Chronic Porphyromonas gingivalis lipopolysaccharide induces adverse myocardial infarction wound healing through activation of CD8⁺ T cells

Abstract

Oral and gum health have long been associated with incidence and outcomes of cardiovascular disease. Periodontal disease increases myocardial infarction (MI) mortality by sevenfold through mechanisms that are not fully understood. The goal of this study was to evaluate whether lipopolysaccharide (LPS) from a periodontal pathogen accelerates inflammation after MI through memory T-cell activation. We compared four groups [no MI, chronic LPS, day 1 after MI, and day 1 after MI with chronic LPS (LPS + MI); n = 68 mice] using the mouse heart attack research tool 1.0 database and tissue bank coupled with new analyses and experiments. LPS + MI increased total CD8⁺ T cells in the left ventricle versus the other groups (P < 0.05 vs. all). Memory CD8⁺ T cells (CD44 ⁺ CD27⁺) were 10-fold greater in LPS + MI than in MI alone (P = 0.02). Interleukin (IL)-4 stimulated splenic CD8⁺ T cells away from an effector phenotype and toward a memory phenotype, inducing secretion of factors associated with the Wnt/β-catenin signaling that promoted monocyte migration and decreased viability. To dissect the effect of CD8⁺ T cells after MI, we administered a major histocompatibility complex-I-blocking antibody starting 7 days before MI, which prevented effector CD8⁺ T-cell activation without affecting the memory response. The reduction in effector cells diminished infarct wall thinning but had no effect on macrophage numbers or MertK expression. LPS + MI + IgG attenuated macrophages within the infarct without effecting CD8⁺ T cells, suggesting these two processes were independent. Overall, our data indicate that effector and memory CD8⁺ T cells at post-MI day 1 are amplified by chronic LPS to potentially promote infarct wall thinning.

NEW & NOTEWORTHY Although there is a well-documented link between periodontal disease and heart health, the mechanisms are unclear. Our study indicates that in response to circulating periodontal endotoxins, memory CD8⁺ T cells are activated, resulting in an acceleration of macrophage-mediated inflammation after MI. Blocking activation of effector CD8⁺ T cells had no effect on the macrophage numbers or wall thinning at post-MI day 1, indicating that this response was likely due in part to memory CD8⁺ T cells.

INTRODUCTION

The prevalence of heart failure in adults in the United States continues to increase to current rates of over 7.5 million cases (1). Incidence of heart failure among patients hospitalized for a myocardial infarction (MI) ranges between 16% and 39% (2–7). Oral health measurements (number of decayed, missing, or filled teeth, mean probing depth, oral hygiene status, and percentage of bleeding sites) have been associated with MI incidence and outcomes. Regression analysis has revealed that periodontal disease increases MI mortality by sevenfold (8–12). Although correlations between oral health and heart disease have been well documented, the mechanisms are not fully understood.

One possible link between oral and cardiovascular health is the fivefold-higher bacterial endotoxin levels detected in the plasma of periodontal patients compared with healthy controls (13). Holmlund et al. (14) determined a relationship between MI and Porphyromonas gingivalis, a bacterium frequently associated with periodontal disease. Antibodies targeting P. gingivalis were higher in subjects with MI, as well as in those with >4 deepened pockets, >20% bleed on probing, and periodontal bone loss. P. gingivalis strain ATCC33277 is commonly identified in over 25% of periodontal disease patients (15, 16). Administration of the endotoxin P. gingivalis lipopolysaccharide (LPS) at subseptic levels induced systemic inflammation similar to conditions seen in periodontal disease patients and was linked to adverse remodeling after MI (17–19).

Previous studies from our group and others have shown that periodontal disease-associated pathogens stimulate an imbalance in MI-induced inflammation and extracellular matrix (ECM) deposition in the left ventricle (LV) (17–20). Chronic LPS increased the expression of proinflammatory cytokines both in circulation and in the infarct, which resulted in accelerated recruitment of macrophages into the infarct starting at post-MI day 1 and inhibition of fibroblast activation and scar formation at day 7 (18, 19). Because macrophages respond to stimulatory signals, including those originating from T cells (21–23), we hypothesized that chronic LPS primes memory T cells to accelerate macrophage activation and contributes to poor cardiac remodeling after MI. Using retrospective (mouse heart attack research tool; mHART), cross-sectional (LPS + MI + MHCi group), and prospective (isolated T cells and macrophages) data, we determined the effect of chronic inflammation on T cells in the heart.

EXPERIMENTAL SECTION

Animal Assurances for Prospective and Cross-Sectional Experiments

All animal procedures were approved by the Institutional Animal Care and Use Committee at the Medical University of South Carolina and the Ralph H. Johnson VA Medical Center in accordance with the Guide for the Care and Use of Laboratory Animals and followed the ARRIVE guidelines (24). Mice (C57BL/6J, 3 and 7 mo old, both males and females) were kept in a light-controlled environment with a 12-h:12-h light/dark cycle and given free access to standard mice chow and water. Echocardiography and MI surgery was performed, as recommended by the guidelines (25, 26).

Analysis of the mHART 1.0 Database and Tissue Bank (Retrospective)

The mHART database consists of data from MI projects using mice that were collected since 2007 (12). Here, we analyzed the data set and tissue bank to investigate T cell-driven mechanisms that could explain periodontal disease effects on MI outcomes (experimental design in Supplemental Fig. S1A; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.13714114). Data in the mHART database were selected based on the following criteria: mouse strain C57BL/6J, age ranging from 3 to 7 mo, males and females, including mice exposed to the periodontal pathogen P. gingivalis LPS chronically (28 days before MI), available echocardiography and plasma data, and existence of paraffin-embedded tissue within the mHART tissue bank. Mice from the periodontal project had been continuously exposed to LPS ATCC 33277 (0.8 μg/g body wt/day; Invivo Gen) by osmotic minipumps (model 2004, 1003 D, and 1007 D; Durect), as previously described (17–19). We chose to use 28 days of LPS as our exposure time due to previous studies where we determined that after 28 days of LPS exposure, there were physiological changes in the LV (17). This was not observed after an acute 1 day of LPS exposure. We defined chronic exposure as 28 days, which is the human equivalent of 3.1 yr of exposure (27).

Sample Selection

The database was accessed on December 20, 2018. Of 2,095 mice evaluated, 68 matched study criteria and tissue from these mice were available to be included in the current analyses. Details on the animals within mHART were previously reported (28). All mice evaluated in this study were enrolled between 2011 and 2014. Both MI surgeries and echocardiography were performed according to established guidelines (25, 26). One investigator (KYDP) collected all of the data and tissue obtained from mHART and oversaw the newly obtained analyses. Echocardiography data collected from mHART encompassed 75% of the data presented in this study, with 60% of histological staining and 100% of the plasma data collected from mHART (Supplemental Fig. S1B). All of the immunofluorescence staining was newly performed using the mHART tissue databank for the MI and LPS + MI groups. New tissue samples were collected for the LPS + MI + MHCi and the LPS + MI + IgG groups.

Coronary Artery Ligation

Coronary artery ligation was performed, as described previously (18, 19, 25, 29–31). Mice were anesthetized with 1%–2% isoflurane in oxygen, intubated, and put on a standard rodent ventilator. The mouse was placed on the surgical table (Rodent Surgical Monitor+) in a supine position with their paws gelled and placed on the mouse designated contact electrodes for electrocardiogram (EKG) reading. Before the surgery, buprenorphine (0.05 mg/kg, Ralph H Johnson VAMC Pharmacy) was administered. An 8-0 suture was used to ligate the left coronary artery, and MI was confirmed by blanching of the LV in addition to visualized changes in the EKG. The chest and ribcage were closed via a 6-0 suture and the animal was given oxygen via nose cone until it was able to get up and move on its own. Mice were then placed in a clean cage on a heating board for 4 h or until they were fully recovered. At the time of tissue collection, echocardiography was performed as described previously and as recommended by the APS guidelines (26, 30). For tissue collection, mice were anesthetized with 1%–2% isoflurane in an oxygen mix. Hearts were removed, and the LV and right ventricle were separated and weighed individually. The coronary vasculature was flushed with cardioplegic solution, consisting of 69 mM NaCl, 12 mM NaHCO₃, 11 mM glucose, 30 mM 2,3-butanedione monoxime, 10 mM EGTA, 0.001 mM nifedipine, 50 mM KCl, and 100 U heparin in 0.9% saline (pH 7.4). The LV was sliced into apex, middle, and base sections, stained with 1% 2,3,5-triphenyltetrazolium chloride (TTC, Sigma), and photographed for evaluation of infarct area. Photoshop (Adobe) was used to quantify percentage of infarct area to total LV area.

Echocardiography

Left ventricular physiology was determined by echocardiography (Vevo 2100, VisualSonics; Toronto, CA) following the recommendation in the “Guidelines for Measuring Cardiac Physiology in Mice” (8, 32, 33). Mice were anesthetized with 1.0%–1.5% isoflurane in 100% oxygen. EKG, heart rate, and body temperature were monitored during the imaging procedure. All images were acquired at heart rates >400 beats/min for physiologically relevant measurements. Measurements were taken from the LV parasternal long-axis (B-mode) and short-axis (M-mode) views. Wall thickness was determined based on M-mode images. For each echocardiographic variable, three images from consecutive cardiac cycles were measured and averaged.

Immunofluorescence

Immunofluorescent staining was performed to evaluate T-cell phenotypes in the LV after MI (n ≥ 4/group). To expose antigen epitopes, heat-mediated antigen retrieval (Target Retrieval Solution, Dako) was performed. Slides were then rinsed, blocking serum was applied (Goat Serum), and slides were incubated for 30 min. Primary antibody for CD3 (Novus NBP2-12159; 1:50) was added and incubated overnight. The following day, slides were washed and incubated with fluorescent secondary antibody (Alexa Fluor 546 goat anti-rat; Invitrogen A11081; 1:400) for 30 min at room temperature. In between primary incubations, slides were blocked with blocking serum for 1 h to inhibit unspecific binding. Then, slides were incubated with primary antibody for CD8 (Cedarlane CLANT280; 1:100) overnight and the next day, fluorescent secondary antibody corresponding to CD8 (Alexa Fluor 633 goat anti-rat; Invitrogen A21094; 1:400) was added and incubated for 30 min at room temperature. Immediately, slides were washed and incubated overnight with primary antibodies for CD27 (Invitrogen MA5-29671; 1:75) and CD44 (Invitrogen MA5-17875; 1:50). The last set of fluorescent secondary antibodies corresponding to CD27 (Alexa Fluor 488 goat anti-rabbit; Invitrogen A11034; 1:400) and CD44 (Alexa Fluor 594 goat anti-rat; Invitrogen A11007; 1:400) were added the following day for 30 min. After the final step, slides were treated with 1X TrueBlack (Biotium) to minimize autofluorescence. Vectashield Mounting Medium for Fluorescence with DAPI (Vector Laboratories H-1200) was added to coverslips, which were placed on slides and sealed with clear nail polish. Fluorescence minus one (FMO) controls were used to determine the exposure time and adjust spillover in a given channel. Images were separated into noninfarcted and infarct tissue. To quantify staining, three to five random images within the infarct region from each slide were captured with a 20× air objective, NA 0.7 on the Leica TCS SP5 confocal microscope (Supplemental Fig. S2). Quantification was performed using the colocalization tool available from Image Pro 10 (Media Cybernetics). CD3 was selected as the parent, and percent colocalization was determined using the other T-cell markers. Data were normalized to the infarct area (bright-field image) to determine percent area of T-cell subtypes. Representative images were taken by ×2 digital zoom of the acquired images. To calculate the CD3⁺CD8⁺CD44⁺CD27⁻ population, we subtracted the percent population that was CD3⁺CD8⁺CD44⁺CD27⁺ from the CD3⁺CD8⁺CD44⁺ population.

Analysis of MertK on macrophages was performed by incubating hearts with primary antibody targeting Mac3 (Cedarlane CL8943AP; 1:100) overnight followed by Alexa Fluor 546 secondary antibody (Invitrogen; 1:400). Following secondary antibody incubation, MertK primary antibody (Invitrogen 14–5751-82; 1:100) was added and slides were incubated overnight. The next morning, slides were washed with PBS and secondary AlexaFluor 488 (Invitrogen; 1:400) was added. Five images were acquired from the infarct region with a 40× air objective, NA 0.7 on either the Leica TCS SP5 or Nikon A1Rsi laser scanning confocal microscope (Supplemental Fig. S3). Quantification was performed using the colocalization tool available from Image Pro 10 (Media Cybernetics). Percent area was calculated using the number of Mac3 or MertK positive cells normalized to the infarct area (bright-field image). Colocalization was used to determine percentage of double-positive Mac3 and MertK staining and normalized to the bright-field image to calculate percent infarct area. Representative images were taken by ×2 digital zoom of the acquired images.

Immunohistochemistry

For macrophage (Mac3) and neutrophil (Ly-6B.2; Cedarlane CL8993AP; 1:100) immunohistochemistry, 64% of the LPS + MI group was obtained from already acquired immunohistochemistry results in the mHART database (19, 32). The rest were newly stained, as described previously (10). As an internal control to confirm interbatch reproducibility, Mac3 was requantified in five samples from the mHART tissue bank and analyzed to show no difference between staining batches (P = 0.62; Supplemental Fig. S4A). The LV middle section was fixed in 10% zinc formalin (Thermo Fisher Scientific), paraffin embedded, and sectioned for histological examination, as described previously (30). To expose antigen epitopes, heat-mediated antigen retrieval (Target Retrieval Solution, Dako) was performed. Slides were then rinsed, blocking serum was applied (Goat Serum), and slides were incubated for 30 min. Primary antibody for either Mac3 or neutrophils was added and incubated overnight. The following day, slides were washed and incubated with anti-rat secondary antibody (Vector Laboratories, Cat. No. PK-6104) for 1 h at room temperature. Slides were stained with the HistoMark BLACK Peroxidase Substrate (Cat. No. KPL 54–75-00) followed by a counterstain with eosin Y (Sigma; Cat. No. H110232). Picrosirius red (PSR) was used to stain for collagen. Slides were incubated in a solution of 0.2% phosphomolybdic acid (Electron Microscopy Sciences; Cat. No. RT 26357-01) for 3 min. The slides were then rinsed and transferred to the solution containing 0.1% Sirius Red in saturated picric acid (Electron Microscopy Sciences; Cat. No. RT 26357-02) for 90 min followed by acidified water for 2 min. For all immunohistochemistry data within mHART and newly acquired samples, ten 40× magnification images were captured from within the infarct region and analyzed. Analysis was performed using Image Pro 10 (Media Cybernetics) to determine the average percent positive per field of view per mouse LV infarct zone.

Splenic CD8⁺ T-Cell Stimulation (Prospective)

CD8⁺ T cells were isolated from spleens of unoperated mice (day 0; no LPS; n = 5/sex; 3.7 ± 0.1 mo of age for secretomics and n = 6 females; 3.7 ± 0.1 mo of age for flow analysis) by magnetic bead isolation (Miltenyi Biotec 130-104-075). Spleens were disaggregated with tweezers. The cell suspension was applied over preseparation filters (30 μm, Miltenyi Biotec 130-041-407) and incubated with red blood cell lysis solution (Miltenyi Biotec 130-094-183) to remove red blood cells. The cell suspension was then applied over a magnetic MS column (Miltenyi Biotec 130-042-201) to collect CD8⁺ T cells (Miltenyi Biotec 130-104-075). Purified cells were plated in a six-well plate (1 × 10⁶ cells/well) and stimulated with interleukin (IL)-4 (1.5 ng/mL; R&D No. 404-ML/CF) in RPMI supplemented with 0.1% fetal bovine serum (FBS) and 1X antibiotics/antimycotic solution for 24 h. Additional groups of CD8⁺ T cells cultured in RPMI with either 0.1% FBS or 10% FBS for 24 h served as controls.

Flow Cytometry

After CD8⁺ T-cell stimulation, cells were centrifuged and the pellet was collected. Cells were then assessed for viability (Supplemental Fig. S4B) and markers of T-cell activation. Cell concentrations were adjusted to 0.5 × 10⁶ cells/100 uL with FcR Blocking reagent (Miltenyi Biotec) and incubated for 20 min at 4°C. Cells were then incubated for 30 min with Live/Dead Fix Yellow (Invitrogen) at 4°C to determine viable population. Primary antibodies (PE Anti-CD3e Miltenyi Biotec 130-120-076 1:50; APC-Vio770 anti-CD8a Miltenyi Biotec 130-120-737 1:50; PE-Vio770 anti-CD27 Miltenyi Biotec 130-113-641 1:50, VioBright FITC anti-CD44 Miltenyi Biotec 130-120-287 1:50, or mix) were added and cells were incubated for 20 min at 4°C in the dark. Cells were washed in PEB (phosphate-buffered saline pH 7.2, 0.5% bovine serum albumin, and 2 mM ethylenediaminetetraacetic acid) centrifuged at 300 g for 5 min, filtered with 5-mL tube with cell-strainer cap, and analyzed on the MACS Quant Analyzer (Miltenyi Biotec). FlowLogic software (Inivai Technologies) was used to analyze flow data.

Secretomics

After stimulation, T cells were centrifuged and conditioned media was collected and snap frozen in liquid nitrogen for proteomic analysis. The conditioned media was reduced, alkylated, and trypsin digested into peptides. The released peptides were prepared using Sep-Pak Vac C18 cartridge (Waters, Milford, MA) and analyzed by liquid chromatography-tandem mass spectrometry using a Q Exactive (Thermo Fisher Scientific, Waltham, MA) together with a 15 cm × 75 μm C18 column (5-μm particles with 100 Å pore size). A target value of 1e6 ions at a resolution setting of 70,000 was set and in MS2 1e5 ions at a resolution setting of 17,500. To exclude low-probability protein identifications, the false discovery rate (FDR) was set at 0.01. SEQUEST was searched for MS/MS spectra using Proteome Discoverer (version 2.1; Thermo Fisher Scientific) against the mouse RefSeq database (November 3, 2013 version) and SwissProt database (accessed January 18, 2017) containing 53,918 or 24,967 sequences, respectively. Samples were run in duplicate and results combined for analysis. Spectral counting was used for peptide quantification and total spectral counts for each sample were used for normalization. A total of 256 unique proteins were identified. Proteins that were expressed in at least 5 samples/group were considered for analysis. Missing values were replaced with 0.001 to calculate the ratio and P value. The mass spectrometry proteomic data were deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD011757 (http://www.ebi.ac.uk/pride/archive/).

Western Blot Analysis

Conditioned media was concentrated (50% volume decrease) by Speedvac centrifugation. Once concentrated, samples were volume loaded (10 μL volume, with 5-μL loading dye) and separated on 4%–12% Criterion XT Bis-Tris gels (Bio-Rad), transferred to a nitrocellulose membrane (Bio-Rad). After blocking with 5% nonfat milk (Bio-Rad), the membrane was probed with anti-amyloid-βA4 precursor protein (Novus NBP2-15575; Supplemental Fig. S4D) at 1:1,000 followed by secondary antibody (Abcam, Ab205718, 1:5,000) and detected with SuperSignal West Femto (Thermo Fisher Scientific). The blots were examined using the ImageQuant analysis software on the luminescent image analyzer.

Bioinformatics

Pathway enrichment of the plasma and secretomic data were performed using Reactome (https://reactome.org/download-data). The false discovery rate (FDR) cutoff of 0.02 was used to determine top pathways within the data sets. The secretome data set was visualized by bubble plot with Z-score on x-axis and FDR q-value on y-axis. Bubble size was dependent on number of proteins identified within this pathway. Z-score was calculated by the following formula, as described by Riddell et al. (34): $Z \mathrm{s}\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{e}=\frac{\left(\mathrm{u}\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{g}-\mathrm{d}\mathrm{o}\mathrm{w}\mathrm{n}\mathrm{r}\mathrm{e}\mathrm{g}\right)}{\sqrt{\mathrm{C}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t}}}$ where the count is the total number of differentially expressed genes in the pathway and upregulated or downregulated refers to number of genes with log fold change values >1.5 and <1.5, respectively.

T-Cell Secretome Stimulation of Bone Marrow Monocytes

Monocytes from the bone marrow of unoperated female mice (n = 4; 4.4 ± 0.1 mo of age) were isolated by magnetic bead isolation, as previously described (29, 35). Purified monocytes were stimulated for 4 h with the secretome of T cells stimulated with: 1) 0.1% FBS or 2) IL-4 + 0.1% FBS. As a control, monocytes were cultured in only RPMI supplemented with 0.1% (negative control) or 10% FBS (positive control). After stimulation, monocytes were assessed for phenotypic changes by flow cytometry including Ly6C (Miltenyi Biotec 130-111-779, 1:25), F4/80 (Thermo Fisher Scientific MF48020, 1:50), CD86 (Miltenyi Biotec 130 105 135, 1:10), and cell viability (Invitrogen Live/Dead Fix Yellow, 1:1,000). To measure phagocytic ability of the stimulated monocytes, fluorescently labeled particles (Invitrogen A10025) were added to the media during stimulation and visualized on the flow cytometer, as described by the manufacturer. Cells were washed and resuspended in cold PBS supplemented with 0.5% bovine serum albumin (BSA) and 2 mM EDTA. Primary antibodies were added, and cells were incubated for 20 min at 4°C in the dark. Cells underwent filtration using a 5-mL tube with cell-strainer cap and analyzed, as described above. Analysis of inflammatory markers Ly6C and F4/80 was gated after removal of cellular debris and nonviable cells.

Electric Cell-Substrate Impedance Sensing

Cell migration was analyzed using electric cell-substrate impedance sensing (ECIS, Applied Biophysics), as described previously (19). Monocytes from male mice (3.4 ± 0.1 mo of age; n = 3 biological samples; plated in triplicate) were isolated and plated in an ECIS 96-well plate (2.0 × 10⁵ cells, triplicates/condition). Cells were equilibrated for 2 h to ensure adherence. After 2 h, cells in the center of the well were wounded for 10 s at 1,200 uA, 4,000 Hz. After wounding, media was changed, the CD8⁺ T-cell secretome (10% final volume) was added, and migration into the center of the well was recorded. The rate of migration was calculated as the slope of the impedance during the first hour after wounding.

Inhibition of CD8⁺ T Cells by MHCi (Cross-Sectional)

The experimental design for the LPS + MI + MHCi experiments is illustrated in Supplemental Fig. S5. Mice were exposed chronically to LPS through Alzet osmotic minipumps for 28 days before MI was induced through permanent ligation of the left anterior descending coronary artery, as described previously (17–19). To inhibit CD8⁺ T-cell activation, MHC-I blocking antibody (MHCi; 0.2 µg/day; eBioscience SF1-1.1.1) was constantly infused by osmotic minipumps implanted subcutaneously 7 days before MI (21 days after LPS infusion) until tissue collection. As a negative isotype control for the MHCi experiments, an IgG antibody (0.2 µg/day; abcam ab97049) was administered to determine any potential off target effects.

Statistics

Data are presented as means ± SE. For two group comparisons, the nonparametric Wilcoxon rank-sum test was used. Multiple group comparisons were analyzed using one-way ANOVA, followed by the Student–Newman–Keuls when the Bartlett’s variation test was passed, or the nonparametric Kruskal–Wallis test, followed by Dunn post hoc test when the Bartlett’s variation test did not pass. Statistical significance was set at P < 0.05.

RESULTS

Chronic LPS Increased T-Cell-Mediated Inflammation

Under nondiseased conditions, there were little to no T cells present in the LV. CD8 can also be expressed by other immune cells including dendritic cells so we used CD3⁺CD8⁺ to identify CD8⁺ T cells (36). T-cell numbers (CD3⁺) were not different in the LV of the LPS group compared with D0 controls (Fig. 1). After MI, T cells increased in the infarct in both MI and LPS + MI groups compared with day 0 with 49 ± 3% being CD8⁺ T cells in the MI group and 55 ± 4% in the LPS + MI group (P = 0.04). This is in line with the previous studies by Yan et al. (37) who showed CD8⁺ T cells made up around 40% of the T-cell population at day 1 post-MI without LPS. To test whether chronic LPS activated a memory response, we measured CD44 and CD27 to distinguish effector and memory CD8⁺ T cells. Both MI and LPS + MI but not LPS alone resulted in an influx of activated (CD44⁺) T cells. LPS + MI intensified the number of CD8⁺CD44⁺ T cells in the infarct compared with all other groups. Assessments for CD8⁺ memory T cells (CD3⁺CD8⁺CD44⁺CD27⁺) and CD8⁺ effector T cells (CD3⁺CD8⁺CD44⁺CD27⁻) demonstrated that with LPS + MI, there was an increase in the recruitment of the CD8⁺ T-cell memory phenotype and the effector population compared with all other groups (Fig. 1). The effector CD8⁺ T cells were the main subpopulation identified as only 15 ± 1% of the CD8⁺ T cells were classified as memory T cells in the LPS + MI group.

Figure 1.

Chronic lipopolysaccharide (LPS)-induced inflammation increased memory CD8⁺ T cells in the infarct at day 1 post- myocardial infarction (MI). Confocal imaging of D0 and LPS hearts showed little to no T cells present. T-cell numbers (CD3⁺) and activated T cells (CD3⁺CD44⁺) increased in the left ventricle (LV) in the MI and LPS + MI groups with CD8⁺ T cells (CD3⁺CD8⁺ and CD3⁺CD8⁺CD44⁺) being significantly upregulated only in the LPS + MI group. The CD3⁺CD8⁺CD44⁺CD27⁻ (effector) and the CD3⁺CD8⁺CD44⁺CD27⁺ (memory) population increased in the LPS + MI group only. n = 6–12/group; Scale bar: 40 μm; females are designated by yellow dots and males by blue dots; multiple group comparisons were analyzed by one-way ANOVA with Student–Newman–Keuls posttest. *P < 0.05 vs. D0; #P < 0.05 vs. LPS; †P < 0.05 vs. MI.

Periodontitis Induces a Systemic Inflammatory Response and Activation of CD8⁺ T Cells

Pathway analysis of plasma analytes after chronic LPS exposure showed that top enriched pathways ranked by FDR were interleukin (IL) signaling (specifically IL-4, -13, and -18), chemokine receptors, peptide ligand-binding receptors, IL-1 processing, inflammasome pathway, class A/1 (Rhodopsin-like receptors), G protein-coupled receptor ligand binding, and RUNX1 and FOXP3 control of Treg development (Table 1). IL-4 and -18 are known regulators of memory CD8⁺ T-cell activation and proliferation (38–40). These data indicate that the circulating endotoxins from periodontitis are the likely driver for increasing memory CD8⁺ T cells in the LV through activation of interleukin signaling.

Table 1.

Plasma analytes show an enrichment in proteins involved with interleukin signaling after chronic LPS exposure

Pathway	Reaction Ratio	FDR
Interleukins signaling	58/490	7.44E-15
Interleukin-4 and -13 signaling	6/46	2.00E-13
Chemokine receptors	7/19	1.02E-08
Interleukin-18 signaling	4/4	3.53E-05
Peptide ligand-binding receptors	7/76	7.03E-05
Interleukin-1 processing	4/5	9.33E-05
Inflammasome pathway	2/4	2.33E-04
Class A/1 (rhodopsin-like receptors)	7/152	0.0038
G Protein-coupled receptor ligand binding	7/179	0.0134
RUNX1 and FOXP3 control of Treg development	3/20	0.0306

Pathways are ranked by their false discovery rate (FDR). Reaction ratio, proportion of reactions found vs. total reactions represented by this pathway; LPS, lipopolysaccharides.

Mapping the T-Cell Secretome Constituents

IL-4 plays a role in activating type 2 inflammation and has been shown to be beneficial after MI (41, 42). Because IL-4 signaling is a known regulator after MI and was one of the top pathways upregulated with chronic LPS, we aimed to dissect whether IL-4 acted as a mediator of the memory response. Isolated splenic naïve CD8⁺ T cells from day 0 (no LPS) mice were cultured in RPMI with either 0.1% FBS (unstimulated), 10% FBS (positive control), or 0.1% FBS + IL-4 (IL-4 stimulated). We used naïve splenic CD8⁺ T cells isolated from day 0 (no LPS) mice to ensure cells were not being preactivated due to chronic LPS administration. Interestingly, 60% of the identified proteins were associated with cellular membrane (23%), protein complexes (25%), extracellular space (5%), or immunological synapse (7%; Fig. 2A). The intracellular proteins observed were not due to a loss in cell viability (Supplemental Fig. S4B), consistent with a previous report and likely result of extracellular vesicles production (43).

Figure 2.

In vitro stimulation of CD8⁺ T cells with interleukin-4 (IL-4) alters the T-cell proteome and memory function. A: mass spectrometry of the conditioned media demonstrated that CD8⁺ T cells secreted proteins within the cellular membrane (23%), protein complexes (25%), extracellular space (5%), and immunological synapse (7%) in addition to proteins within the intracellular compartment (40%). B: bubble plot visualizes pathway changes by plotting z score on x-axis and false discovery rate (FDR) q value on y-axis. FDR cut-off of 0.02 was used to determine top pathways within the data sets. Size of bubble is dependent on number of proteins identified within this pathway. Z score was calculated by the following formula: z score = ((upreg-downreg))/√Count, where the count is the total number of differentially expressed genes in the pathway and upreg and downreg is the number of those genes with log fold change values >1.5 and < 1.5, respectively. n = 5/sex/stimulation. C: flow cytometry of cells after stimulation demonstrated that with IL-4 stimulation, there was an upregulation in CD27 expression resulting in a slight shift from effector (CD44⁺CD27⁻) to the memory phenotype (CD44⁺CD27⁺). n = 6/stimulation (all females); flow cytometry data were analyzed by repeated-measures one-way ANOVA with Student–Newman–Keuls posttest. *P < 0.05 vs. unstimulated; #P < 0.05 vs. 10%.

Out of the 256 proteins identified, 12 proteins were different between the unstimulated and IL-4-stimulated groups (Table 2). Of these 12, 3 had a twofold or greater increase (Plexin-B2, MOB kinase activator 3 C and Rho GTPase-activating protein 26) and 5 had a twofold or greater decrease (Desmoglein-4, Type II inositol-1,4,5-trisphosphate-5-phosphatase, nidogen-2, DENN (differentially expressed in normal cells and neoplasia) domain-containing protein 5B, and Ras guanyl-releasing protein; Supplemental Fig. S4C). In addition, two proteins were detected only in the unstimulated group (Protein phosphatase 2 and amyloid-βA4 protein) and two were detected only in the secretome of IL-4 stimulated cells (aryl hydrocarbon receptor nuclear translocator and Retinoic acid-induced gene 3 protein).

Table 2.

Top 12 proteins in T cells altered by IL-4 stimulation ranked by P value

Pathway	Fold Change	P Value
Plexin-B2	2.3	0.0137
Desmoglein-4	−3.1	0.0173
Type II inositol-1,4,5-trisphosphate-5-phosphatase	−2.1	0.0179
Protein phosphatase 2	0	0.0205
Amyloid-βA4 protein	0	0.0207
MOB kinase activator 3 C	7.3	0.0226
Aryl hydrocarbon receptor nuclear translocator	∞	0.0263
Retinoic acid-induced gene 3 protein	∞	0.0266
Rho GTPase-activating protein 26	5.8	0.0318
Nidogen-2	−2.1	0.0351
DENN domain-containing protein 5B	−4.9	0.0382
Ras guanyl-releasing protein 3	−5.4	0.0386

Fold change, interleukin (IL)-4 stimulated vs. unstimulated cells; 0, not identified in IL-4 stimulation but highly expressed in control; ∞, not identified in control but highly expressed with IL-4 stimulation.

By pathway analysis, changes in the secretome after IL-4 stimulation reflected upregulation of Wnt signaling and oxidative stress-induced senescence, along with downregulation of the β-catenin transactivating complex (decreased formation and increased deactivation biomarkers), epigenetic regulation of gene expression, and RHO GTPase signaling (Table 3, Fig. 2B). Wnt/β-catenin signaling is critical for the differentiation, polarization, and survival of mature T cells and promotes memory T-cell formation (44). Evaluation of memory markers (CD44 and CD27) revealed that IL-4 stimulation increased markers associated with the memory population (CD27⁺; Fig. 2C).

Table 3.

List of pathways enriched in CD8⁺ T cells stimulated with IL-4 compared with unstimulated cells using Reactome

Pathway	z-Score	Reaction Ratio	FDR
β-Catenin: T-cell factors-transactivating complex formation	−0.2774	4/43	0.0029
Signaling in response to Wnt	0.6124	44/71	0.0042
Oxidative stress-induced senescence	0.2582	14/40	0.0157
Deactivation of the β-catenin transactivating complex	0.3333	7/14	0.0157
Epigenetic regulation of gene expression	−0.3536	22/34	0.0165
Signaling by RHO GTPases	−0.2887	12/20	0.0165

Pathways are ranked by their false discovery rate (FDR). Z score indicates pathway upregulation or downregulation compared with unstimulated cells. Reaction ratio indicates the proportion of reactions found vs. total reactions represented by this pathway.

The Secretome of IL-4-Stimulated CD8⁺ T Cells Facilitated Monocyte Activation

To determine whether the proteins secreted by unstimulated cells and IL-4-stimulated cells regulated monocyte physiology, we stimulated bone marrow monocytes from unoperated animals with the CD8⁺ T-cell secretome (Fig. 3A). Monocyte viability decreased in response to the secretome of IL-4-treated T cells compared with monocytes cultured in the secretome from unstimulated CD8⁺ T cells (Fig. 3B). Phagocytosis was not different (P = 0.07) in monocytes that were cultured in the secretome from the IL-4-stimulated group compared with the unstimulated CD8⁺ T cells. No changes were observed in F4/80 (P = 0.55) or CD86 expression (P = 0.34). The secretome of unstimulated CD8⁺ T cells did not affect monocyte migration after wounding (Fig. 3C). The secretome of IL-4-stimulated T cells increased monocyte migration after wounding similar to levels of the positive control (monocytes in 10% FBS). Overall, our data suggest that IL-4-stimulated memory CD8⁺ T cells were able to initiate monocyte migration but decreased monocyte viability compared with unstimulated CD8⁺ T cells.

Figure 3.

T-cell secretome altered viability and migration of bone marrow monocytes. A: experimental design of monocyte stimulation with the CD8⁺ T-cell secretome. B: flow cytometry of in vitro-stimulated bone marrow monocytes. When compared with unstimulated CD8⁺ T cells, the secretome of CD8⁺ T cells stimulated with IL-4 decreased monocyte viability. There were no differences in phagocytosis, F4/80, or CD86 expression between groups. n = 4/stimulation; Paired t test was used. C: cellular migration was measured using electric cell-substrate impedance sensing assay. Monocytes were isolated from male mice; n = 3 biological samples plated in triplicate. Multiple group comparisons were analyzed by one-way ANOVA with Student–Newman–Keuls posttest. *P < 0.05 vs. unstimulated; #P < 0.05 vs. negative control.

Memory T Cells Are Key Regulators of Macrophage-Mediated Phagocytosis in the Infarct after MI

CD8⁺ T cells are activated by antigen-presenting cells through interactions between CD8 and major histocompatibility complex (MHC)-I (45). As proof of in vivo mechanism relevance, we infused a blocking antibody for the (MHC)-I continuously in LPS-exposed mice starting 7 days before MI (21 days after LPS) and lasting through the duration of the MI to prevent CD8⁺ T-cell activation. Mice that received the blocking antibody no longer had elevated total T-cell numbers compared with the MI group. This was due to a reduction in CD8⁺ T-cell numbers in the day 1 infarct (Fig. 4, A and B). LPS + MI + MHCi significantly decreased effector CD8⁺ T cells (CD3⁺CD8⁺CD44⁺CD27⁻) but not CD8⁺ memory T cells (CD3⁺CD8⁺CD44⁺CD27⁺) compared with the LPS + MI group, indicating that initiation of the memory response was independent of MHC-mediated activation. No differences were observed between the LPS + MI + IgG control and the LPS + MI group indicating that the LPS + MI + MHCi was selective in inhibiting CD8⁺ T-cell activation.

Figure 4.

Confocal imaging (A) showed inhibition of CD8⁺ T-cell activation via major histocompatibility complex-I-blocking antibody (MHCi) attenuated total T-cell numbers due to a decrease in the CD8⁺ T-cell population. LPS + MI + MHCi decreased CD44 expression on the CD8⁺ T cells as demonstrated by the lower number of effector (CD3⁺CD8⁺CD44⁺CD27⁻) but not memory (CD3⁺CD8⁺CD44⁺CD27⁺) CD8⁺ T cells compared with the LPS + MI group (B). No differences were observed between the LPS + MI group and the LPS + MI + IgG controls. Overlay images are shown for LPS + MI + MHCi and LPS + MI + IgG groups (for other groups, see Fig. 1). n = 4–12/group; Scale bar: 40 μm; Females are designated by yellow dots and males by blue dots; multiple group comparisons were analyzed by one-way ANOVA with Student–Newman–Keuls posttest; †P < 0.05 vs. MI; ‡P < 0.05 vs. LPS + MI. LPS, lipopolysaccharide; MI, myocardial infarction.

Previously, we reported that with chronic LPS exposure (no MI), there was a slight yet significant decrease in ejection fraction (18). At day 1 post-MI, LPS + MI and LPS + MI + IgG mice were already showing signs of adverse remodeling with reduced wall thickness, as determined by M-mode images in both systole and diastole (Fig. 5A and Table 4). With LPS + MI + MHCi treatment, wall thickness was reduced during diastole. No differences were observed in wall thickness during systole in the LPS + MI + MHCi group compared with MI alone indicating that effector CD8⁺ T cells were likely exacerbating myocyte loss through cytotoxic mechanisms.

Figure 5.

Inhibiting the effector CD8⁺ T-cell phenotype had no effect on lipopolysaccharide (LPS)-induced macrophage recruitment and activation. A: representative image of M-mode short axis echocardiography images. B: immunohistochemistry for neutrophils [polymorphonuclear neutrophils (PMN); n = 4–16/group] and macrophages (Mac-3; n = 4–16/group) in the infarct area indicated an acceleration in macrophage trafficking with LPS + MI that was not effected by LPS + MI + MHCi. LPS + MI + IgG decreased macrophage recruitment compared with LPS + MI. No effect on neutrophils or collagen (picrosirius red staining; PSR; n = 4–16/group) was observed between the groups. Insets show a ×2 digital zoom. Scale bar: 100 μm; C: confocal imaging of the infarct for Mac-3 and phagocytosis receptor, MertK confirmed immunohistochemistry findings. No differences were observed in overall MertK levels between groups. Percent double-positive Mac3 and MertK staining indicated that there was an increase in macrophage-mediated phagocytosis in the LPS + MI group that remained elevated with LPS + MI + MHCi. LPS + MI + IgG administration decreased the number of Mac3⁺MertK⁺ cells (n ≥ 4/group). Scale bar: 40 μm; Females are designated by yellow dots and males by blue dots; multiple group comparisons were analyzed by one-way ANOVA with Student–Newman–Keuls posttest; †P < 0.05 vs. MI; ‡P < 0.05 vs. LPS + MI. MI, myocardial infarction.

Table 4.

Chronic LPS exposure decreases cardiac function as determined by echocardiography

	D0	LPS	MI	Lps + MI	LPS + MI + MHCi	LPS + MI + IgG
n (male/female)	18 (9/9)	20 (10/10)	18 (9/9)	12 (6/6)	5 (2/3)	7 (4/3)
EDD, mm	3.56 ± 0.09	3.69 ± 0.06	4.29 ± 0.12a,b	4.48 ± 0.07a,b	4.63 ± 0.23a,b	4.38 ± 0.12a,b
ESD, mm	2.28 ± 0.07	2.57 ± 0.07a	3.90 ± 0.15a,b	4.19 ± 0.07a,b	4.23 ± 0.23a,b	3.96 ± 0.11a,b
FS, %	36 ± 1	31 ± 1a	10 ± 2a,b	6 ± 1a,b	9 ± 2a,b	9 ± 1a,b
EDV, µL	61 ± 3	59 ± 3	77 ± 5a,b	78 ± 3a,b	83 ± 9a,b	84 ± 8a,b
ESV, µL	20 ± 1	25 ± 1	66 ± 5a,b	66 ± 3a,b	70 ± 8a,b	67 ± 6a,b
EF, %	65 ± 1	57 ± 1a	16 ± 1a,b	16 ± 1a,b	17 ± 3a,b	20 ± 2a,b
Wall thickness
Diastole	0.84 ± 0.02	0.79 ± 0.01	0.79 ± 0.04	0.51 ± 0.03a,b,c	0.59 ± 0.06a,b,c	0.62 ± 0.04a,b,c
Systole	1.25 ± 0.03	1.14 ± 0.02	0.90 ± 0.07a,b	0.58 ± 0.04a,b,c	0.74 ± 0.11a,b	0.68 ± 0.04a,b,c
Infarct size, %	NA	NA	58 ± 3	58 ± 3	58 ± 2	55 ± 1

Values are means ± SE. EDD, end-diastolic dimension; EDV, end-diastolic volume; EF, ejection fraction; ESD, end-systolic dimension; ESV, end-systolic volume; FS, fractional shortening; LPS, lipopolysaccharides; MI, myocardial infarction. ^aP < 0.05 vs. D0; ^bP < 0.05 vs. LPS; ^cP < 0.05 vs. MI.

Immunohistochemistry for neutrophils showed no differences between groups at day 1 after MI (Fig. 5B). As we previously published (18), chronic LPS accelerated macrophage recruitment into the infarct at day 1 post-MI. Administration of MHCi did not affect macrophage numbers compared with LPS + MI (Fig. 5, B and C). MertK expression on Mac3⁺ cells was also upregulated in the LPS + MI group compared with MI alone and no differences were observed with LPS + MI + MHCi administration compared with the LPS + MI group (Fig. 5C) indicating that this initial recruitment of macrophages was not regulated by the CD8⁺ effector T cells. IgG administration decreased macrophage numbers and MertK expression compared with the LPS + MI group (Fig. 5, B and C). At post-MI day 1, no differences were observed in collagen levels between groups.

DISCUSSION

The goal of this study was to investigate the role of T cells in chronic inflammation induced by a periodontal pathogen (LPS) and MI wound healing. The major findings were that 1) chronic LPS activated memory CD8⁺ T cells, 2) IL-4 was able to upregulate memory markers on CD8⁺ T cells in vitro independent of MHC-I activation, and 3) blocking activation of effector CD8⁺ T cells had no effect on the macrophage MertK expression yet prevented excessive infarct wall thinning at post-MI day 1. Overall, our data revealed that with chronic LPS, there was an exacerbation of effector and memory CD8⁺ T-cell infiltration likely contributing to remodeling of the LV after an MI, as described in our proposed model (Fig. 6).

Figure 6.

Schematic of proposed mechanisms mediating the role of chronic periodontitis in adverse post-MI remodeling. Activation of CD8⁺ T cells occurs after antigen-presenting cells (APC) including macrophages engulf the periodontal pathogen and begin to present the antigen on their major histocompatibility complex (MHC) receptors. After a myocardial infarction (MI), both the effector and memory populations increased in the animals exposed to chronic lipopolysaccharide (LPS). MHC-I blocking antibody attenuated the effector CD8⁺ T cells and improved wall thinning during systole indicative of preserved myocyte function; however, likely due to an increase in interleukin (IL) signaling the memory CD8⁺ T cells remained elevated. In vitro analysis demonstrated that memory CD8⁺ T cells decreased macrophage viability and increased chemotaxis. IgG antibody administration (LPS + MI + IgG) decreased macrophage numbers and MertK expression indicating a decrease in phagocytosis likely through the inhibition of LPS-induced Fc receptor (FcR) upregulation. Phagocytosis of the necrotic debris would facilitate in scar formation and thus may improve cardiac wound healing. Image was generated using Biorender.com and published with permission.

P. gingivalis is one of over 300 bacterium responsible for periodontal disease biofilm development (46). Antibody levels against major periodontal pathogens, including P. gingivalis, A. actinomycetemcomitans, T. forsythia, and T. denticola, increase relative risk for MI independent of known cardiovascular risk factors (47). In two separate studies, 25%–35% of thrombus aspirates collected from patients with MI contained periodontal pathogens (20% for A. actinomycetemcomitans, 3% for P. gingivalis, and 2% for T. denticola) (48, 49). Not surprisingly, there was also an increase in macrophage markers (CD14 and CD68) and T cells (CD3) within the thrombi (49, 50). Our data suggest that LPS from P. gingivalis activates the memory response leading to an acceleration in inflammation post-MI.

Despite no differences in IL-4 expression in the left ventricle with LPS compared with day 0, there was an increase in IL-4 signaling pathways. Multiple studies have demonstrated an increase in IL-4 signaling molecules after MI, including IL-4-induced gene-1 and IL-4 receptor α (35, 51). This has led to the hypothesis that although IL-4 may not be endogenously produced in the infarct, multiple cells including macrophages and fibroblasts upregulate these receptors and are likely highly responsive to IL-4 treatment (35, 41, 51). In the setting of eosinophil deficiency, IL-4 therapy has been shown to attenuate adverse cardiac remodeling (52). Daseke et al. (41) showed that administering exogenous IL-4 after MI promoted resolution of inflammation by inhibiting proinflammatory neutrophils and stimulating macrophage-mediated removal of apoptotic neutrophils. Although the authors did not directly address the effects of IL-4 on T cells, macrophage signaling at post-MI day 3 upregulated T-cell differentiation, indicating that IL-4 signaling may also indirectly affect the CD8⁺ T-cell population after MI.

The literature that describes the influence of IL-4 on CD8⁺ T-cell function is inconsistent, whereas some have shown that IL-4 can dampen the cytotoxic activity of CD8⁺ T cells, others have demonstrated the exact opposite effect (33, 53, 54). A majority of the studies evaluated the effect of IL-4 on CD8⁺ T-cell physiology during viral infection; our in vitro study is the first to describe a possible mechanism for IL-4-activated memory CD8⁺ T cells in regulating monocyte physiology without injury or infection. In agreement with our findings, Weinreich et al. (54) demonstrated that IL-4 was necessary for generation of memory CD8⁺ T cells after listeria infection. In response to malaria parasite challenge, IL-4 stimulated CD8⁺ T-cell activation and initiated cells to migrate to nonlymphoid organs and establish memory populations (40). The complex interactions among immune cells within the myocardial environment may be obscuring our understanding of IL-4 signaling. Additional studies evaluating cellular heterogeneity and the immune cell interactome are necessary to understand all of IL-4 actions in post-MI remodeling.

The secretome data were composed of 40% of intracellular proteins, which was initially surprising. This was not due to increased cell death as cell viability was unaltered. One possible explanation is that many of the proteins identified in the secretome were linked to secretory vesicles (55). IL-4 can alter extracellular vesicle release in macrophages and B cells (56, 57). Breast cancer cells internalized exosomes released from IL-4-activated macrophages more efficiently than those released from nonactivated macrophages (57). Secretion of CD8⁺ T-cell extracellular vesicles decreases cell viability in virus-infected cells and mesenchymal tumor stromal cells and prevents tumor invasion and metastasis (58–60). IL-4 stimulation of CD8⁺ T cells may, therefore, stimulate cell death via exosome-mediated mechanisms. Additional studies are needed to fully test this hypothesis.

Previously we demonstrated that during the first 7 days post-MI, CD8⁺ T cells play a protective role by activating macrophage-mediated removal of the necrotic debris (30). Timely removal of necrotic cells is important for scar formation. However, it is also critical that this process is balanced to inhibit off-target effects. With chronic LPS, we found an increase in MertK expression on macrophages in the infarct. Inhibiting effector CD8⁺ T-cell activation with the LPS + MI + MHCi did not affect macrophage numbers or MertK expression. Surprisingly, the LPS + MI + IgG controls decreased macrophages compared with the LPS + MI group despite having no effect on the CD8⁺ T-cell subtypes. One possible explanation is that LPS exposure has been shown to enhance Fc receptors that bind the Fc portion of immunoglobulins and increase FcR-dependent phagocytosis in tissue macrophages (61–63). Our data indicate that CD8⁺ T cells play a role in excessive wall thinning; however, CD8⁺ T cell effects on cardiac remodeling under chronic inflammatory settings are likely independent from macrophage-mediated phagocytosis. Instead, LPS is likely activating macrophages in an attempt to facilitate the removal of the periodontal pathogen.

Our previous studies have also shown an increase in the M1 proinflammatory macrophage phenotype after chronic LPS + MI compared with MI controls (18, 19). In the current study, we did not observe any changes in M1 marker CD86 in monocytes stimulated with the T-cell secretome. Interestingly, we did observe a 1.5-fold increase in macrophage numbers for LPS + MI males compared with females (Fig. 5B). This was not due to differences in T-cell phenotypes (P = 0.314) and did not alter cardiac physiology (P = 0.477 for ejection fraction), as measured by echocardiography. Although IL-4 stimulation of CD8⁺ T cells in vitro increased markers of memory, the cellular classification in vivo is likely more complex due to the many players in the post-MI environment. It should also be pointed out that our in vitro system did not contain LPS and thus only evaluated the effect of IL-4 alone. An important comparison would be to characterize and compare IL-4-treated CD8⁺ T cells and in vivo-derived CD3⁺CD8⁺CD44⁺CD27⁺ to fully understand how our in vivo system compares to our in vitro system. This may also give us additional insight into whether the CD8⁺ T cells were facilitating in the sex differences observed in macrophage numbers within the LPS + MI group.

Chronic LPS increases circulating troponin I, indicative of cardiac injury (17). Elevated troponin I would prime the memory response so that upon ischemic injury, memory cells would become activated. What antigen memory CD8⁺ T cells are targeting is unknown. Bansal et al. (64) reported that 8 wk after MI, cardiac and splenic T cells retained memory for cardiac antigens and upon adoptive transfer, initiated cardiac fibrosis in a nonischemic heart. Hofmann et al. (65) demonstrated that antigen-mediated activation was vital for the protective function of CD4⁺ T cells after MI. Whether this is important for CD8⁺ T-cell function after MI is not known. There has been evidence to suggest that CD8⁺ T cells can become resident memory cells within nonlymphoid tissues and can facilitate local protection against infection by amplifying secondary memory in absence of ongoing T-cell stimulation (66–68). The heart does have resident memory T cells (67); however, their role after MI is still unknown. Our data indicate that in a chronic inflammatory setting, there is an increase in both effector and memory populations. Future studies evaluating whether this increase is due to recruitment alone or the expansion of local resident T cells would give us better insight into their role in cardiac disease pathology.

Using an MHC-I inhibitor before MI decreased the number of effector CD8⁺ T cells but surprisingly had no effect on memory CD8⁺ T cells. Morris et al. (53) illustrated that antigen presentation is required for proliferation of naive CD8⁺ T cells. In contrast, memory CD8⁺ T-cell numbers increased after IL-4 stimulation, which was further, enhanced by MHC class I antigen presentation (53). In response to bacterial infection, memory T cells activate acutely through cytokines such as IL-18 and IL-15 independent of antigen presentation (69). Previously we demonstrated that IL-15 and IL-18 increased in the infarct with LPS + MI compared with MI alone (18). Our current work expands on these findings and demonstrates an increase in antigen-independent activation of CD8⁺ memory T cells in chronic LPS mice after an MI. Increased IL signaling may therefore be facilitating in rapid memory T-cell differentiation and the initial recruitment of immune cells and removal of necrotic debris. This is likely to be beneficial during the acute remodeling process, but will have off-target effects chronically.

Limitations

Because of limitations of the mHART database inventory, we were not able to measure a full-time course of T-cell numbers after LPS exposure alone. Our data show that T cells are present within the LV after 28 days of LPS; however, it is unclear whether 28 days of exposure is necessary or whether T-cell numbers increase earlier. Future studies evaluating that shorter and longer exposure times are needed to benefit our understanding of the cardiac T-cell response.

To mimic the chronic inflammatory response observed in periodontal disease, we infused LPS from the periodontal pathogen P. gingivalis for 28 days before MI. Periodontal disease is characterized by gingivival tissue inflammation mediated by a bacterial biofilm containing hundreds of microbes that are needed to fully recapitulate the pathology (70). Future studies evaluating how P. gingivalis alone versus the multitude of bacteria that are associated with periodontal disease activate the immune system and facilitate in LV remodeling are needed to give us additional insight into the relationship between oral and cardiovascular health.

In summary, our results indicate that effector and memory CD8⁺ T cells at post-MI day 1 are amplified by chronic LPS, leading to excessive infarct wall thinning. Therapeutic strategies to improve responses to MI should include consideration for preexisting comorbidities in their design.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S4: https://doi.org/10.6084/m9.figshare.13714114.

GRANTS

This work was supported by the National Institutes of Health HL148114, HL137319, U54GM115458, HL145817, UL1TR001450, T32GM123055, and U54DA016511; the American Heart Association Innovator Project IPA35260039; and the Biomedical Laboratory Research and Development Service of the Veterans Affairs Office of Research and Development Award I01BX000505 and IK2BX003922. This work was also financially supported, in part, by the 2019 S&R Foundation Ryuji Ueno Award that was bestowed upon K. Y. DeLeon-Pennell by the American Physiological Society.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, the Veterans Administration, or the American Physiological Society.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.L.L. and K.Y.D-P. conceived and designed research; Y.Z., A.C., V.Y.V., K.O., C.G., K.W., J.M., M.T., P.B., D.V.I., M.L.L., and K.Y.D-P. performed experiments; Y.Z., A.C., K.O., M.T., P.B., D.V.I., M.L.L., and K.Y.D-P., analyzed data; Y.Z., A.C., D.V.I., M.L.L., and K.Y.D-P. interpreted results of experiments; Y.Z., M.L.L., and K.Y.D-P. prepared figures; Y.Z., M.L.L., and K.Y.D-P. drafted manuscript; Y.Z., A.C., V.Y.V., K.O., C.G., K.W., J.M., M.T., P.B., D.V.I., M.L.L., and K.Y.D-P., edited and revised manuscript; Y.Z., A.C., V.Y.V., K.O., C.G., K.W., J.M., M.T., P.B., D.V.I., M.L.L., and K.Y.D-P. approved final version of manuscript.