Atherosclerosis impairs naive CD4 T cell responses via disruption of glycolysis

Abstract

Objective:

CD4 T cells are important regulators of atherosclerotic progression. The metabolic profile of CD4 T cells controls their signaling and function, but how atherosclerosis affects T cell metabolism is unknown. Here, we sought to determine the impact of atherosclerosis on CD4 T cell metabolism and the contribution of such alterations to atheroprogression.

Approach and Results:

Using PCR arrays, we profiled the expression of metabolism genes in CD4 T cells from atherosclerotic, apolipoprotein-E knockout (ApoE^−/−) mice fed a Western diet (WD). These cells exhibited dysregulated expression of genes critically involved in glycolysis and fatty acid degradation, compared to those from animals fed a standard laboratory diet. We examined how T cell metabolism was changed in standard laboratory diet or WD-fed ApoE^−/− mice or humans by measuring glucose uptake, activation, and proliferation in CD4 T cells. We found that naive CD4 T cells from WD-fed ApoE^−/− mice failed to uptake glucose and thus, displayed impaired proliferation and activation, compared to CD4 T cells from standard laboratory diet-fed animals. Similarly, as in mice, we observed that naive CD4 T cell frequencies were reduced in circulation of human subjects with high CVD compared to low CVD, as assessed clinically based on medically necessary coronary angiography. Naive T cells from high CVD subjects also showed reduced proliferative capacity.

Conclusion:

These results highlight the dysfunctional changes that occur in CD4 T cell metabolism and immune responses during atherosclerosis. Targeting metabolic pathways within naive CD4 T cells could thus yield novel therapeutic approaches for improving CD4 T cell responses against atheroprogression.

Introduction

Atherosclerosis is a chronic inflammatory disease characterized by lipid-laden plaques filled with monocytes, monocyte-derived cells, and T cells on arterial walls. CD4 T cells are themselves divided into distinct subsets, with each subpopulation displaying unique impacts on atheroprogression^1–3. For example, naive CD4 T cells in the blood are inversely correlated with subclinical atherosclerosis, while memory CD4 T cells are positively correlated⁴. Additionally, helper-1 (Th1), IFNγ-producing CD4 T cells typically worsen atherosclerosis^{5, 6}, whereas Foxp3⁺, regulatory T (Treg) cells protect against the disease^7–9. CD4 T cell subsets are dynamic populations that can switch between one another, depending on their antigenic and cytokine milieu¹⁰. Indeed, IFNγ⁺ Treg cells have been described in atherosclerosis¹¹ and our laboratory has shown that certain Treg cells can switch to T follicular helper cells¹², which are pro-atherogenic. Still, little is known about the specific contributions of naive CD4 T cells to atherosclerosis progression.

A rapidly expanding body of evidence indicates that the metabolic state of CD4 T cells impacts their responses and contributions to disease progression¹³. CD4 T cells rely on oxidative phosphorylation (OXPHOS) to efficiently generate ATP^14–16 but switch to aerobic glycolysis for more rapid ATP generation upon activation^{17, 18}. Aerobic glycolysis and OXPHOS are both required for proper CD4 T cell proliferation and survival, while the former is also required for optimal IFNγ production¹⁹. Inhibition of acetyl-coA carboxylase (ACC1), a key enzyme in fatty acid synthesis, causes inflammatory T cells to switch to Treg cells^{20, 21}. Importantly, both mouse models of obesity and obese human patients show changes in fatty acid metabolism that favor the production of inflammatory T cells over Treg cells²¹. In contrast, the glucose sensor AMP-activated protein kinase (AMPK) regulates the mitochondrial energetics of activated T cells and can mediate the treatment of inflammatory bowel disease via metformin by decreasing the inflammatory T cell/Treg cell ratio^{22, 23}.

While the aforementioned studies reveal that metabolic reprogramming alters the fate, functions, and responses of CD4 T cells, little is known about the metabolic profiles of CD4 T cells in atherogenesis. In this study, we sought to examine the metabolic state of CD4 T cells during atherosclerosis progression and to assess how CD4 T cell metabolism impacts atherosclerosis.

Material and Methods

The authors declare that all supporting data are available within this article and the accompanying online supplementary files.

Human Samples:

Peripheral blood mononuclear cells (PBMCs) were isolated from blood of donors undergoing medically necessary cardiac catheterization at the Cardiac Catheterization Laboratory at the University of Virginia, Charlottesville^{24, 25}. Enrolled participants included males and females of different ages and informed consent was obtained from all subjects. CVD subject characteristics are detailed in Table 1. Studies were performed with the approval of the Institutional Review Board of Health Science Research at the University of Virginia. PBMCs were isolated from whole blood with Ficoll-Paque^™PLUS (GE Healthcare) and SepMate^™ tubes (STEMCELL Technologies, Inc.). PBMCs were frozen in 10% DMSO in FCS until used for CyTOF analysis or in vitro culture assays.

Table 1 –

CVD Subject Characteristics

Individual Characteristics	Low CVD (n=14)	High CVD (n=14)
General Characteristics
Male (%)	50% (7 of 14)	71 4% (10 of 14)
Age (years)	57.57 (10.07)	66.07 (7.81)
Weight (lbs)	189.90 (20.5)	204.90 (49.80)
BMI*	30.14 (5.968)	30.03 (4.081)
Statins	71.4% (10 of 14)	71.4% (10 of 14)
Diabetes	3 of 14 (21.42%)	4 of 14 (28.57%)
BP Systolic (mmHG)	122.4 (16.90)	137.0 (15.75)
BP Diastolic (mmHG)	75.86 (11.04)	70.71 (11.01)
Immune Parameters
hsCRP (mg/L)‡	5.159 (8.962)	3.809 (4.156)
WBC (k/μL)§	6.70 (1.90)	5.66 (1.12)
Hgb (k/μL)ǁ	13.09 (1.216)	13.26 (1.418)
Platelets (k/μL)	230.4 (69.39)	202.8 (52.31)
Lipids
Total Cholesterol (mg/dl)	164.5 (37.36)	173.1 (53.2)
Triglycerides (mg/dl)	93.9 (57.74)	132.5 (74.64)
HDL Cholesterol (mg/dl)#	46.36 (17.44)	40.07 (12.97)
LDL Cholesterol (mg/dl)**	102.7 (27.29)	111.1 (12.25)
Disease Severity
Gensini Score	3.75 (3.58)	42.71 (9.89)
Immune Compartment
Lymphocytes (k/μL)	1.55 (0.49)	1.67 (0.59)
Monocytes (k/μL)	0.63 (0.16)	0.51 (0.22)
Neutrophils (k/μL)	4.31 (1.52)	3.35 (0.56)
Basophils (k/μL)	0.032 (0.017)	0.03 (0.012)

All variables are indicated as mean±SD

^*BMI=body mass index

^†BP= blood pressure

^‡CRP= C-reacbve protein

^§WBC=white blood cells

^ǁHgb = hemoglobin

^#HDL = high density lipoprotein.

^**LDL = low density lipoprotein

Mice:

Apolipoprotein-E knockout (ApoE^−/−) mice (on a C57BL/6J background) were purchased from The Jackson Laboratory (Bar Harbor, Maine; Stock number 002052). Mice were used for experiments at 6 weeks of age and were fed a Western diet (WD; 42% from fat, 0.2% from cholesterol) (Harlan, #TD 88137, Placentia, CA) for up to 20 weeks. Control (C57BL/6J) animals were fed a standard laboratory diet containing 0% cholesterol and 5% calories from fat (Pico lab, #5053, Saint Louis, MO). All mice were housed in microisolator cages in a pathogen free animal facility at La Jolla Institute for Immunology (LJI). Due to the large variation within metabolic processes, these experiments were performed by arbitrarily choosing male mice. All experiments followed the LJI Animal Care and Use Committee guidelines and all studies were approved according to criteria outlined in the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health.

Mass Cytometry:

Antibodies used for mass cytometry are shown in Table 2. Directly conjugated antibodies were purchased from Fluidigm and purified antibodies were ordered from the companies listed in Table 2. Conjugations were performed with the Maxpar X8 Multi-Metal Labeling Kit (Fluidigm) according to the manufacturer’s instructions. Samples were stained for CyTOF mass cytometry, as described previously^{25, 26}, and stained with the antibody cocktail shown in Table 2. Prior to surface staining, one low CVD sample and one high CVD sample, in addition to a staining healthy control, were barcoded with CD45 (low CVD-CD45–89Y; High CVD-CD45–147Sm; Healthy-CD45–115Ln). Cells were resuspended at 2×10⁶ cells/mL in 0.1mLxEQ^™ Four Calibration Beads (Fluidigm). Batch effects and signal drifting were minimized by tuning every 6hrs during long runs. Data was normalized using the Matlab-based NormalizerR2013a_Win64.

Table 2 –

CyTOF Antibodies

Antigen	Clone	Metal Conjugate	Manufacturer	Conjugation
CD45	HI30	89Y	DVS	DVS
CD45	HI30	115ln	Biolegend	In-house
CCR6	G034E3	141Pr	DVS	DVS
CD19	H1B19	142Nd	DVS	DVS
CD127	A019D5	143Nd	DVS	DVS
CD69	FN50	144Nd	DVS	DVS
CD4	RPA-T4	145Nd	DVS	DVS
CD8	RPA-T8	146Nd	DVS	DVS
CD45	HI30	147Sm	Biolegend	In-house
CD14	RM052	148Nd	DVS	DVS
CD45RO	UCHL1	149Sm	DVS	DVS
CD107a	H4A3	151Eu	DVS	DVS
EOMES	WD1928	152Sm	DVS	DVS
CD45RA	205410	153Eu	DVS	DVS
CD3	UCHT1	154Sm	DVS	DVS
PD-1	EH12.2H7	155Gd	DVS	DVS
CXCR3	G025H7	156Gd	DVS	DVS
CD27	L128	158Gd	DVS	DVS
CD39	A1	160Gd	DVS	DVS
T-bet	4B10	161 Dy	DVS	DVS
GITR	621	162Dy	DVS	DVS
Bcl-6	K11291	163Dy	DVS	DVS
CD95	Dx2	164Dy	DVS	DVS
Foxd3	PCH101	165HO	ThermoFisher	In-house
TIGIT	MBSA43	166Er	DVS	DVS
CCR7	G043H7	167Er	DVS	DVS
ICOS	C398.4A	168Er	DVS	DVS
CD25	2A3	169Tm	DVS	DVS
CTLA-4	14D3	170Er	DVS	DVS
CXCR5	RFBB	171Yb	DVS	DVS
Ki67	BL168	172Yb	DVS	DVS
HLA-DR	L243	174Yb	DVS	DVS
CCR4	205410	175Lu	DVS	DVS
CD57	HCD57	176Yb	DVS	DVS
CD11b	ICRF44	209Bi	DVS	DVS

Flow cytometry:

Peri-aortic lymph nodes, peripheral lymph nodes, and spleens harvested from mice were passed through a 40 μm cell strainer and washed with ice cold PBS. Antibodies were purchased from Ebioscience or Biolegend (San Diego, CA). Samples were stained with LIVE/DEAD fixable dye (Life Technologies, Carlsbad, CA). Human CD4 T cells were stained with antibodies against CD4 (clone RPA-T4), CD3 (clone HIT3a), CCR7 (G01), and CD45RA. Mouse cells were stained with antibodies against CD4 (clone RM4–4), TCRβ (clone H57–597), CD44 (Clone IM-7), CD25 (clone PC61), CD69 (clone H1.2F3), and/or PD-1 (clone J43) in cold FACS buffer (2% BSA, 0.01% sodium azide in PBS). All stains were carried out for 30 minutes on ice. For 2-NBDG staining, stimulated CD4 T cells were washed and incubated in glucose free media for 1 hour. 2-NBDG (100 M; ThermoFisher) was added for 30 minutes at 37°C. Cells were then washed and stained as described above. For Glut 1 staining, stimulated CD4 T cells were stained as described above, fixed, permeabilized and stained with a rabbit anti-glut1 (Novus Biologicals), followed by a secondary goat anti-rabbit AF647. For mitochondrial potential, JC1 mitoprobe (Thermofisher) was added to stimulated CD4 T cells for 15 minutes at 37°C. Cells were then washed with PBS and stained as described above. For FCCP control wells, stimulated CD4 T cells were treated with FCCP (50 mM) for 5 minutes. For apoptosis measurements, cells were first stained for surface markers then subsequently stained for Annexin V and 7AAD, using the Apoptosis Detection Kit I (BD biosciences) according to manufacturer’s instructions. Events were acquired using LSRII (BD, Bioscience, San Diego, CA) and data were analyzed using FlowJo Ver 9.8 and 10.0.08 (Tree Star, Ashland, OR).

In vitro CD4 T cell stimulation:

For human samples, naive CD4 T cells were isolated from previously frozen PBMCs using the EasySep^™ Human Naive T cell isolation kit II (STEMCELL Technologies, Inc., Vancouver, Canada) according to the manufacturer’s instructions. Naive CD4 T cells were then stimulated with Dynabeads^™ Human T-Activator CD3/CD28 (ThermoFisher Scientific) according to the recommended concentration for 2 to 3 days. Recombinant human IL-2 (Peprotech) was added at a final concentration of 100 IU/ml. For mouse samples, naive CD4 T cells were isolated from spleens and lymph nodes of standard laboratory diet- and WD-fed ApoE^−/− mice using the EasySep^™ Mouse Naive CD4 T cell isolation kit (STEMCELL Technologies, Inc.). Naive CD4 T cells were stimulated in αCD3-coated plates (2 μg/ml) and soluble αCD28 (1 μg/ml) was added to these wells for 2 days. All cells were cultured in RPMI with 10% FCS, L-glutamine, penicillin and streptomycin, and β-mercaptoethanol (for mouse cells). For the proliferation assays, isolated naive CD4 T cells were labeled with CTV (ThermoFisher) according to the manufacturer’s instructions, prior to stimulation. The percentage of cells divided was calculated using FlowJo v. 10.0.08 (Tree Star, Ashland, OR).

Intracellular staining:

CD4 T cells were stimulated with PMA and Ionomycin (50 ng/ml and 1 μg/ml, respectively) (Sigma-Aldrich) for 5 hrs in the presence of GolgiPlug (BD Biosciences) at 1 μl/ml. Cells were then stained for CD4, TCRβ, CD44, and CD25 as described above. Afterwards, cells were fixed and permeabilized using the eBioscience^™ Foxp3/transcription factor staining kit (ThermoFisher). Cells were then stained with antibodies against IFNγ (clone XMG-1.2) and Tbet (clone ebio4B10).

Seahorse Assay:

Equal numbers of stimulated CD4 T cells were plated in glucose free media (Seahorse Biosciences) containing glutamine (2mM), with or without glucose (11 mM), for OCR and ECAR measurements at basal conditions. 4×10⁵ cells per well were utilized for these experiments. Oligomycin (1 μM), FCCP (1 μM), and 2-Deoxyglucose (100 nM) were added at the indicated time points. Extracellular acidification rates (ECARs) and oxygen consumption rates (OCRs) were recorded using a Seahorse XF-96 machine according to the manufacturer’s instructions. Compounds were bought from Seahorse Biosciences or Sigma. Each ECAR or OCR measurement is an average of at least three wells.

Glucose Consumption Assay:

Stimulated CD4 T cells were washed with glucose free media. Glucose (1 mg/ml; Sigma) was added and cells were incubated for 4 hours. Supernatant was taken and tested for the levels of remaining glucose using a glucose and sucrose assay kit (Sigma) according to the manufacturer’s instructions.

Fatty Acid Oxidation Assay:

Oxidation of exogenous free fatty acids was assayed utilizing the XF Palmitate-BSA FAO substrate with the XF cell mito stress kit according to the manufacturer’s protocol (Seahorse Bioscience). Negatively selected naive CD4 T cells (4×10⁵ cells) were incubated for 30 minutes with fatty acid oxidation assay medium (111mM NaCl, 4.7mM KCl, 1.25mM CaCl₂, 2.0mM MgSO₄, 1.2mM Na₂HPO₄, 2.5mM glucose, 0.5mM carnitine, and 5mM HEPES). To block fatty acid oxidation, naive CD4 T cells were pre-treated with 40μM etomoxir for 15 minutes at 37°C. Afterwards, 34μM BSA or palmitate-BSA (200μM palmitate conjugated with 34μM BSA) was added to the medium and the OCR was assayed under basal conditions and in response to 1μM oligomycin, 1.5μM fluorocarbonyl cyanide phenylhydrazone (FCCP), and 100nM rotenone + 1μM antimycin A.

Arrays and qPCR:

Total cellular RNA from naive or stimulated CD4 T cells from standard laboratory diet- or WD-fed ApoE^−/− mice was extracted by Trizol (ThermoFisher) followed by RNA purification using the Direct-zol^™ RNA miniPrep kit (Zymo Research, Irvine, CA) per the manufacturer’s instructions. RNA purity and quantity were determined by a nanodrop spectrophotometer (ThermoFisher Scientific) and equal amounts of RNA was used to synthesize cDNA via either the RT² First Strand Kit (Qiagen) or iscript cDNA synthesis kit (Bio-Rad). RT² Profiler^™ PCR Arrays (Qiagen) were used to assess the expression of genes involved in AMPK signaling and glucose metabolism in mouse cells according to the manufacturer’s instructions. mRNA expression was measured via real time quantitative PCR using TaqMan Gene Expression system in combination with pre-designed TaqMan primers for G6pc, Gapdh, Gusb, and Pfkfb3 (Applied Biosytems). Data were analyzed and presented on the basis of relative expression method. The ΔΔCT method was used with GUSB as a housekeeping gene.

Statistical Analysis:

All results are expressed as means (± s.e.m). Results were analyzed by unpaired, two-tailed Student’s t-tests. A P value of < 0.05 is considered to be significant. Statistical analyses were performed using GraphPad Prism software version 8 (GraphPad Software, Inc.). The normality and variance of our data were not examined before performing parametric tests.

Results

CD4 T cells from atherosclerotic mice have functional defects

To assess the metabolic state of CD4 T cells during atherosclerosis, we studied CD4 T cells in the well-characterized atherosclerotic apolipoprotein-E knockout (ApoE^−/−) mouse. ApoE^−/− mice were fed a Western diet (WD) for 20 weeks, after which we characterized the phenotypes of naive CD4 T cells. We found a 25% decrease in the percentage of naive CD4 T cells in the peri-aortic lymph nodes of WD-fed mice compared to standard laboratory diet-fed controls (Figure 1A). This was accompanied by a 5-fold increase in surface expression of the exhaustion marker, Programmed Cell Death Protein-1 (PD-1), by naive CD4 T cells (Figure 1B). When isolated from WD-fed mice and stimulated with αCD3/αCD28, naive CD4 T cells displayed a 50% reduction in IFNγ production (Figure 1C). This change was not due to decreased expression of the transcription factor T-bet, as this protein was increased 2-fold in CD4 T cells from WD-fed mice compared to those from standard laboratory diet-fed controls (Figure 1C).

Figure 1: Naive CD4 T cells are exhausted and less responsive to stimulation during atherosclerosis.

(A-B) CD4 T cells from peri-aortic lymph nodes of standard laboratory diet (SLD)- and Western diet (WD)-fed ApoE−/− mice were examined for naive CD4 T cell percentages (A), and the expression of the exhaustion marker PD-1 on naive CD4 T cells (B) using flow cytometry following 20 weeks of WD feeding. (C) Naive CD4 T cells from SLD- and WD-fed mice were stimulated with αCD3 and αCD28 and the ability to produce IFNγ and express T-bet was detected following intracellular staining. Results are expressed as mean ± s.e.m. from one of two experiments (n=4). Statistically significant differences were at * P<0.05, and ** P<0.01 (unpaired Student t-test).

Upon examining the activation and proliferation potential of CD4 T cells from WD-fed mice, we found that T cells from WD-fed mice had a decreased capacity (30% reduction) to upregulate activation markers (CD44 and CD25) and to divide (Figure 2A), compared to CD4 T cells from control, standard laboratory diet-fed mice. This was accompanied by a 5-fold increase in the percentage of apoptotic (AnnexinV⁺7AAD⁺) CD4 T cells from WD-fed mice (Figure 2B). Taken together, these results suggest that naive CD4 T cells from WD-fed mice exhibit suboptimal responses to antigen stimulation.

Figure 2: Naive CD4 T cells from atherosclerotic mice are less proliferative and undergo more apoptosis following stimulation.

Naive CD4 T cells from standard laboratory diet (SLD) fed- and Western diet (WD)-fed mice were stimulated with αCD3 and αCD28. Flow cytometry plots and bar graphs showing the percentages of divided cells (A) and Annexin-V⁺7AAD⁺ cells following αCD3 and αCD28 stimulation of naive CD4⁺ T cells from SLD- or WD-fed mice (B). Results are expressed as mean ± s.e.m. from one of two experiments (n=4). Statistically significant differences were at * P<0.05, and ** P<0.01 (unpaired Student t-test).

CD4 T cells from atherosclerotic mice have impaired glucose metabolism

We next decided to compare the metabolic capacity of CD4 T cells from WD- and standard laboratory diet-fed ApoE^−/− mice. We first examined the ability of stimulated CD4 T cells to catabolize glucose. CD4 T cells from WD-fed mice had a 3-fold reduction in their ability to consume glucose compared to those from standard laboratory diet-fed mice (Figure 3A). Using the Seahorse XF metabolism platform to measure glycolysis via extracellular acidification rates (ECAR), we found that CD4 T cells from WD-fed ApoE^−/− mice have a lower ECAR at basal conditions (Figure 3B) and a reduced capacity to respond to treatment with oligomycin and FCCP (Figure 3B). We also observed a similar pattern with oxidative phosphorylation, when we examined the oxygen consumption rate (OCR) of these cells (Figure 3C). This was accompanied by a 50% reduction in the ability of CD4 T cells from WD-fed ApoE^−/− mice to uptake 2-NBDG (Figure 3D).

Figure 3: CD4 T cells from atherosclerotic mice display reduced glucose metabolism.

Naive CD4 T cells isolated from standard laboratory diet (SLD)- and Western diet (WD)-fed mice were stimulated with αCD3 and αCD28 and the ability to consume glucose (A), undergo glycolysis (B) and oxidative phosphorylation (C), uptake 2-NBDG (D), upregulate glucose transporter Glut1 (E), CD4 T cells mitochondrial potential (F), and glycolytic reserve (G) were determined following αCD3 and αCD28 stimulation. Results are expressed as mean ± s.e.m. from one of two experiments (n=4 or 5). Statistically significant differences were at * P<0.05, *** P<0.001 and **** P<0.0001 (unpaired Student t-test).

To investigate the mechanisms underlying the diminished ability of CD4 T cells from WD-fed ApoE^−/− mice to uptake glucose, we quantified surface-level expression of the glucose transporter Glut1 on these cells following stimulation. We observed a 25% reduction in the expression of Glut1 on CD4 T cells from WD-fed mice following stimulation with anti-CD3 and -CD28 (Figure 3E), which may explain the reduction in glucose uptake by these cells. We next examined the mitochondrial potential of stimulated CD4 T cells from WD-fed mice and noted lowered, albeit statistically insignificant, mitochondrial health (Figure 3F), suggesting that the observed defects in CD4 T cell metabolism result from inhibition of proper glycolysis. To further dissect how glucose metabolism is altered in CD4 T cells from WD-fed mice, we assessed the glycolytic capacity and glycolytic reserve of these cells after stimulation using a GlycoStress test from Seahorse Biosciences. Our results show that both the glycolytic capacity and reserve of CD4 T cells from WD-fed mice were 25 and 60% lower, respectively, compared to cells from standard laboratory diet-fed counterparts (Figure 3G). Taken together, these results suggest that CD4 T cells exhibit altered capacities for proper uptake and catabolism of glucose during atherosclerosis.

CD4 T cells from atherosclerotic mice show alterations in glucose and fatty acid oxidation related genes

Next, we examined the expression of metabolism-related genes in naive CD4 T cells isolated from spleens and lymph nodes of standard laboratory diet- and WD-fed ApoE^−/− mice via gene-expression arrays. We observed a decrease in the enrichment of key glycolysis genes (Pfkfb1–4) and an increase in a key fatty acid β-oxidation gene (Cpt1a), encoding Carnitine palmitoyltransferase I, in naive CD4 T cells from WD-fed ApoE^−/− mice compared to standard laboratory diet-fed controls (Figure 4A). These results indicate that naive CD4 T cells from WD-fed mice are biased towards fatty acid oxidation for meeting their energy demands. Upon stimulation with anti-CD3, CD28, CD4 T cells from WD-fed mice exhibited increased expression of both Gapdh and G6pc, encoding the enzyme Glucose 6 phosphatase (which reverses the first step of glycolysis), and decreased expression of the gene Pfkfb3, which encodes 6-Phosphofructo-2-Kinase, compared to standard laboratory diet-fed ApoE^−/− mice (Figure 4B). This increase in Gapdh and G6pc and attenuation in Pfkfb3 mRNA accumulation was confirmed by qRT-PCR (Figure 4C). As a whole, these data illustrate profound alterations in CD4 T cell metabolism during atherosclerosis.

Figure 4: Naive CD4 T cells show changes in expression of glucose metabolism and fatty acid oxidation genes in atherosclerotic mice.

(A) Gene expression array in sorted naive CD4 T cells from standard laboratory diet (SLD)- and WD-fed mice without stimulation. (B) Gene expression array in CD4 T cells following stimulation with αCD3 and αCD28. (C) Relative mRNA expression levels of G6pc, Gapdh, and Pfkfb3 in isolated CD4 T cells from standard laboratory diet (SLD)- and WD-fed ApoE−/− mice following stimulation with αCD3 and αCD28. Data were analyzed using the 2^-ΔΔCt method and it was normalized to GUSB. Results are expressed as the mean ± s.e.m from one experiment (n=3). Statistically significant differences were at * P<0.05 (unpaired Student t-test).

Naive CD4 T cells from WD-fed ApoE^−/− mice display increased palmitate dependency

To determine the dependence of naive CD4 T cells on fatty acid oxidation, naive CD4 T cells from standard laboratory diet- and WD-fed ApoE^−/− mice were incubated with palmitate-BSA and oxidative phosphorylation was measured via the Seahorse assay. WD-fed ApoE^−/− naive CD4 T cells exhibited increased OCR in response to the free fatty acid palmitate-BSA (Fig. 5A), suggesting that WD feeding elicits increased reliance on palmitate. However, upon performing a mitochondrial stress test using oligomycin/FCCP/Antimycin A and Rotenone, our results showed no significant differences between the kinetics of OCR in naive CD4 T cells from standard laboratory diet- (Fig. 5B) and WD-fed (Fig. 5C) ApoE^−/− mice. These data hint that the mitochondrial redox potential of naive CD4 T cells from both mice groups is similar, despite their differential reliance on fuel sources. Thus, T cells within the lipid-rich environment of the plaque may have adopted a better usage of lipids as a source of energy. In support of this, we examined the expression of genes involved in fatty acid and cholesterol metabolism in T cells from our published work on mouse atherosclerotic aorta using scRNA-Seq²⁷. We found moderate to high expression of genes related to cholesterol metabolism (ABCA1, ABCG1, ACAT1, SOAT1), yet little to no expression of fatty acid metabolism genes CD36, FABP4, or Cav-1 in naive or effector CD4 T cells (Online Figure I). Collectively, our data suggest that tonic TCR signaling within T cells of the lipid-rich atherosclerotic plaque can modulate T cell metabolism and induce functional changes in T cell proliferation, activation, and exhaustion.

Figure 5: Naive CD4 T cells from WD-fed atherogenic mice display increased reliance on palmitate compared to naive CD4 T cells from standard laboratory diet-fed mice.

OCR of palmitate-BSA treated naive CD4 T cells isolated from standard laboratory diet (SLD)- and WD-fed mice (A) at baseline. Fatty acid oxidation assay of naive CD4 T cells isolated from standard laboratory diet-fed (SLD) (B) and WD-fed (C) mice using oligomycin, FCCP, and antimycin/rotenone in response to etomoxir, palmitate-BSA, and palmitate-BSA+Etomoxir. Results are expressed as mean s.e.m. from one experiment. Statistically significant differences were at **P<0.01 (unpaired Student t-test).

Naive CD4 T cell frequencies are reduced in human subjects with severe CVD

To characterize how naive CD4 T cell frequencies or phenotypes change with disease risk, we assayed naive CD4 T cells from peripheral blood mononuclear cells (PBMCs) of subjects undergoing a medically-necessary coronary angiography. CVD severity in each subject is quantified using the well-characterized clinical Gensini score²⁸. In humans, naive T cells are phenotyped as CD45RA⁺CCR7⁺ cells that express low levels of CD95 (Fas). Examining the phenotype of naive CD4⁺ T cells (CD4⁺CD3⁺CD45RA⁺CCR7⁺CD95^lo) revealed a significant reduction (~ 50%) in naive CD4 T cell percentages in individuals with high CVD (Gensini score >30) compared to low CVD (Gensini score<10) (Figure 6A, B). We next examined if CD4 T cells from high CVD patients were defective in their responses to stimulation compared to individuals with low CVD. Stimulation of naive CD4 T cells from high CVD patients with αCD3/αCD28 coated beads revealed a diminished proliferative capacity, as illustrated by a 10-fold less CTV dilution, in comparison to low CVD CD4 T cells (Figure 6C, D). These results suggest that similar to WD-fed ApoE^−/− mice, naive CD4 T cell responses to activation are impaired in high CVD patients. Naive CD4 T cell frequencies inversely correlated with clinical Gensini score, which ranges from 0–100 (Figure 6E). The age of individuals with high CVD were significantly older than those with low CVD. Indeed, age is a well-established risk factor for CVD²⁹. Plotting of naive CD4 T cell frequencies by age revealed no correlation. Therefore, we assert that advanced atherosclerosis severity elicits reduced naive CD4 T cell frequencies.

Figure 6: Human subjects with high CVD have fewer naive CD4 T cells and these cells have reduced proliferative capacity.

PBMC from human subjects with documented low CVD (Gensini<10) and high CVD (Gensini>30) were stained with a T cell-focused panel, and CyTOF mass cytometry was performed. Flow cytometry plots (A) and bar graphs showing the percentages of naive CD4 T cells (CD4+CD3+CD45RA+CCR7+CD95lo) from low (n=14) and high (n=14) CVD individuals normalized to absolute lymphocyte count (k/μL) (B). Results are expressed as mean ± s.e.m. from 14 low and 14 high CVD individuals. Naive CD4 T cells were isolated from PBMCs of low (n=7) and high (n=6) CVD individuals and stimulated with αCD3/αCD28 coated beads for 48hrs. Flow cytometry plots (C) show the percentage of proliferating CD4 T cells in the presence (red) or absence (blue) of Dynabeads using CTV proliferation dye and the percentage of cells divided (D). Results are expressed as mean ± s.e.m. from two independent experiments. Pearson correlation of Gensini score (E) and age (F) by frequencies of CD4⁺ T_N cells of CD45⁺ cells, normalized to lymphocyte counts (k/μL). Statistically significant differences were at * P < 0.05 (unpaired Student’s t-test).

Discussion

In this study, we show that atherogenesis influences the phenotype, function and metabolic potential of CD4 T cells in individuals with severe CVD disease and mouse model of atherosclerosis. The association between metabolism and atherosclerosis is known. CAD patients display an abnormal fatty acid and glucose metabolism^{30, 31}. More recently, it is becoming apparent that metabolic changes occur in macrophages present in the plaque during atherosclerosis. The hypoxic environment of the atherosclerotic plaque induces an increase in the glycolytic potential of plaque-resident macrophages^{32, 33}. This increase in glycolysis results in attenuated cell viability in response to inflammation. Moreover, macrophages from individuals with atherosclerotic coronary artery disease display abnormal glucose utilization and hyperinflammation compared to healthy controls³⁴. Still, the metabolic changes in CD4 T cells during atherosclerosis and their effects on CD4 T cell functions have not been characterized. Herein, we show that CD4 T cells from both humans and mice display defective glycolysis under atherosclerotic conditions, which attenuates the ability of these cells to proliferate and produce IFNγ.

Blocking glycolysis has been previously shown to result in loss of IFNγ production by CD4 T cells^{15, 35, 36}. Additionally, Gapdh has been shown to interact with mRNA transcripts for IFNγ and block translation of the protein¹⁹. Our results are consistent with these findings and demonstrate that CD4 T cells from ApoE^−/−, WD-fed mice display a reduction in IFNγ production, increased expression of Gapdh, and a reduction in glycolysis. These results suggest that atherosclerosis disrupts glycolysis in CD4 T cells and affects their ability to perform effector functions.

In the Multi-Ethnic Study of Atherosclerosis (MESA) cohort, a decrease in naive CD4 T cells is associated with subclinical atherosclerosis⁴. Our results show that naive CD4 T from atherosclerotic subjects display signs of exhaustion via increased expression of PD-1, which has been shown to alter the metabolic phenotype and function of CD4 T cells³⁷. Indeed, stimulation of PD-1 results in a metabolic switch from glycolysis to lipolysis in CD4 T cells, by increasing expression of Cpt1a. Consistent with these findings, our gene array results suggest a decrease in the expression of glycolysis-related genes and an increase in fatty acid oxidation genes, including Cpt1a, by CD4 T cells. We therefore hypothesize that PD-1 ligation is at least partially responsible for this switch.

Changes in the metabolic phenotypes of T cells have been documented in other inflammatory diseases. In patients with rheumatoid arthritis, CD4 T cells exhibit reduced aerobic glycolysis, failure to generate ATP, and increased levels of apoptosis, as a result of decreased Pfkfb3 expression³⁸. In this study, we observe a concomitant reduction in expression of Pfkfb3 in CD4 T cells from atherosclerotic mice, which could explain the increase rates of apoptosis observed in these cells. Moreover, when metabolic programs do not meet energy requirements of the cell, cells will readily undergo apoptosis. In obese patients and in mouse models of obesity, CD4 T cells show altered fatty acid and glucose metabolism, which biases the production of Th17 cells over Treg cells^{21, 39}. This effect is driven by an upregulation of lipid kinase acetyl CoA carboxylase. Similarly, in multiple sclerosis, inhibition of carnitine palmitoyltransferase 1 (Cpt1), a mitochondrial enzyme responsible for shuttling fatty acids from the cytosol to the mitochondria for β-oxidation⁴⁰, reduces disease severity⁴¹. We show here that similar metabolic shifting occurs in CD4 T cells in atherosclerosis.

Atherosclerosis is characterized by an increase in effector memory-like CD4 T cells within atherosclerotic ApoE^−/− mice and humans with coronary artery disease⁴². Moreover, in peripheral blood, there is a positive correlation between disease severity and memory-like CD4 T cells⁴. Cholesterol signaling during atherosclerosis has been shown to alter T cell functioning and activation in mice^{43, 44}, including phenotypic switching of Treg cells into T follicular helper (Tfh) or T effector cells¹². While naive CD4 T cells that fail to divide or become anergic could be hypothesized to initially benefit atheroprogression, it may be more likely that this scenario indicates a disruption in the ability of the adaptive immune system to respond to stimuli. Thus, one can predict that, in individuals with severe CVD, immune responses to other insults such as viruses, bacteria, or cancer cells, would likewise be defective. Supporting this notion is a positive correlation between breast cancer and hypercholesterolemia in breast cancer patients^{45, 46}. Whether naive CD8 T cells are similarly affected to CD4 T cells during atherosclerosis remains to be explored.

The results obtained from this work provide preclinical evidence for defects in CD4 T cell function and metabolism during atherosclerosis and may aid the design of new therapeutic targets for optimizing T cell responses against the disease.

Highlights.

• Naive CD4 T cells from atherosclerotic ApoE^−/− mice fed a western diet display lower capacity to proliferate and produce IFNγ following stimulation compared to cells from standard laboratory diet-fed mice.

• CD4 T cells from atherosclerotic mice have a defect in glucose metabolism and show differential expression of metabolism genes.

• Naive CD4 T cells in patients with high CVD have reduced capacity to respond to stimulation and uptake glucose.

• The lipid-rich environment of the atherosclerotic plaque changes the metabolic potential of CD4 T cells.

Abbreviations

CVD

Cardiovascular disease

ApoE^−/−

Apolipoprotein E knockout mice

WD

Western diet

Treg

Regulatory T cells