Inhibition of NFAT Signaling Restores Microvascular Endothelial Function in Diabetic Mice

Cardiovascular complications are the major cause of morbidity and mortality in patients who have diabetes. Based on current trends, the rising incidence of diabetes (expected to reach 700 million people worldwide by 2025) will undoubtedly equate to increased cardiovascular mortality (1). Central to the development of diabetic vascular complications is endothelial dysfunction, which appears well before any clinical sign and, importantly, is potentially reversible. It is a generalized phenomenon that takes place in both large blood vessels and the microvasculature (2). Impaired endothelial function is characterized by a decreased production or bioavailability of nitric oxide (NO), leading in the larger vessels to increased vasoconstriction, inflammation, and thrombosis and predisposing to atherosclerosis while contributing in the microvasculature to the development of retinopathy, nephropathy, cardiopathy, skin lesions, and cerebrovascular dysfunction (3). The molecular mechanisms leading to endothelial dysfunction in diabetes are still not clear.

Previous studies from our group demonstrate that hyperglycemia is an effective stimulus for activation of the calcium (Ca²⁺)/calcineurin-sensitive transcription factor nuclear factor of activated T cells (NFAT) c3 in conduit and medium-sized resistance arteries (4–6). This activation involves the local release of extracellular nucleotides (i.e., uridine triphosphate and uridine diphosphate) acting on purinergic P2Y2/4 and P2Y6 receptors, leading to increased intracellular Ca²⁺ and subsequent activation of calcineurin and NFAT (4). High glucose also reduces nuclear export of NFATc3 by inhibiting glycogen synthase kinase-3β and c-Jun N-terminal kinase, contributing to increased NFATc3 nuclear accumulation (4).

In a series of follow-up publications primarily focusing on the smooth muscle, we showed that once activated, NFAT increases the expression of inflammatory mediators in the arterial wall, such as osteopontin (OPN), cyclooxygenase-2, interleukin (IL)-6, vascular cell adhesion molecule 1, tissue factor, and allograft inflammatory factor-1 (5–9). More recently, we demonstrated a role for NFAT in diabetes-driven atherosclerotic plaque formation via regulation of inflammatory mediators and antioxidant defenses (i.e., NOX4 and catalase) (6,10). However, it is still unknown whether NFAT plays a role in diabetes-driven endothelial dysfunction, particularly in microcirculatory beds, where vascular complications of diabetes are commonly found. Interestingly, in cultured bovine aortic endothelial cells, blockers of calcineurin/NFAT signaling (i.e., cyclosporine A and tacrolimus [FK506]) have been shown to upregulate endothelial NO synthase (eNOS) expression (11). We have shown that the skin microcirculation provides a representative model of generalized microvascular function and is useful for in vivo experimental mechanistic studies of endothelial function in small animals (12,13).

In this article, we hypothesize that NFAT is expressed in skin microvessels, that hyperglycemia may activate NFAT in skin microvessels, and that in vivo blockade of NFAT with A-285222 can attenuate microvascular dysfunction in diabetic mice. We also explore whether eNOS expression may be modulated by NFAT signaling in this context.

An expanded version of this section is available in the Supplementary Data. Adult male diabetic Akita (Ins2^+/−) mice were used according to the study protocol shown in Fig. 1A as well as crossbred with FVBN 9x-NFAT-luciferase reporter (NFAT-luc) mice (4,14,15) to generate Akita/NFAT-luc mice and nondiabetic wild-type (WT)/NFAT-luc littermates. Akita mice have a point mutation of the insulin 2 gene, which leads to pancreatic β-cell apoptosis, hypoinsulinemia, and severe hyperglycemia. Nonobese Akita mice develop type 2 diabetes phenotypes, including peripheral and hepatic insulin resistance and cardiac remodeling, and as such are a good model for studying vascular complications of subjects with type 1 and lean type 2 diabetes (16). Microvascular function was assessed noninvasively and longitudinally using a laser Doppler imager in combination with iontophoresis of endothelium-dependent and -independent vasodilators (acetylcholine [ACh] and sodium nitroprusside [SNP], respectively) and localized heating as previously described (12,13). To standardize basal perfusion, blood vessels were preconstricted with iontophoresis of 1% phenylephrine (PE) for 5 min (current of 100 μA), prior to iontophoresis of ACh or SNP. To determine the contribution of endothelium-derived NO to ACh-mediated vasodilatation, microvascular responses to ACh were assessed 30 min after i.p. injection of the nonselective inhibitor of NO synthase, l-N^G-nitro-l-arginine methyl ester (l-NAME; Sigma Chemicals) (20 mg/kg) (12). On a separate day and to determine the maximum microvascular dilator capacity, a hyperemic response was initiated by localized heating of the skin using a specially designed heating probe (SH02 skin heating unit and SHP3 probe; Moor Instruments Ltd, Axminster, U.K.). Perfusion was measured continuously as the temperature of the heating probe was increased to 42°C (1°C/min) and maintained for >10 min, which was sufficient for maximum vasodilatation to plateau.

Mice were euthanized by cervical dislocation after anesthesia with 3% isoflurane in oxygen (2 L/min) or, for NFAT-luc experiments, anesthetized by i.p. injection of 7.5 mg ketamine hydrochloride and 2.5 mg xylazine per 100 g body weight and euthanized by exsanguination through cardiac puncture. All animal protocols were performed in accordance with U.K. Home Office regulations (project license PIL60/4265) and the Malmö/Lund Animal Care and Use Committee (permits M78–10, M29–12, and M9–15) and abided by the Guide for the Care and Use of Laboratory Animals published by Directive 2010/63/EU of the European Parliament. Luciferase activity was measured in the skin, aortic arch, abdominal aorta, and carotid arteries in tissue homogenates as previously described (14). Skin sections were used for histology (hematoxylin and eosin), immunohistochemistry (eNOS), or quantification of NFATc3 nuclear accumulation using confocal microscopy as previously described (14,17,18). Organ culture of skin biopsies and dissociation of dermis and epidermis were done according to established protocols (19–22). Skin fractions were used for gene expression, absolute quantification of NFAT isoforms, and Western blot analysis. Plasma cytokines, endothelial activation markers, and total NO were measured in blood samples taken from the tail vein; albumin and creatinine were measured from urine. Total NO-related species (nitrite, nitrate, and S-nitrosothiols) were measured 4 weeks after treatment with A-285222 (Post-2) using a gas-phase chemiluminescence reaction of NO with ozone using a Sievers NO analyzer model 280i (Analytix Ltd, Boldon Colliery, U.K.). Blood pressure was assessed using a noninvasive tail-cuff blood pressure system. A-285222 inhibits all NFAT family members and was provided by Abbott Laboratories (Abbott Park, IL). This drug belongs to a series of 3,5-bis(trifluoromethyl)pyrazole derivatives demonstrated to maintain NFAT in a phosphorylated state, blocking its nuclear import and subsequent transcription, without affecting nuclear factor-κB or AP-1 activation or calcineurin phosphatase activity (23). Results are expressed as means ± SEM unless specified otherwise.

While no significant differences in the responses to PE at baseline were found between nondiabetic WT and Akita mice, microvascular responses to iontophoresis of ACh were impaired in Akita mice (Fig. 1B), with less vasodilation to ACh than in nondiabetic WT mice. Diabetic Akita mice were treated with the NFAT blocker A-285222 (0.29 mg/kg body weight) or vehicle (saline) for 4 weeks, and microvascular function was assessed before (Pre) (Fig. 1C), directly after (Post-1) (Fig. 1D), or 4 weeks after the end of the treatment (Post-2) (Fig. 1E). Before commencement of treatment, baseline measurements of microvascular function were not significantly different among the groups (Fig. 1C). Treatment with A-285222 resulted in improved ACh responses compared with vehicle-treated controls at both time points post-treatment (Fig. 1D and E). Figure 1F summarizes maximum responses to ACh at each time point, showing a loss of endothelial function in Akita mice that is not observed in mice treated with A-285222. Treatment had no effect on blood glucose or body weight (Fig. 1G and H).

Pretreatment with the NO synthase inhibitor l-NAME 30 min prior to iontophoresis of ACh completely abrogated the improved response observed after treatment with A-285222 and had no impact on the responses measured in vehicle-treated mice (Fig. 2A). The dotted line in Fig. 2A indicates the ACh response in mice after treatment with A-285222 in the absence of l-NAME treatment (corresponding to Post-1 data in Fig. 1F). Iontophoresis of the NO donor SNP, in contrast, eliminated the differences in microvascular responses observed after treatment among the groups, effectively vasodilating the vessels of vehicle-treated mice to the same extent as those that had been treated with A-285222 (Fig. 2B). Treatment with A-285222 had no effect on skin microvascular responses to ACh (regardless of whether animals were treated or not with l-NAME prior to iontophoresis of ACh) and no effect on the responses to SNP on blood glucose levels or body weight in control nondiabetic WT mice (Supplementary Figs. 1 and 2A and B). No significant differences were found in the responses to PE in WT and Akita mice with and without A-285222 (data not shown).

Skin microvascular responses to localized heating, which are largely dependent on NO (24), were comparable between NFAT-treated and saline-treated groups when measured prior to initiating the treatment (Fig. 2C). Four weeks’ treatment with the NFAT blocker A-285222 resulted in an improved vasodilator response compared with vehicle-treated mice (Fig. 2D and E). The longitudinal change in maximum vasodilator response (Pre to Post-1) indicates a loss of heat-induced vasodilation in vehicle-treated diabetic mice over 4 weeks (negative delta value), whereas A-285222 treatment improves the vasodilator response to heat (positive delta value), such that the difference between the groups was significantly different (Fig. 2F). Treatment with the NFAT blocker A-285222 had no effect on skin microvascular responses to localized heating in control nondiabetic WT mice (Supplementary Fig. 2C–F).

To investigate NFAT isoforms expressed in mouse skin and, in particular, in skin microvessels, we collected skin from the flank of the mice corresponding to the area used for microvascular endothelial function measurements (Fig. 3A). Because vessels in the mouse skin are predominantly located in the dermis (arrowheads in Fig. 3B), we enzymatically dissociated the dermis from the epidermis for measurements of NFAT isoform expression and NFAT-dependent transcriptional activity. Absolute quantification of NFAT isoform mRNA expression showed that NFATc3 is the predominant isoform in both skin fractions (Fig. 3C), followed by NFATc1, NFATc2, and NFATc4, which were expressed at 6-, 17-, and 630-fold lower copy numbers in dermis when compared with NFATc3. A tendency toward lower NFAT isoform expression was observed in the dermis of Akita mice when compared with control WT mice, but differences were only statistically significant for NFATc4, which is overall expressed at very low levels (Fig. 3D). Using confocal immunofluorescence microscopy of skin whole mounts, we found abundant NFATc3 staining in microvascular endothelial cells, visualized by CD31-positive immunostaining (Fig. 3E). Despite comparable NFAT mRNA levels in dermis and epidermis, the dermis accounted for most of the NFAT-dependent transcriptional activity measured in whole skin samples (Fig. 3F). Treatment with dispase causes separation at the dermal–epidermal basement membrane (21), yielding pure fractions as evidenced by exclusive expression of the endothelial cell marker CD31 in the dermis fraction in Western blot experiments (Fig. 3G).

In samples collected at termination (Post-2) from the same mice used for measurements of microvascular function, we found that in vivo treatment with A-285222 significantly decreased NFATc3 nuclear accumulation in skin microvessels (Fig. 4A and B). NFAT-dependent transcriptional activity in the skin of Akita mice gradually increased with longer durations of high glucose exposure (Supplementary Fig. 3A). Akita/NFAT-luc mice develop pronounced and sustained hyperglycemia from 4 weeks postnatal, failing to gain as much body weight as WT/NFAT-luc littermates, as well as proteinuria from 8 weeks postnatal and further increasing as the mice became older (Supplementary Fig. 3B–D).

In a separate cohort from that used for monitoring time-dependent effects of diabetes in the model (Supplementary Fig. 3), Akita/NFAT-luc and WT/NFAT-luc littermates were treated with A-285222 (0.29 mg/kg body weight) or vehicle (saline) for 4 weeks, and NFAT-dependent transcriptional activity was determined in skin and various arteries. In intact whole skin, a tendency to increased NFAT-luc activity (P = 0.06) was observed in Akita/NFAT-luc mice when compared with nondiabetic WT/NFAT-luc mice, while activity was lower in diabetic mice treated with A-285222 (Fig. 4C). Further analysis of the dermis fraction, where vessels are located, showed a significant diabetes-induced increase in luciferase mRNA, which was completely blocked by treatment with A-285222 (Fig. 4D). The effect of the blocker is unlikely due to effects on glucose metabolism, since neither blood glucose nor body weight was affected by the treatment (Supplementary Fig. 4A and B). In line with previous work from our laboratory in which a streptozotocin-induced diabetes model was used (5,6), NFAT-luc activity was significantly enhanced in the aortic arch of Akita/NFAT-luc mice when compared with WT/NFAT-luc controls, and it was completely abolished by A-285222 (Supplementary Fig. 4C). A similar pattern was observed in the abdominal aorta and carotid arteries (Supplementary Fig. 4D and E). No effects of the treatment were observed in the nondiabetic mice (Fig. 4C and D and Supplementary Fig. 4C and D).

To elucidate the mechanism underlying the effects of NFAT signaling inhibition on microvascular function, we measured eNOS expression in the dermis of Akita/NFAT-luc and WT/NFAT-luc mice after treatment with A-285222 or vehicle. Although we could not detect diabetes-driven changes in eNOS expression at the mRNA level (Fig. 5A), both eNOS and phosphorylated eNOS (p-eNOS) protein expressions were lower in Akita mice compared with nondiabetic WT mice (Fig. 5B). Interestingly, a clear upregulation of eNOS mRNA expression was observed in diabetic mice directly after treatment with A-285222. Although not significant, eNOS mRNA seemed to remain elevated at 9 weeks after the end of the treatment (Fig. 5A), but no significant changes were detected at the protein level (Fig. 5B). Consistent with the elevated eNOS mRNA levels in the dermis, higher circulating levels of NO were found after treatment with A-285222 (Post-2) in the same Akita mice that exhibited improved endothelial function, when compared with vehicle-treated Akita mice (Fig. 5C). Immunohistochemistry confirmed expression of eNOS protein in the wall of skin microvessels in these mice (Fig. 5D).

To assess whether the effects on NFAT transcriptional activity and eNOS expression induced by systemically administered A-285222 could be attributed, at least in part, to local effects of the blocker on the skin, we organ-cultured intact skin samples from NFAT-luc mice using a protocol that has been shown to preserve both skin function and structure (19). Culture of the skin with A-285222 (1 µmol/L) for 24 h resulted in significantly reduced luciferase mRNA and a concomitant increase in eNOS mRNA expression measured in the dermis fractions (Fig. 5E and F). Organ culture of the skin had per se an impact on p-eNOS expression in the dermis, with significantly increased expression levels at 24 h of culture, but levels no different from those measured in freshly dissociated dermis at 48 h (Supplementary Fig. 5). At 48 h, eNOS protein expression was significantly higher in samples that had been cultured with A-285222 compared with controls, while no significant changes were found in p-eNOS protein expression (Fig. 5G).

The improved microvascular responses to ACh and localized heat observed in Akita mice after treatment with A-285222 were accompanied by significantly reduced plasma levels of the proinflammatory cytokines IL-6 and OPN and the endothelial activation marker soluble intercellular adhesion molecule 1 directly after the intervention (Post-1), while other cytokines or markers (e.g., IL-1α, IL-10, and E-selectin) were unaffected (Fig. 6A–F). The inhibitory effect of the blocker on IL-6 and OPN remained evident 4 weeks after the intervention (Post-2). With the exception of IL-1α, all other plasma molecules increased during the study protocol in the vehicle-treated mice as they aged (Fig. 6A–F). Despite elevated levels of circulating IL-6 and OPN, we could not detect changes in the levels of IL-6 or OPN mRNA in the dermis of diabetic or WT mice treated with A-285222 (Supplementary Fig. 6).

In addition to increased levels of circulating NO (Fig. 5C), plasma endothelin-1 (Et-1) was significantly reduced 4 weeks after treatment with A-285222 when compared with levels in vehicle-treated Akita mice (Fig. 7A). At the same time point, both systolic and diastolic blood pressure were significantly lower in A-285222–treated Akita mice compared with vehicle-treated mice (Fig. 7B and C). There was no significant difference in heart rate between groups (445 ± 51 vs. 424 ± 43 bpm for vehicle vs. A-285222; P = 0.767). Treatment with the NFAT blocker A-285222 had no effect on plasma levels of inflammatory cytokines or Et-1 in control nondiabetic WT mice (Supplementary Fig. 7).

As the study progressed, many of the Akita mice appeared ill, and ∼40% either died or had to be euthanized before completion of the study because of failure to thrive (Fig. 7D). Even though Kaplan-Meier curves were not statistically different between the two groups, all mice treated with A-285222 survived until termination of the study (P = 0.0817) (Fig. 7D). Figure 7E shows survival data from an independent study conducted for another project, which included 33 Akita mice (C57BL/6J background; mixed sexes) that were treated with A-285222 or vehicle for 4 weeks as in the current study. In line with the data in Fig. 7D, all A-285222–treated mice survived until termination of the study, whereas a substantial number of vehicle-treated mice died or had to be euthanized (Fig. 7E). Differences between survival curves were statistically significant (P = 0.0299) (Fig. 7E) and remained significant even if only male mice are included in the analysis (P = 0.0085; N = 16).

The findings from the present in vivo study show that diabetes is associated with an increase in NFAT-dependent transcriptional activity and microvascular dysfunction, which can be significantly improved by blocking of NFAT activation. The improvement in microvascular responses was associated with a reduction in vascular NFATc3 nuclear accumulation and NFAT-dependent transcriptional activity in the skin, and it was mediated largely through increased NO availability. Our results point at the NFAT signaling as a novel pathway involved in the development of microvascular dysfunction in diabetes. Blocking of hyperglycemia-induced NFAT transcriptional activity with A-285222 leads to an improvement in microvascular function, which, in combination with the observed reduction in circulating proinflammatory cytokines and Et-1 levels, may contribute to lower blood pressure and increased survival of the mice.

We believe our findings of improvements in the skin microcirculation following blocking of NFAT activation have clinical significance. Measurement of skin microvascular responses to iontophoresis of ACh and localized heating using laser Doppler imaging has proven a successful strategy to monitor and predict diabetic complications in clinical studies. Skin microvascular function is significantly impaired in patients with diabetic retinopathy compared with those without complications (25,26) and with young children with diabetes before clinical signs of microangiopathy (27). Furthermore, skin microvascular responses are strongly correlated to insulin sensitivity and coronary heart disease and its risk factors (28).

Most NFAT-related studies in the context of skin have focused on the role of NFATc2 in epidermal keratinocytes in skin diseases such as psoriasis (29–31). The absolute quantification experiments unexpectedly revealed that NFATc3 is the predominant NFAT isoform in both dermis and epidermis (Fig. 3C). This is interesting considering that NFATc3 seems to be the dominant NFAT isoform in all other vascular beds that we have examined so far, including cerebral, retinal, carotid, aorta, hepatic, renal, femoral, mesenteric arteries, and portal vein (L.M.B. and M.F.G., unpublished observations, and refs. 14,18,32), also considering that NFATc3 is readily activated by elevated extracellular glucose (4–6,32).

The predominance of this particular NFATc3 isoform in the dermis, where the microvessels are mainly located and the increased NFAT-transcriptional activity observed in the Akita mice was seen, suggest that NFAT activation could be contributing to the observed deterioration in microvascular responses. Worth noting, though, is that differences in baseline microvascular responses to ACh between WT and diabetic mice were noticeable already at 7–12 weeks of age, while we could only dissect the difference in whole-skin NFAT-luc activity beyond 12 weeks of age. This could be due to limitations of the technique when using whole-skin homogenates compared with the better resolution obtained when using isolated dermis (see comparison between Fig. 4C and D) but also most likely due to engagement of other hyperglycemia-induced factors recognized as drivers of endothelial dysfunction early on in the process—perhaps increased oxidative stress (33,34) and inflammation (35).

Despite potentially multiple pathways driving the deterioration in microvascular responses in the diabetic mice, we found a clear improvement in skin microvascular function after inhibition of NFAT signaling with A-285222. Treatment effects could be attributed to changes in NFAT activity, as demonstrated by the reduction in NFATc3 nuclear accumulation in skin microvessels and by reduced NFAT-dependent transcriptional activity, as assessed by NFAT-luc activity and luciferase mRNA expression in the dermis. Importantly, the decreased NFAT activation and improved microvascular function were not due to a glucose-lowering effect of the NFAT blocker.

In agreement with our previous findings demonstrating that the skin microvascular response to iontophoresis of ACh in mice is mediated through NO (12), we show in this study that the improvement in microvascular responses to ACh immediately following 4 weeks’ inhibition of NFAT with A-285222 and maintained for 4 weeks after the end of treatment is due to enhanced production of NO in the diabetic mice compared with vehicle-treated mice. This was evidenced by abrogation of the improved microvascular response following inhibition of NO production by l-NAME. The improvement in microvascular responses to ACh comes from endothelium-derived NO and not from a generalized improvement in smooth muscle responsiveness. Further support for increased NO bioavailability after treatment with A-285222 comes from the finding of improved sustained maximum vasodilator response to localized heating, which is largely mediated through an NO-dependent pathway (24). In addition to a significant decrease in NFAT activation in the dermis of diabetic mice after treatment with A-285222, as evidenced by decreased luciferase mRNA, we found significant upregulation of eNOS expression, which further supports the role of NFAT signaling in mediating microvascular responses via increased NO bioavailability. Collectively, these data suggest that inhibition of NFAT with A-285222 leads to enhanced NO bioavailability in the skin of Akita mice, rather than affecting the downstream smooth muscle response to NO.

While our data support the idea that inhibition of NFAT transcriptional activity increases NO bioavailability in the skin microvasculature and while positive eNOS immunostaining was found in skin microvessels, we cannot exclude the possibility of A-285222 affecting other potential NO-producing cells in the skin. Mowbray et al. (36) demonstrated the presence of enzyme-dependent NO production in all cell types of human skin, including keratinocytes, melanocytes, and immune cells, but did not attribute storage site to any particular cell.

Treatment with A-285222 was also associated with significantly elevated plasma NO in diabetic mice, which raises the question of whether the improved microvascular response was the result of systemic or local effects of the drug on the skin. While our findings cannot exclude systemic effects (discussed in the next section), the organ culture experiments showing significantly decreased NFAT-dependent transcriptional activity and concomitantly increased eNOS mRNA and protein expression in the dermis after overnight culture of whole skin with A-285222 certainly support the idea that at least part of the improved microvascular function observed in vivo could be attributed to local effects of the drug on NFAT activity in the skin. We did not observe any effects of NFAT inhibition on the levels of p-eNOS, suggesting a regulatory role in the expression but not the activity of eNOS.

Our findings are at odds with findings from other studies. Yang et al. (37) reported vascular endothelial growth factor–induced activation of calcineurin/NFAT signaling in endothelial progenitor cells, which led to increased eNOS protein expression and NO production, while inhibition of calcineurin with cyclosporine A or inhibition of NFAT with 11R-VIVIT decreased eNOS protein levels and NO production. This discrepancy could be due to cell-specific differences between circulating endothelial progenitor cells (obtained by culture of mononuclear cells from human peripheral blood) and microvascular cells in situ regarding the NFAT isoform engaged in the regulation of NO (NFATc1 in their study) or differences in the transcriptional program engaged by the different stimuli (vascular endothelial growth factor vs. hyperglycemia in our study) or combinations thereof. We and others have previously demonstrated that NFAT isoforms coexpressed in the same cell type can be differentially regulated by the same stimuli and that much of what determines the efficiency of the stimulus is the ability to deliver an appropriate pattern of intracellular calcium signaling in combination with precise regulation of NFAT export kinases (17,38–40). In pancreatic acinar cells, for example, we showed that NFATc3 but not NFATc1 is readily activated by cholecystokinin or ACh (41). Another example of differential control of NFAT isoforms is illustrated by the dramatic NFATc3 activation in the endothelium of retinal microvessels after acute hyperglycemia and the complete lack of NFATc2 response to elevated extracellular glucose in retinal endothelial cells (32). Our data cannot rule out that the effects of A-285222 on microvascular function can be mediated by the other NFATc isoforms, despite NFATc3 being the predominant isoform in this preparation and one readily inhibited by treatment with A-285222.

In another elegant study demonstrating the importance of inositol 1,4,5-triphosphate receptor 1 (IP3R1) for activation of eNOS in endothelial cells and maintenance of blood pressure, Yuan et al. (42) proposed that the increased eNOS expression and phosphorylation downstream of IP3R1 is mediated via calcineurin/NFATc3 signaling. While they provide solid evidence for the involvement of IP3R1 in eNOS activation, the involvement of NFATc3 is less well supported. Conclusions are based on confocal microscopy measurements of NFATc3 nuclear accumulation in a limited number of human umbilical vein endothelial cells, a cell type of questionable relevance for blood pressure control, without further confirmation of changed NFAT-dependent transcriptional activity as a consequence of the nuclear translocation. Because calcineurin interacts not only with NFAT but also with a large number of substrates (43), calcineurin inhibitors (i.e., cyclosporine A used by Yuan et al. [42] and FK506) are ambiguous tools for dissecting the potential involvement of NFAT signaling. Instead, the A-285222 compound used in this study has been shown to maintain NFAT in a phosphorylated state without affecting calcineurin phosphatase activity (23).

Our findings in the skin have potential translational importance to the regulation of blood pressure. The skin can be considered one of the largest organs in the body, and, by virtue of its large NO storage capacity, the potential for regulation of blood pressure by the skin microcirculation has come under recent interest (44). Indeed, our data would support such an idea since changes in NFAT transcriptional activity and eNOS expression after treatment with A-285222 were associated with a significant reduction in blood pressure, although it is also possible that the blood pressure–lowering effects of NFAT blocking could arise from changes in vascular function in areas other than the skin or in combination with changes in the skin. Nieves-Cintrón et al. (45) and Amberg et al. (46) have demonstrated a role of NFATc3 in mediating angiotensin II–induced hypertension via downregulation of the β1 subunit of large conductance, calcium-activated K⁺ channels and Kv1.2 channels in arterial smooth muscle. Interestingly, treatment with A-285222 resulted not only in increased eNOS expression in the skin but also in elevated circulating NO levels. This is in line with what others have shown, namely, that serum total NO-related products correlate directly with those of the upper dermis (36).

Even though the organ culture experiments support the idea of a local direct effect of A-285222 on NFAT activity and eNOS production, it is possible that parallel systemic effects of the blocker or local effects elsewhere than in the skin could lead to changes in circulating molecules, which are capable of affecting microvascular function and blood pressure. We did indeed measure an effect of A-285222 treatment on plasma levels of inflammatory cytokines and Et-1.

In conclusion, in this in vivo study we have shown, with the advantage of longitudinal measurements of microvascular function at different time points in the same animals, that hyperglycemia-induced NFAT activation is associated with endothelial dysfunction in the microcirculation, which can be markedly improved by therapeutic blocking of NFAT activation. The improvement in endothelial function is largely mediated through increased availability of NO, and this, combined with a more anti-inflammatory phenotype and reduction in blood pressure, might underlie the increased survival of A-285222–treated mice. Further studies should explore the potential of NFAT inhibition as a novel therapeutic modality for the treatment of diabetic microvascular dysfunction.

Funding. This work was supported by the British Heart Foundation (project grant PG/12/58/29767), the Swedish Heart and Lung Foundation (20160872), the Swedish Research Council (2018-02837, 2014-03352, and EXODIAB 2009-1039), the Swedish Society for Medical Research, the Swedish Foundation for Strategic Research (LUDC-IRC #15-0067), the Swedish Diabetes Association (Diabetesfonden), The Crafoord Foundation, Albert Påhlsson Foundation, and Knut and Alice Wallenberg Foundation. Support was also provided by the Innovative Medicines Initiative Joint Undertaking SUMMIT (115006), comprising funds from the European Union’s Seventh Framework Programme (FP7/2007-2013) and European Federation of Pharmaceutical Industries and Associations companies’ in-kind contribution.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. E.G.-V., A.D.M., L.M.B., A.A., J.R.G., A.-M.D.A., and A.V.Z. performed the experimental work. All authors contributed to the design of the study, the analysis and interpretation of the data, and the writing of the manuscript. R.J.M., M.F.G., and F.K. secured funding for the study. M.F.G. and F.K. conceived of and supervised all parts of the study and wrote the article. All authors critically revised and approved the final version of this article. M.F.G. and F.K. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.