TLR2 modulates antibodies required for intestinal ischemia/reperfusion-induced damage and inflammation

Abstract

In multiple clinical conditions including trauma and hemorrhage, reperfusion magnifies ischemic tissue damage. Ischemia induces expression of multiple neoantigens including lipid alterations which are recognized by the serum protein, β2-glycoprotein I (β2-GPI)². During reperfusion, binding of β2-GPI by naturally occurring antibodies (Ab) results in an excessive inflammatory response which may lead to death. As β2-GPI is critical for intestinal ischemia/reperfusion (IR)-induced tissue damage and Toll-like receptor 2 (TLR2) is one of the proposed receptors for β2-GPI, we hypothesized that IR-induced intestinal damage and inflammation requires TLR2. Using TLR2^−/− mice, we demonstrate that TLR2 is required for IR-induced mucosal damage, as well as complement activation and pro-inflammatory cytokine production. In response to IR, TLR2^−/− mice have increased serum β2-GPI compared to wildtype mice but β2-GPI is not deposited on ischemic intestinal tissue. In addition, TLR2^−/− mice also did not express other novel antigens suggesting a sequential response. Unlike other TLRs, TLR2^−/− mice lacked the appropriate Ab repertoire to induce intestinal IR tissue damage or inflammation. Together, these data suggest that in addition to the inflammatory response, IR-induced injury requires TLR2 for naturally occurring Ab production.

INTRODUCTION

Although the mortality rate for mesenteric ischemia/reperfusion (IR) has decreased in recent years, it remains at 40–60% (1, 2). Cellular damage induced by the lack of blood flow to the intestine (mesenteric ischemia) is significantly enhanced upon return of blood flow (reperfusion) and frequently results in systemic inflammation. During reperfusion, both a cellular and a humoral innate response is required and inhibition of either the humoral cascade or the cellular infiltrate attenuates IR-induced tissue damage (3, 4). The inflammatory infiltrate of neutrophils and macrophages releases significant levels of free radicals, cytokines and eicosanoids including PGE2 and LTB4 (5). Importantly, the release of PGE2 is necessary but not sufficient for intestinal IR-induced injury (6).

The humoral response includes naturally occurring Ab (NAb) recognition of newly expressed neoantigens and generation of an excessive inflammatory response including complement activation (reviewed in (7)). Multiple groups identified neoantigens by administering mAb to IR-resistant, Ab-deficient Rag-1^−/− mice (8–10). Using this model, several intracellular antigens including DNA, non-muscle myosin (NMM), and annexin IV (Ann IV) have been identified (9, 11–13). In conjunction with anti-phospholipid mAb, Ab to the serum protein, β2-glycoprotein I (β2-GPI) also restored tissue damage in Rag-1^−/−, IR-resistant mice (10). Although multiple neoantigens have been identified, the mechanism of expression of these neoantigens remains unknown.

Recent studies also indicate a significant role for Toll-like receptors (TLRs) in IR-induced tissue damage and inflammation (6, 14). As pathogenic receptors, TLRs recognize distinct components of the microbe, with TLR2 recognizing Gram positive bacterial lipoproteins and lipoteichoic acid while TLR4 recognizes lipopolysaccharide from Gram negative bacteria (15). Although TLRs recognize commensal microflora to maintain intestinal homeostasis (16), these pathogen recognition receptors also induce inflammation after tissue damage (17). Upon activation, most TLRs including TLR2 and TLR4 signal through the common MyD88 pathway. Recently we demonstrated that MyD88 has a critical role in intestinal IR-induced tissue damage (6). As a regulator of complement activation, TLR4 is critical in IR-induced tissue injury, C3 production and the cellular response in the intestine, kidney, brain, lung and heart (6, 18–23). Similarly, TLR2 plays a role in renal, cerebral, and myocardial IR (18, 24, 25). A recent publication indicated that TLR2 is required for the cellular response to intestinal IR (26). However, the role of TLR2 in antibody deposition and complement activation remains unclear.

As both TLR2 and TLR4 use a similar signal transduction pathway through MyD88, we hypothesized that similar to TLR4, TLR2 is critical to initiation of IR-induced pathology. Using TLR2^−/− mice, we demonstrate that TLR2 is required for both the humoral and the cellular response during IR-induced injury. TLR2 plays a role in activation of the cellular infiltrate. Unlike TLR4 or TLR9 deficient mice (27), TLR2^−/− mice also lack the appropriate Ab repertoire to initiate intestinal IR-induced damage or inflammation. In addition, despite the presence of the proteins, TLR2 but not TLR4 is required for neoantigen exposure indicating a dual role for TLR2 in IR-induced injury and inflammation. Thus, although both TLRs are required, TLR2 has a unique role in intestinal IR compared to TLR4.

MATERIALS AND METHODS

Mice

C57Bl/6 (wildtype control), TLR2^−/− and Rag1^−/−mice were obtained from Jackson Labs and bred in the Division of Biology at Kansas State University with food and water access ad libitum. The TLR2^−/− mice were backcrossed to the C57BL/6 background for at least 9 generations and maintained as specific pathogen free (Helicobacter species, mouse hepatitis virus, minute virus of mice, mouse parvovirus, Sendai virus, murine norovirus, Mycoplasma pulmonis, Theiler’s murine encephalomyelitis virus, and endo- and ecto-parasites). Research was conducted in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and was approved by the Institutional Animal Care and Use Committee at Kansas State University.

Ischemia/reperfusion

Animals were subjected to IR as described previously (6). Briefly, ketamine (16mg/kg) and xylazine (80 mg/kg) or isofluorane (2%) anesthetized mice were subjected to a laparotomy 30 min prior to applying a small vascular clamp (Roboz Surgical Instruments) to the superior mesenteric artery. Ischemia was confirmed by blanching of the intestine and absence of pulsations distal to the clamp. Covering the bowel with surgical gauze moistened with warm normal saline prevented intestinal desiccation. After 30 min of ischemia, 2 hr of reperfusion was induced by removing the clamp and confirming the return of pulsatile flow to the superior mesenteric artery. All mice received buprenorphine (0.06mg/kg) for pain. Some experiments reconstituted Rag-1^−/− or TLR2^−/− mice by i.v. injection of 200 μl whole sera or 100 μg of Protein G purified Ab from TLR2^−/− or wildtype (C57Bl/6) mice 30 minutes prior to ischemia. Sham treated animals underwent the same surgical intervention except for vessel occlusion. All procedures were performed with the animals breathing spontaneously and body temperature maintained at 37°C using a water-circulating heating pad. Additional ketamine and xylazine or isofluorane was administered immediately prior to sacrifice. After sacrifice, blood and 2 cm sections of the small intestine 10 cm distal to the gastroduodenal junction were harvested for histological evaluation as well as eicosanoid and cytokine determination.

Histology and Immunohistochemistry

Immediately after removal, mid-jejunal specimens were fixed in 10% buffered formalin phosphate and embedded in paraffin, sectioned transversely (8μm), and H&E stained. The mucosal injury score (SMI) was graded on a six-tiered scale similar to that of Chiu et al (28). Briefly, the average damage score of the intestinal section was determined by the average scores of two blinded observers (trained in evaluating intestinal injury). Each observer graded 75–150 villi on a scale of 0–6. Normal villi were assigned a score of zero; villi with tip distortion were assigned a score of 1; a score of 2 was assigned when Guggenheims’ spaces were present; villi with small regions of disruption of the epithelial cells were assigned a score of 3; a score of 4 was assigned to villi with large regions of exposed but intact lamina propria with epithelial sloughing; a score of 5 was assigned when the lamina propria was exuding; last, villi that displayed hemorrhage or were denuded were assigned a score of 6. Photomicrographs were obtained from H&E stained slides using a 20X, 0.5 Plan Fluor objective on Nikon 80i microscope and images acquired at room temperature using a Nikon DS-5M camera with DS-L2 software.

An additional 2 cm intestinal section was immediately snap-frozen in O.C.T. freezing medium and 8 μm sections were transversely cut and placed on slides for immunohistochemistry. Following acetone fixation, the nonspecific binding was blocked for 30 min by incubating with 10% sera in phosphate buffered saline (PBS). After washing in PBS, the tissues were incubated with Ab for 1 hr at room temperature or ON at 4° C. The C3, IgM, MBL-c, and β2-GPI deposition, and Ann IV and NMM expression on the tissue sections was detected by staining with a purified rat-anti-mouse anti-C3 (Hycult Biotechnologies) or anti- IgM Ab or anti-MBL-c Ab followed by a Texas-red conjugated donkey-anti-rat IgG secondary Ab (Jackson Immunoresearch) or anti-Annexin IV or anti-NMM (AbCam) followed by a Texas-red conjugated donkey-anti-rabbit IgG secondary Ab or anti- β2-GPI Ab (Millipore) followed by Texas-red conjugated donkey-anti-mouse IgG secondary Ab. The detection of β2-GPI was performed similarly with the exception of all solutions being made with 1% BSA in PBS and 2% Rag-1^−/− sera used for blocking with an additional 1 hour of blocking in only 1% BSA in PBS prior to the sera block. Each experiment contained serial sections stained with the appropriate isotype control Ab. All slides were mounted with ProLong Gold (Invitrogen). A blinded observer obtained images at room temperature using a Nikon eclipse 80i microscope equipped with a CoolSnap CF camera (Photometrics) and analyzed using Metavue software (Molecular Devices).

Eicosanoid and Cytokine Determination

The ex vivo generation of eicsanoids in small intestine tissue was determined as described previously (29). Briefly, fresh mid-jejunum sections were minced, washed and resuspended in 37°C oxygenated Tyrode’s buffer (Sigma, St. Louis, MO). After a 20-minute incubation at 37°C, supernatants were collected and supernatants and tissue were stored at −80°C until assayed. The concentration of leukotriene B₄ (LTB₄) and prostaglandins E₂ were determined using an enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI). Cytokine analysis of the same intestinal supernatants was determined using a Milliplex MAP immunoassay kit (Millipore) and read on a Milliplex Analyzer (Millipore). The tissue protein content was determined using the bicinchoninic acid assay (Pierce, Rockford, IL) adapted for use with microtiter plates. Eicosanoid and cytokine production was expressed per mg protein per 20 min.

Real-time PCR

Total RNA was isolated from the jejunum using TRIzol reagent (Invitrogen) according to manufacturer’s instructions. RNA integrity and genomic DNA contamination was assessed using a BioAnalyzer (Agilent) and quantity determined by Nanodrop evaluation. Only samples with no DNA contamination and RIN values greater than 7.0 were used for cDNA synthesis. Total RNA (1 ug) was reverse transcribed using qScript first strand cDNA synthesis kit (Quanta Biosciences) using random primers. Quantitative real-time PCR was performed in 25 ul volumes using a Mini-Opticon real time thermal cycler (Bio-Rad) and Perfecta SYBR Green Fastmix reagent (Quanta Biosciences) using the following protocol: 3m at 95°C; 50 cycles of 10s at 95°C, 20s at Tm (Table 1), 10s at 72°C; melt curve starting at 65°C, increasing 0.5°C every 5s up to 95°C. After amplification, Cox-2 and β2-GPI Ct values were normalized to 18s rRNA and then ΔΔCt fold change relative to Sham-treated wildtype mice was determined as described previously (30). Melt-curve analysis of the PCR products ensured amplification of a single product.

Table 1.

Real Time PCR Primer Sequences

Gene	Tm (°C)a	Sequence
B2-Glycoprotein I	57	FWD: CGGATGACCTACCATTTGCT
		REV: GGGACACATCTCAGGGTGTT
Cox-2	55	FWD: ATTCAACACACTCTATCACTGGC
		REV: TGGTCAAATCCTGTGCTCATACAT
Ribosomal 18sb	58	FWD: CTGGTAATTCATCTCTCTGCTCTG
		REV: GCGACCAAAGGAACCATAAC

^aAnnealing temperature

^bHouse-keeping gene to which genes of interest were normalized

β2-GPI Western Blot Analysis

Immediately after Sham or IR treatment, blood was collected from wildtype and TLR2^−/− mice and allowed to clot on ice at least 30 min to collect whole serum after centrifugation. Sera (3 μL per lane) was loaded in non-reducing SDS buffer and run on a 10% SDS-PAGE gel followed by transfer to a PVDF membrane. The membrane was blocked with a 5% milk in TBS with 0.1% Tween-20 solution for 1 hour followed by overnight incubation with an HRP-goat-anti-human β2-GPI ab (Bethyl Laboratories) diluted in block solution at 1:2500. The Blots were visualized with SuperSignal West Pico ECL Substrate (Thermo Scientific) and developed with an ECOMAX film developer (Protec Medical Devices).

Anti-β2-GPI ELISA

Blood was collected from untreated wildtype and TLR2^−/− mice, allowed to clot on ice at least 30 min and whole serum collected by centrifugation. Polystyrene microtiter plates (NUNC Immuno) were coated overnight with 8 μg/mL mouse β2-GPI (purified as in (31)) in 0.02 M carbonate buffer, pH 9.6. All wells were blocked for 1 hour with 20% FBS in PBS. Serum and standard samples were incubated for 1 hour with shaking, washed with PBS with 0.05% Tween-20 then incubated with HRP conjugated donkey-anti mouse IgG ab (Jackson Immuno) and developed with TMB 1-Component substrate (KPL). Purified anti-mouse β2-GPI monoclonal Ab (FC1) (10) was used to obtain a standard curve.

Antibody Isotype Analysis

Blood was collected from untreated wildtype and TLR2^−/− mice, allowed to clot on ice at least 30 min and whole serum collected by centrifugation. A Milliplex Mouse Immunoglobulin Isotyping Immunoassay (Millipore) was used to determine serum antibody isotypes from each strain of mice, according to manufacturer’s instructions, read on a Milliplex Analyzer (Millipore), and analyzed using Milliplex Analyst software (Millipore).

Statistical Analysis

Data are presented as mean ± SEM and were compared by one-way analysis of variance with post hoc analysis using Newman-Kuels test (Graph Pad/Instat Software Inc. Philadelphia, PA). The difference between groups was considered significant when P < 0.05.

RESULTS

Previous studies indicated that complement and TLR4 interact to mediate inflammation and tissue damage in response to mesenteric IR (6, 14). In addition, intestinal damage required the adapter protein, MyD88 (6). Therefore, we hypothesized that mesenteric IR-induced damage would be attenuated in TLR2^−/− mice. We tested the hypothesis by subjecting wildtype and TLR2^−/− mice to intestinal Sham or IR treatment. As Sham treatment was not significantly different between the strains of mice, the scores were pooled (Fig. 1A, B, D). As expected, after IR, wildtype C57Bl/6 sustained significant tissue injury (Fig 1A, C). In contrast, TLR2^−/− mice sustained minimal intestinal IR-induced damage which was not significantly different from Sham treatment (Fig. 1A, D, E).

Figure 1. IR-induced intestinal damage requires TLR2 expression.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment and mid-jejunal sections H&E stained and A) scored for mucosal damage (75–150 villi per section). B–E) Representative microphotographs of H&E stained sections. * = p ≤ 0.05 compared to Sham treatment, Φ = p ≤ 0.05 compared to wildtype IR treatment. Each bar is representative of 7–10 animals per group.

Previous studies indicated that TLR4 and MyD88 activation results in Cox-2-mediated PGE₂ production which is required for IR-induced tissue damage (6, 14). As TLR2 also signals through MyD88, we examined the intestinal eicosanoid production in response to IR in TLR2^−/− mice. Similar to intestinal damage, Cox-2 transcript was elevated in response to IR in wildtype mice but not in TLR2^−/− mice (Fig. 2A). Correlating with increased Cox-2 expression, the intestines of wildtype but not TLR2^−/− mice produced significant quantities of PGE₂ (Fig 2B). As demonstrated previously, intestinal IR also increases the chemotactic eicosanoid, LTB₄ in wildtype mice (32–35). Similar to injury, IR did not induce LTB4 in TLR2−/− mice (Fig. 2B). These data suggest that TLR2 is required for the IR-induced eicosanoid response. Similar to eicosanoid production and previous studies with TLR4^−/− and MyD88^−/− mice (14), intestinal production of CXCL1 (KC; keratinocyte-derived chemokine) increased in response to IR in wildtype but not TLR2^−/− mice (Fig. 2B). Despite a lack of CXCL1 and LTB4 production, immunohistochemistry demonstrated an IR-induced macrophage infiltration in TLR2^−/− mice similar to that found in intestines of wildtype mice after IR (Data not shown). However, the TLR2^−/− inflammatory infiltrate was not activated to produce cytokines IL-6, IL-12p40 and TNFα whereas these cytokines were significantly elevated in intestines from wildtype mice (Fig. 2C). Together, these studies suggest that similar to TLR4, TLR2 is critical to the cellular innate immune response to intestinal IR.

Figure 2. Eicosanoid and cytokine production is significantly decreased in TLR2^−/− mice.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment. A) Mid-jejunal cox-2 transcription was determined by qRT-PCR analysis. Intestinal sections were analyzed ex vivo for B) PGE₂ and LTB₄ and C) IL-6, IL-12p40, TNF-α, and CXCL1 (KC) production. * = p ≤ 0.05 compared to Sham treatment, Φ = p ≤ 0.05 compared to wildtype IR treatment. Each bar is representative of 4–10 animals per group.

Tissue damage in response to intestinal IR also requires complement activation and Ab recognition of ischemia-induced neoantigens (36). TLR4 deficient mice produce the appropriate Ab but decrease complement C3 production and deposition as a method of regulating the inflammatory response (14). To determine if TLR2^−/− mice regulate tissue damage and inflammation in a similar manner, we initially found no change in complement C3 transcription in response to IR between wildtype and TLR2^−/− mice (data not shown). We examined complement initiation on the IR treated intestinal tissue by examining intestinal deposition of IgM and complement components, MBL-C and C3, by immunohistochemistry. As indicated in Figure 3, IgM, C3 and MBL-C were not deposited in response to Sham treatment. However, significant deposits of IgM and both MBL-C and C3 were present in wildtype mice after mesenteric IR (Fig. 3). In contrast, little to no IgM, C3 or MBL-C were deposited after similar treatment of the TLR2^−/− mice (Fig. 3). Although TLR2 does not regulate complement production, these data suggest that the absence of TLR2 significantly decreases initiation of complement activation compared to wildtype mice.

Figure 3. Complement components and IgM are not deposited in response to IR in TLR2^−/− mice.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment. Intestinal sections were stained for A) IgM B) C3 and C) MBL-C by immunohistochemistry. Microphotographs (200x) are representative of 3–4 animals stained in at least 3 independent experiments.

Previous studies indicated that the appropriate Ab repertoire is required for recognition of IR-induced neoantigens, complement activation, and eicosanoid production (14, 37). Importantly, TLR4 was not required for production of the appropriate Ab repertoire to induce damage in Ab-deficient, IR-resistant, Rag-1−/− mice (8, 14, 37) . To determine if TLR2^−/− mice have the proper Ab repertoire, we injected Rag-1^−/− mice with sera or Ig purified from C57Bl/6 or TLR2^−/− mice and subjected the mice to IR. As indicated in Figure 4A, after IR, Rag-1^−/− mice administered sera or Ab obtained from C57Bl/6 mice sustained significant intestinal damage with no difference between administering sera or purified Ab. In contrast, no significant increase in IR-induced intestinal damage was observed when Rag-1^−/− mice were injected with sera or Ab from TLR2^−/− mice (Fig. 4A). In addition, Rag-1^−/− mice reconstituted with Ab from C57Bl/6 but not TLR2^−/− mice secreted significant PGE₂ (Fig. 4B) and LTB₄ (Fig. 4C) in response to IR. Importantly, eicosanoid production by Rag-1^−/− mice treated with Ab from TLR2^−/− mice was similar to un-reconstituted Rag-1^−/− mice after Sham or IR treatment (Fig. 4B, C). As there was no significant difference between treatments with either sera or purified Ab, subsequent studies were performed with purified Ab.

Figure 4. TLR2^−/− mice lack the pathogenic Ab required for IR-induced intestinal damage and eicosanoid production.

Rag1^−/− mice were subjected to Sham or IR treatment with or without administration of antibodies from either wildtype or TLR2^−/− mice. A) Mid-jejunal sections were H&E stained and scored for mucosal damage. B) Intestinal sections were analyzed ex vivo for B) PGE₂ and C) LTB₄. * = p ≤ 0.05 compared to Rag1^−/− IR treatment, Φ = p ≤ 0.05 compared to Rag1^−/− IR + C57BL/6 ab treatment. Each bar is representative of 5–10 animals per group.

Previous studies suggested that antibodies produced by TLR2^−/− mice produce different isotypes. The plasma from wildtype and TLR2^−/− mice contained lower quantities of total IgM Ab isotype (Fig. 5A). Isotypes IgG3 and IgG2b were significantly lower in the TLR2^−/− plasma as well (Fig. 5A). To examine the damaging antibodies, we evaluated the anti-β2-GPI IgG and IgM. Sera pooled from wildtype mice contained 49.3 ± 8.1 ng/ml anti-β2-GPI IgG, whereas sera obtained from TLR2^−/− mice contained only 22.1 ± 8.5 ng/ml of anti- β2-GPI IgG (Fig. 5B). Sera from Rag-1^−/− mice contained no detectable Ab and was used as a negative control. Importantly, the TLR2^−/− mice also had significantly less anti-β2-GPI IgM compared to the wildtype C57Bl/6 mice (Fig. 5B). Together these data suggest that the TLR2^−/− Ab repertoire is not sufficient to induce damage in response to IR.

Figure 5. The Ab repertoire is altered in TLR2^−/− mice.

Whole plasma or sera from wildtype and TLR2^−/− mice was analyzed by ELISA for A) total antibody isotypes and B) anti-β2-GPI specific IgG and IgM. Φ = p ≤ 0.05 compared to wildtype plasma or sera. Each bar is representative of 3–5 different lots of plasma or sera.

To determine if providing the appropriate Ab was sufficient to restore injury to the TLR2^−/− mice, we injected Ab purified from wildtype mice into TLR2^−/− mice. Similar to previous results, wildtype mice but not TLR2^−/− mice sustained injury in response to IR. Surprisingly, administration of wildtype Ab was not sufficient to restore IR-induced injury in TLR2^−/− mice (Fig. 6A). Similarly, PGE₂ and LTB₄ were not elevated in response to IR when wildtype Abs were injected into TLR2^−/− mice prior to IR (Fig. 6B, C). This lack of response may be due to a lack of neoantigen expression and/or may be due to TLR2-induced eicosanoids, cytokines or free radicals from the inflammatory cell response.

Figure 6. Pathogenic Ab are not sufficient to restore IR-induced injury in TLR2^−/− mice.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment with or without administration of antibodies from wildtype mice. A) Mid-jejunal sections were H&E stained and scored for mucosal damage. B) Intestinal sections were analyzed ex vivo for B) PGE₂ and C) LTB₄. * = p ≤ 0.05 compared to Sham treatment, Φ = p ≤ 0.05 compared to wildtype IR treatment. Each bar is representative of 4–10 animals per group.

As antigens stimulate Ab production, we examined the ability of TLR2^−/− mice to locally express β2-GPI transcript within the intestine and systemic β2-GPI protein expression. In response to IR, local intestinal β2-GPI mRNA was not increased in wildtype mice but was significantly elevated in TLR2^−/− mice with a 6 fold increase compared to wildtype Sham levels (Fig. 7A). Under normal conditions, β2-GPI protein is present in C57Bl/6 and TLR2^−/− sera at similar levels (Fig 7B). In response to IR in C57Bl/6 mice, β2-GPI sera levels do not change; however, IR induces a significant release of β2-GPI into the sera of TLR2^−/− mice. Thus, despite the presence of β2-GPI, production of the injurious Ab repertoire and in particular, anti-β2-GPI Ab appears to require TLR2.

Figure 7. TLR2 deficiency increases IR-induced β2-GPI transcription and serum levels.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment. A) Mid-jejunal β2-GPI transcription was determined by qRT-PCR analysis. * = p ≤ 0.05 compared to Sham treatment. Each bar is representative of 4–6 animals per group. B) β2-GPI serum protein production was determined by western blot analysis after Sham or IR treatment. C) A representative blot is shown. Graph and blot are representative of 3 total blots with 3 different sets of serum.

Finally, we examined intestinal deposition and surface expression of three of the known IR-induced neoantigens, β2-GPI, NMM and AnnIV. Intestinal sections from Sham treated mice of either strain showed minimal staining for each of the neoantigens (Fig. 8). In contrast, intestinal sections from IR-treated, wildtype or TLR4^def mice expressed all three neoantigens while the IR-treated, TLR2^−/− intestines expressed little if any of the three neoantigens on the cell surface (Fig. 8). When quantitated using Image J, expression on TLR2^−/− intestine was significantly lower than that of the wildtype or TLR4^def intestines for each neoantigen examined (Fig. 8). The staining of TLR4^def intestines was not significantly different when compared to wildtype intestines. These data suggest that despite significant levels of β2-GPI present in the serum, intestinal deposition of β2-GPI and surface expression of neo-antigens, NMM and AnnIV requires TLR2.

Figure 8. IR does not induce neoantigen expression in TLR2^−/− mice.

Wildtype and TLR2^−/− mice were subjected to Sham or IR treatment. Intestinal sections were stained for A) β2-GPI B) Non-Muscle Myosin and C) Annexin IV by immunohistochemistry and relative fluorescence measured by ImageJ analysis. Microphotographs are representative of 3–4 animals stained in at least 3 independent experiments. * = p ≤ 0.05 compared to Sham treatment, Φ = p ≤ 0.05 compared to wildtype IR treatment. Each bar is representative of 3–4 animals per group.

DISCUSSION

Together, these data suggest that the absence of TLR2 compromises the IR-induced cellular and humoral innate responses by significantly decreasing the release of pro-inflammatory cytokines and eicosanoids, decreasing deposition of complement initiators, attenuating neoantigen surface expression and altering the Ab repertoire. We propose the following model which will require testing in the future. The process of IR requires β2-GPI and anti-β2-GPI Ab to induce endothelial damage and inflammation (10, 31, 38). As expected TLR2 plays a role in the inflammatory process, however, we demonstrate that TLR2 is critical to multiple components of intestinal IR-induced injury and inflammation. Specifically, TLR2 is essential for i) production of the pathogenic naturally occurring Ab, ii)inflammatory cytokine production and PGE2 production and iii) expression of neoantigens which are recognized by the naturally occurring Ab. Then in a TLR2 independent manner, Ab bind to the cell surface (iv) and activate complement (v). Thus, TLR2 plays a critical role in multiple inflammatory processes which occur during IR-induced tissue damage.

Our data demonstrate that TLR2 is required for IR-induced intestinal damage using a time course of 30 min ischemia and 2 hr reperfusion. Using a 45 min ischemic period followed by 60 min reperfusion, a recent study also demonstrated that the neutrophilic infitration, Cox-2 and TNF mediated at least a portion of IR-induced intestinal damage in a TLR2 dependent manner (26). These studies differ from those performed on young (4 wk) mice in which wildtype mice sustained minimal intestinal injury (0.67 injury score on a 0–4 scale) and TLR2^−/− mice exhibited increased injury (39). The time course in the previous study (39) was 60 mins of ischemia followed by 90 min reperfusion, which others have indicated is a non-recoverable surgery (40). The discrepancy in the results may be due to animal age, time course, or technical differences of using a bulldog clamp with unknown pressure vs a vascular clamp or may be explained by incomplete clamping which would correlate with the low injury score reported for wildtype mice. In contrast, our data are consistent with IR experiments performed in heart (41, 42) and kidney (25, 43, 44). Multiple studies have demonstrated that the absence or blockade of TLR2 protects against IR-induced myocardial damage and inflammation (41, 42). Similarly, TLR2 is critical to IR-induced renal damage which is mediated by both MyD88 dependent and independent mechanisms (25, 43, 44). In contrast to injury, the data presented here agree with the previous intestinal IR-induced study that demonstrates significantly reduced production of inflammatory cytokines in TLR2^−/− mice compared to wildtype mice (39). Similar results in renal and myocardial IR indicate that multiple cell types require TLR2 for the inflammatory response and tissue damage (42, 44).

Previous studies demonstrated that complement is critical to intestinal, renal and myocardial IR-induced damage although renal IR-induced injury appears to utilize the alternative pathway while intestinal and myocardial IR-induced damage is dependent on the classical and MBL pathways of complement activation (45). The current study demonstrates that in response to IR, complement deposition within intestines of mice also requires TLR2 expression. To our knowledge, this is the first study in intestinal IR that examined the interactions of TLR2 and complement. In renal IR, the alternative complement component, factor B was shown to interact with TLR2 (46). The current study did not examine the specific pathways involved but IgM, C3 and MBL depositions correlated with injury suggesting the classical and MBL pathways were activated. As Ab is required for IR-induced injury in the intestine (8) and heart (47) but not kidney (48), it appears that different mechanisms of complement activation are active in reperfusion injury of different organs.

Previous studies indicate that TLR4 and TLR9 deficient mice contain the appropriate Ab to induce IR injury in Rag-1^−/− mice (6, 27). In contrast, TLR2^−/− mice do not contain the necessary pathogenic Ab for IR-induced tissue damage, despite a similar number of total Ab. A requirement for TLR stimulation to produce auto-reactive antibodies has been demonstrated in autoimmune disease. In the AM14 mouse models of SLE, immune complexes ligate both IgM and TLR7 or TLR9 to induce autoreactive B cells to produce high quantities of Ab (49, 50). Additional studies in the MRL/lpr mouse model of SLE demonstrated that kidney mesangial cells require immune complexes containing HMGB1 and TLR2 for increased cytokine production and human blood cells require TLR2 for increased autoreative Ab production (51, 52).

The specific subclass of Ab produced also varies in the absences of TLR2. Similar to the results of others (53), our data demonstrate that TLR2^−/− mice contain fewer IgG antibodies compared to IgG levels in wildtype mice. Lartigue et al showed that in a mouse model of Systemic lupus erythematorsus, both TLR2 and TLR4 deficient mice had fewer IgG and these results correlated with less severe disease (53). In addition, the overall natural IgM levels were not significantly different (53). Importantly, the absence of either TLR in the SLE model resulted in decreased self-reactive autoantibodies from multiple subclasses of IgG (53). While they did not examine IgM, our data indicate that although the overall IgM concentration was not significantly different, the anti-β2-GPI specific IgM antibodies decreased in the absence of TLR2. It is possible that other endogenous antigens induce antibodies in a TLR2 dependent manner as well.

Our data also correlate with clinical studies showing induction of significantly more inflammatory cytokines following myocardial infarction in patients who have increased levels of autoreactive Ab (54). In vitro studies provided more cell type specific data demonstrating that Ab induce human monocyte derived macrophages to secrete inflammatory cytokines in a TLR2-dependent manner (54). In addition, pathogenic anti-phospholipid Ab stimulate an inflammatory response by multiple cell types including monocytes, fibroblasts and endothelial cells (55, 56). Importantly, blocking of TLR2 inhibits the Ab-induced cytokine release (54).

Within the last 15 years, multiple IR-induced neoantigens have been identified. Based on Ab recognition and peptide inhibition, at least three neoantigens appear to be required for intestinal IR-induced damage (9, 10, 13, 38, 57). However, the specific role of each neoantigen is unclear at this time. Our data also indicate that despite increased serum β2-GPI, TLR2^−/− mice contain significantly less anti-β2-GPI Ab and decreased neoantigen is expressed in response to IR. This may be accounted for by multiple different hypotheses. As both TLRs have been proposed to be a receptor for β2-GPI, it is possible that TLR2 is a primary receptor and β2-GPI is required for other neoantigen expression. Another possibility is that crosstalk between TLR2 and TLR4 results in the differential neoantigen expression. In an LPS stimulation of endothelial cells, stimulation of TLR4 increases activation of TLR2 (58). As similar process may occur in IR and subsequently enhance neoantigen expression. Similarly, it is possible that TLR2 and TLR4 are critical on different tissues and processes and similar to myocardial IR crosstalk occurs between multiple signaling pathways (59). Finally, it is possible that TLR2 is required for recognition of other endogenous antigens which also results in complement activation. Among others, these endogenous antigens may include high mobility group protein B1 (HMGB1) (60) or the gut microbiome as antibiotic treatment down regulates TLR2 and TLR4 resulting in attenuated IR-induced injury (61). Future studies will be required to determine if TLR2 is directly or indirectly critical for β2-GPI deposition and exposure of other neoantigens and the specific interactions between TLR2 and TLR4 during intestinal IR.