Hyperglycaemia does not affect antigen-specific activation and cytolytic killing by CD8+ T cells in vivo

Metabolism is of central importance for T cell survival and differentiation. It is well known that T cells cannot function in the absence of glucose, but it is less clear how they respond to excessive levels of glucose. In the present study, we investigated how increasing levels of glucose affect T-cell-mediated immune responses. We examined the effects of increased levels of glucose on CD8+ T-cell behaviour in vitro by assessing activation and cytokine production, as well as oxygen consumption rate (OCR), extracellular acidification rate (ECAR) and intracellular signalling. In addition, we assessed in vivo proliferation, cytokine production and cytolytic activity of cells in chemically induced diabetic C57BL/6 mice. Elevated levels of glucose in in vitro cultures had modest effects on proliferation and cytokine production, while in vivo hyperglycaemia had no effect on CD8+ T-cell proliferation, interferon γ (IFNγ) production or cytolytic killing.


Introduction
Glucose is one of the important nutrients availabletoTcells, and is mostlytakenupviaGlut1 in these cells [1].Glut1 is up-regulated upon activation, which leadstoincreased glucose uptake and glycolysis to promote growth, proliferation, cell survival anddifferentiation [2].Asaresultofthis,Glut1deficiency in Tcells decreases effector cell expansion and the ability to induce inflammatory disease in vivo [1].Recent studies have clarified how T cellsup-regulate their anaerobic glycolysis during a rapid effector response, and how this type of rapid but low efficiency generation of energy must be replaced by engagement of the mitochondria and fatty acid oxidation [3] or the ability to sustain high levels of ATP generation through elevated glycolysis [4] for the cellstodifferentiate into long-lived memory T cells.Incontrast,Foxp3 + Treg favours fatty acid oxidation [5,6],and induction of anergy in effector T cellsreduces their metabolism [7].The metabolism of T cellsisadrugable target, and indeed the mammalian target of rapamycin (mTOR) is at the centre of the cell response to nutrient availability anddictates cell decisions to grow anddifferentiate [8][9][10].
We were interested in how an abundance of glucose, as is the case in diabetes, affects the adaptive immune system.As competition for resources can lead to suppression of immune responses [11],while the elevated presence of glucose has been reported to both boost the immune responses to tumours [11] and enhance the survival of mice after administration of lethaldoses of influenza virus [12],itseemed likelythatelevatedlevelsofglucose could enhance immune responses.Inorder to provide sufficient levels of glucose, many cell culture media contain 'diabetic'levelsofglucose, with concentrations often in the 12-15 mM range or even higher, which is well above the levelsseeninhealthy people(below 6 mM in the fasting state and below 7.8 mM 2hp o s t p r a n dial).On the other hand,p a t i e n t sw i t hdiabetes have numerous and more serious infections than the healthy control subjects [13,14],anddecreased responses to vaccination [15,16] indicating that elevated glucose levels do not boost immune responses in vivo.
Here, we investigated how increasing levelso fg lucose in vitro, varying from a low but physiologicallyn o r m oglycaemic concentration of 5.5 mM (1 g/l)uptoanemphaticallyhyperglycaemic environment of 25 mM (4.5 g/l), affected T-cell behaviour.Wehavealso investigated the in vivo effects of hyperglycaemia (ranging between 15 and 25 mM), on OVA-specific CD8 + T-cell proliferation, cytokine production and cytolytic killing in streptozotocin (STZ)-induceddiabetic C57BL/6mice.

Methods
Mice OT-I were bred at the University of Cambridge and maintained under specific pathogen-free conditions.Male C57BL/6mice(Charles River) were used between 6 and10weeks of age.Mice were housed in IVC with free access to standard chow and water.The present study was carried out in accordance with U.K. Home Office Regulations (project licence number 80/2442and 70/8442).

STZ-induced diabetes
Male C57BL/6m i c ew e r eg i v e nS T Z (Sigma, 40 µg/gb o dyw e i g h t )dissolved in citrate buffer (pH4 .5)in t r a p e r itoneallyf o r5d ays.Diabetes normally developed within 10-14 days with no signs of STZ-inducedl ymphopaenia (Supplementary Figure S1).Glycosuria was detected using Diastix strips (Bayer Diagnostics) anddiabetes confirmed by a blood glucose measurement of >13.3 mM, using a Breeze2 blood glucose meter (Bayer).

Antibodies and flow cytometry
Cellswereresuspended in FACSbuffer(PBS+0.5% BSA) filtered through 30-µm CellTrics filters (Partec), incubated with Fc block (eBioscience), stained with antibody, washed and resuspended in PBS.7AAD (BD Bioscience) was used to assess cell death.Data were collected on a Cyan Cytometer (DAKO)and analysed using FlowJo(T reeStarInc.).For intracellular cytokine staining, the cellswerestimulated with PMA(50 ng/ml)and ionomycin (2000 ng/ml)for 5 h.Brefeldin A (5 µg/ml)w a sa dded for the last 3 h.After surface marker staining, the cellsw e r ew a s h ed,fi x ed, permeabilized (intracellular staining kit, eBioscience), and stained for detection of cytokine.

T-cell activation for functional assays
Cellswereisolated from spleen andlymph nodes and cultured in low glucose (5.5 mM) DMEM with 10% FBS, 1% penicillin-streptomycin, and β-mercaptoethanol supplemented with additional glucose as indicated.Lymphocytes(2 × 10 5 )werestimulated as appropriate (see below) for 3 days in the presence of the indicated glucose concentrations at 37 • C with 5% CO 2 .OT-I cellswerestimulated either with the OVApeptideSIINFEKL or the lower affinity altered peptide ligand SIIGFEKL (both from Sigma) as indicated.Proliferation was assessed by CFSE staining (5 µM).After gating on CD8 + Tcells, the percentage of proliferating cellsineachpopulation was determined.Supernatantcytokine analysis was performed with cytometric bead array (eBioscience) as recently described [17],a n d ATP content in cultures was assessed using the CellTiter-Glo R Luminescent Cell Viability Assay (Promega) in accordance with the manufacturers' instructions.The cellswerecultured in 96-well plates at a concentration of 2.5 × 10 4 cellsperwell in 100 µl of the indicated culture medium.For analysis, the supernatants were transferred into a 384-well Optiplate (PerkinElmer) andluminescence read using a Mithras LB 940 (Berthold Te chnology).

Measurements of T-cell metabolism
Naïve OT-I CD8 + Tcellswereisolated using MACSbeads(Miltenyi) according to the manufacturer's instructions.For studies of activated cells, OT-I splenocytes were cultured for 5days in the presence of 10 ng/ml SIINFEKL peptide and1 0U / mlI L-2( P e p r o T e c h ) .N a ïve cellsw e r es e e ded in a 96-well seahorse plate at 3 × 10 5 cellsp e rw e ll,a n d activated cellswereseeded at 1.5 × 10 5 cellsperwell,and analysed using the Mitostress kit (Agilent Technologies) according to the manufacturers' instructions.Seahorse assay medium (Agilent Technologies) was supplemented with the indicated glucose concentration, 1 mM glutamine and1mM pyruvate.Oligomycin was administered at 1.5 µM, FCCPat1 µM and rotenone/antimycin A at 1 µM (all from Agilent Technologies).Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured using a XF96 Seahorse analyser.ATP turnover was derived from the difference in OCRs between basal respiration and inhibition after oligomycin administration according to the manufacturers' instructions.

Zap phosphorylation assay
CD8 + Tc e llsw e r es o r t e d using MACS( Miltenyi), seeded in V-bottom plates (2 × 10 5 /well)a n d incubated with stimulating antibodies (anti-CD3, clone 145-2C11,2µg/ml and anti-CD28,clone 37.51, 10 µg/ml)fortheindicated time with the indicated concentration of glucose at 37 • C.Afterstimulation, the cellswereimmediatelyfixed in 4% PFA for 30 min, then washed in PBS and stored in ice-cold methanol at -20 • C,s t a i n e d with anti p-Zap319a n d detected with anti-rabbit IgG Fab2 Alexa 647 (Molecular Probes).

In vivo CTL assay
Male C57BL/6 mice were immunized with SIINFEKL peptidea t2 5 µg/dose emulsified in IFA (Sigma) sc in the left haunch.Ten days later, targets were injected.Syngeneicsplenocytes were either peptide-pulsed (100 nM,30 min, 37 • C)and subsequently labelled with 10 µMCFSE, or non-pulsed andlabelled with 1 µMCFSE.The splenocyte populations were then mixed at 50:50,and10 7 cellswereinjected in the tail vein.Twenty-four hours later, the inoculum draining and control sideinguinallymph nodes were collected,and the ratio of CFSE hi compared with CFSE intermediate cellscompared with non-immunized controlstocalculate % of specific killing of peptide-pulsed targets.

Statistical analysis
Differences between groups were tested using the Student's t test, significant P-values are indicated with *(P≤0.05),**(P≤0.01),***(P≤0.001)or****(P≤0.0001).Comparison of multiplegroupsintheSeahorseassayswasperformed using two-way ANOVAf o llowed by Dunnett'sm u ltiplec o m p a r i s o nt e s t .A ll data analyses were performed using GraphPad Prism 7 software.

Results
OT-I cell proliferation to high affinity, but not low-affinity peptide, is increased when glucose levels are raised Weassessed the in vitro proliferation of OT-I cells, which are CD8 + Tcellsreactivetoovalbumin peptide257-264 (SIINFEKL) presented on C57BL/6 MHC class I molecule H2K b .Increasing levelsofglucose resulted in increased proliferation of these cellsinresponsetotheircognatepeptide(Figure1a, top left panel,withrepresentativeCFSE traces in the top right panel).However, this proliferative change with increasing levelsofglucose were not seen with the low-affinity peptide ligand SIIGFEKL (Figure 1a, bottom left panel)ormedium alone (Figure 1a, bottom right panel), indicating that increased glucose did not alter the threshold for activation.Cellsc u ltured in an excess of culture medium did not grow more in higher concentrations of glucose, as reflected in the ATP content in cultures at different time points (Figure 1b).In contrast with increased proliferation seen in high glucose cultures, we saw a decrease in interferon γ (IFNγ)production in cultures with glucose levelsof25 mM (Figure 1c, left panel), and no IFNγ produced in response to the altered peptide ligand at any glucose concentration (Figure 1c, right panel).Production of GM-CSF, TNF, IL-10, IL-17 andIL-2appeared unaffected,withatrend towardsincreased production at moderate levelso fg lucose (10-15 mM)a n d a decrease at high levels( 2 0-25 mM)( F i g u r e1d).To control the changes in osmolarity caused by increased glucose concentrations, we included 20 mM mannitol,asugarwithsimilar molecular weight to glucose but not metabolized by cells, added to a 5.5-mM glucose base medium.

Elevated levels of glucose do not alter OCR in naïve or activated OT-I cells
To assess whether increased glucose concentration changed the metabolic activity of the OT-I cells, we assessed their OCRs (Figure 2a,b, left panels) and ECAR (Figure 2a,b, middlep a n e ls) in response to drugs that affect the electron transport chain [3].Oligomycin inhibits the ATP synthase stopping mitochondrial ATP generation, FCCPisa protonophore which uncouples ATP synthesis from the electron transport chain by letting H + ions into the matrix independent of the ATP synthase whilerotenone/antimycin A inhibit the complex I and III respectively, leading to complete shut down of the electron transport chain.Wefound that increasing the levelsofglucose modestlyincreases the ECAR of naïve cellsinadose-dependent manner (Figure 2a, middlepanel)butdoes not affect the alreadyhigher ECAR of activated cells (Figure 2b, middlepanel).The ATP turnover, determined by the difference in OCR between basallevelsand the levelsafteroligomycin inhibition of the ATP synthase, were unaffected by glucose concentration in both naïve cells(Figure2a,rightpanel)and activated cells(Figure2b,rightpanel).Furthermore, the immediate activation of T cellsasdetermined through Ca 2+ fluxing (Figure 2c) andZap-70 phosphorylation (Figure 2d)was also unaffected by glucose concentration.

Hyperglycaemia does not affect OVA-specific proliferation, IFNγ production or CTL killing in vivo
To ass ess w het her any of t he mo dest differences recorded in vitro were of importance in vivo,weperformed experiments in C57BL/6 mice, which had been rendereddiabetic using low-dose STZ injection.Control or hyperglycaemic C57BL/6micewereimmunized with OVAintheleft haunch, and the proliferation of injectedCFSE-labelledOT-I cells was assessed in the inoculum-draining left inguinall ymph nodea n d the control right inguinall ymph node.There was no differenceinhowwell the transferredOT-I cellsproliferated in the diabetic hosts compared with control hosts (Figure 3a).To further assess the properties of the activatedOT-I cells, the cellsfromthelymph nodes were restimulated briefly in vitro with PMAand ionomycin, andIFNγ production was recorded.Therewasnodifference between the groups (Figure 3b).The inoculum-draining and non-draining lymph nodes from immunized mice were also restimulated with SIINFEKL peptide for measurement of production of other cytokines.There was no difference in the production of IL-2, IL-17, IFNγ, GM-CSF, TNF, IL-6orIL-1α (Figure 3c).Wealso plotted the levelsofcytokine and proliferation against measured blood glucose level at the end of the experiment, but found no correlation in any experiment (results not included).In vivo cytolytic T lymphocyte (CTL) assays demonstrated no difference in the capacity for OVA-specific CTL cytotoxicity in diabetic hosts compared with controls (Figure 3d).

Discussion
Wehaveinvestigated how levelsofglucose in the diabetic range affect T-cell responses both in vitro and in vivo.We find that hyperglycaemia has modest effects on proliferation and cytokine production in vitro, which could simply reflect the fact that an in vitro culture has to adapt to the amount of nutrient availableinthewell.When the cells are cultured in excess volumes of media, as in the cultures prepared to assess ATP content, no difference in the accumulation of ATP could be detected in cultures with higher levelso fglucose.To support this, we find that OCR and A TPturnoverofbothna ïve and activatedOT-I cellsareunaffected by the hyperglycaemic conditions, and that initial intracellular activation events after T-cell receptor (TCR) ligation are unaltered by hyperglycaemia.Interestingly, we find that naïve OT-I cells demonstrate increased ECAR in hyperglycaemic conditions, and it remains to be determined if this has any biological significance.In vivo CD8 + T-cell proliferation and cytokine production was unaffected in diabetic C57BL/6, as was in vivo cytolytic killing.This finding is in contrast with a previous study, which demonstrated greater survival of tumour cellsinSTZ-induceddiabetic mice [18].Itishoweverpossiblethattheelevated glucose levelsindiabetic mice affect not only CTL but also the tumour cells, and this may contribute to their greater survival.Animportantpointtomakehereisthedifference between STZ protocols.Many groups administer one high dose of 200 µg/gbodyweight [18][19][20] and may see a resulting down-regulation of immune responses.STZ is a glucosamine-nitrosourea that causes DNA damage, and is particularlytoxictoβ-cellsasitistakenupviathe Glut2 transporter, which is expressed in β-cellsand to a lower extent in kidney, liver and small intestine.However, at high doses STZ canbetoxictoothercell types as well,whichisdemonstrated by the lymphopenia seen in high-dose treated mice [19].Theinjection protocol used in our studyusesrepeatedlow dose injections of 40 µg/gbodyweight, which avoids off target effects, and no lymphopenia was recorded as shown in the Supplementary Data (S1).
Health complications such as changes in immune reactivity in diabetes are caused by a complex network of interacting mechanisms, and it is difficulttodetermine which effects are caused by excess glucose itself, and how that effect is exerted.Hyperglycaemia has effects on the innate immune system in that it can inhibit neutrophil migration, phagocytosis, superoxideproduction and microbial killing [21,22] anddecrease the production of antimicrobial peptides [23].Neutrophilshavebeenreported to take up less antigen in a hyperglycaemic host [24],whichcould indirectly lead to depressed T-cell responses, as they may not receive optimal antigen presentation.Hyperglycaemia also affects the ability of tolerogenic DC to induce generation of antigen-specific tolerance in T cells [25],and there are reports that hyperglycaemia can induce expression of proinflammatory cytokines like IL-17 in CD4 + Tcells [26,27].All these effects on immune cells may contribute to altered immune status in diabetic patients.
Inthepresentstudy, we demonstrate that antigen-specific proliferation and killing by OT-I cellsareunaffected by hyperglycaemia in vivo,indicating that an abundance of glucose does not in itselfeithersuppressorboostshort-term T-cell responses.Itremainstobedetermined whether long-term effects of hyperglycaemia may alter antigen presentation to T cells, or the maintenance of the T cellsthemselves, thus affecting the formation and maintenance of T-cell memory.

Figure 1 .
Figure 1.Effects of increasing levels of glucose in culture medium on OT-I cell proliferation and cytokine production (a) Proliferation was assessed by CFSE dilution in OT-I cells in response to SIINFEKL (left panel), the low-affinity altered peptide SIIGFEKL (middle panel) and no peptide control (right panel).(b) ATP content at different time points in response to stimulation with anti-CD3 and anti-CD28 antibody in different concentrations of glucose or mannitol control.(c)IFNγ production was assessed using intracellular staining in cultures with SIINFEKL peptide (left) and low affinity altered peptide (right).(d) Cytokine production in OT-I cultures in response to SIINFEKL peptide in the presence of increasing concentrations of glucose or 25 mM mannitol (m) as an osmolarity control was assessed using cytokine bead array.The results are representative of at least three experiments.Differences between groups were tested using the Student's t test, significant P-values are indicated with *(P≤0.05),**(P≤0.01),***(P≤0.001) or ****(P≤0.0001).If no P-value is indicated, there was no significant difference between the groups.

Figure 2 .
Figure 2. Effects of increasing levels of glucose on OT-I cell metabolic activity and intracellular signalling (a) OCR was determined for naïve OT-I cells in response to compounds that target different parts of the mitochondrial electron transport chain ((a), left panel).Oligomycin, FCCP and rotenone/antimycin A were administered at the indicated time points (indicated in the figure with arrows numbered 1, 2 and 3 respectively) followed by four separate measurements for each condition.The ECAR was determined after addition of FCCP for maximum activation ((a), middle panel).Each data point indicate the average of the four measurements in one well.Eleven replicates per condition were assessed.ATP turnover was calculated from the difference in OCR between the basal and oligomycin stimulated conditions ((a), right panel).(b) OCR was determined for activated OT-I cells as described above ((b), left panel), as was ECAR ((b), middle panel) and ATP turnover ((b), right panel).(c)C a 2+ fluxing and (d) Zap-70 phosphorylation were determined in cells after activation in culture media with different concentrations of glucose using flow cytometry.The data are representative of at least two experiments.Differences between groups were determined through two-way ANOVA followed by Dunnett's multiple comparison test with P-values below 0.05 considered significant.

Figure 3 .
Figure 3. Hyperglycaemia does not affect immune responses in vivo in STZ-induced diabetic C57BL/6 mice (a) In vivo proliferation of CFSE-labelled OT-I cells was assessed in response to an inoculum containing OVA in control mice compared with mice rendered diabetic through injection of STZ.The left panel shows the result of one experiment with each dot representing one mouse, assessing either the inoculum-draining lymph node (dILN) or the control inguinal lymph node (cILN).The right panels show representative FACS plots.(b)IFNγ production in cells from the experiment described in (a), stimulated ex vivo with PMA for 4 h.The left panel shows the result of one experiment with each dot representing one mouse, the right panels show representative FACS plots.(c) Cytokine production in lymphocytes from mice immunized with OVA as above and restimulated in vitro with SIINFEKL peptide.(d) In vivo killing assay comparing the antigen-specific killing of SIINFEKL-pulsed syngeneic splenocytes during a 24-h period in mice immunized with the same peptide 8 days previously.The left panel shows the result of one experiment with each dot representing one mouse; the right panels show representative FACS plots.The results are representative of at least three experiments, and differences between groups were tested using the Student's t test.