IL-6 enhances CD4 cell motility by sustaining mitochondrial Ca²⁺ through the noncanonical STAT3 pathway

Significance

In this manuscript, we identify a pathway involving mitochondrial Ca²⁺ and mitochondrial Stat3 that strengthens the motility fitness of activated CD4 T cells. Considering that motility is essential for CD4 T cells to find and reach their targets and mediate pathogenesis, these findings could lead to the development of therapeutic strategies to intervene in CD4 T cell–mediated inflammatory diseases.

Abstract

Interleukin 6 (IL-6) is known to regulate the CD4 T cell function by inducing gene expression of a number of cytokines through activation of Stat3 transcription factor. Here, we reveal that IL-6 strengthens the mechanics of CD4 T cells. The presence of IL-6 during activation of mouse and human CD4 T cells enhances their motility (random walk and exploratory spread), resulting in an increase in travel distance and higher velocity. This is an intrinsic effect of IL-6 on CD4 T-cell fitness that involves an increase in mitochondrial Ca²⁺. Although Stat3 transcriptional activity is dispensable for this process, IL-6 uses mitochondrial Stat3 to enhance mitochondrial Ca²⁺-mediated motility of CD4 T cells. Thus, through a noncanonical pathway, IL-6 can improve competitive fitness of CD4 T cells by facilitating cell motility. These results could lead to alternative therapeutic strategies for inflammatory diseases in which IL-6 plays a pathogenic role.

Interleukin 6 (IL-6) is a pleiotropic cytokine that has been implicated in inflammatory diseases, autoimmune disorders, obesity, and in cancer (1, 2). The pathogenic role of this cytokine in inflammatory diseases is evident by the success of an anti–IL-6 receptor blocking antibody, tocilizumab, in the treatment of rheumatoid arthritis and other autoimmune diseases and, more recently, of cancer in combination with some types of immunotherapy (3, 4). IL-6 regulates CD4 T-cell differentiation, function, and survival (5). Th1 differentiation is inhibited by IL-6 through increased SOCS1 expression (6). IL-6 also mediates the differentiation of naïve CD4 T cells into effector Th2 cells producing IL-4 and IL-13 (7). IL-6, together with TGF-β, promotes the differentiation of naïve CD4 T cells into effector T helper 17 cells. In contrast, IL-6 inhibits regulatory T cell (Treg) differentiation (8). IL-6 promotes CD4 T follicular helper (9) cell differentiation and induces IL-21 production by naïve CD4 T cells (10). Thus, IL-6 plays an important role in determining the function of effector CD4 T cells by regulating cytokine gene expression.

IL-6 regulates gene expression primarily through the activation of the Stat3 transcription factor (11). In the canonical Stat3 signaling pathway, phosphorylation at Tyr⁷⁰⁵ of Stat3 by JAK/Tyk2 kinases promotes its dimerization and translocation from the cytosol to the nucleus, where Stat3 dimers bind to specific DNA promoters and modulate the expression of target genes (12). Several studies identified a noncanonical Stat3 pathway, through which Stat3 is imported into mitochondria (13, 14). The presence of mitochondrial Stat3 (mitoStat3) has now been described in cancer cells, neuronal cells, cardiac cells, and T cells (13–17). MitoStat3 enhances Complex I of the mitochondrial electron transport chain (ETC), leading to increases in mitochondrial membrane potential (MMP) coupled to increased oxidative phosphorylation and mitochondrial adenosine triphosphate (ATP) synthesis (13, 14, 18). We have shown that during the activation of CD4 T cells, IL-6 promotes the accumulation of Stat3 in mitochondria, and it increases MMP (17). However, the increased MMP triggered by IL-6 in activated CD4 T cells is not reflected in increased mitochondrial respiration (17). Instead, IL-6 uses MMP to enhance and sustain mitochondrial Ca²⁺, and this effect is dependent on Stat3 but independent of Stat3 transcription activity (17).

Mitochondrial Ca²⁺ is known to play an important role in cell migration and motility (9). In Zebra fish embryos, mitochondrial Ca²⁺ regulates migration of pluripotent stem cells (19). In breast and ovarian cancer cells, as well as in endothelial cells and lymphoma cells, alteration of mitochondrial Ca²⁺ uptake impairs cell migration (9, 20, 21). Mitochondrial Ca²⁺ also contributes to sperm motility (22, 23). Recent studies indicate that mitochondrial Ca²⁺ contributes to the actin cytoskeleton dynamics (20). In addition, since they are highly dynamic organelles, mitochondria can contribute to creating high Ca²⁺ microdomains during cell migration (9)

Chemotaxis-triggered migration is essential for CD4 T cells to move from secondary lymphoid tissues to the sites of infection or inflammation (24). Basal motility of T cells, defined as random walk and exploratory spread, is also relevant for activated CD4 T cells within these inflammatory sites to search for and reach their cell targets (25). Similarly, basal motility can contribute to T cells finding malignant cells within tumors (25). Although no studies have investigated the role of mitochondrial Ca²⁺ in CD4 T-cell migration or motility, it is known that mitochondria redistribute to the adhesion zone during migration of lymphocytes along the endothelium, and the mitochondrial protein Miro-1 couples mitochondria with microtubules in T cells (26–28). Here, we show that the increase in mitochondrial Ca²⁺ triggered by IL-6 during activation of CD4 T cells promotes their motility. IL-6–induced motility in CD4 T cells is mediated by mitochondrial Stat3, but it is independent of the transcriptional activity of Stat3. Thus, our studies reveal an intrinsic mechanism by which IL-6 enhances CD4 T-cell immune response. These findings could lead to the development of alternative therapeutic approaches for treatment of chronic inflammatory diseases caused by pathogenic CD4 T cells.

Results

The Presence of IL-6 During Activation Enhances CD4 T-Cell Motility.

Since mitochondrial Ca²⁺ is involved in cell motility and the presence of IL-6 during activation of CD4 T cells is known to increase mitochondrial Ca²⁺ (29), we investigated the effect of IL-6 on CD4 T-cell motility. CD4 T cells were activated with anti-CD3 and anti-CD28 monoclonal antibodies (mAbs) in the presence or absence of IL-6 for 2 d, extensively washed, resuspended in medium alone, and loaded onto ICAM-coated microscope chambers, and images were captured under bright field in the confocal microscope over 10 min. While CD4 T cells activated without IL-6 barely moved, CD4 T cells activated in the presence of IL-6 traveled a longer linear distance (Fig. 1A and SI Appendix, Fig. S1 A and B). Interestingly, in addition to the total distance, the velocity of CD4 T cells activated in the presence of IL-6 was markedly higher than the velocity of those few CD4 T cells activated without IL-6 that were able to move (Fig. 1B and SI Appendix, Fig. S1B). IL-6 had no effect on the expression of LFA-1, the receptor for ICAM (SI Appendix, Fig. S1C). To further investigate the increased motility of CD4 T cells activated in the presence of IL-6, after activation and extensive washes, cells were placed into transwell chambers in medium alone. After 5 h, the cells that migrated through the transwell were counted. While only a few CD4 T cells activated without IL-6 migrated, a higher number of CD4 T cells activated in the presence of IL-6 were able to go through the transwell (Fig. 1C). Since these assays were performed in medium alone, to eliminate the possibility that IL-6 during activation may have caused an up-regulation of chemokine production, we examined the supernatant for the presence of CXCL10, CCL19, and CCL5 chemokines known to induce migration of T cells (30). Neither CXCL10 nor CCL19 were under the level of detection. CCL5 level was detected, but the levels between CD4 T cells activated with or without IL-6 were comparable (SI Appendix, Fig. S1D). Thus, the effect of IL-6 on motility does not seem to be indirectly caused by inducing chemokine production. Instead, IL-6 causes an intrinsic change in CD4 T cells during activation that increases the motility of CD4 T cells even when IL-6 is no longer present, and the cells move faster and further.

Fig. 1.

IL-6 increases motility of mouse CD4 T cells. (A and B) CD4 T cells were activated with anti-CD3/CD28 mAbs in the presence or absence of IL-6 (50 ng/mL). After 48 h, cells were harvested, washed, and loaded onto chambers on a temperature-controlled stage. Live images were captured using a confocal microscope (bright field) every 30 s for 10 min. Distance (A) and velocity (B) of individual cells are shown. (C) CD4 T cells were activated as in A for 48 h, washed, and cells were seeded into transwell chambers in medium alone. After 5 h, cells that migrated in the wells were counted (n = 3). (D) CD4 T cells were activated as in A, and after 48 h they were washed and placed into transwell chambers in the presence of CCL19 (100 ng/mL) that was added to the lower chamber. After 5 h, cells that had migrated in the wells were counted. (E) CCR7 expression by CD4 T cells activated for 48 h in the absence or presence of IL-6 as determined by flow cytometry analysis. (F and G) CD4 T cells were activated as in A for 48 h, washed, and loaded onto chambers on a temperature-controlled stage in the presence of CCL19 (100 ng/mL). Live microscopy images were captured as described in A. Distance (F) and velocity (G) for individual cells are shown. Error bars show ± SD; *P < 0.05, as determined by Student’s t test. Results are representative of two to three experiments.

To better address the intrinsic effect that IL-6 has on the ability of the cells to move, we also examined its effect on chemokine-induced migration, in which the levels of chemokine are the same in both cell types. Following 2 d activation in the presence or absence of IL-6, CD4 T cells were washed and placed in transwell chambers to which CCL19 was added. As expected, relative to the number of cells that go through the transwell in the absence of CCL19 (Fig. 1C), a higher number of cells migrated through the transwells when CCL19 was present (Fig. 1D). Nevertheless, even in the presence of CCL19 during the migration assay, a higher number of migrating cells were recovered from CD4 T cells activated with IL-6 (Fig. 1D). The cell surface expression of CCR7, the receptor for CCL19, was comparable in CD4 T cells when IL-6 was present or absent (Fig. 1E); therefore, the effect of IL-6 is not due to differential expression of the CCL19 receptor. We further evaluated the motility of these cells by adding CCL19 in the chambers at the time the cells were imaged by live microscopy. Even in the presence of CCL19, CD4 T cells activated with IL-6 traveled a longer distance (Fig. 1F). In addition, the velocity during the migration with CCL19 was also higher in CD4 T cells activated with IL-6 (Fig. 1G). Thus, the presence of IL-6 during activation of CD4 T cells causes an intrinsic effect in activated CD4 T cells that enhances their motility even in the absence of chemokines or cytokines.

Mitochondrial Ca²⁺ Is Necessary for IL-6 to Enhance CD4 T-Cell Motility.

Considering the role of mitochondrial Ca²⁺ in cytoskeleton dynamics and cell migration and (20) the effect on IL-6 mitochondrial Ca²⁺, we investigated whether the increased motility of CD4 T cells activated in the presence of IL-6 was mediated by increased mitochondrial Ca²⁺. Mitochondrial Ca²⁺ uptake is mediated by the mitochondrial Ca²⁺ uniporter (MCU), and mitochondrial Ca²⁺ efflux to the cytosol is mediated by the mitochondrial Na⁺/Ca²⁺ antiporter (mNCLX) (31). We therefore activated CD4 T cells in the presence of IL-6 for 2 d. During the last 4 h of activation, cells were treated with CGP-37157, an inhibitor of mNCLX, or vehicle. Cells were then washed extensively, and an equal number were placed into transwells with medium alone. The number of cells that migrated through the transwell was determined. Inhibition of mitochondrial Ca²⁺ release with CGP-37157 markedly reduced the number of CD4 T cells that went through the transwell (Fig. 2A). In addition, we examined the effect of Ru360, an inhibitor of MCU, that also decreases mitochondrial Ca²⁺ in CD4 T cells (17) (SI Appendix, Fig. S2). Treatment with Ru360 reduced the number of IL-6–activated CD4 T cells going through the transwells (Fig. 2B). To further demonstrate the need of mitochondrial Ca²⁺ for the motility of CD4 T cells activated in the presence of IL-6, we used live microscopy as described in the previous section. CD4 T cells were activated in the presence of IL-6, treated with Ru360, and placed on the warm chamber in medium alone at the confocal microscope. Treatment with Ru360 caused a significant reduction of the distance that IL-6–activated CD4 cells travel (Fig. 2C). Importantly, Ru360 also lowered the velocity of the motile cells (Fig. 2D). Thus, pharmacological inhibition of mitochondrial Ca²⁺ blocks the enhancement in travel distance and speed of activated CD4 T cells caused by IL-6.

Fig. 2.

IL-6–induced motility of mouse CD4 T cells during activation requires mitochondrial Ca²⁺. (A) CD4 T cells were activated with anti-CD3/CD28 mAbs in the presence of IL-6 for 48 h, and during the last 4 h, they were treated with (10 μM) CGP-37157 or vehicle. Cells were washed and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (B) CD4 T cells were activated for 48 h in the presence of IL-6, and for the last 18 h, they were treated with (10 μM) Ru360 or vehicle. Cells were then washed and placed in transwell chambers as in A. (C and D) CD4 T cells were activated in the presence of IL-6, treated with Ru360 or vehicle, washed, and loaded onto chambers on a temperature-controlled stage, in medium alone. Live images were captured under a confocal microscope. Distance (C) and velocity (D) of individual cells are shown. (E) Expression of MCU in CD4 T cells activated in the absence or presence of IL-6 as determined by Western blot analysis. β-actin was used as a control. (F) Expression of MCU in freshly isolated CD4 T cells from cMCU KO mice by Western blot analysis. (G) Mitochondrial Ca²⁺ in CD4 cells from WT and cMCU KO mice activated in the presence of IL-6 for 48 h was determined by staining with Rhod-2 AM and flow cytometry analysis. (H) CD4 T cells from WT and cMCU KO mice were activated with anti-CD3/CD28 mAbs in the presence of IL-6 for 48 h, washed, and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (I and J) CD4 T cells from WT mice were stained with carboxyfluorescein succinimidyl ester (CFSE) or CellTrace Violet and activated in the absence or presence of IL-6, respectively. After 48 h, cells were washed and coinjected i.v. (5 × 10⁶ cells total) at equal ratio into WT recipient mice (n = 5). After 16 h, the presence of donor cells in the spleen was examined by flow cytometry analysis (I). The number of each labeled donor cell type is shown (J). (K) CD4 T cells from WT or cMCU KO mice were stained with CFSE or CellTrace Violet, respectively, activated in the presence of IL-6 for 48 h, washed, and injected into WT recipient mice (n = 4) as described in J. After 16 h, CD4 T cells were analyzed by flow cytometry. Numbers of CD4 T cells labeled with CFSE or CellTrace Violet in the spleen are shown. (L–N) CD4 T cells were activated with anti-CD3/CD28 mAbs in the absence (MED) or presence of IL-6 (50 ng/mL) for 48 h, cytospinned, stained for actin (L, red), vinculin (M, green), or phospho-MLC (P-Ser19, red) (N), and visualized at the confocal microscope. DAPI (blue) was used as a nuclear marker. Error bars show ± SD; *P < 0.05, as determined by Student’s t test. Results are representative of two or three experiments.

Since the use of pharmacological inhibitors to reduce mitochondrial Ca²⁺ could have limitations such as low cell permeability (32), we also used MCU-deficient CD4 T cells to further demonstrate the role of mitochondrial Ca²⁺ in IL-6–induced CD4 T-cell motility. We first examined whether IL-6 had an effect on the expression of MCU, but the levels of MCU in CD4 T cells activated in the absence or presence of IL-6 were comparable (Fig. 2E). To obtain MCU-deficient CD4 T cells, we used T-cell conditional MCU knockout mice (cMCU KO) generated by crossing MCU^f/f mice with CD4-Cre transgenic mice. Western blot analysis for MCU showed the reduction in the levels of MCU in cMCU KO CD4 T cells (Fig. 2F). Since T cell cMCU KO mice have not previously been reported, we first analyzed the frequency of mature CD4 and CD8 T cells in the spleen, lymph nodes, and thymus by flow cytometry analysis. No difference in the CD4 and CD8 T-cell populations were observed between cMCU KO mice and wild-type (WT) mice (SI Appendix, Fig. S3A). We confirmed that MCU contributes to mitochondrial Ca²⁺ in CD4 T cells by examining mitochondrial Ca²⁺ by flow cytometry in WT and cMCU KO CD4 T cells activated in the presence of IL-6 for 2 d. Lower levels of mitochondrial Ca²⁺ were present in cMCU KO CD4 T cells compared with WT CD4 T cells (Fig. 2G), verifying that MCU contributes to mitochondrial Ca²⁺ entry in activated CD4 T cells.

To investigate whether MCU could have an effect on T-cell activation, CD4 T cells from WT and cMCU KO mice were activated in the presence or absence of IL-6 and after 2-d expression of activation markers was assayed by flow cytometry. No difference in the surface expression of CD44 and CD69 was observed (SI Appendix, Fig. S3B). We also examined proliferation of activated CD4 T cells by CFSE staining, but no differences were found between WT and cMCU KO CD4 T cells (SI Appendix, Fig. S3C). Moreover, analysis of cell death during activation by flow cytometry shows comparable levels of cell survival between WT cells and cMCU KO CD4 T cells (SI Appendix, Fig. S3D). Thus, mitochondrial Ca²⁺ is dispensable for activation and proliferation of CD4 T cells. We then examined the motility of cMCU KO CD4 T cells activated in presence of IL-6 using the transwell chamber with only medium. The number of migrating MCU-deficient CD4 T cells was substantially lower than the number of WT CD4 T cells (Fig. 2H), further demonstrating that IL-6 improves the fitness of activated CD4 T cells to migrate further and more rapidly by sustaining mitochondrial Ca²⁺.

To investigate the role of mitochondrial Ca²⁺ in the migratory capacity of activated CD4 cells in vivo, we first examined the effect that the presence of IL-6 during activation had on the in vivo migratory capacity of these cells using cell adoptive transfer. WT CD4 T cells were activated with anti-CD3/CD28 mAbs in presence of IL-6 or absence of IL-6 and labeled with CellTrace Violet or CFSE, respectively. After 2 d, cells were extensively washed, and equal numbers of CD4 T cells activated with or without IL-6 were mixed (SI Appendix, Fig. S4A) and injected intravenously (i.v.) into WT host mice (Fig. 2I). After 16 h, spleens were harvested and the presence of transferred CD4 T cells was examined by flow cytometry analysis. The number of transferred CD4 T cells that had been activated with IL-6 was higher than the number of transferred CD4 T cells activated without IL-6 (Fig. 2J), showing superior in vivo trafficking of CD4 T cells activated in the presence of IL-6. We then examined the role of mitochondrial Ca²⁺ in the enhanced-trafficking effect of IL-6. WT and cMCU KO CD4 T cells were labeled with different dyes and activated in the presence of IL-6. After 2 d, cells were washed, and equal numbers of WT and cMCU KO CD4 T cells were mixed and injected into WT host mice (SI Appendix, Fig. S4B). After 16 h, the presence of both types of donor cells in the spleen was examined by flow cytometry. A lower number of cMCU KO CD4 T cells relative to the number of WT CD4 T cells was found (Fig. 2K). The proliferation rate of the transferred WT CD4 cells in vivo was not higher than cMCU KO CD4 cells (SI Appendix, Fig. S4C). Thus, during the activation of CD4 T cells, mitochondrial Ca²⁺ is essential for IL-6 to improve the motility and trafficking of these cells both in vitro and in vivo.

IL-6 Reduces Cytoskeleton Stiffness of CD4 T Cells during Activation.

Mitochondrial Ca²⁺ contributes to the motility of the cells by reducing stiffness of the cells (9, 20). Impairment of mitochondrial Ca²⁺ by silencing MCU expression has been shown to increase actin cytoskeleton stiffness and to induce cell polarization loss, essential parameters that compromised cell motility. We therefore examine microfilament components such as myosin light chain (MLC) and F-actin in cells activated in the absence or presence of IL-6 by confocal microscopy. Analysis of F-actin fibers by staining with rhodamine phalloidin showed higher density of F-actin fibers and actin bundles distributed within the cytoplasm in CD4 T cells activated without IL-6, indicative of increased cytoskeleton stiffness in these cells relative to cells activated in the presence of IL-6, where fewer actin bundles could be detected (Fig. 2L). Loss of mitochondrial Ca²⁺ has also been found to alter the cell distribution of vinculin (9, 33), another membrane cytoskeleton protein involved in focal adhesion plaques and associated with actin. Confocal microscopy analyses of vinculin showed a more diffused distribution of vinculin in the cytosol in cells activated with IL-6 relative to the accumulation of vinculin in focal areas in cells without IL-6 (Fig. 2M), further supporting the idea that IL-6 decreases stiffness in CD4 T cells. The presence of phospho-MLC is another marker of cell stiffness, and it was found to be increased in MCU-deficient cells (20). Analysis of phospho-MLC (Ser19) by immunostaining and confocal microscopy (Fig. 2N) or by Western blot analysis (Thr18/Ser19) (SI Appendix, Fig. S5) showed reduced levels of phospho-MLC in CD4 T cells activated in the presence of IL-6. Thus, correlating with its ability to increase mitochondrial Ca²⁺, IL-6 reduces cytoskeleton stiffness of CD4 T cells and facilitates CD4 T-cell motility.

IL-6–Induced Mitochondrial Ca²⁺ and Motility in CD4 T Cells Is Mediated by MitoStat3.

IL-6 is one of the major activators of the Stat3 transcription factor, and Stat3 controls expression of IL-6–regulated genes through the canonical pathway (34). In addition to its role as a transcription factor, Stat3 has also been found in mitochondria, where it associates with Complex I of the ETC activity, increases Complex I activity, and thereby increases MMP and ATP production (13, 14, 35–37). In addition, since high MMP promotes the uptake of Ca²⁺ into the mitochondria, mitochondrial Stat3 can also increase mitochondrial Ca²⁺. IL-6–induced mitochondrial Ca²⁺ requires Stat3 but does not require Stat3 transcriptional activity (17, 36). Thus, IL-6 could use mitochondrial Stat3 (mitoStat3) to increase MMP and mitochondrial Ca²⁺ and thereby enhance CD4 T-cell motility. We first investigated the ability of IL-6 to increase motility of CD4 T cells from mice expressing a mutant Stat3 (mut-Stat3) carrying the dominant negative mutation found in autosomal dominant hyperimmunoglobulin E syndrome (mut-Stat3 mice) (38). mut-Stat3 inhibits DNA binding and transcriptional activity of Stat3 without altering the total levels of Stat3 (38). We have previously shown that inhibition of the Stat3 canonical pathway in mut-Stat3 CD4 cells does not prevent the IL-6–mediated increase in MMP and mitochondrial Ca²⁺ (17). To investigate the effect of mut-Stat3 on CD4 T-cell motility, CD4 T cells from WT and mut-Stat3 mice were activated with or without IL-6 for 2 d, washed, and placed on transwell chambers with medium alone. The number of mut-Stat3 CD4 T cells activated with IL-6 that migrated through the transwell was comparable to the number of WT CD4 T cells (Fig. 3A). Thus, IL-6 does not require Stat3 transcriptional activity during activation to promote cell motility.

Fig. 3.

Impaired CD4 T-cell motility in Stat3 S⁷²⁷A KI mice. (A) CD4 T cells from WT and mut-Stat3 mice were activated with anti-CD3/CD28 mAbs in the presence or absence of IL-6 for 48 h, washed, and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (B and C) WT and Stat3 S⁷²⁷A KI CD4 T cells were activated in the presence or absence of IL-6 for 48 h, and MMP was determined by tetramethylrhodamine, ethyl ester (TMRE) staining (B). Percentage of TMRE^high cells as defined in B is shown (n = 3) (C). (D and E) Mitochondrial Ca²⁺ in CD4 T cells from WT and Stat3 S⁷²⁷A KI mice activated as in B was assayed by staining with Rhod-2 AM and flow cytometry analysis (D). Percentage of TMRE ^high cells as defined in D is shown (E) (n = 3). (F) CD4 T cells from WT and Stat3 S⁷²⁷A KI mice were activated as in A and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (G) CD4 T cells from WT or Stat3 S⁷²⁷A KI mice were stained with CFSE or CellTrace Violet, respectively, activated in the presence of IL-6 for 48 h, washed, and injected into WT recipient mice (n = 4). After 16 h, CD4 T cells were analyzed by flow cytometry. Numbers of CD4 T cells labeled with CFSE or CellTrace Violet in the spleen are shown. (H and I) WT and Stat3 S⁷²⁷A KI CD4 T cells were activated as in A in the presence of IL-6 for 48 h, cytospinned, stained for actin (H) or phospho-MLC (I), and visualized at the confocal microscope. DAPI (blue) was used as a nuclear marker. Error bars show ± SD; *P < 0.05, as determined by Student’s t test. Results are representative of two to three experiments.

Phosphorylation of Stat3 on Ser⁷²⁷ is not essential for transcription activity, but it has been shown to promote mitoStat3 function (14). We therefore used Stat3 knockin (KI) mice in which Ser⁷²⁷ is replaced by Ala to prevent phosphorylation (mS⁷²⁷A Stat3 KI mice) (39). We first examined phosphorylation of Stat3 on Ser⁷²⁷ and verified that IL-6 increases phospho-Ser⁷²⁷ in WT CD4 T cells and that no phospho-Ser⁷²⁷ was found in CD4 T cells from mS⁷²⁷A Stat3 KI mice (SI Appendix, Fig. S6A). We then investigated whether the mS⁷²⁷A mutation of Stat3 could have a broad effect on activation of CD4 T cells. WT and mS⁷²⁷A CD4 T cells were activated in the presence or absence of IL-6, and after 2 d, we examined the expression of activation markers. No differences in the surface expression of CD44, CD69, or CD62L were observed between WT and mS⁷²⁷A KI CD4 T cells independently whether IL-6 was present or not (SI Appendix, Fig. S6B). In addition, there was no difference in cell expansion upon activation between WT and mS⁷²⁷A KI CD4 T cells (SI Appendix, Fig. S6C). Thus, mitoStat3 seems to be dispensable for activation of CD4 T cells. We then asked whether mitoStat3 is required for IL-6 to increase MMP and mitochondrial Ca²⁺ during activation of CD4 cells, using WT and mS⁷²⁷A KI CD4 T cells. Analysis of MMP using TMRE staining and flow cytometry showed that IL-6 failed to increase MMP in mS⁷²⁷A KI CD4 T cells (Fig. 3 B and C). In addition, IL-6 also failed to increase mitochondrial Ca²⁺ in mS⁷²⁷A KI CD4 T cells (Fig. 3 D and E). The mechanism by which phosphorylation of Stat3 on Ser⁷²⁷ regulates Stat3 mitochondrial activity remains unclear, but it has been proposed that Ser⁷²⁷ phosphorylation is required for accumulation of Stat3 in mitochondria in some cells (35, 40, 41). We examined whether subcellular localization of Stat3 could be affected in mS⁷²⁷A KI CD4 T cells activated with IL-6 by immunostaining for Stat3 together with MitoTracker as a mitochondrial marker and confocal microscopy. We found that S727A-Stat3 accumulated in mitochondria of CD4 T cells activated in the presence of IL-6, although to somewhat reduced levels relative to WT cells, by immunostaining for Stat3 together with MitoTracker as a mitochondrial marker and analysis by confocal microscopy (SI Appendix, Fig. S6 D and E).

We then examined whether the ability of IL-6 to promote cell motility is dependent on mitoStat3. CD4 T cells from WT and mS⁷²⁷A KI mice were activated for 2 d in the presence or absence of IL-6, washed, and placed in medium alone in transwells. The number of mS⁷²⁷A KI CD4 T cells activated with IL-6 that migrated through the transwell was substantially lower than the number of WT CD4 cells (Fig. 3F). Moreover, we also examined the enhanced-trafficking effect of IL-6 on mS⁷²⁷A KI CD4 T cells in vivo. WT and mS⁷²⁷A KI CD4 T cells were labeled with different dyes, activated in the presence of IL-6, and after 2 d cells were washed and equal numbers were mixed and injected into WT host mice, as described in Results. A lower number of mS⁷²⁷A KI CD4 T cells relative to the number of WT CD4 T cells were found in the spleen (Fig. 3G). Thus, together the results indicate that during activation of CD4 T cells, IL-6 uses predominantly mitoStat3 to improve fitness of CD4 T cells leading to superior motility and trafficking in vitro and in vivo.

In addition to IL-6, other cytokines can also activate Stat3. To further support the role of Stat3 on CD4 cell migration, we examine the effect of IL-21 and IL-27, two cytokines known to regulate Stat3, and with receptors expressed on CD4 T cells (42, 43). CD4 T cells were activated in the presence or absence of IL-21 or IL-27, and after 2 d, cell motility was assayed using the transwell system as described in Results. Both IL-21 and IL-27 increased the motility of CD4 T cells (SI Appendix, Fig. S7A). Since we propose that the effect on cell motility is due to increased mitochondrial Ca²⁺ mediated by Stat3, we also examined the effect of these cytokines on MMP and mitochondrial Ca²⁺. Interestingly, although the effect was not as marked as the effect of IL-6, both IL-21 and IL-27 increased MMP (SI Appendix, Fig. S7B) and mitochondrial Ca²⁺ (SI Appendix, Fig. S7C).

Since IL-6 failed to increase mitochondrial Ca²⁺ in mS⁷²⁷A KI CD4 T cells, we investigated whether the impaired migration of these cells in the presence of IL-6 was due to increased cytoskeleton stiffness. WT and mS⁷²⁷A KI T CD4 cells were activated in the presence of IL-6, and after 2 d, actin fibers and phosphorylation of MLC2 were examined by confocal microscopy. Interestingly, the actin bundles within the cytosol observed in mS⁷²⁷A KI T CD4 cells (Fig. 3H) resemble the actin distribution in WT CD4 cells activated in the absence of IL-6 (Fig. 2L). In addition, mS⁷²⁷A KI CD4 T cells also contained more phosphorylated MLC (Fig. 3I). Thus, mitoStat3 regulates CD4 T cells motility by reducing cell stiffness through mitochondrial Ca²⁺.

Motility of Human Naïve and Memory CD4 T Cells upon Activation Requires the Elevated Levels of Mitochondrial Ca²⁺ Induced by IL-6.

Considering that this alternative IL-6 signaling pathway affecting CD4 T-cell motility may be clinically relevant for human health, we investigated whether IL-6 could also regulate motility of human CD4 T cells through the mitoStat3/mitoCa²⁺ axis. We used CD4 T cells isolated from peripheral blood mononuclear cells (PBMC) of healthy volunteers. Since the cytokine production in human CD4 T cells often differs from mouse CD4 T cells, we first examined their ability to produce IL-6 after activation with anti-CD3 and anti-CD28 mAbs. As expected, no IL-6 was detected in activated mouse CD4 T cells (SI Appendix, Fig. S8). Interestingly, however, high levels of IL-6 were produced by activated human CD4 T cells (SI Appendix, Fig. S8). Considering that the frequency of memory/activated CD4 CD45RO T cells in PBMC from healthy subjects is relatively high, we examined the production of IL-6 in naïve CD4 CD45RA T cells (CD4 RA) and memory/activated CD4 CD45RO T cells (CD4 RO) (44) from a number of healthy volunteers. CD4 RA and CD4 RO cells were activated with anti-CD3/CD28 Abs for 24 h, and IL-6 production was examined by enzyme-linked immunosorbent assay (ELISA). While the levels of IL-6 were almost undetectable in CD4 RA cells, high levels of IL-6 were produced by CD4 RO cells in all tested individuals, and there was heterogeneity in the actual IL-6 levels being produced among individuals (Fig. 4A).

Fig. 4.

IL-6 increases human CD4 T-cell motility through mitochondrial Ca²⁺. (A) Isolated CD4 RA and CD4 RO cells from healthy individuals were activated with anti-human CD3 and CD28 Abs and IL-6 production after 24 h of activation was determined by ELISA (n = 7). (B and C) Isolated CD4 RA and CD4 RO cells were activated with anti-CD3/CD28 mAbs for 48 h, stained for TMRE (B) or Rhod2-AM (C), and analyzed by flow cytometry. (D and E) Isolated CD4 RO cells from healthy volunteers were activated with anti-CD3/CD28 mAbs in the absence or presence of an anti–IL-6 Ab (10 µg/mL). After 48 h, cells were stained with TMRE (D) or Rhod2-AM (E) and analyzed by flow cytometry. (F and G) Isolated CD4 RA cells from healthy volunteers were activated with anti-CD3/CD28 mAbs in the absence or presence of a human IL-6 (50 ng/mL). After 48 h, cells were stained with TMRE (F) or Rhod2-AM (G) and analyzed by flow cytometry. (H) CD4 RA were activated in the presence or absence of IL-6 for 48 h, washed, and cells were placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (I) CD4 RA were activated in the presence IL-6 for 48 h, treated with Ru360 or vehicle for the last 24 h, washed, and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (J) CD4 RA were activated in the presence IL-6 for 48 h, treated with atovaquone or vehicle for the last 20 h, washed, and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (K) Isolated CD4 RO were activated in the presence or absence of anti–IL-6 for 48 h, washed ,and cells were and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). (L) CD4 RO were activated for 48 h, treated with atovaquone or vehicle for the last 20 h, washed, and placed into transwell chambers in medium alone. Migrated cells were counted after 5 h (n = 3). Error bars show ± SD; *P < 0.05, as determined by Student’s t test. Results are representative of two to three experiments.

Since human CD4 RO cells were able to produce IL-6, we examined MMP and mitochondrial Ca²⁺ during activation of isolated CD4 RA and CD4 RO cells. Correlating with their ability of producing IL-6, activated CD4 RO cells showed higher MMP compared with activated CD4 RA cells (Fig. 4B). In addition, analysis of mitochondrial Ca²⁺ by flow cytometry also showed high mitochondrial Ca²⁺ in activated CD4 RO cells compared with CD4 RA cells (Fig. 4C). To examine whether IL-6 produced by CD4 RO cells contributed their higher MMP and mitochondrial Ca²⁺ in an autocrine manner, CD4 RO cells were activated in the absence or presence of a blocking anti–IL-6 Ab. Blocking IL-6 during activation of CD4 RO cells reduced both MMP (Fig. 4D) as well as mitochondrial Ca²⁺ (Fig. 4E). Thus, IL-6 produced by CD4 RO cells acts in an autocrine manner to increase mitochondrial Ca²⁺ during activation. We then examined the effect of IL-6 on CD4 RA activated in the absence or presence of IL-6. After 2 d, MMP and mitochondrial Ca²⁺ were examined. The presence of IL-6 during activation increased MMP (Fig. 4F) and mitochondrial Ca²⁺ (Fig. 4G) of CD4 RA cells. In addition, we examined whether IL-6 produced by CD4 RO cells is sufficient to enhance mitochondrial Ca²⁺ of CD4 RA cells during activation. Isolated CD4 RA cells were activated by themselves or together with purified CD4 RO cells in the presence or absence of anti–IL-6 Ab. After 2 d, mitochondrial Ca²⁺ was examined in CD4 RA cells by gating on CD45 RA. The presence of CD4 RO cells increased mitochondrial Ca²⁺ in CD4 RA cells, but this increase was blocked with anti–IL-6 Ab (SI Appendix, Fig. S9). Together, these results show that that this alternative signaling pathway of IL-6 also takes place in both human naïve and memory CD4 cells.

We then investigated whether the presence of IL-6 during activation of CD4 RA cells could promote their motility. CD4 RA cells were activated in the presence or absence of IL-6 for 2 d, washed, and placed in transwell chambers in medium only. The number of migrated CD4 RA activated in the presence of IL-6 was higher than the number of cells activated in the absence of IL-6 (Fig. 4H). To determine whether the effect of IL-6 on cell motility was mediated by mitochondrial Ca²⁺, CD4 RA cells were activated with IL-6 for 2 d and treated with Ru360. Fewer migrated cells were recovered when cells were treated with Ru360 compared to vehicle (Fig. 4I). Thus, the presence of IL-6 during activation of human CD4 RA cells enhances their motility by sustaining mitochondrial Ca²⁺.

To investigate the role of mitoStat3 on the effect of IL-6 on motility of human CD4 cells, we used pharmacological inhibitors. However, most available Stat3 inhibitors target the ability of Stat3 to mediate transcription, either by blocking translocation of Stat3 to the nucleus or its ability to bind DNA, but they have no effect on mitoStat3 (17, 18, 45, 46). Accordingly, we have shown that Static, an inhibitor of Stat3 dimerization (47), does not prevent IL-6 from increasing MMP in mouse CD4 cells (17). Similarly, treatment with STA21 (inhibitor of Stat3 dimerization and DNA binding) (48) and FLLL32 (an inhibitor of Stat3 Tyr⁷⁰⁵ phosphorylation) (49) had no effect on MMP in mouse CD4 cells activated in the presence of IL-6 (SI Appendix, Fig. S10 A and B). Atovaquone is a small molecule used as anti-malaria drug for its effect on mitochondria respiration (50), but it also acts as an inhibitor of Stat3 (51), and recently it has been suggested to inhibit mitoStat3 (52). Interestingly, unlike the inhibitors of the Stat3 canonical pathway, atovaquone caused a major reduction on MMP in mouse CD4 cells activated with IL-6 (SI Appendix, Fig. S10C). No effect of atovaquone on the expression of activation markers such as CD44, CD62L, and CD69 was observed (SI Appendix, Fig. S10D), indicating that this was not an overall effect on activation of CD4 cells. Since atovaquone has been shown to reduce Ser⁷²⁷ as well as Tyr⁷⁰⁵ phosphorylation of Stat3 (51), we tested the effect of atovaquone on phosphorylation of Stat3 in CD4 cells activated with IL-6. In addition to Tyr⁷⁰⁵ phosphorylation, atovaquone also reduced the levels of phospho-Ser⁷²⁷ Stat3 (SI Appendix, Fig. S10E). To further verify that the effect of atovaquone on MMP was partially mediated by Stat3, we tested its effect in Stat3-deficient CD4 cells activated in the presence of IL-6. Atovaquone reduced mitochondrial Ca²⁺ in WT CD4 cells but had no effect on Stat3-deficient CD4 cells (SI Appendix, Fig. S10F), supporting the idea that most of the effect of atovaquone was mediated through the inhibition of Stat3.

We then examined the effect of atovaquone on human CD4 cells. Atovaquone also reduced mitochondrial Ca²⁺ in human CD4 RA cells activated with IL-6 (SI Appendix, Fig. S10G). We then investigated the effect of atovaquone on the motility of human CD4 RA cells activated in the presence of IL-6 using the transwell chambers. Atovaquone reduced the number of cells that migrated through the transwell compared to vehicle (Fig. 4J), supporting the role of mitoStat3 in mediating the effect of IL-6 on motility of human CD4 cells.

Since IL-6 produced by CD45 RO cells sustains their mitochondrial Ca²⁺ in an autocrine manner, we investigated whether IL-6 could enhance their motility also in an autocrine manner. CD4 RO cells were activated with anti-CD3/CD28 Abs alone or in the presence of a blocking anti-IL6 Ab. A total of 2 d after activation, cells were washed and placed into the transwell chambers with medium alone. A large number of CD4 RO cells migrated through the transwell, but this number was markedly lower when cells were activated in the presence of anti–IL-6 Ab (Fig. 4K). In addition, treatment with atovaquone reduced the number of migrated CD45 RO cells (Fig. 4L). Thus, the ability to produce autocrine IL-6 during activation confers CD4 RO cells with a superior capacity to move in the absence of chemokines just by increasing mitochondrial Ca²⁺ through Stat3.

Discussion

IL-6 is a key immunomodulator, primarily through its effects on CD4 and CD8 T cells. The beneficial effect of blocking IL-6 signaling in different inflammatory diseases such as rheumatoid arthritis demonstrates the contribution of this cytokine to the pathogenic aspect of the immune response. Similarly, blocking IL-6 suppresses the harmful effects of T-cell immunotherapies in cancer treatment, further demonstrating the essential role of this cytokine in keeping the balance of T-cell immune response. However, IL-6 also plays an important role in antibody production and protective immunity through the regulation of specific cytokine gene expression. IL-6 by itself induces IL-21 gene expression in both CD4 and CD8 T cells to contribute to antibody response. IL-6 also promotes IL-4 and IL-13 gene expression in CD4 T cells in the context of allergic airway inflammation (53). Together with TGF-β, IL-6 promotes the expression of IL-17 in inflammatory disease. Regulation of gene expression of these cytokines by IL-6 is predominantly mediated directly or indirectly through Stat3 acting as a transcription factor. Here, we describe that the presence of IL-6 during activation of CD4 T cells improves their ability to move even when IL-6 or chemokines are not present. This effect of IL-6 is mediated by increasing mitochondrial Ca²⁺ and is independent of Stat3 transcriptional activity but uses mitochondrial Stat3.

For most tissues, cell motility is not essential because of the static stage of the cells with no need for the cells to move or migrate. In contrast, some cells from the immune system are constantly moving and migrating. Specifically, motility is key for T cell–mediated immune response. T cells migrate through the lymphatics and the blood stream and reach tissues. This type of migration is primarily directional and mediated by chemokines. However, naïve and activated CD4 and CD8 T cells are constantly moving without a specific direction searching for antigen-presenting cells or antigen-specific target cells. While there are a number of studies examining T-cell migration in response to chemokines, little is known about factors that could affect the motility characteristics (distance and velocity) of these cells to move within a given microenvironment. Here, we show that the presence of IL-6 during activation of CD4 T cells improves their fitness and motility by increasing the distance that the cells travel as well as the velocity of the cells. Interestingly, this function of IL-6 is dependent on mitochondrial Ca²⁺. In addition to pharmacological approaches, here we used MCU-deficient mice to address the contribution of mitochondrial Ca²⁺ to CD4 T-cell function. Interestingly, our studies show that MCU seems to be dispensable for activation and proliferation of CD4 cells, but it is necessary for IL-6 to enhance motility of CD4 T cells. While we detect a major reduction in the migration of MCU-deficient CD4 T cells, migration is not totally abrogated, and some mitochondrial Ca²⁺ levels remain in these cells. Similar results on mitochondrial Ca²⁺ in other cells/tissues in MCU-deficient mice have led to the speculative proposal for the presence of a compensatory mechanism for MCU deficiency, including the potential effect through MCUb (54). Mitochondrial Ca²⁺ has been shown to be essential for sperm motility and migration of other cell types (22, 23). The specific mechanism by which mitochondrial Ca²⁺ regulates cell migration has not been fully dissected, but it augments the cytoskeleton dynamics and thereby decreases cell stiffness, an important mechanical aspect for cell motility (20, 55–57). Here, we show that IL-6 also decreases cytoskeleton stiffness in CD4 T cells, and this can facilitate the motility of the cells.

Previous studies have associated IL-6 with migration of human CD4 Tregs and CD4 RO cells by up-regulating chemokine receptor expression or chemokine production (58–60). Here, we show that the presence of IL-6 during activation increases the motility of CD4 cells without chemokines, since the cells are extensively washed and incubated in medium alone. Furthermore, our live imaging studies show that the effect of IL-6 is primarily on motility, since the cells previously activated in the presence of IL-6 when placed in medium alone travel a longer distance (nondirectional) and at faster speeds. When these cells are placed in medium containing CCL19, they are still capable of similar distances and speeds. We also show no effect of IL-6 on chemokine receptors or chemokine expression on CD4 T cells, and the effect of IL-6 on CD4 T-cell motility did not require Stat3-mediated transcription. It has been reported that IL-6 and IL-8 could be acting as direct chemokines in some subpopulations of nonactivated human CD4 cells (58). However, in our studies, while IL-6 is present during activation, it is not present during the motility assays (transwells or live imaging). When CD4 T cells were treated only with IL-6 without activation, no sign of motility could be detected. Instead, the studies here indicate that late during activation, the presence of IL-6 helps to sustain mitochondrial Ca²⁺, and this improves the mechanical fitness of the cells that enhances the motility.

The function of Stat3 as a transcription factor has been extensively characterized in CD4 and CD8 cells (61, 62). In contrast, very few studies have investigated the regulation and function of mitoStat3 in these and other cells from the immune system despite numerous studies in other cell types (18, 36, 63). We have previously shown that Stat3 translocates to the mitochondria during activation of CD4 cells in the presence of IL-6 (17). In this study, we have investigated the function of mitoStat3 using CD4 cells from S⁷²⁷A KI mice. Unlike Stat3 KO mice, these mice are healthy, and Stat3 transcriptional activity is not impaired in these mice, suggesting that mitoStat3 is not as essential as nuclear Stat3 for most functions (39, 64). Similarly, while Stat3-deficient CD4 cells have impaired proliferation and survival upon activation, these functions were not affected in CD4 cells from S⁷²⁷A KI mice. However, IL-6 failed to sustain mitochondrial Ca²⁺, and, more importantly, motility of CD4 cells activated in the presence of IL-6 was severely compromised. Thus, the functions of mitoStat3 in CD4 cells seem to be more limited than the role of canonical Stat3.

Considering the pathogenic role of CD4 cells and IL-6 in a number of inflammatory diseases such as rheumatoid arthritis, the results from our study provide two alternative pathways that could be used as targets for treatment of these inflammatory diseases. Inhibition of mitoStat3 could be a strategy to reduce mitochondrial Ca²⁺ and CD4 migration. Moreover, considering the low toxicity reported for some inhibitors of mitochondrial Ca²⁺, these could also be developed as therapeutics. In addition to inflammatory diseases, these two pathways could be also targeted to minimize the cytokine release syndrome caused by some immunotherapies as well as acute respiratory distress syndrome.

Material and Methods

Animals Studies.

C57BL/6J mice were purchased from Jackson Laboratories. MCU^f/f mice (65), Stat3^f/f mice (66), mut-Stat3 mice (38), and mS727A Stat3 KI mice (67) have been previously described. More details are provided in SI Appendix.

Purification and Activation of CD4 Cells.

Mouse CD4 T cells were obtained from lymph nodes and spleens using negative selection depletion. CD4 cells were activated with plate-bound anti-CD3 (5 μg/mL) and soluble anti-CD28 (1 μg/mL) mAbs in the presence or absence of IL-6 (50 ng/mL). For human CD4 T cells, blood was obtained from healthy volunteers, from the Colorado Children Hospital Blood bank, or from healthy volunteers at the University of Vermont. All volunteers provided informed consent. This study was approved by the University of Vermont Institutional Review Board. CD4 cells were isolated and activated with anti-CD3 antibody (10 μg/mL) and anti-CD28 Ab (4 μg/mL) in the presence or absence of human IL-6 (50 ng/mL). More details regarding other cytokines or pharmacological inhibitors are provided in SI Appendix.

Transwell Migration Assay.

For transwell migration, CD4 cells 1 to 2 × 10⁴ were washed extensively with phosphate-buffered saline and placed into the transwell. Medium alone or containing 100 ng/mL CCL19 was added to the wells. The number of cells in the wells that migrated from the transwells after 4 h was counted by trypan blue staining. More details are provided in SI Appendix.

Live Microscopy Analyses.

For live microscopy, glass-bottom microwell dishes were coated with ICAM-1Fc (2 μg/mL). CD4 cells at 1.5 × 10⁵ per dish were allowed to migrate. Cell migration in the presence or absence of CCL19 (100 ng/mL) was tracked by time-lapse microscopy with a Zeiss LSM-510 inverted microscope. The migration speed and distance of individual cells were determined with ImageJ. More details are provided in SI Appendix.

In Vivo Migration Experiments.

CD4 T cells from WT, MCU KO, or mS727A Stat3 KI mice, labeled with CFSE or CellTrace Violet, activated in vitro for 48 h with or without IL-6 as described in the Purification and Activation of CD4 Cells section and were injected at a 1:1 ratio i.v. into WT hosts. After 16 h, hosts mice were harvested, and spleen cells were stained with anti-CD4 Ab followed by flow cytometry analysis. More details are provided in SI Appendix.

Flow Cytometry Analysis.

MMP analysis was performed by staining CD4 T cells with TMRE as previously described (17). Mitochondrial calcium analysis was performed by staining with Rhod-2 AM. Different antibodies were used for analysis of activation markers. Samples were then analyzed by flow cytometry using the Cytek Northern Lights (Cytek) or LSRII (BD Biosciences) flow cytometers. More details are provided in SI Appendix.

Western Blot Analysis.

Western blot analysis was performed using whole-cell extracts and specific antibodies. Details about the methods and specific antibodies are provided in SI Appendix.

Confocal Microscopy Analysis.

Analysis of F-actin, vinculin, pMLC2 (S19), and Stat3 was performed by confocal microscopy. CD4 cells after activation were fixed, permeabilized, and stained with specific antibodies or with rhodamine phalloidin to label F-actin. Scanning confocal microscopy analyses were performed using a Zeiss LSM 780 Confocal Laser Scanning microscope. More details are provided in SI Appendix.

Cytokine production.

The IL-6 concentrations in culture supernatants were examined by ELISA. More details are provided in SI Appendix.

Statistical Analysis.

The significance of differences between two groups was determined using GraphPad Prism version 8.0, by standard Student’s t test. The significance of differences among more than two groups was determined by one-way or two-way ANOVA. Standard P < 0.05 was used as the cutoff for significance. More details are provided in SI Appendix.

Data Availability

All study data are included in the article and/or supporting information.