Possible New Strategies for the Treatment of Congenital Hyperinsulinism

Objective Congenital hyperinsulinism (CHI) is a rare disease characterized by persistent hypoglycemia as a result of inappropriate insulin secretion, which can lead to irreversible neurological defects in infants. Poor efficacy and strong adverse effects of the current medications impede successful treatment. The aim of the study was to investigate new approaches to silence β-cells and thus attenuate insulin secretion. Research Design and Methods In the scope of our research, we tested substances more selective and more potent than the gold standard diazoxide that also interact with neuroendocrine ATP-sensitive K+ (KATP) channels. Additionally, KATP channel-independent targets as Ca2+-activated K+ channels of intermediate conductance (KCa3.1) and L-type Ca2+ channels were investigated. Experiments were performed using human islet cell clusters isolated from tissue of CHI patients (histologically classified as pathological) and islet cell clusters obtained from C57BL/6N (WT) or SUR1 knockout (SUR1-/-) mice. The cytosolic Ca2+ concentration ([Ca2+]c) was used as a parameter for the pathway regulated by electrical activity and was determined by fura-2 fluorescence. The mitochondrial membrane potential (ΔΨ) was determined by rhodamine 123 fluorescence and single channel currents were measured by the patch-clamp technique. Results The selective KATP channel opener NN414 (5 µM) diminished [Ca2+]c in isolated human CHI islet cell clusters and WT mouse islet cell clusters stimulated with 10 mM glucose. In islet cell clusters lacking functional KATP channels (SUR1-/-) the drug was without effect. VU0071063 (30 µM), another KATP channel opener considered to be selective, lowered [Ca2+]c in human CHI islet cell clusters. The compound was also effective in islet cell clusters from SUR1-/- mice, showing that [Ca2+]c is influenced by additional effects besides KATP channels. Contrasting to NN414, the drug depolarized ΔΨ in murine islet cell clusters pointing to severe interference with mitochondrial metabolism. An opener of KCa3.1 channels, DCEBIO (100 µM), significantly decreased [Ca2+]c in SUR1-/- and human CHI islet cell clusters. To target L-type Ca2+ channels we tested two already approved drugs, dextromethorphan (DXM) and simvastatin. DXM (100 µM) efficiently diminished [Ca2+]c in stimulated human CHI islet cell clusters as well as in stimulated SUR1-/- islet cell clusters. Similar effects on [Ca2+]c were observed in experiments with simvastatin (7.2 µM). Conclusions NN414 seems to provide a good alternative to the currently used KATP channel opener diazoxide. Targeting KCa3.1 channels by channel openers or L-type Ca2+ channels by DXM or simvastatin might be valuable approaches for treatment of CHI caused by mutations of KATP channels not sensitive to KATP channel openers.


INTRODUCTION
Congenital hyperinsulinism (CHI) is a rare heterogeneous genetic disorder, but the most frequent cause of severe, persistent hypoglycemia in neonates, infants and children. The main reasons for developing CHI are defects in important genes regulating pancreatic b-cell function. To date, mutations in 14 essential genes controlling insulin secretion have been reported including ABCC8 and KCNJ11. ABCC8 and KCNJ11 genes encode the K ATP channel subunits SUR1 and Kir6.2, respectively, and mutations in these genes represent the most prevalent cause of CHI. Defects in these genes are responsible for the failure of b-cells to respond to normal regulatory mechanisms, leading to inappropriate and excessive insulin release despite low blood glucose concentrations resulting in frequent episodes of hypoglycemia (1,2). There are some excellent reviews giving detailed information about molecular mechanisms underlying the pathophysiology of CHI (1)(2)(3)(4)(5).
Based on histopathological observations, three distinct forms of CHI are described: focal, diffuse and atypical. In focal CHI affected b-cells are localized only in small specific parts of the pancreas. Conversely, in diffuse CHI all pancreatic b-cells seem to be affected (6). If the histology of the tissue does not fit in one of the forms, it is regarded as an atypical form of CHI. It is characterized by a mosaic-like assembly of hyper-functional islets spread over the pancreas (7).
Persistent hypoglycemia is responsible for seizures and finally for severe brain damage (8). Thus, it is necessary to diagnose CHI rapidly and to start as early as possible with a suitable treatment. Treatment options include medical therapy and surgical intervention (9). First-line drug for treating CHI is the K ATP channel agonist diazoxide (10). However, numerous side effects of diazoxide limit its use. Some of the most common undesired effects are Na + and fluid retention, hypertrichosis and loss of appetite. Life threatening side effects also occur including cardiac failure, pulmonary hypertension, hyperuricemia, bone marrow suppression, and anemia (11)(12)(13)(14)(15)(16). Additionally, diazoxide is only effective when K ATP channels are functional (10). Alternatives to the therapy with diazoxide and novel medications include glucagon, somatostatin analogues, nifedipine, GLP1-receptor antagonists [exendin- ], and sirolimus [ (17)(18)(19)(20)(21)(22), reviewed in (3)]. Many of these drugs act by lowering the Ca 2+ influx into b-cells (23)(24)(25). The aforementioned drugs also have numerous undesirable effects, which may be a reason for reconsidering their therapeutic usefulness: gastrointestinal symptoms, formation of gall stones, suppression of pituitary hormones, necrotizing enterocolitis, hypotension, immune suppression, thrombocytosis, impaired immune response, and many more (26)(27)(28)(29)(30)(31). Recently, a new full human monoclonal antibody to the insulin receptor XMetD (also known as XOMA 358 or RZ358) has been proposed as a novel therapeutic strategy (32)(33)(34)(35). First results in a Phase 2a clinical trial exhibited an improved glycemic control in patients with persistent hypoglycemia (36).
In patients that cannot be treated sufficiently with drugs, surgical treatment is indicated. While partial pancreatectomy is beneficial for patients with focal CHI (37,38), in case of diffuse and drug-unresponsive CHI, near-total pancreatectomy is usually required (39,40). Due to different post-operative complications like recurrent hypoglycemia, pancreatic exocrine insufficiency and diabetes, patients with diffuse CHI are far from being cured after surgery (41,42). In order to reduce the development of diabetes postsurgically, a 70 to 90% resection of pancreas have been considered; however, the outcome is still unpredictable (39,43).
Taken together, it is of great importance to explore new pharmacological options for CHI therapy in order to maintain euglycemia and reduce severe side effects from current medical and surgical treatment. Aim of this study was to find new strategies, which are able to silence b-cells by inhibiting extensive Ca 2+ influx into the cell. For this purpose, new and approved drugs interacting with K ATP channels and with K ATP channel-independent targets have been tested on islet cell clusters obtained from biopsies of CHI patients and islet cell clusters from WT and SUR1 -/mice.

Cell and Islet Preparation
Human islets of Langerhans were obtained from different biopsies of children undergoing pancreatic surgery. Ethics approval for the study involving human participants was approved by the ethic commission of the Universitätsmedizin Greifswald (BB 050/13). Written informed consent was provided by the legal guardians of the children for the study. The islets were taken from biopsies of eight CHI patients. Genetic studies showed that seven patients had mutations in the ABCC8 gene encoding the SUR1 subunit of K ATP channels. In one biopsy no mutation was found for eight genes tested ( Table 1: 2). According to postsurgical evaluation of the biopsies by the Department of Pathology at the University Hospital Greifswald, the tissue was identified as pathological and assigned to the CHI type (mosaic, diffuse or focal). Islets of these pathological samples were isolated by injecting collagenase (2-4 mg/ml) into the biopsy material and by handpicking islets after digestion at 37°C. Afterward, islets were cultured in a CMRL 1066 medium with 5.5 mM glucose supplemented with 10% fetal calf serum, 100 U/ml of penicillin, 100 mg/ml of streptomycin, 10 mM HEPES, and 2 mM L-glutamine. Next day, the samples were shipped to Tübingen and/or Münster for further analysis.
Mouse islets of Langerhans were isolated from adult C57BL/ 6N (WT) mice or SUR1 knockout (SUR1 -/-) mice on a C57BL/ 6N background. The mice were bred in the animal facility of the Department of Pharmacology at the University of Tübingen. The principles of laboratory animal care (NIH publication no. 85-23, revised 1985) and German laws were followed. The animal study was reviewed and approved by the Regierungspräsidium Tübingen ( § 4 Abs. 3 TierSchG). Islets were isolated and cultured as previously described (58). For experiments, human or mouse islet cell clusters of similar size were used, obtained by dispersing islets by trypsin treatment. Human and mouse islet cell clusters were kept in cell culture up to 3 days.

Measurements of the Mitochondrial Membrane Potential (DY)
DY was measured by Rh123 fluorescence at 480 nm excitation wavelength as described in (59). One data point was measured every 3 s. The effects were evaluated by averaging the values of the last 60 s of each interval before solution change. At the end of each experiment FCCP (0.5 µM) was added to evaluate maximal mitochondrial depolarization. Rh123 fluorescence corresponds to the proton gradient across the inner mitochondrial membrane and thus to ATP production. A decrease in fluorescence indicates a hyperpolarization and an increase in ATP production and vice versa.

Patch-clamp recordings
Hamster cDNA encoding for the SUR1 E1507K protein together with WT human cDNA for Kir6.2 was expressed in a stably transfected HEK-293 cell line (60). Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% FCS, glutamine, 100 U/ml of penicillin, and 100 mg/ml of streptomycin. Expression was induced by addition of doxycycline (300 µM) and cells were used for characterization of channel activity from 24 to 72 h. Patch-clamp recordings were done in the inside-out configuration. K ATP currents were measured at a membrane potential of -50 mV (pipette voltage, +50 mV); inward currents are shown as downward deflections. Patch pipettes had a resistance of 6-8 MW. Currents were recorded with an EPC-9 patch-clamp amplifier using Patchmaster software (HEKA, Lambrecht, Germany). Analyses to estimate mean current were done offline in IgorPro 7 (Wavemetrics, Inc., Lake Oswego, OR). With diazoxide or NN414 so many channels open simultaneously that a single channel evaluation of open probability (Po) was not possible. We therefore evaluated the mean current for 20 s before the end of a maneuver.

Statistics
Each series of experiments with islet cell clusters from mice was performed with at least three independent mouse preparations. The number of preparations for recordings with human islet cell clusters varied and is indicated for every series. Box plots were generated using Graphpad Prism 8. Boxes correspond to the interquartile range, the line within the box to the median, and the cross to the mean. Whiskers correspond to the maximum and minimum values. Statistical significance of differences was assessed by Student's t test. Multiple comparisons were made by ANOVA followed by Student-Newman-Keuls test. P values ≤ 0.05 were considered significant.

Effects of Nifedipine on the Cytosolic Ca 2+ Concentration in Human Islet Cell Clusters
Oscillations of the cytosolic Ca 2+ concentration ([Ca 2+ ] c ) are driven by fluctuations of the membrane potential and [Ca 2+ ] c is the trigger signal for insulin secretion. Consequently, [Ca 2+ ] c is a very robust surrogate parameter for insulin secretion. It can be determined easily and online with few cell material, which is an enormous advantage when working with human tissue. Glucosestimulated insulin secretion in human pancreatic b-cells is completely suppressed by pharmacologic blockage of L-type Ca 2+ channels (61). This mechanism should also be functional in CHI islet cells. Thus, as control we tested the effect of the Ltype Ca 2+ channel blocker nifedipine on [Ca 2+ ] c in human islet cell clusters isolated from tissue of a patient with diffuse CHI ( Table 1: 7). Figure 1A

K ATP Channel Openers
Effects of Diazoxide on [Ca 2+ ] c of Human Islet Cell Clusters K ATP channels (SUR1/Kir6.2) of pancreatic b-cells play a crucial role as they couple cellular metabolism to electrical activity. In electrically inactive b-cells, K ATP channels are open. CHI is characterized by permanently active b-cells and thus opening of these channels is one strategy to treat it. Diazoxide is an opener of K ATP channels that is already established in CHI therapy. We tested the effect of diazoxide on [Ca 2+ ] c as a control. In human islet cell clusters isolated from two patients, one with focal and one with mosaic form of CHI ( Table 1: 1 and 4), 250 µM diazoxide clearly decreased the mean fluorescence ratio ( Figures 1C, D). These results show that the channels of these patients are in principle functional and can be influenced by the K ATP channel opener, although mutations in the ABCC8 gene were reported to be the cause of CHI. Obviously, the complex regulation of the channels is disturbed, e.g., the sensitivity to MgATP (1).

Effects of NN414 on [Ca 2+ ] c of Human and Mouse Islet Cell Clusters and Mitochondrial Membrane Potential of Mouse Islet Cell Clusters
The diazoxide analogue NN414 is suggested to be a selective agonist of pancreatic b-cell K ATP channels, and is 100-fold more potent than diazoxide (62). Therefore, it has been proposed as useful drug for the treatment of diseases with excessive insulin secretion (62). NN414, in a concentration of 5 µM, completely abolished oscillations of [Ca 2+ ] c and reduced [Ca 2+ ] c to basal levels ( Figure 2A). Application of 5 µM NN414 to human CHI islet cell clusters taken from three different forms (focal, diffuse, and atypical mosaic, Table 1: 1, 2, and 4) significantly lowered the mean fluorescence ratio ( Figure 2B).
To test whether NN414 specifically interferes with K ATP channels, we studied the effects of this compound on [Ca 2+ ] c with islet cell clusters of WT mice and mice lacking functional K ATP channels (SUR1 -/mice) (63). As expected, 5 µM NN414 abolished [Ca 2+ ] c oscillations in islet cell clusters of WT mice ( Figure 2C, black trace) and reduced the mean fluorescence ratio ( Figure 2D, left part). By contrast, NN414 hardly affected [Ca 2+ ] c oscillations and did not decrease the mean fluorescence ratio in islet cell clusters obtained from SUR1 -/mice (Figures 2C, gray trace, 2D, right part).
Some K ATP channel openers affect mitochondrial function in addition to their direct influence on K ATP channels (64). To address this point, comparative experiments with islet cell clusters of the two mouse genotypes were performed evaluating a possible effect of NN414 on the mitochondrial membrane potential. Figure 2E shows typical recordings of DY for a WT and a SUR1 -/islet cell cluster. The switch from 0.5 to 10 mM glucose is accompanied by a strong decrease in Rh123 fluorescence reflecting hyperpolarization of DY and thus ATP production upon the increase of glucose concentration (65,66). This maneuver was performed in each cell cluster to test for glucose responsiveness. At the end of each experiment, the uncoupler FCCP was applied to evaluate maximal mitochondrial depolarization. Neither in WT nor in islet cell clusters from SUR1 -/mice NN414 (5 µM) exerted any effect on DY ( Figure 2F).

Diazoxide and NN414 Open K ATP Channels Carrying a CHI Mutation
Mutations in the K ATP channel subunits are the most common cause of CHI. However, they do not necessarily lead to diazoxide unresponsiveness. Response to diazoxide is even observed in patients in whom non-response would be predicted (67). Moreover, focal CHI is clinically heterogeneous and responsiveness or resistance to diazoxide was reported for patients with the same mutation in K ATP channels (68). It is unclear whether these clinical effects are due to interference of diazoxide with the mutated channels or off-target effects. As an example illustrating the efficacy of K ATP channel openers on mutant K ATP channels, we used SUR1 E1507K /WT Kir 6.2 channels since the SUR E1507K mutation leads to CHI (69), but patients with this Glu to Lys mutation respond well to diazoxide (45). SUR1 E1507K /WT Kir 6.2 channels were expressed in HEK-293 cells, and diazoxide was tested in comparison to NN414. Figure 3A shows that when inside/out patches from cells expressing SUR1 E1507K /WT Kir 6.2 channels were pulled into nucleotide-free medium, numerous channels were activated as nucleotides inhibiting channel activity by antagonism on their WT pores were washed away (see start of the experiment before ATP application). Addition of 1 mM ATP rapidly inhibited

VU0071063 Silences Islet Cell Clusters in a K ATP Channel-Dependent and -Independent Manner
Recently, Raphemot et al. discovered a novel xanthine derivative, VU0071063 that directly and selectively activates K ATP channels (70). They found that VU0071063 is more potent and activates K ATP channels with a faster kinetic than diazoxide. These findings encouraged us to test its effect on changes in [Ca 2+ ] c on human islet cell clusters from CHI patients. Administration of VU0071063 (30 µM) to islet cells clusters from pancreatic tissue with a focal or diffuse lesion (  Figure 4A shows a typical example. The mean fluorescence ratio clearly changed ( Figure 4B). Due to the limited pathological material, which explains the low number of experiments, we did not perform a statistical test with this data.
In the human islet cell clusters, a drop of [Ca 2+ ] c was noticed directly after withdrawal of VU0071063 ( Figure 4A, asterisk). Presumably, this drop is due to ATP-dependent sequestration of Ca 2+ into the ER (71). This suggests that VU0071063 affects additional targets besides K ATP channels. To evaluate this assumption, [Ca 2+ ] c of islet cell clusters from WT mice and SUR1 -/mice was measured. VU0071063 (30 µM) suppressed Ca 2+ oscillations and lowered [Ca 2+ ] c in both genotypes ( Figures  4C-F). Note that the effect was weaker in the cells of the knockout mice. Like in human islet cell clusters, the drug further reduced [Ca 2+ ] c after its removal in both WT and SUR1 -/mouse islet cell clusters ( Figure 4C, asterisks). This points to alterations in mitochondrial metabolism, which can cause changes in K ATP channel activity independent of any direct interaction with the channel proteins (72). As the mitochondrial membrane potential is for the most part directly linked to ATP production (65), we

K ATP Channel-Independent Drugs
In the following part we present drugs and potential strategies, which could be effective in CHI patients non-responsive to K ATP channel openers.

K Ca 3.1 Channel Openers as a Potential Approach
In addition to K ATP and voltage-gated K + channels, pancreatic b-cells express K + channels regulated by the cytosolic Ca 2+ concentration (K Ca ) (72). Depending on their single channel conductance, there are three groups whose existence has been detected in pancreatic b-cells (73)(74)(75)(76). It has been demonstrated that the K Ca channels of intermediate conductance (K Ca 3.1, SK4) play an important role in the K + current (K slow ) that contributes to b-cell hyperpolarization at the end of a burst phase with electrical activity (66,74,77,78). Previous results from Düfer et al. (74) demonstrated that activation of K Ca 3.1 channels hyperpolarized the membrane potential of pancreatic b-cells from WT mice. Since about 50% of K slow is K ATP current (79), the sulfonylurea-insensitive K Ca component could be even more significant in b-cells lacking functional K ATP channels, which resembles the situation in CHI channelopathies.
To verify this assumption, we evaluated the effect of the K Ca 3.1 opener DCEBIO on islet cell clusters isolated from SUR1 -/mice. DCEBIO (100 µM) effectively abolished the glucose-induced oscillations of [Ca 2+ ] c ( Figure 5A) and reduced the mean fluorescence ratio ( Figure 5B). Next, we tested the effect of DCEBIO on human islet cell clusters from tissue of pancreatectomies. DCEBIO was tested on human islet cell clusters isolated from pancreatic tissues with mosaic and diffuse forms of CHI (  Figure 5C) and significantly decreased the mean fluorescence ratio ( Figure 5D).

Effect of Dextromethorphan on [Ca 2+ ] c of Human Islet Cell Clusters
Dextromethorphan (DXM) is a known antagonist of NMDA receptors. Active NMDA receptors can activate other ion channels, like Ca 2+ -activated K + channels or K ATP channels and thus potentiate K + outflow (80). A block of NMDA receptors leads to prolonged depolarization and increases insulin secretion (81). Lesser-known is its ability to directly inhibit L-type Ca 2+ channels. Carpenter et al. found that DXM moderately inhibits L-type Ca 2+ channels, thereby lowering [Ca 2+ ] c . This effect was observed with permanently depolarized cells under stimulating glucose concentrations (82). Since permanent depolarization is a characteristic of CHI b-cells, DXM may offer a possibility to rescue, i.e., silence the overstimulated cells.
The measurement in Figure 6A shows a recording of [Ca 2+ ] c of a permanently depolarized islet cell cluster from a SUR1 -/mouse in the presence of 10 mM glucose and application of 100 µM DXM. The drug significantly lowered the mean fluorescence ratio ( Figure 6B). The mean fluorescence, application of DXM rapidly reduced [Ca 2+ ] c in an islet cell cluster from a patient with diffuse CHI ( Figure 6C). The mean fluorescence ratio measured in islet cell clusters obtained from two patients with focal and two patients with diffuse CHI ( Table 1: 3, 5, 6 and 7) was lowered ( Figure 6D).

Statins as a Potential Strategy to Silence Human Islet Cell Clusters
Lipid-lowering statins are inhibitors of the enzyme HMG-CoAreductase, which plays a significant role in cholesterol synthesis by converting HMG-CoA to mevalonate. For these drugs it has been reported that they increase the risk of type 2 diabetes (83). Different studies have been conducted in order to enlighten the mechanism how the statins impair insulin secretion. In the study using b-cells isolated from rats, Yada et al. (84) showed that simvastatin in a concentration of 3 µg/ml acutely blocked L-type Ca 2+ channels, thus lowering insulin secretion. Furthermore, Yaluri et al. demonstrated that simvastatin diminished glucosestimulated insulin secretion and [Ca 2+ ] c in MIN6 b-cells via multiple mechanisms (85).
Hence, we considered simvastatin as a potential therapeutic strategy to treat CHI. In order to confirm that simvastatin shows its effect when functional K ATP channels are lacking, we measured [Ca 2+ ] c in islet cell clusters from SUR1 -/mice ( Figure 7A). Simvastatin in a concentration of 7.2 µM [according to the concentration of 3 µg/ml that was used in the study of Yada et al. (84)] rapidly decreased the glucosestimulated Ca 2+ oscillations and diminished the mean fluorescence ratio ( Figure 7B). Further, we tested simvastatin on human islet cell clusters. Figure 7C shows a representative measurement of [Ca 2+ ] c in an islet cell cluster isolated from pancreatic tissue of a patient with focal CHI. The mean fluorescence ratio markedly declined ( Figure 7D). The biopsy material was obtained from two patients with focal and two patients with diffuse CHI ( Table 1: 5 -8).

Possible KATP Channel-Dependent Strategies to Treat CHI
In the present study islet cell clusters isolated from biopsies of CHI patients were used to search for new strategies to treat the disease. Pancreatic islet cell clusters isolated from either focal, diffuse or atypical pancreatic tissue were used as material. We are aware that these are distinct diseases. Due to the limited material we did not separate our results according to the different CHI forms. Noteworthy, the aim of the study was not to suggest novel drugs for CHI treatment but to optimize existing approaches, to search for novel targets and concepts for future drug development. Islet cell clusters were sensitive to the L-type Ca 2+ channel blocker nifedipine. This maneuver resulted in a decrease of [Ca 2+ ] c as expected from numerous observations with murine b-cells and insulin-secreting tumor cell lines as well as the restricted number of studies with human b-cells. This shows that the biopsy material is suitable to receive reliable and reproducible results. This is also confirmed by the results observed with diazoxide used as gold standard in CHI treatment. Noteworthy, nifedipine has been used for the treatment of diazoxide-unresponsive CHI (19,29), but due to reported hypotension in patients with mutations in the ABCC8 gene, it is not commonly recommended for the treatment of CHI (29,86). Diazoxide is usually effective in all forms of CHI including severe cases caused by mutations in the genes encoding K ATP channels (ABCC8 and KCNJ11), e.g., in (2,67,87,88). Our sample cohort was derived from patients harboring ABCC8 missense mutations (patients 1, 6, and 7), a mutation affecting splicing (patient 3) and two nonsense mutations (patients 4,5). The missense and splicing mutations studied possibly allow production of SUR1 proteins albeit at reduced function or level. For nonsense mutation c.3970G>T p.(Glu1324*) detected in exon 32 of patient 4, nonsense-mediated RNA decay (NMD) has been predicted in silico, however, clinically the patient was reported to be diazoxide-responsive at dosage 8mg/kg/d. This may suggest escape of NMD with this particular mutation and generation of a truncated SUR1 protein lacking the last encoded six exons but retaining residual channel function. For one patient (patient 2) no mutation was found in the K ATP channel genes (ABCC8 and KCNJ11) or in other CHI genes tested. This is not unusual in clinical routine. As this patient displayed a partial response to diazoxide (see Table 1) it was included in the study despite the unknown genetic background. Serious adverse drug effects are a consequence of the non-selectivity of diazoxide for pancreatic K ATP (SUR1/Kir6.2) channels (89). Besides pancreatic

NN414
As one strategy to improve CHI therapy we tested K ATP channel openers more specific for b-cells. In comparison to diazoxide, NN414 is reported to be a selective SUR1 agonist, 100-fold more potent than diazoxide, suggesting that the drug is effective at much lower concentrations (62). Early, prediabetic stages of type 2 diabetes mellitus (T2DM) are normally characterized by compensatory hypersecretion of insulin. K ATP channel openers have been suggested as beneficial medication to counteract excessive hormone release in prediabetic patients as insulin hypersecretion may cause or contribute to the development of glucose intolerance and b-cell degeneration in T2DM (93). NN414 has been used in numerous in vitro and in vivo studies to achieve b-cell rest, thereby preserving b-cell function and preventing apoptosis (94,95). An animal in vivo study revealed a significant potential of NN414 in the treatment of disorders resulting from excessive insulin release (96). Alemzadeh et al. showed in a 6-week study that NN414 reduced hyperinsulinemia and improved glucose responsiveness in Zucker obese rats in a dose-dependent manner. NN414 entered human clinical trials for the treatment of T2DM. In healthy subjects, it inhibited insulin release, was well tolerated, and did not induce clinically relevant changes in safety parameters besides side effects on the gastrointestinal tract (97). NN414 was advanced in phase 2 of clinical trials where it showed a tendency to improve b-cell secretory function in diabetic patients (98,99). The clinical trial was stopped because of elevated liver enzymes in treated patients (99,100). The SUR1 selectivity, the low doses, and the reproducible Ca 2+ -lowering effect observed in our study in islet cell clusters from biopsy material (Figures 2A, B) suggest to consider NN414 as a potential alternative to diazoxide for the treatment of CHI with at least partially functioning K ATP channels. Of note, this paper is not intended to characterize different types of CHI with respect to their diazoxide responsiveness or to recommend general treatment of all CHI types with NN414. Liver enzymes have to be monitored during therapy with NN414, but moderate elevation of their plasma concentration is no criterion to exclude the drug, although it would be desirable to develop NN414 analogues without this side effect. Notably, increased concentrations of circulating liver enzymes is one of the most reported side effects for octreotide that is used off-label as second-line therapeutic in the long-term management of CHI and for sirolimus that is proposed for patients resistant to diazoxide and octreotide (28,101,102), reviewed in (3).

VU0071063
Recently, a novel K ATP channel activator, VU0071063 was discovered (70). VU0071063 is reported to be more selective for SUR1/Kir6.2 channels than for SUR2A/Kir6.2 and SUR2A/ Kir6.1 channels. It has been demonstrated that it opens SUR1/ Kir6.2 channels with a higher potency than diazoxide (70).  VU0071063 was shown to activate K ATP channels expressed in HEK-293 cells and to reduce glucose-stimulated Ca 2+ influx in murine b-cells (70). Our data demonstrate at a first glance a beneficial characteristic of VU0071063 in human islet cell clusters isolated from CHI patients ( Figures 4A, B), supporting the idea of a direct activation of K ATP channels in pancreatic islets. By contrast, the observation that removal of VU0071063 from the solution initiated a transient drop in [Ca 2+ ] c suggests that the drug does not selectively interfere with K ATP channels, but also with ATP production (71). This assumption is supported by the following observations: 1) VU0071063 strongly and reversibly depolarized DY in both, WT and SUR1 -/islet cell clusters, which points to inhibition of ATP formation. . These differences clearly show that the effects of NN414 are, in contrast to those of VU0071063, caused by a specific interference with K ATP channels. In summary, our data with K ATP channel agonists demonstrate that these drugs can be effective in different forms of CHI caused by mutations in K ATP channels. As shown by patch-clamp experiments (Figure 3), diazoxide and NN414 act as direct channel openers in mutated K ATP channels with a dominant mutation comparable to WT channels. With respect to specificity, dosage, and expected side effects, NN414 seems superior to diazoxide. VU0071063 is unsuitable because of its multiple and yet not completely understood mode of action and therefore potential adverse side effects.
Possible K ATP Channel-Independent Strategies to Treat CHI Current Second-And Third-Line Therapy Regimen Targeting K ATP Channel Independent Pathways There are mutations in ABCC8 or KCNJ11 genes known to disrupt the expression of K ATP channels at the cell surface (4,103,104). In this case openers, e.g., diazoxide, are ineffective in the treatment of CHI (51,105). For these patients it is indispensable to find drugs targeting mechanisms, which can  induce b-cell rest and inhibit insulin release independent of K ATP channels. Currently available alternatives to diazoxide therapy are somatostatin analogues (octreotide, octreotide-LAR, and lanreotide), sirolimus and exendin-(9-39) (3). Octreotide, a short-acting synthetic somatostatin analogue, inhibits insulin secretion by binding to and activating somatostatin receptors 2 and 5 (SSTR2 and SSTR5) (106). Activation of SSTRs shows multifactorial modulation of bcells, which involves inhibition of the adenylate cyclase/cAMP pathway, activation of G protein-activated inwardly rectifying K + (GIRK) channels, decrease in Ca 2+ influx via P/Q-type Ca 2+ channels and inhibition of exocytosis (24,107,108). Infants respond well to initial doses of octreotide, but tachyphylaxis after a few doses makes it not suitable for the long-term treatment. Long-acting somatostatin analogues (octreotide-LAR and lanreotide) have similar effects as octreotide but have the advantage that they are given once every 4 weeks, which improves therapy compliance and quality of life (21,109,110). However, due to a similar mechanism of action as octreotide, long-acting somatostatin analogues show similar side effects (111). Sirolimus, a mammalian target of rapamycin (mTOR) inhibitor, is an immunosuppressive and anti-proliferative agent that has been used in patients with diffuse CHI, unresponsive to diazoxide and octreotide therapy (22,112). It suppresses insulin release by different mechanisms, which have not been fully elucidated (30). It has been proposed that downregulation of mTOR pathway decreases insulin production in pancreatic bcells and b-cell growth and may restore ketogenesis (112,113). Furthermore, upregulation of liver gluconeogenesis by sirolimus contributes to insulin resistance (114). However, severe and lifethreatening side effects reported for the above-mentioned drugs restrict their use.

Opening of K Ca 3.1 Channels
Beside K ATP channels, Ca 2+ -activated K + channels of intermediate conductance (K Ca 3.1, SK4) contribute to K slow , the hyperpolarizing current that terminates bursts of action potentials in b-cells (74,78). Accordingly, K Ca 3.1 channels may become predominant regulators of membrane potential and insulin secretion when functional K ATP channels are absent, which makes these channels ideal as drug targets in CHI. Our data show that an opener of K Ca 3.1 channels, DCEBIO, was able to strongly reduce [Ca 2+ ] c in SUR1 -/islet cell clusters. Furthermore, DCEBIO was highly effective in silencing human islet cell clusters obtained from pancreatic tissue of CHI patients ( Figure 5). These experiments provide valuable support for the idea of targeting K Ca 3.1 channels in the treatment of CHI. To follow this strategy would of course require the search for new K Ca 3.1 channel openers with high selectivity for b-cells since unspecific K Ca 3.1 channel openers are expected to exert severe side effects in numerous organs (e.g., lung, cells of the hematopoietic system, and salivary glands) (115)(116)(117).

DXM as L-type Ca 2+ Channel Antagonist
The block of NMDA receptors by DXM is expected to increase insulin secretion (81). However, DXM has a higher affinity for Ltype Ca 2+ channels than to its known target, the NMDA receptor (118). As mentioned before, DXM moderately inhibits L-type Ca 2+ channels, thereby lowering [Ca 2+ ] c in permanently depolarized b-cells under stimulating glucose concentrations (82,119). This is exactly what we observe in our experiments with depolarized islet cell clusters isolated from pancreatic tissue of CHI patients and with depolarized islet cell clusters from SUR1 -/mice ( Figure 6). Thus, DXM could be an alternative strategy for the treatment of CHI especially in the diffuse form of CHI. In patients with focal lesions, where only a distinct portion of b-cells, i.e., the focal ones, is dysregulated, the drug could cause an undesired increase in [Ca 2+ ] c and insulin secretion in healthy islets. The benefits of this drug are that it is already available as a pharmaceutical and that inhibition of L-type Ca 2+ channels by e.g., nifedipine is already a proved treatment of CHI (19,29). Considering side effects of nifedipine, like hypotension, the moderate effect of DXM on L-type Ca 2+ channels could be of advantage (82).

Simvastatin
It was shown that simvastatin lowers insulin secretion by blocking L-type Ca 2+ channels (84,85). The effect of simvastatin on [Ca 2+ ] c was comparable with that of nifedipine in the insulin-secreting cell line MIN-6 (85). This interaction might contribute to the increased risk to develop diabetes mellitus under a cholesterol-lowering therapy with statins (83). With respect to CHI patients, this side effect could constitute a suitable therapeutic approach. Our results obtained from experiments with islet cell clusters isolated from SUR1 -/mice and human islet cell clusters isolated from patients with CHI indeed point toward a possible beneficial effect of simvastatin in the treatment of CHI ( Figure 7). Noteworthy, statins are widely used and well tolerated in the long-term therapy. In contrast to nifedipine, which affects the cardiovascular system, statins are safe with respect to blood pressure or heart rate (120). Additionally, it is proposed that statins induce hepatic gluconeogenesis in human liver cells by activation of the pregnane X receptor (PXR) (121,122), which could also counteract hypoglycemic conditions in CHI patients.

CONCLUSIONS
There is a clear need to develop novel approaches to prevent hypoglycemia in CHI patients and to establish better therapies with less side effects for the different forms of CHI. In this study, we had access to biopsy material of CHI patients and give suggestions, which drugs or targets should be studied in future.
Promising results were obtained for NN414, DCEBIO, DXM and simvastatin.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors upon request.

ETHICS STATEMENT
Ethics approval for the study involving human participants was approved by the ethic commission of the Universitätsmedizin Greifswald (BB 050/13). Written informed consent was provided by the legal guardians of the children for the study. The animal study was reviewed and approved by the Ethics Committee of Regierungspraesidium Tuebingen.

AUTHOR CONTRIBUTIONS
JS researched the data and wrote and edited the manuscript. TH, AG, and JK researched the data, contributed to the discussion, and edited the manuscript. WB, CW, and UL provided human bcells and contributed to discussion. IW supported the manuscript as an expert in CHI genetics and contributed to the discussion. MD and PK-D contributed to discussion and study design and edited the manuscript. GD designed the study, wrote and edited the manuscript, and contributed to discussion.
GD is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by grants from the DFG (DR225/11-1 to GD and INST 211/647-1 FUGG to MD).