Costimulatory CD226 Signaling Regulates Proliferation of Memory-like NK Cells in Healthy Individuals with Latent Mycobacterium tuberculosis Infection

It is now widely accepted that NK cells can acquire memory, and this makes them more effective to protect against some pathogens. Prior reports indicate memory-like NK cells (mlNKs) in murine model of Mycobacterium tuberculosis (Mtb) as well as in healthy individuals with latent TB infection (LTBI). The increased expression of CD226 was evident in mlNKs from LTBI+ people after stimulation with γ-irradiated Mtb (γ-Mtb). We thus evaluated the contribution of costimulatory CD226 signaling in the functionality of mlNKs in LTBI+ people. We found that blockade of CD226 signaling using the antibody- or CRISPR/Cas9-mediated deletion of the CD226 gene in NK cells diminished the proliferation of mlNKs from LTBI+ people. Blocking CD226 signaling also reduced the phosphorylation of FOXO1 and cMyc expression. Additionally, cMyc inhibition using a chemical inhibitor reduced proliferation by mlNKs from LTBI+ people. Moreover, blocking CD226 signaling reduced glycolysis in NK cells, and the inhibition of glycolysis led to reduced effector function of mlNKs from LTBI+ people. Overall, our results provide a role for CD226 signaling in mlNK responses to Mtb.


Blockade of CD226 Signaling Reduces Proliferation by mlNKs from LTBI+ People
The co-stimulatory CD226 signaling contributes to NK cell function and tumor clearance [32,33]. Previous report indicates that CD226 signaling contributed to the expansion of Ly49H+ NK cells after MCMV infection and to the optimal differentiation of Ly49H+ memory NK cells [35]. It has been previously reported that CD226 expression was more on murine memory-like NK cells in a mouse model of TB [17] and these memory-like NK cells from LTBI+ donors proliferated and showed enhanced degranulation and IFN-γ production [17] (Figure S1A-D). We hypothesized that CD226 signaling contributes to the proliferation of mlNKs from LTBI+ people. To investigate this, we first stimulated peripheral blood mononuclear cells (PBMC) from LTBI+ (n = 5) people with γ-irradiated Mtb (γ-Mtb) for 5 days and assessed CD226 expression on mlNKs ( Figure 1A-C). CD56+CD27+ mlNKs from LTBI+ donors (n = 5) proliferated in response to γ-Mtb stimulation ( Figure 1D-H) compared to LTBI− (n = 6) people ( Figure S1A-H) in accord with published report by others [17]. MFI of CD226 was higher on mlNKs after γ-Mtb stimulation, while frequencies of CD226+ mlNKs were comparable ( Figure 1A-C). Next, we inhibited CD226 signaling using CD226 blocking antibody and the NK-cell-specific deletion of CD226 gene using CRISPR/Cas9 ( Figure S2A-C). Both approaches led to significantly reduced proliferation of mlNKs from LTBI+ people while isotype antibody had no measurable effect on NK cell proliferation ( Figure 1I,J). Blocking CD226 antibody treatment also led to lesser IFN-γ production, degranulation as well as activation profile measured by CD69 and HLA-DR expression by mlNKs from LTBI+ people compared to isotype-antibody-treated counterparts ( Figure 1K,L and Figure S1I-K). Blockade of CD226 with antibody also led to reductions in protein levels of IFN-γ, TNF-α, MIP-1β and MCP1 in the culture supernatants on day 5 when measured by ELISA ( Figure S3A-E). Together, these results suggest a role of CD226 signaling in the proliferation of mlNKs from LTBI+ people (n = 5). on day 5 when measured by ELISA ( Figure S3A-E). Together, these results suggest a role of CD226 signaling in the proliferation of mlNKs from LTBI+ people (n = 5).  . Mean ± s.d. two-sided Student's t-test or ANOVA. *** p < 0.001, **** p < 0.0001.

Inhibition of CD226 Signaling Leads to Reduction in FOXO1 Phosphorylation in NK Cells from LTBI+ People
Recent report demonstrated that the activation of CD226 signaling led to the phosphorylation of FOXO1 and its proteasomal degradation [33]. FOXO1 is a negative regulator of NK cell function [34] thus CD226 signaling results in NK cell activation, by removing a negative regulator, FOXO1. We asked whether similar scenario occurs in NK cells from LTBI+ people when CD226 signaling is engaged. PBMC taken from LTBI+ people (n = 5) and stimulated with γ-Mtb were treated with blocking CD226 antibody or isotype matched antibody and NK cells were purified on day 2 to assess the phosphorylation of FOXO1. We noted the increased phosphorylation of FOXO1 in the isotype-antibody-treated group (Figure 2A-C). Conversely, blocking CD226 led to the lower phosphorylation of FOXO1, suggesting the increased presence of negative regulator FOXO1 in this group (Figure 2A-C). Overall, in agreement with a previous report in NK cells [33], these results indicate that blocking CD226 signaling reduced the phosphorylation of FOXO1 in NK cells from LTBI+ people (n = 5).

Inhibition of CD226 Signaling Leads to Reduction in FOXO1 Phosphorylation in NK Cells from LTBI+ People
Recent report demonstrated that the activation of CD226 signaling led to the phosphorylation of FOXO1 and its proteasomal degradation [33]. FOXO1 is a negative regulator of NK cell function [34] thus CD226 signaling results in NK cell activation, by removing a negative regulator, FOXO1. We asked whether similar scenario occurs in NK cells from LTBI+ people when CD226 signaling is engaged. PBMC taken from LTBI+ people (n = 5) and stimulated with γ-Mtb were treated with blocking CD226 antibody or isotype matched antibody and NK cells were purified on day 2 to assess the phosphorylation of FOXO1. We noted the increased phosphorylation of FOXO1 in the isotype-antibodytreated group (Figure 2A-C). Conversely, blocking CD226 led to the lower phosphorylation of FOXO1, suggesting the increased presence of negative regulator FOXO1 in this group (Figure 2A-C). Overall, in agreement with a previous report in NK cells [33], these results indicate that blocking CD226 signaling reduced the phosphorylation of FOXO1 in NK cells from LTBI+ people (n = 5).

Inhibition of CD226 Signaling Reduces cMyc Levels in NK Cells from LTBI+ People
We next explored molecules involved in the proliferation of mlNKs that may be affected by CD226 blockade. cMyc is previously reported to be a target of XBP1s and integrates signals from endoplasmic reticulum (ER) stress to drive NK cell responses after MCMV infection and tumors [36]. PBMC from LTBI+ people (n = 5) were stimulated with γ-Mtb and treated with isotype antibody or CD226 blocking antibody. NK cells were purified at day 1 and day 2 post treatment. When we measured ER stress related genes (day 1 post treatment) such as Ire1α, Xbp1, we did not observe significant differences between isotype antibody and CD226 blocking antibody regimens ( Figure S4B). While the expression levels of Chop, Atf4 and Bip were elevated in blocking CD226 antibody groups ( Figure S4B). We also measured molecules associated with Wnt signaling [37] and found no differences in Tcf7 gene expression ( Figure S4A). Dll1 expression was elevated while Jag1 expression was reduced upon CD226 blocking antibody groups ( Figure S4A). Axin 2 expression was reduced in both isotype and CD226 blocking antibody groups ( Figure S4A). Notably, Myc gene which is associated with cellular proliferation was higher in isotype antibody group, while Myc expression was reduced when CD226 signaling was blocked ( Figure 3A). We also validated these results using Western blotting ( Figure 3B-D) performed on day 2 post treatment. Overall, these results indicate lower cMyc levels in NK cells after CD226 blockade.

Inhibition of CD226 Signaling Reduces cMyc Levels in NK Cells from LTBI+ People
We next explored molecules involved in the proliferation of mlNKs that may be affected by CD226 blockade. cMyc is previously reported to be a target of XBP1s and integrates signals from endoplasmic reticulum (ER) stress to drive NK cell responses after MCMV infection and tumors [36]. PBMC from LTBI+ people (n = 5) were stimulated with γ-Mtb and treated with isotype antibody or CD226 blocking antibody. NK cells were purified at day 1 and day 2 post treatment. When we measured ER stress related genes (day 1 post treatment) such as Ire1α, Xbp1, we did not observe significant differences between isotype antibody and CD226 blocking antibody regimens ( Figure S4B). While the expression levels of Chop, Atf4 and Bip were elevated in blocking CD226 antibody groups ( Figure  S4B). We also measured molecules associated with Wnt signaling [37] and found no differences in Tcf7 gene expression ( Figure S4A). Dll1 expression was elevated while Jag1 expression was reduced upon CD226 blocking antibody groups ( Figure S4A). Axin 2 expression was reduced in both isotype and CD226 blocking antibody groups ( Figure S4A). Notably, Myc gene which is associated with cellular proliferation was higher in isotype antibody group, while Myc expression was reduced when CD226 signaling was blocked ( Figure 3A). We also validated these results using Western blotting ( Figure 3B-D) performed on day 2 post treatment. Overall, these results indicate lower cMyc levels in NK cells after CD226 blockade.

Inhibition of cMyc Using a Chemical Inhibitor Reduces Proliferation by mlNKs from LTBI+ People
We investigated whether the inhibition of cMyc using a chemical inhibitor of cMyc (MYCi975) [38] impacts the proliferation of CD56+CD27+ mlNKs from LTBI+ people. To do this, PBMC from LTBI+ people (n = 5) were pretreated with cMyc inhibitor or mock for 30 min and stimulated with γ-irradiated Mtb. The inhibitor and mock were replenished on alternate days and the proliferation of CTV-labeled CD56+CD27+ mlNKs was assessed 5 days later by flow cytometry. The culture supernatants on day 5 were assessed for the quantification of cytokines ( Figure S5A-E, n = 5). We observed a significant reduction in the proliferation and effector function of CD56+CD27+ mlNKs from LTBI+ people (n = 5) when cMyc was inhibited compared to mock-treated groups ( Figure 3E-H and Figure S6A). cMyc inhibition also led to a reduction in the protein levels of IFN-γ, TNF-α, MIP-1β, MCP1 and IL-10 in the culture supernatants on day 5 when measured by ELISA ( Figure S5A-E, n = 5). Overall, these results indicate a role for cMyc in the proliferation of CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation.

Inhibition of CD226 Signaling Lowered Glycolysis in NK Cells, and Glycolysis Inhibition
Reduced Proliferation by mlNKs from LTBI+ People cMyc has been shown to regulate glycolysis in various cells [39][40][41][42]. Since CD226 blockade reduced cMyc levels, we probed whether the inhibition of CD226 signaling modulates the glycolytic metabolism of NK cells from LTBI+ people. PBMC from LTBI+ donors (n = 4) were stimulated with γ-Mtb and treated with isotype antibody or CD226 blocking antibody. NK cells were purified at day 2 post treatment and subjected to measure glycolysis (extracellular acidification rate) by Seahorse analyzer. We found reduced glycolysis in NK cells from LTBI+ people when CD226 signaling was inhibited ( Figure 4A and Figure S6C-E). Glycolysis inhibition has been shown previously to impact NK cell proliferation and IFN-γ production [43]. To determine if the inhibition of glycolysis impacts the function of CD56+CD27+ mlNKs from LTBI+ people, PBMC from LTBI+ people (n = 5) were stimulated with γ-Mtb and treated with mock or 2DG (glycolysis inhibitor) [43]. The proliferation and function of CD56+CD27+ mlNKs were assessed 5 days later. We observed a significant reduction in the proliferation of CD56+CD27+ mlNKs from LTBI+ people when glycolysis was inhibited compared to mock-treated groups ( Figure 4B,C and Figure S6B). We also observed a reduction in IFN-γ production and degranulation by CD56+CD27+ mlNKs from LTBI+ people from 2DG-treated groups compared to mock treatment ( Figure 4D,E). Together, these results indicate a role for glycolysis in the proliferation of CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. After this, 5 × 10 5 purified NK cells (purity > 90%) were plated to perform the glycolysis assay using seahorse analyzer (n = 4). (B) Chemical blockade of glycolysis using 2DG (0.5 or 1 mM every three days) after stimulation with γ-Mtb for 5-day time period was complete. (B-E) Proliferation, frequencies of IFN-γ+ cells and CD107a+ cells among CD3−CD45+CD56+CD27+ cells unstimulated or stimulated with γ-Mtb and treated with 2DG (n = 5). Mean ± s.d. ANOVA. * p < 0.05, *** p < 0.001, **** p < 0.0001.

Discussion
In this report, we show that co-stimulatory signaling through CD226 contributes to the function of mlNKs in LTBI+ people. Our results show that the blockade of CD226 signaling using a blocking antibody or CRISPR/Cas9-mediated CD226 deletion reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Furthermore, our results revealed that blocking CD226 signaling reduced the phosphorylation of FOXO1 and lowered cMyc levels. We also showed that the inhibition of cMyc using a chemical inhibitor reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Finally, we show that CD226 blockade lowered glycolysis in NK cells and inhibiting glycolysis reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Overall, our results suggest a role of costimulatory signaling through CD226 in the expansion capacity and function of CD56+CD27+ mlNKs from LTBI+ people and could imply that defects in CD226 signaling may influence mlNKs function in human and possibly Mtb infection outcomes. After this, 5 × 10 5 purified NK cells (purity > 90%) were plated to perform the glycolysis assay using seahorse analyzer (n = 4). (B) Chemical blockade of glycolysis using 2DG (0.5 or 1 mM every three days) after stimulation with γ-Mtb for 5-day time period was complete. (B-E) Proliferation, frequencies of IFN-γ+ cells and CD107a+ cells among CD3−CD45+CD56+CD27+ cells unstimulated or stimulated with γ-Mtb and treated with 2DG (n = 5). Mean ± s.d. ANOVA. * p < 0.05, *** p < 0.001, **** p < 0.0001.

Discussion
In this report, we show that co-stimulatory signaling through CD226 contributes to the function of mlNKs in LTBI+ people. Our results show that the blockade of CD226 signaling using a blocking antibody or CRISPR/Cas9-mediated CD226 deletion reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Furthermore, our results revealed that blocking CD226 signaling reduced the phosphorylation of FOXO1 and lowered cMyc levels. We also showed that the inhibition of cMyc using a chemical inhibitor reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Finally, we show that CD226 blockade lowered glycolysis in NK cells and inhibiting glycolysis reduced proliferation by CD56+CD27+ mlNKs from LTBI+ people in response to γ-Mtb stimulation. Overall, our results suggest a role of costimulatory signaling through CD226 in the expansion capacity and function of CD56+CD27+ mlNKs from LTBI+ people and could imply that defects in CD226 signaling may influence mlNKs function in human and possibly Mtb infection outcomes.
Prior report showed a critical role of the ER stress sensor IRE1α and its substrate transcription factor XBP1 in promoting NK cell effector responses to fend off virus infection and to clear tumors [36]. These antiviral and anti-tumor effects were mediated by cMyc, a molecule implicated in cellular proliferation [36]. We noted in LTBI+ people that the expression of Ire1α, Xbp1 did not vary between isotype and CD226 blocking antibody treatment regimens suggesting ER stress may not be involved in the regulation of mlNK responses in LTBI+ persons. Multiple reports have advocated a role for Wnt signaling in proliferation by memory-like NK cells [11,14]. Our results show the equivalent expression of multiple Wnt signaling genes (Axin2, Tcf7) in CD226-antibody-treated versus isotypetreated NK cells from LTBI+ people. cMyc regulates proliferation in diverse cell types [36,37,[44][45][46]. In line with these studies, we report reduced mlNKs proliferation with CD226 blockade due to diminished cMyc levels. Our finding of the inhibition of CD56+CD27+ mlNKs proliferation and effector function from LTBI+ people after treatment with cMyc inhibitor is also consistent with cMyc promoting proliferation and effector function in NK cells as also noted by others [36,47]. However, it is unclear how CD226 signaling promotes cMyc expression in mlNKs cells which we are currently investigating. One likely effect involves effects on aspect of immunometabolism. In fact, some reports do indicate a critical role of cMyc in modulating metabolic process [39][40][41]. Some show that cMyc promotes glycolysis [39][40][41], but others suggest its role in oxidative phosphorylation and mitochondrial metabolism [42]. It was not established if CD226 signaling modulates metabolism in NK cells. Recently, CD226 signaling was reported to promote glycolysis in endothelial cells [48]. Similarly, we noted diminished glycolysis in NK cells from LTBI+ people after CD226 blockade. It is possible that inhibiting CD226 signaling exerts its negative effect on glycolytic process via reduction in cMyc activity. In addition to NK cells, CD226 signaling can also regulate CD8 T cell function [49]. In fact, blocking PD-1 and inhibiting CD226 function compromise tumor clearance by CD8 T cells [50][51][52]. It may be worthwhile to see whether CD226 signaling regulates cMyc expression/glycolytic pathway in CD8 T cells that are actively clearing the tumors. We speculate that the outcome is likely to be positive. Overall, the results from our current NK cell study suggest that CD226 signaling may also play a role in regulating the metabolism of NK or even CD8 T cells in other contexts such as infection or tumor settings.

Limitations
Due to limited numbers of CD56+CD27+mlNKs, our experiments involved probing the protein levels of cMyc, FOXO1 or glycolysis assays in total NK cells after stimulation of the PBMC from LTBI+ people with γ-Mtb and treatments with isotype or blocking CD226 antibodies. We noted reduced CD56+CD27+ mlNKs proliferation from LTBI+ people after stimulation with γ-Mtb and treatments with cMyc inhibitor or glycolysis inhibitor (2DG). Thus, we speculate that CD226 inhibition also reduces FOXO1, cMyc levels or glycolysis in CD56+CD27+ mlNKs from LTBI+ people. Due to human primary cell studies with in vitro nature, the findings from this study require further studies with murine models with CD226 gene deletion where we believe that the absence of the CD226 gene may lead to multiple defects in NK and T cell compartment leading to increased susceptibility to Mtb.

Ethics Statement
All human samples studies were approved by the Institutional Review Board of the University of Texas Health Science Center at Tyler with protocol number 2021-033. For this study we collected blood samples from healthy volunteers for the isolation of PBMC.

Sample Collection
All donors were well-informed about the study before the acquisition of the informed consent. For the study blood from healthy persons positive to QuantiFERON-TB Gold tests and healthy persons negative to QuantiFERON-TB Gold tests were collected. All subjects were between 25 and 55 years old, including both genders.

c-Myc and Glycolysis Inhibition
PBMC from LTBI+ donors were cultured in 24-well plates (2.5 × 10 6 cells/well) with RPMI-1640 (Cytiva) complete media with 10% AB heat-inactivated human serum (Sigma-Aldrich, Burlington, MA, USA) at 37 • C and 5% CO 2 . The cells were treated with MYCi975 (6 µM, 12 µM or 18 µM, MedChemExpress, Monmouth Junction, NJ, USA) or 2-DeoxyGlucose (2DG, 0.5 mM and 1 mM, Sigma-Aldrich) or control DMSO for 30 min. After this time, the cells were stimulated with γ-Mtb (10 µg/mL) for 5 days and analyzed by flow cytometry. cMyc inhibitor was supplemented every other day while 2DG was supplemented every three days. For intracellular staining, surface staining was followed by cellular permeabilization with Fixation/Permeabilization solution according to the manufacturer's protocol (BD biosciences). The cells were washed with Perm/Wash buffer and were stained for FITC anti-human IFN-γ (clone: 45.B3, BioLegend) and APC anti-human CD107a (clone H4A3, BioLegend). In these assays, Brefeldin A (BD Biosciences) was added 4 h prior to surface staining, and CD107a antibody was added 2 h before Brefeldin A. The cellular acquisition was performed on LSRFortessa X-20 and data analysis was conducted using FlowJo v10 software (TreeStar, Ashland, OR, USA).

Metabolism Assays Using Seahorse Analyzer
Human PBMC from LTBI+ donors were treated with blocking anti-human CD226 clone 11A8 (10 µg/mL, BioLegend) or isotype control antibody (10 µg/mL, BioLegend) for 1 h at 37 • C and 5% CO 2 . The cells were stimulated with γ-Mtb or left untreated for 48 h at 37 • C and 5% CO 2 . NK cells were enriched using human NK isolation kit (Miltenyi Biotec, Gaithersburg, MD, USA. purity more than 90%). NK cell glycolytic status or extracellular acidification rate (ECAR) was measured via extracellular flux assay using Seahorse XFe96 Analyzer (Agilent Technologies, Santa Clara, CA, USA). NK cells were washed and resuspended in Agilent XF assay medium supplemented with 1 mM L-Glutamate and then plated for the assay at a density of 5 × 10 5 NK cells/well in a 96-well Seahorse plate coated with Poly-L-Lysine. During the assay glucose (10 mM), oligomycin (1 µM) and 2DG (50 mM) were injected according to the manufacturer's protocol. Basal glycolysis, glycolytic capacity and glycolytic reserve were calculated using Seahorse Wave 2.6.3 Desktop Software, Agilent Technologies, USA.

CRISPR/Cas9-Mediated Deletion of CD226 Gene
NK cells isolated from PBMC were kept with human IL-15 (10 ng/mL) (BioLegend) for 24 h. After this, 2-5 × 10 6 NK cells were resuspended in 90 microliters Nucleofector TM solution and 10 microliters sgRNA-Cas9 complex was added (0.3 pmol of each sgRNA and 40 pmol spCas9). Four CD226 sgRNAs (Synthego, Redwood City, CA, USA) designed from CD226 human gene (CD226+69947022, CD226+69947028, CD226+69947029, and CD226-69947061) were used. spCas9 protein was also procured from Synthego. NK cells were electroporated using Amaxa TM human NK cell nucleofector TM kit (Lonza, Bend, OR, USA) according to the manufacturer's protocol. After electroporation, NK cells were maintained with 10 ng/mL human IL-15 for 5 days, and the deletion of CD226 was verified by Western blotting and flow cytometry. CD226 deleted or control NK cells were stained with CTV and co-cultured with PMBC isolated from the same donor that were stained with CFSE. Proliferation was measured by flow cytometry after the stimulation of co-cultured cells with γ-irradiated Mtb.

Western Blotting
The protein expression and phosphorylation of c-Myc and FOXO1 were determined by Western blotting using whole cell protein extracts from 10 6 NK cells lysed with RIPA buffer with protease and phosphatase inhibitors (G-Biosciences, St Louis, MO, USA). The samples were run in 10% SDS-PAGE gel and transferred onto nitrocellulose membranes, which were blocked with 5% skimmed milk in TBST for 1h. After this, the membranes were incubated overnight at 4 • C with primary antibodies for cMyc (Cell signaling, 1:1000 dilution), pcMyc (Cell signaling, 1:1000 dilution), FOXO1 (Cell signaling, 1:1000 dilution), pFOXO1 (Cell signaling, 1:1000 dilution), β-actin (Cat# 4967, Cell signaling, 1:1000 dilution). The membranes were incubated for 1h with secondary antibody conjugated with HRP and proteins bands were visualized by Clarity ECL substrate (Bio-Rad, Hercules, CA, USA) followed by capturing with ChemiDoc™ Imaging System (Bio-Rad).

Statistical Analysis
GraphPad Prism 7 software was used for experimental analysis. The unpaired or paired t-test or ANOVA was used for determining the statistical significance at * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 and **** p ≤ 0.0001.