The Multi-Functional Calcium/Calmodulin Stimulated Protein Kinase (CaMK) Family: Emerging Targets for Anti-Cancer Therapeutic Intervention

The importance of Ca2+ signalling in key events of cancer cell function and tumour progression, such as proliferation, migration, invasion and survival, has recently begun to be appreciated. Many cellular Ca2+-stimulated signalling cascades utilise the intermediate, calmodulin (CaM). The Ca2+/CaM complex binds and activates a variety of enzymes, including members of the multifunctional Ca2+/calmodulin-stimulated protein kinase (CaMK) family. These enzymes control a broad range of cancer-related functions in a multitude of tumour types. Herein, we explore the cancer-related functions of these kinases and discuss their potential as targets for therapeutic intervention.


Introduction
Ca 2+ is a major second messenger in cells and is essential for a variety of important signalling pathways. Alterations in intracellular Ca 2+ signalling regulate many biological processes, including gene transcription, exocytosis, the cell cycle, migration and apoptosis. Cytoplasmic Ca 2+ concentrations rise in response to a variety of stimuli, which activate Ca 2+ -channels in the plasma membrane, or by release from intracellular stores. It is increasingly being realised that disruption of normal Ca 2+ signalling contributes to the development of tumourigenic phenotypes [1], and aberrant Ca 2+ signalling has been implicated in each of the hallmarks of cancer originally identified by Hanahan and Weinberg [2]. Ca 2+ signals in the form of spikes or oscillations and is tightly regulated. The decoding of this is achieved by several downstream effectors, including calmodulin (CaM). Binding of Ca 2+ dramatically changes the conformation of CaM and increases its affinity for a large number of CaM-binding proteins, including the multifunctional CaM kinases (CaMKK, CaMKI, CaMKII and CaMKIV). These multifunctional kinases are widely expressed and control a variety of cancer related functions in a range of cancer types. Their potential as targets for anti-cancer therapeutic intervention have recently begun to be appreciated. Binding of Ca 2+ /CaM to CaMKKα prevents phosphorylation at S52, S74, T108, and S458, but enhances phosphorylation at S475 [25]. Additionally, phosphorylation of CaMKKα on S74, T108 and S458 negatively regulates activity [24,27,28] and phosphorylation of S458 blocks Ca 2+ /CaM binding. There are four CaMKII isoforms-CaMKIIα, CaMKIIβ, CaMKIIγ, and CaMKIIδ, with each isoform sharing a similar structure. CaMKII consists of a unique N-terminal domain (grey), adjacent to a catalytic domain (red) which contains an ATP binding region, and a regulatory domain (blue) containing overlapping autoinhibitory and calmodulin (CaM) binding regions. Phosphorylation sites are indicated by green balls. All isoforms also contain a C-terminal association domain (brown), which is involved in the formation of CaMKII multimers.
CaMKII requires Ca 2+ /CaM for initial enzyme activity. Binding of CaM molecules to adjacent subunits within a holoenzyme allows autophosphorylation at T826 of one or both of these subunits to occur [58]. Autophosphorylation of T286 in CaMKIIα (T287 in CaMKIIβ, γ, and δ) occurs rapidly and alters the affinity for Ca 2+ /CaM and enzyme activity [59,60]. CaMKII phosphorylation at T286 generates an autonomously active kinase that remains active even following CaM dissociation. Phosphorylation at T286 can also regulate the function of CaMKII by altering its binding to specific subcellular locations [61][62][63]. Once CaMKII is Ca 2+ -independent, and Ca 2+ /CaM is no longer bound, secondary sites within the CaM-binding site can be phosphorylated (T305/306 in CaMKIIα, and T306/307 in CaMKIIβ, γ, and δ) [64,65]. Phosphorylation at these sites prevents CaM binding and CaMKII subsequently becomes insensitive to changes in Ca 2+ /CaM [66].
Other forms of post-translational modification have also been demonstrated to alter CaMKII activity. Specifically, a pair of methionine residues (M281/282), present in the β, γ, and δ, but not α, isoforms [67], can become oxidised and produce a conformational change in CaMKII, similar to that produced following T286 phosphorylation, leading to an autonomous activation of CaMKII [68]. However, not all CaMKII phosphorylation sites modulate Ca 2+ /CaM binding and kinase activity. Phosphorylation at several sites that have no direct effect on kinase activity, but alter molecular targeting, have been identified in vivo, including T253, S279 and S314 [69][70][71][72].

CaMKIV
Ca 2+ /calmodulin-stimulated protein kinase IV (CaMKIV) is encoded by the CAMK4 gene. Two different isoforms (CaMKIVα and CaMKIVβ) are produced by alternative processing [73,74]. The CaMKIV expression pattern is similar to that of CaMKKβ, with both primarily being expressed in the brain, however, CaMKIV is also present in immune cells, the testes and ovaries [9,16,75,76]. It is particularly enriched in cerebellar granule cells, and subsequently, has previously been referred to as CaMK-Gr. The CAMK4 gene also encodes calspermin, a Ca 2+ /CaM binding protein of unknown function that is exclusively expressed in spermatids in the testes [76]. CaMKIV has been implicated in the regulation of homeostatic plasticity, neurite outgrowth, fear memory, immune and inflammatory responses, the regulation of cyclic AMP element binding protein (CREB) and cell proliferation [34,[77][78][79][80][81].
CaMKIVβ is identical to CaMKIVα, except for the addition of 28 amino acids at the N-terminus. Similar to the other CaMK family members, CaMKIV has a catalytic domain adjacent to an autoregulatory domain containing an autoinhibitory region that overlaps with the CaM-binding region ( Figure 4). CaMKIV requires Ca 2+ /CaM to become active, as well as phosphorylation of the conserved T in the activation loop (T200 in human CaMKIV and T196 in the rat enzyme) by CaMKK [82], which generates an autonomously active kinase. Following this, CaMKIV then autophosphorylates at S12 and S13, which enable Ca 2+ /CaM independent activity. Double S11A/S12A phospho-null mutants lack Ca 2+ /CaM-dependent basal activity and are unresponsive to CaMKK [83], indicating that S12 and S13 mediate an intrasteric inhibition, and are essential for full activation of CaMKIV. Similar to T305/6 in CaMKII, phosphorylation of CaMKIV within the CaM binding region (S336 in human CaMKIV and S332 in the rat enzyme) prevents CaM binding and inactivates CaMKIV [84].

The Role of CaMK Family Members in Cancer
Due to their role as mediators of Ca 2+ -signalling, it is not surprising that the CaMK family members have been identified as being overexpressed or aberrantly activated in a wide variety of cancer types. Additionally, each have also been implicated in controlling cellular processes important in cancer-related functions, such as cell proliferation, apoptosis, motility and invasion.

CaMKK
Despite the restricted expression of CaMKKβ in normal cells, it is overexpressed in several types of cancer compared to adjacent benign tissue (Table 1), including gastric tumours [85], hepatocellular carcinoma [86], high-grade gliomas [87], ovarian [88] and prostate cancer [89][90][91]. Androgen stimulation of androgen-dependent and castration-resistant prostate cancer cell lines results in upregulation of CaMKKβ [89,91,92], indicating direct regulation by the androgen receptor. Subsequent studies have shown that CaMKKβ is a key effector of the androgen receptor in stimulating glycolysis through the activation of AMPK and phosphofructokinase (PFK), which drives anabolism and proliferation of prostate cancer cells [91]. Furthermore, CaMKKβ expression increases with human prostate cancer progression and Gleason score [89,91], and with tumour progression in a transgenic adenocarcinoma of the mouse prostate (TRAMP) mouse model of prostate cancer [89], suggesting that it may also play a role in prostate cancer progression.
Increased CaMKKβ expression also correlates with poor patient survival in additional tumour types (Table 1). High CaMKKβ expression is associated with poor disease-free survival in hepatic cancer patients [86], and poor overall survival in glioma patients [87]. These studies suggest that CaMKKβ may be a useful prognostic marker for liver and brain cancers. Prostate cancer progression (n = 5) Increased protein expression in prostate cancer compared to normal prostate and with increasing Gleason score [89] Normal prostate and prostate cancer TMA (n = 80), cancer, adjacent normal and metastases TMA (n = 95) Increased protein expression in primary prostate cancer and bone metastasis compared to normal prostate [90] Gastric Gastric adenocarcinoma and normal oesophagus TMA (n = 98) Increased protein expression in gastric tumours compared to normal oesophagus [85] Liver Hepatocellular carcinoma transcriptome profile microarray (n = 247); matched normal and tumour (n = 22).
Increased expression in liver cancer and CAMKK2 high correlates with poor disease-free survival. CaMKKβ protein upregulated in tumour compared to adjacent normal tissue. [86] Glioma Human glioma and normal brain tissue (n = 147 for expression and n = 101 for methylation); Whole genome mRNA expression microarray (n = 305 diffuse glioma samples, n = 151 methylation microarray, n = 275 GBM) CAMKK2 mRNA and protein is more highly expressed in high-grade gliomas compared to low-grade. Increased expression and CAMKK2 high correlates with poor overall survival. CAMKK2 is differentially methylated between low and high grade glioma. [87] Ovarian High grade serous papillary ovarian cystadenocarcinoma and high-grade ovarian carcinoma with mucinous features (n = 4) Increased protein expression in high grade serous papillary cystadenocarcinoma and high-grade ovarian cancer with mucinous features compared to non-malignant stromal tissue. [88] CaMKI AML TCGA AML database (n = 186) CAMK1D high correlates with poor overall survival [93] Endometrial cancer Endometrial carcinoma and normal endometria (n = 31 and n = 20) Protein expression is associated with PCNA-labeling, stage, histological grade, the presence of invasion and outcome [94] Breast cancer Primary breast ductal carcinoma (n = 35) PNCK mRNA is more highly expressed in a subset (8/23) of human breast tumours compared to benign breast tissue [95] ccRCC ccRCC and adjacent normal tissue (n = 92) and primary ccRCC tissue (n-248) PNCK mRNA and protein expression higher in tumour compared to normal. Patients with PNCK high have shorter overall survival [96]  Protein expression is associated with PCNA-labeling, stage, histological grade, the presence of invasion and outcome [94] Colon cancer Paracancerous tissue, well-differentiated and poorly differentiated colon cancer (n = 5, n = 6, n = 6) CaMKII protein expression increased in colon cancer compared to paracancerous tissue, and increased with poor differentiation [99] Breast cancer GOBO Breast Cancer Database (n = 1881); Normal, primary and metastatic breast cancer TMA (n = 40, n = 70, and n = 10) CAMK2 high associated with worse overall and distant metastasis free survival. Total CaMKII protein and T286/7 phosphorylation is increased in primary breast cancer and metastases [100] Osteosarcoma Chondroblastic, osteoblastic and fibroblastic subtypes (n = 114) Phosphorylation of αCaMKII at T286 is increased in osteosarcoma compared to normal osteoblasts and mesenchymal stromal cells [101] Primary osteosarcoma tumours (n = 4) Phosphorylation of αCaMKII at T286 is increased in osteosarcoma [102] Lung cancer Oncomine databases (n = 187, n = 226, n = 130) CAMK2G high associated with worse overall survival [103] GWAS in NSCLC patients (n = 354) Rs10023113 in CAMK2D associated with survival [104] Gastric cancer Non-metastatic and metastatic gastric cancer tissues (n = 10, and n = 10) Phosphorylation at T286 is increased in metastatic compared to non-metastatic tissue [105] CaMKIV AML TCGA AML database (n = 186) CAMK4 high correlates with poor overall survival [93] HCC Normal liver, chronic hepatitis, cirrhosis, and HCC (n = 4, n = 6, n = 4, n = 12) CaMKIV protein expression and activation increased in HCC compared to normal liver and cirrhosis [106] CaMKKβ positively regulates cancer cell proliferation, migration and invasion in a variety of cell types in vitro ( Table 2). Downregulation of CaMKKβ expression using small interfering RNA (siRNA) or pharmacological inhibition inhibits prostate cancer cell proliferation [89,91,92,107], migration and invasion [92]. Conversely, CaMKKβ overexpression in LNCaP prostate cancer cells increases cell migration, further supporting a role for CaMKKβ overexpression in prostate cancer progression [92]. By contrast, Shima et al. showed that CaMKKβ overexpression in LNCaP cells decreases cell proliferation and tumour growth in vivo [90], indicating that further examination of the role of CaMKKβ in prostate cancer cell proliferation is required. Downregulation or pharmacological inhibition of CaMKKβ also decreases proliferation, migration and invasion of glioma [87], gastric [85,108] and liver cancer [86] cells and expression of a dominant negative CaMKK mutant suppresses medulloblastoma cell migration [109], suggesting that CaMKK activity is essential for this process. Taken together, these studies highlight the importance of CaMKK in controlling cancer cell proliferation and metastatic processes in a range of cancer types, indicating that its role in these functions is not cell-type specific, and also suggesting that CaMKKβ may be a valid anti-cancer target for a variety of cancer types.

CaMKI
The various CaMKI isoforms have been shown to be overexpressed in several different cancer subtypes (Table 1). CaMKI is more highly expressed in stage III and IV endometrial carcinomas, when compared to stage I and II, and is associated with proliferating cellular nuclear antigen (PCNA)-labelling, clinical state, histological grade, and the presence of invasion, indicating that CaMKI may play a role in endometrial carcinoma progression [94]. Expression of several of the specific CaMKI isoforms have also been examined in specific cancer types. For example, CaMKIγ/PNCK is overexpressed in a subset of primary breast cancers compared to benign mammary tissue [95], and in clear renal cell carcinoma compared to adjacent non-tumour tissues [96]. Furthermore, a significant correlation between PNCK expression and Fuhrman grade, tumour size, T and N stage was observed in these clear cell renal carcinoma samples, and high levels of PNCK is an independent predictor for poor patient survival [96]. CAMKID has also been identified as a potential prognostic marker in acute myeloid leukaemia (AML), as high levels of CAMKID mRNA are associated with significantly worse AML patient survival [93].
Unsurprisingly, CaMKI has been implicated in a variety of cancer-related cellular processes ( Table 2). CaMKI controls the progression of MCF-7 breast cancer of cells through G 1 [110], potentially by regulating cdk4 and retinoblastoma protein (Rb) phosphorylation, as overexpression of the kinase-inactive CaMKI K49A mutant prevents cdk4 activation and Rb hyperphosphorylation in WI-38 fibroblasts [17]. Indeed, knockdown of CaMKI expression in MV-4-11 and Kasumi-1 AML cells significantly decreases cell proliferation and inhibits leukaemia development and prolongs the survival of AML xenografted mice [93]. By contrast, CaMKI overexpression does not affect leukaemia development, but overexpression of the kinase-inactive CaMKI K49E mutant significantly decreases AML colony formation and increases mouse survival in AML transplantation studies [93], demonstrating the importance of CaMKI activity in controlling cell proliferation. Similar to expression of a dominant negative CaMKK mutant, expression of a dominant negative CaMKI mutant significantly decreases medulloblastoma cell migration [109], highlighting the importance of the CaMKK-CaMKI cascade in this cellular process. Murine CKLiK is upregulated by the haematopoietic cell-specific ETS family transcription factor, PU.1, in murine erythroleukemia cells, and is involved in apoptosis [122], suggesting that the human homolog of CKLiK may function in a similar fashion. These studies demonstrate that CaMKI can control cell proliferation, migration, and survival in medulloblastoma, breast and haematopoietic cancers, and therefore, may be a potential anti-cancer target for these cancer types.

CaMKII
Tumour cells express an entirely different spectrum of CaMKII isozymes than adult neuronal tissue or their non-transformed tissue counterparts. Eight distinct β, γ and δ CaMKII isozymes have been identified in human mammary tumour and neuroblastoma cell lines [123]. In addition to expression of these novel tumour variants, CaMKII is also overexpressed in a variety of cancer types (Table 1). CaMKII is more highly expressed in stage III and IV endometrial carcinomas when compared to stage I and II, and is associated with PCNA-labelling, clinical state, histological grade, and the presence of invasion, indicating that similar to CaMKI, CaMKII may play a role in endometrial carcinoma progression [94]. CaMKII is overexpressed in colon cancer compared to adjacent normal tissue and increases with poor differentiation [99], and in primary breast cancer compared to adjacent normal breast, as well as in lymph node metastases [100], indicating that CaMKII may play a role in cancer progression. In regards to specific isoforms, CaMKIIγ is increased in AML patient blasts compared to normal peripheral blood cells [98], in lung cancer compared to normal lung tissue [103] and is upregulated at diagnosis in chronic myeloid leukaemia (CML) [97]. Additionally, CaMKIIγ expression is increased in CML blasts resistant to tyrosine kinase inhibitors [97], glioma cells resistant to the Fas agonistic antibody (CH-11) [124] and also in TRAIL-resistant melanoma cells [125], suggesting that it may play a role in chemoresistance.
The single-nucleotide polymorphism (SNP) rs10023113 in CAMK2D is associated with poor survival of early-stage non-small cell lung cancer patients [104], and high expression of CAMK2G in lung cancers is associated with significantly worse overall survival in 3 different cohorts [103]. Additionally, breast cancer samples that express high levels of CAMK2 exhibit significantly worse overall and distant metastasis-free survival compared to patients with low CAMK2 expression [100]. These studies suggest that CaMK2 genes are potential prognostic biomarkers for a range of cancer types.
Furthermore, not only is CaMKII expression increased in a variety of cancer types, enhanced autophosphorylation at T286 is also frequently observed (Table 1). Increased T286/7 phosphorylation of CaMKII is noted in TRAIL-resistant melanoma cells [125], lung cancer oncospheres [103], osteosarcoma [101,102], leukaemia stem cells [126], AML patient blasts [98], primary and metastatic breast tumours compared to normal adjacent breast tissue [100] and in metastatic gastric cancers compared to non-metastatic tissues [105], indicating that enhanced CaMKII activity is a feature of cancers, and is also associated with metastatic disease.
Furthermore, CaMKIIγ has been implicated in tumourigenesis in a variety of cancer types. CaMKIIγ deletion suppresses T cell lymphomagenesis in vivo [116]. CaMKIIγ knockout mice develop fewer tumours, that are smaller than their wild-type counterparts, in a dextran sodium sulfate (DSS) and azoxymethane (AOM) model of colitis-associated tumourigenesis. Furthermore, only knockout in colonic tissue-resident cells, and not in bone marrow-derived immune cells, is involved in this suppressive effect [132]. By contrast, CaMKIIγ overexpression increases colon proliferation rates, decreases cell death and increases distal colitis-associated cancer compared to control mice [132]. CaMKIIγ −/− mice exhibit increased tumour number and volume in diethylnitrosamine (DEN) and DEN followed by tumour promotor models of hepatic cancer [133]. Additionally, CaMKIIγ knockdown inhibits lung cancer tumourigenicity, and overexpression enhances tumourigenicity in vitro and in vivo [103].
Differentially phosphorylated CaMKII also controls different cellular functions. For example, dephosphorylation of CaMKII at T253 controls the metaphase-anaphase transition in neuroblastoma (SHSY5Y) and breast cancer (MDA-MB-231) cell lines [117]. Overexpression of wild-type or a T286 phospho-mimic mutant (T286D) of CaMKIIα increases cancer cell proliferation [117], migration and invasion [100]. By contrast, overexpression of a T253 phospho-mimic mutant (T253D) prevents cancer cell proliferation [117]. Taken together, these studies highlight the importance of CaMKII in controlling cancer cell proliferation and metastatic processes in a range of cancer types, indicating that its role in these functions is not cell-type specific but broadly applicable, and indicates that CaMKII may be a valid anti-cancer target for a variety of cancer types.

CaMKIV
Despite the restricted expression of CaMKIV in normal tissues, CaMKIV is overexpressed in several different types of cancer (Table 1). CaMKIV expression and activity is increased in hepatocellular carcinoma [106], small cell lung cancer [111], and is significantly associated with clinical stage, myometrial invasion and clinical outcome in endometrial carcinoma [134]. Furthermore, high CAMK4 expression is associated with significantly worse overall survival for AML patients [93], and in endometrial carcinoma [135].
As CaMKIV is expressed in immune cells, it is not surprising that it has been shown to regulate haematopoietic stem cell homeostasis [16]. Additionally, CaMKIV has been implicated in cell proliferation and cell cycle regulation. Decreasing CaMKIV expression inhibits AML development in vitro and in vivo [93], and decreases hepatic cancer cell proliferation [86]. By contrast, overexpression of CaMKIV in AML cells arrests cells in G 0 /G 1 , in an activity dependent manner [121], suggesting that CaMKIV may have cell line dependent effects, even within the same cancer subtype.

The CaMK Family Are Potential Anti-Cancer Therapeutic Targets
Due to the importance of CaMK family members in controlling cancer-related functions, their suitability as anti-cancer targets have begun to be explored. Several pharmacological inhibitors that inhibit the activity of these enzymes have been developed, and their usefulness as anti-cancer treatments in a variety of cancer types has been examined.

STO-609
7-Oxo-7H-benzimidazo[2,1-a]benz[de]isoquinoline-3-carboxylic acid (STO-609, Figure 5) is a selective CaMKK antagonist that inhibits autophosphorylation of both CaMKKα and CaMKKβ [136], without any significant effect on CaMKI and CaMKIV. However, at doses~100-fold higher than the half maximal inhibitory concentration (IC 50 ) for CaMKK, CaMKII and myosin light chain kinase (MLCK) are also inhibited. Additionally, STO-609 has been demonstrated to bind to and activate the aryl hydrocarbon receptor (AhR) [137], indicating that STO-609 may not be as CaMKK selective as initially believed. As CaMKK controls cancer cell proliferation, migration and survival in a variety of cancer types in vitro (Table 2), inhibiting CaMKK activity may be a valid anti-cancer therapeutic strategy in these cancer types. Indeed, STO-609 decreases AML [93], prostate [89,91], gastric [108], hepatocellular [86] and ovarian cancer cell proliferation [88] in vitro, and induces apoptosis in ovarian [88] and gastric cancer lines [108]. Furthermore, treatment with STO-609 significantly reduces tumour burden in prostate and hepatocellular cancer mouse models in vivo (Table 3). Systemically administered STO-609 decreases tumour growth, both as a single agent and additively in combination with AR inhibition, in a C4-2B prostate cancer xenograft model [91], and in the DEN-induced hepatocellular carcinoma mouse model [86], demonstrating that CaMKK inhibition is a valid strategy for the treatment of prostate and hepatic cancer, and based on the in vitro studies, may potentially be suitable for other cancer types, including AML, gastric and ovarian cancers.
Intraperitoneal doses of STO-609 up to 300 µM/kg, or daily injections of 30 µM/kg for 4 weeks, have been shown to be well-tolerated in C57Bl/6 J mice, and did not induce parameters of liver or kidney toxicity [138]. However, the primary limitation of the use of STO-609 as an anti-cancer treatment is its poor solubility, therefore improved derivatives that increase solubility, whilst increasing efficacy will need to be developed. Table 3. Effects of pharmacological inhibitors of CaMK family members on tumour burden in in vivo animal models of cancer. AR androgen receptor; CML chronic myeloid leukaemia; DEN diethylnitrosamine; HCC hepatocellular carcinoma; MNU N-methyl-N-nitrosourea; NOD-SCID non-obese diabetic severe combined immunodeficient; NSG nod scid gamma.  (Figure 5), are membrane permeable CaMKII inhibitors that are competitive for CaM [140]. They were originally described as CaMKII specific inhibitors, as they do not affect other CaM-dependent enzymes, such as AMPK and MLCK [141].
Whilst in vivo studies using these inhibitors do not describe any side-effects, KN-62 and KN-93 have been widely used to examine a variety of heart and brain-related functions. For example, KN-62 and KN-93 can depress the rate of beating of cultured myocytes [155] and afterdepolarisation in the heart [156], respectively, and can block long-term potentiation (LTP) in rat hippocampus [157], indicating that they are likely to affect cardiac, learning and memory processes. Due to these potential side effects, as well as the range of off-target effects using these inhibitors, these inhibitors are unlikely to be translated into clinical use without modification or the use of alternative more cancer-selective modes of drug delivery (e.g., nanoparticles or liposomes). Long inhibitory peptides based on the autoinhibitory domain of CaMKIIα have been developed. The N-terminus of this peptide contains the autophosphorylation site forming the basis for peptide substrates such as autocamtide-2 and -3 [158] and the development of a non-phosphorylatable analogue of autocamtide-2 generated the peptide inhibitors AIP (KKALRRQEAVDAL) [140] and AC3-I (KKALHRQEAVDCL) [159]. AIP competes with the active site of CaMKII, inhibiting activity regardless of how CaMKII was activated, and inhibits CaMKII with over 100-fold selectivity relative to protein kinase C (PKC), PKA and CaMKIV. However, altered selectivity can also occur when peptides are fused to GFP or modified by lipids to increase membrane permeability. For example, myristoylated AIP has been shown to have effects unrelated to CaMKII inhibition [160], indicating the presence of an additional non-CaMKII-related target, and green fluorescent protein (GFP)-AC3-I can inhibit cellular actions of protein kinase D1 (PKD1) [161]. AIP significantly attenuates glioma [119,162] and breast cancer [100] migration and invasion. Furthermore, AC3-I treatment arrests ovarian carcinoma cells in G 2 and stops proliferation [163]. These studies indicate that CaMKII inhibition may be a viable therapeutic option for the treatment of metastatic disease.

CaMKIIN Derived Peptides (CaMKIINtide)
Small endogenous proteins that inhibit CaMKII (CaMKIIN) have been identified in rat brain [164] and there are two known rat isoforms-CaMKIINα [164] and CaMKIINβ [165]. Human CaMKIIN has also been identified in human dendritic cells [166]. CaMKIIN is implicated in the control of progression of cells through S phase. Overexpression of hCaMKIIN in colon cancer cells decreases cell proliferation, whereas silencing increases cell growth rates [166,167]. Furthermore, hCaMKIIN overexpression in ovarian cancer cells decreases proliferation and tumourigenicity [168,169], and also reduces migration and colony formation [169]. This implies a potential application of hCaMKIIN in the treatment of colon and ovarian cancers.
Identification of the core inhibitory domain of CaMKIIN led to the generation of a 28 amino acid peptide inhibitor called CaMKIINtide [164]. CaMKIINtide only inhibits activated CaMKII [164], and does not inhibit PKC, PKA, CaMKI, CaMKIV or CaMKK [164]. CaMKIINtide has also been modified to increase potency [170,171]. A cell permeable form, antCaNtide, decreases medullary thyroid cancer [112], and AML cell proliferation [121] indicating that the development of additional CaMKII specific inhibitors may provide viable therapeutic options for the treatment of haematological and thyroid cancers.
Berbamine has been shown to reduce tumour burden in several different animal models (Table 3). Berbamine decreases liver cancer [139] and CML [126] tumour burden in vivo and decreases tumour burden and significantly increases survival in an N-methyl-N-nitrosurea (MNU)-induced lymphoma model [116].
The main limitation for the use of berbamine clinically as an anti-cancer agent is its short plasma half-life and poor bioavailability at the tumour site after systemic administration. To circumvent this, lipid-based nanoparticles loaded with berbamine have been developed and have been shown to decrease primary tumour growth in a B16F10 mouse melanoma model and also suppress the incidence of lung metastases in vivo [181]. This highlights that newer mechanisms of drug delivery may be useful clinically to increase the cancer-specificity of these drugs, without enhancing toxicity.

Concluding Remarks and Perspectives
The CaMK family members, particularly CaMKII, are attractive anti-cancer targets as they are overexpressed in a plethora of cancer types, compared to adjacent normal tissue, and are vital in the modulation of cancer cell proliferation, survival, invasion and migration. Several targeted and broad-acting CaMK inhibitors have demonstrated pre-clinical anti-cancer efficacy in vivo. Many of the previously believed 'CaMK-specific' inhibitors have been shown to have a variety of off-target effects, which further limits their clinical usefulness. By contrast, CaMKIIN-tide, and its modified derivatives, based on the endogenous CaMKII inhibitory protein (CaMKIIN) have no described off-target effects, and are the most promising lead compounds for further development described here-in, however, their usefulness in vivo remains to be investigated. Additionally, as the CaMK family members are also involved in highly important, non-cancer related functions, direct inhibition using these existing CaMK inhibitors are likely to produce a range of deleterious side-effects if used clinically. Therefore, to be useful therapeutically, cancer-specific inhibitors or more cancer-specific modes of drug delivery would be required to be developed. One such potential strategy would be to encapsulate these CaMK inhibitors in nanoparticles or liposomes that specifically target cancer cells for the delivery of the inhibitor. Whilst this has been examined in other disease models for CaMKIIN inhibitory peptides, such as heart disease and asthma [182,183], they remain to be tested in cancer.