intestinal epithelial-specific TLR4 KO mouse line (Tlr4^ΔIEC) [12], with epithelial deletion of TLR4 and unimpaired immune cell expression TLR4. In characterizing this mouse line, we are given the unique opportunity to rigorously define the regulatory role of epithelial TLR4 on the intestinal barrier under physiological conditions. As such, we aimed to characterize the potential intestinal differences of this Tlr4^ΔIEC line compared to wild-type (WT) mice, using structural, molecular, and electrophysiological assessments.

Materials and Methods

Animal Husbandry

Male and female WT C57BL/6 (n = 5–7) and intestinal epithelial conditional TLR4 KO C57BL/6 (Tlr4^ΔIEC, n = 6–8) mice aged 8–12 weeks were housed in ventilated cages in groups of 3–6 animals per cage with a 12-hour light/dark cycle and access to irradiated standard mouse chow and sterile water. Mice were euthanized via CO₂ exposure and cervical dislocation prior to dissection, in accordance with ethical approval of the University of Adelaide Animal Ethics Committee (M-2019-020) and the University of Adelaide Institutional Biosafety Committee (IBC approval number 14254). The study complied with the National Health and Medical Research Council (Australia) Code of Practice for Animal Care in Research and Teaching (2014).

Breeding Strategy and Genetic Confirmation

The intestinal epithelial conditional TLR4 KO C57BL/6 (Tlr4^ΔIEC) mouse model was created by following a transgenic Vil1-cre/Tlr4loxP breeding strategy (The Jackson Laboratory, Bar Harbor, ME, USA). By crossing a homozygous Tlr4loxP/Vil1-cre WT with a homozygous Tlr4loxP/hemizygous Vil1-cre, this breeding strategy resulted in 1 in 2 offspring being the desired conditional KO. Conditional KO of TLR4 was confirmed via polymerase chain reaction (PCR) analysis as per protocols provided by the Jackson Laboratory [13] for TLR4flox [14] and Vilcre genes [15], where DNA was extracted from mouse ear notches using the Nucleospin Tissue DNA extraction kit and used at a working concentration of 20 ng/μL (Machery-Nagel, Duren, Germany). Primer sequences used for confirmation of genotype were as follows: Vil-cre forward GCTTTCAAGTTTCATCCATGTTG, Vil-cre WT reverse TTCATGATAGACAGATGAACACAGT, Vil-cre mutant reverse GTCTTTGGGTAAAGCCAAGC, TLR4-floxed forward TGACCACCCATATTGCCTATAC, and TLR4-floxed reverse TGATGGTGTGAGCAGGAGAG. Cycling conditions for Vil-cre were as follows: denaturing at 95°C for 3 min, then 95°C for 5 s, and then 60°C for 30 s. The final 2 steps were repeated for 40 cycles. The mutant band, representing the presence of hemizygous Vil-cre was at 85 base pairs, compared to WT Vil-cre at 119 base pairs. Cycling conditions for TLR4 flox were 94°C for 2 min, then the following steps were repeated for 10 cycles: 94°C for 20 s, 65°C for 15 s (decreasing by 0.5°C each cycle), and 68°C for 10 s. Following from this, samples were cycled 28 times at 94°C for 16 s, 60°C for 15 s, and then 72°C for 10 s. The final step was 2 min at 72°C. Homozygous TLR4-floxed samples produced a band of 285 base pairs, heterozygous samples were at 234 base pairs and 285 base pairs, and WT TLR4 produces a band at 234 base pairs. For visualizations, samples were run in 4% agarose and visualized using Midori Green Advance DNA stain (Nippon Genetics, Tokyo, Japan). Conditional KO of intestinal epithelial TLR4 (Tlr4^ΔIEC) resulted in PCR showing mutant Vil-cre (85 base pairs) and homozygous TLR4-floxed (285 base pairs).

Further confirmation of successful KO was conducted by real-time PCR, where small intestinal tissue for both WT and Tlr4^ΔIEC was harvested and scraped to separate the submucosal layer for epithelial-specific TLR4 analyses. For both strains, the whole small intestine was dissected and opened longitudinally. Using a rounded scalpel, a light feather-like scraping was undertaken to separate out an epithelial-dominant sample. Ten nanograms of RNA extracted from these scrapings for both WT and Tlr4^ΔIEC mice was reverse transcribed using the iScript cDNA Synthesis Kit (#1708890; Bio-Rad, Gladesville, NSW, Australia) as per manufacturer's instructions. RT-PCR was performed using the Rotor-Gene 3000 (Corbett Research Sydney, Mortlake, NSW, Australia). Amplification mixes contained 1 μL of cDNA sample (100 ng/μL), 5 μL of SYBR green fluorescence dye (QuantiTect; Qiagen, Hilden, Germany), 3 μL of RNase-free water (Macherey Nagal, Duren, Germany), and 0.5 μL of each forward and reverse primers (50 pmol/μL), to make a total reaction volume of 10 μL. Primer details are as presented in Table 1. Thermal cycling conditions were: 95°C for 10 min, 40 cycles of 95°C for 10 s, 59°C for 30 s, and 72°C for 45 s and a final melt step of 60–95°C changing 1°C per step, holding for 5 s each. Samples were run in triplicate, including negative controls (no cDNA template). Experimental threshold (CT) values were calculated by the Rotor Gene 6 programme. CT values were used to quantify relative mRNA expression of TLR4 and β-actin using the ∆C_T method, where relative expression = 2^{−(CT TLR4 − CT β-actin)} [16].

Table 1.

Real-time PCR primer sequences for TLR4 and β-actin

	Forward primer	Reverse primer
TLR4 (Merck)	5'-CTCTGCCTTCACTACAGAGAC-3' Tm 58.3°C	5'-TGGATGATGTTGGCAGCAATG-3' Tm 69.1°C
β-Actin (Integrated DNA Tech-5'-CTCTTCCAGCCTTCCTTCCT-3' nologies) Tm 56.4°C	5'-AGCACTGTGTTGGCGTACAG-3' Tm 57.9°C

TLR4, Toll-like receptor 4; PCR, polymerase chain reaction.

Ex vivo Electrophysical Assessments

Ussing chambers (EM-CSYS-8 with DM-MC8 voltage clamp/electrode input; Physiologic Instruments, San Diego, CA, USA) were used to assess intestinal electrophysiology in WT and Tlr4^ΔIEC mice as previously described [16]. Briefly, segments of ileum and colon were dissected from mice and flushed with ice-cold 1× phosphate-buffered saline (PBS). One cm segments were opened longitudinally along the mesenteric attachment line, mounted into 0.1 cm² aperture sliders (P2303A; Physiologic Instruments), and inserted into chambers filled with a glucose-fortified Ringers solution consisting of (in millimolar): NaCl 115.4, KCl 5, MgCl₂ 1.2, NaH₂PO₄ 0.6, NaHCO₃ 25, CaCl₂ 1.2, and glucose 10, bubbled with carbogen gas (95% O₂, 5% CO₂) and warmed to 37°C [16]. Ileal segments had the mucosal side bathed in mannitol-fortified Ringers (10 mM) to maintain osmotic balance. Once mounted, tissue was voltage-clamped to zero potential difference, establishing baseline readings. Tissue was allowed to equilibrate for 20 min before short circuit current (Isc, marker of net ion transport/secretion), and transepithelial electrical resistance (marker of barrier integrity) was measured using Acquire and Analyse Revision II (v2.3; Physiologic Instruments).

Histopathological Analyses

Hematoxylin and Eosin Staining

Mouse ileum and colon samples were fixed in 10% formalin and embedded into paraffin wax blocks. Formalin-fixed paraffin embedded blocks were sectioned (4 μm) and mounted on SuperFrost White slides (Menzel-Gläser, Braunschweig, Germany). Slides were then fixed on a 37°C heat block, for a minimum of 1 h. Standard hematoxylin and eosin staining procedures were followed [6]. In brief, slides were dewaxed in 3× washes in 100% histolene for 5 min each and then rehydrated with graded ethanol as previously described [6]. Slides were then stained in Lille-Mayers hematoxylin for 5 min and rinsed until clear in running tap water. Slides were then quickly dipped twice in 1% acid alcohol (5 mL HCl + 500 mL 70% ethanol) and washed in running tap water until clear. Tissue was then placed in Scott's Tap Water (in millimolar) (MgSO₄ 166.2; NaHCO₃ 23.7 in 1 L dH₂O) for 2 min and washed. Counterstaining with alcoholic eosin (Sigma-Aldrich, St Louis, MO, USA) occurred for 2 min, and slides were washed with running tap water until clear. Slides were then treated with 90% ethanol (30 s) and 100% ethanol (30 s) and finally cleared with 3× washes in 100% histolene for 5 min each. Slides were coverslipped with D.P.X neutral mounting medium (Sigma-Aldrich).

Slides were imaged using an Olympus BX51 light microscope and Olympus DP20 microscope camera (Olympus, Tokyo, Japan). A modified version of well-established intestinal injury scoring criteria [17] was used to quantify possible differences between groups, with 1 representing the presence and 0 representing the absence of the pathophysiological marker. The criteria included: disruption of brush border and surface enterocytes, crypt loss/architectural disruption, disruption of crypt cells, infiltration of polymorphonuclear cells and lymphocytes, dilation of lymphatics and capillaries, edema, villous fusion, and villous atrophy. The latter 2 criteria were not assessed in colon samples; therefore, the total possible score for ileum samples was 8 and for colon samples was 6. Furthermore, ileum villus height and crypt depth and colonic mucosal thickness were assessed by a blinded researcher (E.E.B.) using the Olympus cellSens Standard imaging program (Olympus).

Alcian Blue and Periodic Acid–Schiff

Alcian blue and periodic acid–Schiff was used to quantify goblet cells in the mucosal of the ileum and colon from formalin-fixed paraffin-embedded tissue blocks. Tissue was cut into 4-μm sections and placed on SuperFrost White slides (Menzel-Gläser). Slides were dewaxed in 3× washes in 100% histolene for 5 min each and then rehydrated with graded ethanol. Slides were stained in alcian blue for 5 min (194 mL dH₂O + 6 mL 100% acetic acid + 2 g alcian blue) before being incubated in 0.5% periodic acid for 5 min. Slides were again washed with dH₂O and then incubated in Schiff's reagent for 15 min before being counterstained with hematoxylin for 30 s. Slides were cleared in 100% histolene (3 × 5 min) and coverslipped with D.P.X neutral mounting medium. Slides were imaged using the Nanozoomer Digital slide scanner, with goblet cell counts conducted by a blinded researcher (K.R.S.).

Immunofluorescence of Tight Junction Proteins and Immune Cells

To investigate molecular determinants of barrier function, immunofluorescence for ZO-1 and occludin was performed. Briefly, the intestine was removed and immediately flushed with ice-cold 1× PBS. Segments of the ileum and colon were fixed in 10% neutral buffered formalin, processed, and embedded into paraffin wax. Tissues were then cut into 4-μm sections and placed onto FLEX IHC microscope slides (Flex Plus Detection System, #K8020; Dako, Næstved, Denmark) and heated on a heat pad. Slides were deparaffinised via 3× washes with 100% histolene and rehydrated with graded ethanol (100% ethanol for 30 s, 90% ethanol for 30 s and 70% ethanol for 30 s) [16]. Antigen retrieval was via the PT Link bath (pre-treatment module, #PT101; Dako) using an EDTA/Tris buffer consisting of (in millimolar): Tris 9.9 and EDTA 1.3 and 0.5 mL Tween 20 in 1.5 L dH₂O, pH = 9 at 97°C for 20 min.

Tissue samples were stained using the DakoCytomation Autostainer (AutostainerPlusTM; Dako, serial number: AS1271F1104). The primary antibodies used were as follows: ZO-1 (61–7300; Invitrogen, Carlsbad, CA, USA, 0.25 mg/mL, 1:100 dilution), occludin (33–1500; Invitrogen, 0.5 mg/mL, 1:200 dilution), and CD11b (ab133357; Abcam, Cambridge, UK, 1:1,000 dilution). Primary antibody was diluted in 5% normal horse serum (Sigma-Aldrich), 1× PBS for tight junction analyses and 1% bovine serum albumin (BSA) for immune cell analyses (Sigma-Aldrich). The secondary antibodies were AlexaFluor 488 anti-mouse (occludin), AlexaFluor 488 anti-rabbit (CD11b), and AlexaFluor 568 anti-rabbit (ZO-1) (Thermo Fisher, Waltham, MA, USA). The secondary antibody was diluted in 1× PBS, 1% BSA, and 2% fetal bovine serum. DAPI (1 μg/mL) (Sigma-Aldrich) counterstaining was utilized to visualize the nucleus of cells in sample, with 1× PBS as the diluent. A protein block of 10% normal horse serum for tight junctions and 4% BSA for immune cells was used to reduce nonspecific antibody binding during the procedure.

Post-staining, a drop of Fluoroshield (Sigma-Aldrich) was applied to each slide and coverslipped. Slides were then stored in the dark at 4°C to await imaging. Slides were imaged using the Nikon A1 Confocal Microscope using a ×40 objective. Fluorescent staining was quantified via % area stain on the Fiji Image J program as previously described [18].

Statistics

All data were compared using Prism version 8.0 (GraphPad Software, San Diego, CA, USA). Data were first assessed was for normality using the Shapiro-Wilk test. Parametric data were analyzed using a one-way ANOVA or t test and presented as mean ± standard error of mean (SEM). Nonparametric data were analyzed using a Kruskal-Wallis test and presented as median and range. p values <0.05 were deemed significant.

Results

Epithelial TLR4 Does Not Control Intestinal Barrier Function, Tight Junction Integrity, or Immune Cell Infiltration in Healthy Mice

Successful conditional KO of TLR4 on intestinal epithelial cells was confirmed by both genotyping and RT-PCR analysis, with 8-fold higher TLR4 expression in WT intestinal epithelial-dominant scrapings than in Tlr4^ΔIEC scrapings (mean ± SEM): WT 0.29 ± 0.02 versus Tlr4^ΔIEC 0.04 ± 0.02, p = 0.012.

No difference between WT and Tlr4^ΔIEC was observed in the ileum or colon in baseline short-circuit current: ileum (mean ± SEM): WT 113.3 ± 69.8 versus Tlr4^ΔIEC 75.3 ± 34.9 mA p = 0.607, colon: WT 28.4 ± 21.1 versus Tlr4^ΔIEC 36.7 ± 15.5 mA p = 0.752 (Fig. 1a, b). Similarly, there was no difference in baseline transepithelial tissue resistance: ileum (mean ± SEM): WT 22 ± 7.2 versus Tlr4^ΔIEC 20 ± 5.6 Ω × cm², p = 0.831, colon WT 30.8 ± 3.6 versus Tlr4^ΔIEC 45.1 ± 9.5 Ω × cm², p = 0.191 (Fig. 1c, d).

Fig. 1.

Intestinal electrophysiology is not dependent on TLR4 expression. Baseline short-circuit current (Isc, mA) for the ileum (a) and colon (b) and transepithelial tissue resistance (Ω × cm²) for the ileum (c) and colon (d) samples in WT and Tlr4^ΔIEC mice. No difference between groups (WT n = 5–7 and Tlr4^ΔIECn = 6–8, p > 0.05). Data presented as mean ± SEM. WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

ZO-1 and occludin staining was evident at the apicolateral border of epithelial cells of villous and crypt structures of the ileum and colon (Fig. 2, 3). Quantification of tight junction protein staining (% area stain) showed no differences for ZO-1: ileum (mean ± SEM) WT 1.49 ± 0.155 versus Tlr4^ΔIEC 1.17 ± 0.07 p = 0.09; colon WT 1.36 ± 0.26 versus Tlr4^ΔIEC 1.12 ± 0.18 p = 0.47. Similarly, there was no difference in occludin expression: ileum (mean ± SEM) WT 1.07 ± 0.12 versus Tlr4^ΔIEC 0.95 ± 0.13 p = 0.53; colon WT 1.26 ± 0.26 versus Tlr4^ΔIEC 1.02 ± 0.16 p = 0.45. Positive CD11b staining (Fig. 4) was evident in both the ileum mucosa and colon mucosa and submucosa. CD11b-positive immune cells in the ileum were mildly decreased in WT mice: (% area stain, mean ± SEM) WT 0.14 ± 0.02 versus Tlr4^ΔIEC 0.09 ± 0.01 p = 0.04. However, there were no differences in CD11b-positive immune cells in the colon between strains: (% area stain, mean ± SEM) WT 0.53 ± 0.08 versus Tlr4^ΔIEC 0.49 ± 0.08, p = 0.73.

Fig. 2.

Epithelial TLR4 deletion does not influence ZO-1 expression. ZO-1 expression (red) in WT and Tlr4^ΔIEC mice ileum and colon with DAPI (blue) counterstain of nuclei; WT ileum (a), Tlr4^ΔIEC ileum (b), WT colon (c), Tlr4^ΔIEC colon (d), magnification ×40. No difference in ZO-1 expression for either (e) ileum or (f) colon (% area stain, n = 6, p > 0.05). Data presented as mean ± SEM. ZO-1, zonular occludin 1; WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

Fig. 3.

Occludin expression is independent of epithelial TLR4 expression. Occludin expression (green) in WT and Tlr4^ΔIEC mice ileum and colon with DAPI (blue) counterstain of nuclei; WT ileum (a), X ileum (b), WT colon (c), Tlr4^ΔIEC colon (d), magnification ×40. No difference in occludin expression for either (e) ileum or (f) colon (% area stain, n = 6, p > 0.05). Data presented as mean ± SEM. WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

Fig. 4.

Immune cell infiltration does not depend on epithelial TLR4 expression. CD11b expression (green) in WT and Tlr4^ΔIEC mice ileum and colon with DAPI (blue) counterstain of nuclei; WT ileum (a), Tlr4^ΔIEC ileum (b), WT colon (c), Tlr4^ΔIEC colon (d), magnification ×40. Difference in CD11b immune cells in (e) ileum (*p = 0.04), no difference in (f) colon (% area stain, n = 6, p > 0.05). Data presented as mean ± SEM. WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

Epithelial TLR4 Deletion Does Not Affect Intestinal Morphometry

No histological differences were observed between WT and Tlr4^ΔIEC mice in the ileum or colon (Fig. 5a–d), with no change in villus height (p = 0.49), ileum crypt depth (p = 0.66), or colonic crypt depth (p = 0.52) (Fig. 5e–g, p > 0.05). Furthermore, to ensure a comprehensive and translational assessment of intestinal structure, a tissue injury score was assessed. No evidence of microscopic injury was detected in either WT or Tlr4^ΔIEC ileum or colon tissue (ileum p = 0.617, colon p = 0.529). Finally, no difference in goblet cell abundance in ileum villi, ileum crypt, and colon crypt between WT and Tlr4^ΔIEC was observed (Fig. 6a–g, p > 0.05 for all groups).

Fig. 5.

Epithelial TLR4 deletion does not affect intestinal architecture. H&E stain of the ileum and colon from WT and Tlr4^ΔIEC mice. WT ileum (a), Tlr4^ΔIEC ileum (b), WT colon (c), Tlr4^ΔIEC colon (d), magnification ×20. No difference in ileum villus height (e, µm), ileum crypt depth (f, µm), or colon crypt depth (g, µm) (n = 6, p > 0.05). Data presented as mean ± SEM. H&E, hematoxylin and eosin; WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

Fig. 6.

Goblet cell abundance is not affected by TLR4 expression in the intestinal epithelium. AB-PAS stain of the ileum and colon from WT and Tlr4^ΔIEC mice to visualize goblet cells (dark purple/blue stain). WT ileum (a), Tlr4^ΔIEC ileum (b), WT colon (c), Tlr4^ΔIEC colon (d), magnification ×20. No differences in goblet cells in ileum villi (e), ileum crypt (f) or colon crypts (g) (# of goblet cells per structure, WT n = 8, Tlr4^ΔIECn = 7, p > 0.05). Data presented as mean ± SEM. AB-PAS, alcian blue and periodic acid–Schiff; WT, wild-type; SEM, standard error of mean; TLR4, Toll-like receptor 4.

Discussion/Conclusion

TLR4 has received a significant amount of attention for its homeostatic control and therapeutic applications based on its assumed regulation of mucosal barrier function. This is the first study to present the baseline intestinal characteristics of epithelial-specific TLR4 KO mice (Tlr4^ΔIEC) and has found that contrary to expectations, the absence of epithelial TLR4 did not alter intestinal homeostasis. These data highlight the importance of site-specific TLR4 investigation and underscore the limitations of extrapolating evidence from disease models to healthy states.

The concept that TLR4 is involved in homeostatic control is clearly outlined in studies investigating the nervous control of intestinal tissue [19, 20] and immune tolerance [21]. However, TLR4 involvement in homeostatic control of barrier function has been inferred from disease modeling studies. Shi et al. [22] found that intestinal injury in response to dextran sulfate sodium induced colitis was significantly aggravated in a global TLR4 KO mouse model. This suggests TLR4 expression is a protective component of the intestines and supports the healthy functioning of the intestinal barrier [22]; however, this is contradictory to earlier findings which show that TLR4 overexpression leads to impaired intestinal epithelial cell differentiation and barrier dysfunction [23]. Since not all studies distinguish between site-specific expression of TLR4, contradicting evidence is expected. Our study is one of the first to entirely focus on how intestinal epithelial TLR4 expression influences the healthy state of the intestines.

As outlined, our data suggest that epithelial TLR4 is not essential to the regulation of the intestinal environment, namely the role of tight junction protein expression, goblet cell populations, and functional tissue permeability in healthy development. While minor nonsignificant differences have been noted between groups, this is most likely due to stochastic variation. Previous research conducted by our group has demonstrated intestinal permeability and morphology changes in a positive control of epithelial tissue disruption. In mice treated with the chemotherapy irinotecan, baseline intestinal conductance (a measure of intestinal permeability, the opposite of tissue resistance) was significantly increased (53.19 ± 6.46 s/cm², +105.62% relative to WT controls; p = 0.0008) [16]. Furthermore, irinotecan-treated small intestinal and colonic tissue showed severe damage, including villous blunting and crypt degeneration [24]. There is a substantial difference between these positive-control outcomes and the minor changes in the present study, therefore suggesting that current minor variability in the Tlr4^ΔIEC versus WT data is inconsequential. Our current findings indicate that there may be compensatory mechanisms controlling the gastrointestinal microenvironment in the absence of epithelial TLR4. A possible mechanism could be that immune TLR4 is responsible for modulating barrier function and intestinal homeostasis. This aligns with the higher TLR4 expression on immune cell populations and acknowledges the profound immune infiltrate of the gut [25]. TLR4 is expressed on a range of immune cells including, macrophages, myeloid cells, and dendritic cells [3] and has proven roles in dendritic cell maturation [26] and immune tolerance [27]. Considering that immune TLR4 has been shown to control immune system functioning and development of a healthy microbiome, it could be also deduced that immune TLR4 aids in controlling intestinal permeability and barrier function in mice. This notion is supported by our data, which showed no difference in intestinal characteristics between the Tlr4^ΔIEC and WT mice. To confirm this role of exclusive immune TLR4 signaling, future work could be conducted in conditional mice where there is a deletion of immune TLR4 expression.

An alternative mechanism possibly responsible for this compensation could include the recognition of pathogens and tolerance of the commensal microbiota via different TLRs, notably TLR2 [8]. Upon ligand binding, TLR2 activates the MyD88-dependent inflammatory pathway [28]. This pathway is also activated in response to TLR4 activation; therefore, a distinct overlap in TLR4 and TLR2 signaling exists [28]. The similarity between TLR2 and TLR4 is best shown in healthy states, with an early study finding that activation of TLR2 and TLR4 primes dendritic cell tolerance to commensal organisms [29]. Furthermore, the combination of TLR2 and TLR4 signaling is indicated in the healthy control of spontaneous and serotonin-induced contractile responses of mouse ileum [20]. It is the commonality between TLR2 and TLR4 signaling pathways, which may explain why intestinal homeostasis was maintained in our Tlr4^ΔIEC model. A previous research study has investigated disruption of TLR pathways, including TLR2 KO, not MD-2, and has shown in intestinal models of chemotherapy-induced intestinal mucositis that deletion of TLR2 alone increased intestinal inflammation and damage, suggesting TLR2 is a potential therapeutic target [30]. While this evidence indicates importance of TLR2, rather than TLR4, in intestinal regulation, unfortunately, epithelial deletion of TLR2 has not been previously studied in either healthy or diseased states. Therefore, future research could be centered on a TLR2 epithelial-specific KO mouse model, to further investigate this complex relationship between different TLR expression and intestinal function.

Overall, our findings support the use of this Tlr4^ΔIEC mouse line in the investigation of gastrointestinal disease, where TLR4 may be of interest. These mice showed no difference in baseline intestinal characteristics compared to WT, therefore displaying no inherent variability of intestinal function caused by genetic modification of intestinal epithelial TLR4. This is a promising sign for the ongoing viability of this model as the retention of normal intestinal function suggests that the Tlr4^ΔIEC model is reliable. This could allow for future disease models in the Tlr4^ΔIEC mice to dissect the contribution of epithelial TLR4 to disease development. Translationally, the use of these Tlr4^ΔIEC mice in models of gastrointestinal disease will provide much greater insight into the site-specific contribution of TLR4. This would allow for the guiding of future therapeutics, including nanoparticle delivery systems allowing epithelial TLR4 to be augmented in a manner that prevents any systemic effect [31]. This could possibly include TLR4 agonist or antagonist delivery to the site-specific area, meaning that only the intestinal epithelial population of TLR4 would be altered, leaving immune and nervous TLR4 functioning uninterrupted. This is especially important where site-specific TLR4 expression shows distinct and potentially contradicting mechanisms, for example, following cancer treatments [6, 32]. However, while these results shed further insight into the mechanistic roles of epithelial TLR4, they should be approached with caution as this study utilized a small sample size where further functional data, such as metabolic and absorptive capacity, were not assessed. A further limitation of the current study is the exclusion of other related KO models, such as immune-specific TLR4 or TLR2 KO. As data presented in the Tlr4^ΔIEC mice revealed no differences in intestinal functioning, it can be deduced that epithelial TLR4 is likely to have a minor role in intestinal homeostasis. Consequently, future research including these alternative KO models would greatly enhance this field of knowledge.

In conclusion, TLR4 is an important immunosurveillance protein to many areas of current medical research, including inflammatory gastrointestinal diseases and chemotherapy-induced gastrointestinal toxicity [5, 6]. While there is a large body of research surrounding the dual roles of TLR4 in both healthy states and disease, currently, there is very little distinction of cell-specificity in research outcomes. This oversight has the potential to influence the translation of results to clinical practice. To facilitate the emergence of research that considers cell-specific TLR4 expression, a well-validated intestinal epithelial TLR4 conditional mouse model (Tlr4^ΔIEC) must exist. The current study verified that Tlr4^ΔIEC mice are not fundamentally altered prior to future disease modeling studies. These results both support the use of this model in future studies and has presented novel insights into the role of intestinal epithelial TLR4 in homeostatic control.

Statement of Ethics

The study design was approved by the Animal Ethics Committee of the University of Adelaide and complied with the National Health and Medical Research Council (Australia) Code of Practice for Animal Care in Research and Teaching (2014) (animal ethics approval No.: M-2019-020).

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

H.R.W. was funded by the National Health and Medical Research Council CJ Martin Biomedical Research Fellowship, (2018–2022). E.E.C. was funded by the Australian Government RTP Scholarship, (2019–2022). K.R.S. was funded by the Australian Government RTP Scholarship (2017–2021) and Lion's Medical Research Foundation Scholarship (2017–2020).

Author Contributions

All the authors fulfill the ICMJE criteria for authorship. Elise E. Crame is the primary researcher associated with this research and substantially contributed to research design, data acquisition, analysis, and interpretation. She is the primary author of this research, including the drafting and revising process. Joanne M. Bowen, Janet K. Coller, and Hannah R. Wardill provided significant contribution to the study design and conception and had substantial involvement in writing, drafting, and the revision process. Kate R. Secombe provided substantial contribution to data acquisition, analysis, and interpretation. Maxime François and Wayne Leifert provided substantial contribution to data acquisition, analysis, and interpretation.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.