Introduction

Stroke is one of the leading causes of death and disability in the world1, 2. The disease itself and associated morbidity have caused significant social and economic impacts on society and individuals worldwide. The prevalence of stroke is expected to increase and our aging population is especially vulnerable to stroke insults. The clinical trials of anti-excitotoxic therapies (AET) have failed to benefit stroke patients3, thus diminishing the initial excitement of translating research from bench to bedside and using glutamate receptor blockers in treating stroke patients. Even though the mechanisms underlying cerebral ischemia are beginning to be better understood, there is still no clinical or experimental treatment that has shown improved outcome for stroke patients. To ease personal and societal burden of stroke, continuous efforts have been directed towards searching for new therapeutic targets in stroke. This review provides a current view on one of the non-glutamate mechanisms of stroke that mediates through TRPM7 channels from a recent in vivo study 4.

A major event during cerebral ischemia is a concomitant massive release of the excitatory neurotransmitter glutamate, which results in intracellular calcium overload and eventual cell death5. The excitotoxicity in ischemia has been in the centre of stroke research for a long period of time. Triggered release of excessive glutamate causes cell death following ischemia, which is associated with an increase of the intracellular calcium (Ca2+) concentration6, 7, 8. Thus, identifying the source of the excessive Ca2+ influx and/or release from the intracellular Ca2+ stores during ischemia has been a research focus. Traditionally, Ca2+-permeable NMDA (N-methyl-D-aspartic acid), AMPA (DL-alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) receptors8 and L-type voltage-dependent Ca2+ channels9 were considered as the major calcium entry paths and the causes of Ca2+ overload during ischemia. This Ca2+ overload, or a broad spectrum of ion imbalance, during ischemia is considered to initiate a wide range of sequential events that lead to irreversible damage to protein synthesis, mitochondria, cytoskeleton and plasma membrane, and to eventual cell death. In an ideal scenario, the interruption of this Ca2+ overload in ischemia is thought to be clinically beneficial for stroke patients. Blocking these receptor channels prevents the intracellular Ca2+ overload and provides significant neuroprotection in the laboratory. Some of the findings from the bench have been translated into many clinical trials in stroke treatment. However, the results of clinical trials testing AET, which include NMDA and AMPA receptor blockers, turned out to be ineffective and even with unwanted side effects10, 11, 12, 13, 14. This may be due to multiple factors and it will not be the focus of this review15. Because of the limitation of the glutamate mechanism and the unfavourable outcomes of AET trials, stroke researchers have been seeking for alternative, non-glutamate related therapeutic targets that cause ionic imbalance and cell death. Some of these channels include: acid-sensing ion channels16, 17, transient receptor potential (TRP) channels4, 7, 18, 19, 20, 21, and hemichannels22, 23, 24, volume-regulated anion channels25, sodium-calcium exchangers26, 27 and non-selective cation channels28.

Based on the recommendations from the Stroke Therapy Academic Industry Roundtable (STAIR) committee, it is important to validate the preclinical development in proof of concept starting with in vivo rodent models as experimental animal stroke models29. Recent in vivo studies aimed at identifying the non-glutamate mechanisms for stroke have demonstrated the involvement of acid-sensing ion channels16, 17 first, and then the TRPM7 (transient receptor potential melastatin 7) channel4, 7, 18, 19, 20, 21. In this review, we will mainly focus on the current understanding of the molecular, biophysical, and pharmacological properties of TRPM7 as well as its physiological and pathophysiological roles and its therapeutic potential in stroke.

Classification, structures and distributions

Classification

The TRP superfamily is comprised of a group of non-selective cation channels30, 31, 32, 33. Its nomenclature was originated from the first found member of this superfamily, which was identified in a Drosophila phototransduction mutant showing transient receptor potential to a continuous light34. Currently, about 30 mammalian TRP channels have been discovered and named according to their sequence homologous structures. They are classified into six subfamilies: 1) TRPC (canonical), 2) TRPM (melastatin), 3) TRPV (vanilloid), 4) TRPA (ankyrin), 5) TRPML (mucolipin) and 6) TRPP (polycystin). Different TRP channels are activated by different physical and chemical stimuli. The diverse gating mechanisms of TRP channels make them good cellular signal integrators critical for physiological and pathological functions30, 31, 32, 33.

TRPM7 belongs to the melastatin-related subfamily of TRP channels, which is comprised of eight members (eg, TRPM1-8). It was suggested that TRPM7 may also form heteromers with TRPM2 as application of TRPM7 siRNA also down-regulated TRPM2 channel mRNA in an in vitro study21. This is important as TRPM2 has also shown to play a role in oxidative stress-mediated cell death, which is a cellular condition shown in stroke.

Gene and protein structures

In human, TRPM7 gene is located on chromosome 15 in the q21.2 region, and encoded by 39 exons that spans about 127 kb of DNA sequence. The mouse TRPM7 gene is 95% identical to human gene35. It is located on chromosome 2 on cytoband F2 and it is also encoded by 39 exons that spans about 85 kb of DNA sequence.

TRPM7 is a large protein (1864 amino acids in human; 1863 amino acids in mouse) with a predicted molecular weight of approximately 212 kDa. Each subunit has six transmembrane (TM) spanning domains (S1–S6) with a re-entrant pore-forming loop (known as P-loop) between the fifth (S5) and sixth (S6) segments32, 33 (Figure 1). The N-terminus has another hydrophobic region (H1) and four regions of TRPM subfamily homology domain (MHD), but their biological significance is largely undefined. The C-terminus contains a TRP box of ∼25 highly conserved residues, which may interact with phosphatidylinositol 4,5-bisphosphate (PIP2), a positive regulator of some TRP channel36. A coiled-coil domain close to the C-terminus may mediate subunit-subunit interactions and tetrameric assembly of TRPM737. The most unique structural feature of the channel is the enzymatic domain located at the end of C-terminus. In TRPM7, the distal C-terminus has an atypical serine/threonine protein kinase domain that is homologous to a family of α-kinases38. Although this kinase domain does not seem to affect channel activity directly33, 39, 40, it may be important for the regulation of channel function by Mg2+ nucleotides41.

Figure 1
figure 1

Schematic diagram showing proposed transmembrane topology of TRPM7. (A) The putative membrane topology of a single subunit of TRPM7 is shown. Each subunit has six transmembrane (TM) spanning domains (S1–S6) with a re-entrant pore-forming loop between the fifth (S5) and sixth (S6) segments. The intracellularly located N-terminus has another hydrophobic region (H1) and four regions of TRPM subfamily homology domain (MHD). The intracellularly located C-terminus contains a TRP box of ∼25 highly conserved residues (TRP) and a coiled-coil domain (CCD). The distal C-terminus has an atypical serine/threonine protein kinase domain. As indicated in the figure, TRPM7 is a non-selective cation channel that conducts both monovalent ions (eg, Na+ and K+) and divalent ions (eg, Ca2+, Mg2+ and other trace metal ions). (B) Representative current-voltage (I-V) relationship of TRPM7.

Full size image

Tissue and cellular distribution

TRPM7 channel mRNA is ubiquitously expressed in almost all tissues33, 42, 43. Recently, real-time quantitative RT-PCR analyses with either Taqman or SYBR Green were used to create comparative distribution profiles of TRPM channels in selected human tissues, including brain, pituitary, heart, lung, liver, fetal liver, skeletal muscle, stomach, intestine, spleen, peripheral blood mononuclear cells, macrophages, pancreas, prostate, placenta, cartilage, bone and bone marrow42. TRPM7 mRNA has the highest expression in heart, pituitary, bone, and adipose tissue42. Similar distribution patterns of TRPM7 were also observed in mouse tissue samples44. Compared to other TRP members, TRPM7 mRNA expression levels were significantly higher in most tissues.

The TRPM7 protein shown by immunofluorescent labeling is strongly expressed at the plasma membrane in N1E-115 neuroblastoma cells45, and in vascular smooth muscle cells46. In N1E-115 neuroblastoma cells, HA-tagged TRPM7 antibodies were localized in membrane ruffles45. Similarly, protein expression of TRPM7 is also shown in cell bodies and processes of hippocampal neurons with immunostaining4, 47, 48. In superior cervical ganglion neurons, TRPM7 is exclusively localized within cholinergic vesicles49.

Biophysical properties, regulatory mechanisms, pharmacology

Biophysical properties

TRPM7 is a non-selective cation channel that displays several biophysical features that make this channel distinguishable from other TRP members. TRPM7 channel has a reversal potential of approximately 0 mV, and a prominent outward rectification30, 33, 43, 50 (Figure 1B). At negative membrane potentials, TRPM7 conducts a small inward current by transporting divalent cations (eg calcium and magnesium) down their concentration gradients35. TRPM7 current density is usually under 20 pA/pF35, 50. At positive membrane potential, TRPM7 conducts a strong outward current as intracellular cations experience strong driving force to exit the cell. This outwardly rectifying property is entirely due to a voltage-dependent block of monovalent cation influx by extracellular divalents. For instance, in the absence of divalent cations, TRPM7 conducts inward monovalent cations35. Consequently, its I-V relation becomes quasi-linear suggesting the lack of voltage-dependent gating and channel selectivity (Figure 1B).

Unlike many other Ca2+ permeating channels, TRPM7 is characteristically more permeable to a series of trace metal ions. Using equimolar divalent ion substitution approaches, Monteilh-Zoller and colleagues reported a permeation profile for TRPM7 in a sequence of: Zn2+≈Ni2+>>Ba2+>Co2+>Mg2+≥ Mn2+≥Sr2+≥Cd2+≥Ca2+ 51. TRPM7 allows entry of these divalent ions even with physiological levels of extracellular Ca2+ and Mg2+. TRPM7 is constitutively active and this feature makes TRPM7 a good candidate for both sensing the extracellular concentration of divalents and maintaining intracellular Mg2+ homeostasis during ischemic episodes that lead to intense neuronal activity47.

TRPM7 channel activity is regulated by extracellular pH. A decrease in extracellular pH (acidic) strongly potentiated current activity of the recombinant TRPM7 channel expressed in HEK-293 cells (∼10-fold increase at pH 4.0, and 1–2 fold increase at pH 6.0)52 and in CHOK1 cells (∼12-fold increase at pH 4.0)53. However, the TRPM7-like current in the FaDu cell line was insensitive to the acidic condition (pH 5.0)54, while the TRPM7-like inward current in human cervical epithelial HeLa cells was increased at pH 4.055. Protons likely compete with Ca2+ and Mg2+ for their binding sites, thus increase the inward current by releasing the divalent cation block52. Point mutation of Glu1047 (E1047Q) of TRPM7 channels eliminated the proton-enhanced inward current activity, indicating the residue may be involved in the pH sensitivity of the channels53. Although the effects of protons on endogenous TRPM7 remains controversial, TRPM7 can be regulated in acidic pathophysiological conditions, including ischemic stroke56.

TRPM7 is not a mechanosensitive channel, however, shear stress in vascular smooth muscle cells doubled the number of TRPM7 channels near the plasma membrane57. Further studies are required to deduce the mechanism of stress-induced regulation of TRPM7 channels.

Regulatory mechanisms

The heterologously expressed foreign and native TRPM7 channels are constitutively active, and their activities can be modulated by several extracellular and intracellular factors43, 50. TRPM7 activities can be tonically inhibited by intracellular Mg2+, and Mg-complexed nucleotides, MgATP, MgGTP, and such tonic inhibition is usually less than 10% of maximal conductance35, 41, 50. Whole-cell patch-clamp recordings have shown that either adding Mg2+ chelators (eg HEDTA or Na-ATP) intracellularly or omitting Mg2+ and Mg2+-complexed nucleotides in intracellular solutions increased activation of TRPM740, 41, 43. These inhibitory effects may be mediated by binding to C-terminal kinase domain39, 43, which in itself is not essential for the activation of TRPM733, 39, 40. Compared to the wild-type, phosphotransferase-deficient mutant channels (K1648R and G1799D) demonstrated a reduced sensitivity to inhibition by Mg2+ at intermediate concentrations close to the IC5039. Moreover, Demeuse and colleagues41 reported differential sensitivity to Mg2+ and Mg2+-nucleotides inhibition in the wild-type, phosphotransferase deficient point mutant (K1648R), and the Δ-kinase truncation mutant. These findings lead to a hypothetical model: only Mg2+-nucleotides bind to the kinase domain but this domain interacts with the Mg2+ binding site, which is responsible for regulating the channel activity.

Activation of TRPM7 can be regulated via PIP2, which is a substrate of phospholipase C (PLC)58. The C2 domain of PLC is directly associated with the kinase domain of TRPM7. When carbachol, an agonist for Gαq-linked muscarinic type 1 (M1) receptors, was used, PLC-beta was activated. This activation of PLC-beta led to the hydrolysis of localized PIP2, which caused a rapid decreased ITRPM7.

TRPM7 may also be regulated by phosphorylation. A variant of TRPM7 with a missense mutation (T1482I) is found in a subset of patients with Guamanian amyotrophic lateral sclerosis (ALS-G) and Parkinsonism-dementia (PD-G)59. When recombinant TRPM7s with T1482I mutation were heterologously expressed in HEK-293 cells, these channels were functional but showed increased sensitivity to Mg2+ inhibition and reduced phosphorylation compared to wild-type59. Based on the computer analysis of the secondary structure, both Thr-1482 in fish, amphibian, avian, and primate species, and Ser-1482 in murine species are the potential substrates for autophosphorylation by the C-terminus serine/threonine α-kinase domain in TRPM7. Ile-1482 mutation found in these patients, however, cannot be phosphorylated. TRPM6, the closest member to TRPM7, also regulates TRPM7 via cross-phosphorylation, and alters Mg2+ homeostasis regulation60.

Pharmacological properties

There are currently no selective pharmacological tools (both agonists and antagonists) that can specifically modulate TRPM7 channels61. Gene deletion in global TRPM7 knockout animal is confirmed to be embryonically lethal62, 63. Thus, inability to modulate the TRPM7 channels pharmacologically creates a huge obstacle for investigating the physiological and pathophysiological roles of TRPM7 in stroke. There are some successful studies of gene silencing using small interfereing RNA (siRNA) to knockdown the TRPM7 in either mRNA and/or protein expression in central neurons4, 21, 47, peripheral neurons49, vascular endothelial cells64, vascular smooth muscle cells65, gastrointestinal tract interstitial pacemaker cells66 and human epithelial cells67. TRPM7 can be blocked non-specifically by trivalent ions, such as Gd3+ ((IC50 ∼1.4–2.5 μmol/L) and La3+ (IC50 ∼17 μmol/L)21 and 2-Aminoethoxydiphenyl borate (2-APB) (IC50 ∼50 μmol/L), which is a well known non-specific blocker of many TRP channels68. Recently, it has been reported that inhibitors of 5-lipoxygenase (5-LOX), NDGA (nordihydroguaiaretic acid, IC50 ∼6.3 μmol/L), AA861(IC50 ∼6.0 μmol/L), and MK886 (IC50 ∼8.6 μmol/L), can suppress the TRPM7 current in HEK-293 cells69. Application of these molecules also prevented some of the phenomena (eg cell death) associated with TRPM7 when exposed to low extracellular divalent cations and other apoptotic stimuli. These effects seem to be independent of their actions on 5-LOX, and the expression level and cellular concentrations of TRPM7 at the plasma membrane were not affected. Previously, studies have reported less tissue damage during cerebral ischemia and myocardial ischemia-reperfusion injury with 5-LOX inhibition70, 71. Although drawing a connection between cellular protective effects during ischemic injury with 5-LOX and blockade of TRPM7 with 5-LOX would be premature, these findings emphasize the importance of future follow-up studies. Thus, there is a pressing need of specific pharmacological agents for studying the physiological and pharmacological roles of the TRPM7 channels in vivo and potential therapeutic uses.

Physiological functions

Our current knowledge of the physiological functions of TRPM7 channels has recently been improved, even with limited molecular and specific pharmacological tools. Under physiological conditions, several lines of evidence suggest the role of TRPM7 in cell survival and proliferation39, 62, 72. The early embryonic lethality in global TRPM7 knockout mice hints at the requirement of TRPM7 in cell survival and proliferation as embryonic development involves extensive cell proliferation62. In the same study, TRPM7 gene was selectively deleted in developing thymocytes. These T-cells did not differ in its ability of uptake Mg2+ or maintaining global cellular Mg2+, but showed defective thymopoiesis. A more recent study showed that knocking out of TRPM7 kinase domain homozygously resulted in embryonic lethality63, while heterozygous knockout mice were viable, but exhibited abnormal homeostasis63. TRPM7 knockout in chicken DT40 B cells caused growth arrest and eventual cell death in culture39, which may be linked to a regulation of Mg2+ homeostasis72. Supplementing TRPM7 knockout cells with a high Mg2+ containing medium, but not Ca2+ or Zn2+, could restore normal cell growth and survival in culture. Knockdown of TRPM7 with RNA interference reduced Ca2+ and Mg2+ influxes, and decreased cell proliferation in human osteoblast-like cells72, and retinoblastoma cells73. TRPM7 dependence for proliferation and differentiation was also shown in zebrafish mutants as they displayed severe growth retardation and general alterations in skeleton development74.

Several studies have shown the importance of TRPM7 in cell adhesion. Over-expression of TRPM7 in HEK-293 cells lead to cell rounding, loss of adhesion and cell death35. Consistent with these findings, knockdown of TRPM7 in HEK-293 cells increased cell adhesion75. Over-expression of TRPM7 may produce cell rounding by stimulating the activity of the Ca2+-dependent protease m-calpain. TRPM7 has also been implicated in cell motility76. Knockdown of TRPM7 by RNA interference reduced the number of high Ca2+ micro-domains induced by platelet-deprived growth factor (PDGF) and disrupted the turning of migrating WI-38 fibroblasts.

It has also been suggested that TRPM7 is involved in the neurotransmitter release by mediating Ca2+ influx49. In primary rat superior cervical ganglion neurons, TRPM7 is localized in the synaptic vesicles and interacts with synaptic vesicular snapin, synapsin 1 and synaptotagmin 1. Furthermore, there were some correlations between TRPM7 expression levels and quantal sizes, amplitudes and decay times of the excitatory postsynaptic potential (EPSPs). When TRPM7 specific siRNA was used to suppress endogenous TRPM7 in PC12 cells, acetylcholine-secreted-synaptic-like vesicle fusion was inhibited.

Pathophysiological relevance in cerebral ischemia and stroke

Unregulated monovalent or divalent cation influx is implicated in several different cellular mechanisms (eg, excitotoxicity, apoptosis, and oxidative stress) underlying neural cell death during ischemic periods of stroke61. Since cation channels are the main pathways for cation influx from extracellular space, they are closely involved in neuronal cell death. Conventionally, Ca2+ permeable NMDA and AMPA receptor channels are widely accepted as the main pathways of Ca2+ entrance during ischemia as well as the promising therapeutic targets7, 8, 77. Numerous clinical trials testing AETs in stroke patients, however, yielded disappointing outcomes. The shortcomings of AET led researchers to consider other non-glutamate dependent mechanisms10, 11, 12, 13, 14, such as non-specific cation channels including acid-sensing ion channels16, 17, TRP channels4, 7, 18, 19, 20, 21, hemichannels22, 23, 24, volume-regulated anion channels25, sodium-calcium exchangers26, 27 and non-selective cation channels28.

Previous studies demonstrated pathophysiological involvement of TRPM7 in stroke from in vitro data. When primary cultured cortical neurons were subjected to oxygen-glucose deprivation (OGD) for a prolonged period, there was an increase in ROS production, a Ca2+ influx mediated by TRPM7 and cell deaths21. When the primary mouse cortical neurons were transfected with siRNA vector directed against TRPM7, the TRPM7 mRNA expression was suppressed, ROS-mediated activation was inhibited and subsequent cell death under anoxia was reduced. Such effects were consistently shown with cocktail of blockers for glutamate NMDA and AMPA receptor and L-type calcium channels (MK-801, CNQX, and nimodipine), indicating the independent role of TRPM7 in mediating intracellular Ca2+ elevation and subsequent cell death during the prolonged anoxia. In another study, the contribution of TRPM7 channels in cell membrane depolarization, intracellular Ca2+ accumulation and cell swelling during the initial period of brain ischemia is also observed in native CA1 neurons of brain slices78.

TRPM7 in vivo studies have been scarce for a period of time because both the knockout model and selective pharmacological agents are not available. Eventually, a report demonstrated in vivo changes in TRPM7 channels during focal ischemia. Jiang and colleagues48 studied the interaction of nerve growth factor (NGF) with TRPM7 channels using both in vivo cerebral ischemia-reperfusion and in vitro OGD models. NGF, a neurotrophic factor, showed neuroprotective effects during ischemia. In their in vivo model, middle cerebral artery occlusion (MCAO) was performed on rats for 1 h and it was followed by reperfusion that lasted for 5, 10, 20, and 30 h. Both mRNA and protein levels of TRPM7 were up-regulated compared with pre-ischemia, peaking at 20 h after the reperfusion with about 2–3 fold increase. Given that there are increases in both mRNA and protein levels of TRPM7, up-regulation of the channels may be another mechanism that increases TRPM7-like current. Interestingly, these expression levels of TRPM7 were close to the normal level when 500 ng of NGF was applied 30 min before ischemia. The effects of NGF on TRPM7, however, disappeared when NGF was introduced after K252a, which is an inhibitor for the NGF-activated TrkA pathway. When wortmannin, which is an inhibitor for phosphatidylinositol-3 kinase (PI-3K) signal pathway, was applied, NGF effects were also abolished. These findings indicate that TRPM7 may be involved in neuronal cell damage in vivo during ischemia.

Recently, Sun and colleagues4 demonstrated that suppression of TRPM7 channels in vivo reduced neuronal cell death and preserved functions after global cerebral ischemia. The study used virally mediated gene silencing with shRNA to knockdown TRPM7 channels in hippocampal CA1 pyramidal neurons of adult rat brains. The viral vectors were delivered in vivo using stereotaxic microinjection to CA1 area. First, the authors showed that infecting adult hippocampal CA1 neurons in vivo was feasible by using the adeno-associated viral vectors (AAV serotype-1). Secondly, suppression of TRPM7 channels was convincingly demonstrated by measuring: 1) mRNA level in conjunction with the Laser Capture Microdissection for infected hippocampal CA1 cells; 2) protein level with both Western Blot and immunohistochemistry in conjunction with Laser Confocal microscope; and 3) functional level with electrophysiology. Thirdly, the injected viral vectors and transient suppression of TRPM7 channels in the adult rat brains in vivo showed no ill effects on cell survival, neuronal and dendritic morphology, neuronal excitability, or synaptic plasticity. Finally, they showed that following fifteen minutes of global cerebral ischemia induced by occluding both common carotid and vertebral arteries, TRPM7 suppression reduced hippocampal CA1 neuronal death in vivo and preserved functional outcomes after stroke. The survived neurons preserved their morphological integrity and fine structures, and even maintained their electrophysiological properties (LTP) and hippocampal-dependent behaviours, such as fear-associated and spatial-navigation memory tasks. This is the first in vivo evidence showing the important role of TRPM7 channel in mediating ischemic neuronal cell death in stroke.

Working model of TRPM7 activation during cerebral ischemia

During the initial phase of an ischemic attack, a strong NMDA receptor activation leads to a large influx of Ca2+, and the resulting Ca2+ directly stimulates (i) production of nitric oxide (NO) by neuronal nitric oxide synthase (NOS) and (ii) production of superoxide (O2−) from mitochondria17, 18, 19. When NO and O2− are combined, highly reactive species peroxynitrite (ONOO−) form. Along with other factors, such as decreases in pH and extracellular divalents, that are associated with ischemic episodes, ONOO− enhances TRPM7 activation. This completes the lethal positive feedback loop of free radical production. In this model, the failure of AET could be partly explained: AET could delay the process but insufficient to ultimately prevent lethal TRPM7 activation (Figure 2).

Figure 2
figure 2

Working model of TRPM7 activation during cerebral ischemia. (A) During the early phase of an ischemic attack, an increase in the extracellular glutamate activates NMDA receptors. Ca2+ influx due to activated NMDA receptors stimulates: (i) production of nitric oxide (NO) by nitric oxide synthase (NOS) and (ii) production of superoxide (O2−) from mitochondria. NO and O2− combine to produce highly reactive species peroxynitrite (ONOO−). Along with factors, such as decrease in pH, that are associated with ischemia, these free radicals promote sustained activation of TRPM7, which leads to further Ca2+ build-up in the intracellular space. (B) Consequences of unchecked Ca2+ influx. Increased intracellular Ca2+ concentration may lead to excitotoxicity, oxidative stress, inflammatory processes and eventual cell death.

Full size image

TRPM7 channels may also be involved in the ischemic lethal process by conducting metal ions other than Ca2+. For instance, Zn2+, which is the most permeable trace ion through TRPM7, is highly toxic to cells if its concentration exceeds the physiological level79, 80 and has been implicated in cerebral ischemia. After the brief global ischemic insults, a delayed increase in intracellular Zn2+ is observed before cell death in some selective hippocampal CA1 neurons81. Increases in intracellular Zn2+ and neuronal cell death were prevented with the application of the membrane-impermeable zinc-chelator calcium-EDTA (calcium-ethylenediaminetetraacetic acid) before the ischemia. Recently, it has been shown that Zn2+-induced neurotoxicity may be mediated by TRPM782. Both Zn2+-mediated neurotoxicity and neuronal injury associated with oxygen-glucose deprivation (OGD) were reduced by non-specific blockers (Gd3+ and 2-APB) and knockdown of TRPM7 by siRNA. Overexpression of TRPM7 in HEK-293 cells led to increase in intracellular Zn2+ accumulation and Zn2+-mediated cell deaths.

Clinical potentials and therapeutic perspectives

To date, therapeutic intervention for stroke is very scarce. The only approved treatment of acute ischemic stroke by the US Food and Drug Administration (FDA) is the tissue plasminogen activator (tPA), which relieves vascular occlusion by dissolving clots83. Although tPA is a potent treatment for stroke, the usage and effectiveness of tPA are still limited by its short therapeutic window, and intrinsic toxicity. With disappointing preliminary clinical results from drugs targeting glutamate-induced excitotoxicity, considerable efforts have been put into searching for alternative targets.

Several lines of evidence support our hypothesis that TRPM7 is involved in ischemic stroke77. Even though findings from cellular and animal studies are compelling, TRP channels should be studied in their native cellular environment, as the specific cellular environment and expression levels seem to be important for the normal physiological functions. It would be necessary to validate their diverse physiological and pathophysiological functions using in vivo animal models. For in vivo studies, developing tissue-specific or inducible TRPM7 knockout models will be useful as the conventional TRPM7 knockout mouse is not viable62, 63.

Without the development of specific pharmacological modulators of TRPM7, we do not expect to see any preliminary clinical trials in the near future. At the initial stage of most drug development, potential therapeutic targets are first identified, and experimental high-throughput screening (HTS) is used to narrow down the drug candidates that bind to the targets and changes their activities84. Hence, in order to design specific and potent inhibitors, it is important to understand their molecular or structural properties in detail84, 85. This is especially true for TRPM7, which seems to have conflicting roles in cell death and cellular survival4, 21, 62. It may be the case that TRPM7 function would have to be regulated separately, either enhanced or depressed, in different tissues for the prevention of cerebral ischemia and stroke. For instance, Touyz and colleagues86 have shown that reduced Mg2+ influx in cultured vascular smooth muscle cells (VSMCs) of the spontaneously hypertensive rat (SHR) is associated with down-regulation of TRPM7. Furthermore in the normotensive Wistar-Kyoto rat, TRPM7 expression and activity in VSMCs of the SHR were attenuated by angiotensin II. Since hypertension is a well-known risk factor for cerebral ischemia and stroke, it suggests that TRPM7 channels may enhance cerebral ischemia and stroke by regulation of Mg2+ homeostasis87. However, either supplement or depletion of Mg2+ showed no effect on hypertension, thus questioning the role of TRPM7 in vasculature regulation87. In contrast, reduction of TRPM7 expression prevented cerebral ischemia and stroke4 indicating TRPM7 may play a differential role in vascular smooth muscle cells and neurons. Further investigation is required to evaluate pathophysiological roles of TRPM7 channels in different cell types. Therefore, designing activity-dependent antagonists that preferentially target TRPM7 during stroke is critical in the future study. Understanding its temporal, spatial expressions and interactions with other proteins may aid the development of selective drugs for modulating TRPM7 activity.

When specific TRPM7 channel modulators are developed, these might be added in combination therapy for cerebral ischemic stroke in future. Now it is suspected that, although the recruitment of NMDARs is the key event in the early phase of cell death cascades in cerebral ischemia, there is also progressive recruitment of other non-selective cation channels, such as TRPM7, in the later stages. Based on this hypothesis, combination therapy, which includes drugs that are applied at empirically determined time points for each target, will be more effective in providing neuroprotection and potentially facilitating the recovery of function.

TRPM7 genetic variants may be related to various human diseases. For instance, heterologously expressed T1482I TRPM7 variant was found in a subset of ALS-G (Guamanian amyotrophic lateral sclerosis) and PD-G (Parkinsonism dementia) patients59. The mutation resulted in an increase in sensitivity of the channels to intracellular Mg2+-mediated inhibition, thus the patients were more vulnerable to the diseases59. No test has been performed related to the risk assessment of ischemic stroke. Romero and colleagues88, recently conducted a prospective, nested case-control investigation to evaluate the associations of TRPM7 gene variations with the risk of ischemic stroke, and showed that 16 tag-single-nucleotide TRPM7 polymorphisms from 259 Caucasian men had no direct association with the risk assessment of ischemic stroke. However, no test has been done to suggest whether these TRPM7 mutants have neither dysfunction nor abnormality of the expression level of the channels. Thus, these human studies lead to no conclusion between TRPM7 activation and cerebral stroke. Further study is needed to fully understand the biophysical properties of the TRPM7 polymorphisms. Such information will dramatically contribute to our current understanding of the pathophysiological role of TRPM7 in cerebral ischemia and stroke.

Conclusions

With the failure of using NMDA and AMPA antagonists in clinical trials for stroke treatment, other non-glutamate mechanisms to ischemic cell death have been rigorously investigated. Such disappointing clinical outcomes may originate from the insufficient understanding of non-glutamate mechanisms and their molecular cascades involved in stroke, problems in drug development, delivery of drugs, side effects of drug, or limited time windows for treatment. While the compelling findings from both in vitro and in vivo studies indicate the involvement of TRPM7 channels in ischemic neuronal injury, further extensive preclinical testing is required to assess the therapeutic potential of the TRPM7 blockade in stroke.