Judging a Protein by More Than Its Name: GSK-3

See allHide authors and affiliations

Science's STKE  18 Sep 2001:
Vol. 2001, Issue 100, pp. re12
DOI: 10.1126/stke.2001.100.re12


As knowledge of cellular signal transduction has accumulated, general truisms have emerged, including the notion that signaling proteins are usually activated by stimuli and that they, in turn, mediate the actions of specific agonists. Glycogen synthase kinase-3 (GSK-3) is an unusual protein-serine kinase that bucks these conventions. This evolutionarily conserved protein kinase is active in resting cells and is inhibited in response to activation of several distinct pathways, including those acting by elevation of 3′ phosphorylated phosphatidylinositol lipids and adenosine 3′-5′-monophosphate (cAMP). In addition, GSK-3 is distinctly regulated by, and is a core component of, the Wnt pathway. This review describes the unique characteristics of this decidedly oddball protein kinase in terms of its diverse biological functions, plethora of targets, role in several human diseases, and consequential potential as a therapeutic target.


As genes and proteins are discovered, they are often ascribed names based on function--their function at the time of discovery. For gene products that are subsequently found to be pleiotropic in action, this can lead to somewhat inappropriate or even misleading monikers. Glycogen synthase kinase-3 (GSK-3) was first identified as a consequence of its phosphorylation activity toward glycogen synthase, the rate-limiting enzyme of glycogen metabolism. Although its original name has stuck, this enzyme's sphere of influence extends well beyond intermediary metabolism. GSK-3 is an old gene with homologs identified in every eukaryotic species examined. In many of these species, GSK-3 has been implicated in fundamental processes including cell fate determination, metabolism, transcriptional control, and, in mammals, oncogenesis and neurological diseases. In the decade since the cloning of the first genes encoding GSK-3, this molecule has surfaced in a wide range of biological scenarios, and its study continues to yield surprises. Part of the reason for its pleiotropic behavior may relate to the fact that, unlike most protein kinases, GSK-3 is primarily regulated by inactivation. In resting cells, the enzyme displays high activity. After cellular stimulation, this activity is turned off. Further, many of the cellular targets of GSK-3 are inhibited by GSK-3-mediated phosphorylation. As a consequence, when GSK-3 activity is suppressed, this catalog of substrate proteins becomes activated. This review brings together the present knowledge of this enzyme: its physiological roles, means of regulation, and involvement in human disease.

Regulation of GSK-3 by the Wnt Pathway

In multicellular organisms, GSK-3 functions in several distinct signaling pathways. Activation of any of these pathways leads to inactivation of GSK-3 kinase activity by one of at least three distinct mechanisms: (i) inhibition by the action of Dishevelled, (ii) NH2-terminal domain serine phosphorylation, or (iii) tyrosine phosphorylation. Although the molecular details of the first mechanism remain unclear, much knowledge has been leveraged by genetic analysis of this pathway in simpler organisms, resulting in considerable understanding of the biological processes impacted by GSK-3 in several species.

Wnt-1 is the archetypal member of a family of secreted glycoproteins related to the wingless (wg) protein of Drosophila melanogaster [for reviews, see (1-3)] (Fig. 1). Wnt-1 (originally termed int-1) was first described as a gene activated by insertion of mouse mammary tumor virus; deregulated expression of the gene causes oncogenic transformation, measured by the ability to form foci in soft agar and mitogenesis in certain cell types (4-6) and mammary tumors in p53-deficient mice (7). There are currently more than 15 identified Wnt-related genes, six of which are normally expressed in mammary tissue (8). Misexpression of certain Wnts (1 and 10b) in mammary glands results in abnormal cellular organization and glandular function (9, 10), and Wnt7b expression is increased up to 30-fold in 10% of breast tumors (11).

Fig. 1.

Components of the Wnt pathway. (Left) Under resting conditions, GSK-3 is active. The fraction of the kinase associated with Axin phosphorylates the bound β-catenin, targeting it for degradation and maintaining a low level of soluble β-catenin. (Right) After Wnt binding to its receptor, Frizzled, GSK-3 activity is suppressed, alleviating the destruction of β-catenin, which leads to its accumulation and interaction with the TCF/LEF DNA binding proteins and regulation of transcription.

Much of the progress in understanding the signaling pathway induced by Wnts has derived from genetic analysis of the Wnt1 ortholog in Drosophila, termed wingless [(12-14); reviewed in (2, 14-17)]. Maternal loss of wingless function in flies is characterized by ectopic bristles on the ventral cuticle. Later (zygotic) phenotypes include perturbations in cell fate determination. Genetic screens for enhancers or suppressors of these phenotypes have revealed several gene products that play roles in processing and transmission of the Wingless signal. Wingless binds serpentine receptors termed Frizzled (18-22), which, in turn, propagate an unknown signal via a PSD-95/Dlg/ZO-1 (PDZ) domain containing a molecule called Dishevelled (23-26). Dishevelled, in turn, induces the inactivation of Zeste-white3 (also known as Shaggy), the fly homolog of GSK-3 (27-31). Expression of mammalian GSK-3 in Drosophila embryos that are mutant for Zeste-white3/Shaggy functionally rescues these embryos. A critical downstream target of this kinase in Wingless signaling is Armadillo [(32, 33); reviewed in (34)]. In unstimulated conditions (absence of Wingless), Zeste-white3 phosphorylates two serines on Armadillo, which target the molecule for ubiquitination and degradation by the proteasome (35, 36). In the "basal" state, then, cytoplasmic Armadillo protein levels are kept low by continual destruction that is driven by phosphorylation. In the presence of Wingless, a Dishevelled-mediated signal inhibits Zeste-white3, reducing the phosphorylation of Armadillo, resulting in stabilization and accumulation of the protein. The increased concentration of Armadillo facilitates its interaction with nuclear proteins, including DTCF (also known as Pangolin), Teashirt, and Groucho, causing changes in gene expression (both positive and negative) (37-40).

Mammals harbor homologous gene products in the Wnt pathway, which are conserved at both the sequence and functional levels (Fig. 1, Table 1). There are three mammalian Dishevelled proteins (41), two homologs of Zeste-white3/shaggy [GSK-3α and GSK-3β (42)], β-catenin, which is the homolog of Armadillo, and the TCF (also known as LEF) family, which is similar to DTCF (43-47). Many of these proteins are functionally interchangeable in simpler metazoans and all are similarly regulated (29, 30, 48-56). Several additional components of the Wnt/Wg pathway have been identified without the aid of Drosophila genetics, namely, adenomatous polyposis coli (APC), GSK-3 binding proteins (GBPs), and Axin. APC is a tumor-suppressor locus encoding a large protein commonly deleted in familial colorectal cancer [(57-59); reviewed in (60)]. In cells, APC is found in a complex with β-catenin (61), and cells mutant for APC exhibit stabilized β-catenin (62-64). APC also plays a role in exporting β-catenin from the nucleus, providing a means to inactivate the pathway (65). As well as phosphorylating β-catenin, GSK-3 targets APC. In vitro, this phosphorylation reduces the affinity of APC for microtubules, which presumably affects the localization of the complex (66). A substantial fraction of colorectal tumors that are wild type for APC exhibit stabilizing mutations in β-catenin, and such mutations have also been observed in some melanomas, childhood hepatoblastomas, and fibromatoses (67-69). β-catenin stabilization is therefore associated with several human malignancies.

Axin acts an organizing protein or "scaffold" (70, 71). Axin 1 is the product of the mouse fused locus, and mutations in this gene cause axis duplication, a phenotype observed upon disruption of the Wnt pathway in Xenopus (50-56). A related gene termed Axil/Conductin is also present in mammals (72, 73). Axin and Axil (also known as Conductin) contain independent binding sites for β-catenin, GSK-3, and APC in addition to an NH2-terminal RGS (regulator of G protein signaling)-like domain and a COOH-terminally located DIX (Dishevelled homologous) domain implicated in dimerization (73-77). In essence, this large complex acts as a machine to recruit, phosphorylate, and destroy β-catenin.

The role of GSK-3 in the Wnt pathway is unusual in signaling pathways. The Wnt pathway has, in effect, commandeered a subset of GSK-3 molecules from other functions and evolved a regulatory structure that takes advantage of the enzyme for a selective function without compromising specificity. The critical molecule facilitating this signal insulation is Axin. Analogous to the Ste5 and JIP (Jun NH2-terminal kinase interacting protein) scaffolds that tether various mitogen-activated protein kinase (MAPK) components [reviewed in (78)], Axin orchestrates the degradation of β-catenin by bringing together the primary players. Axin acts as a platform to lure β-catenin into the proximity of GSK-3. As discussed below, GSK-3 has many other functions in a cell, yet its function in the Wnt pathway is isolated from its other tasks. Thus, the key role of the Axin complex is to sequester a fraction of the total cellular GSK-3 molecules for regulation of β-catenin. In support of this idea, cellular concentrations of Axin are substantially lower than those of GSK-3, and expression of more Axin in cells increases the rate of β-catenin destruction, presumably by recruiting more GSK-3 to the task of β-catenin phosphorylation. Non-Axin-associated GSK-3 displays no activity toward β-catenin and is therefore irrelevant to this pathway. Perhaps the most striking indication of the separate "life" of GSK-3 in Wnt signaling is the fact that in mice engineered to lack GSK-3β, there is no obvious defect in Wnt biology (79). GSK-3β knockout cells do not compensate by increasing levels of GSK-3α. Instead, Axin switches from binding both GSK-3α and β to only binding GSK-3α. Because the concentration of GSK-3α in cells is higher than Axin, the latter molecule is fully loaded with the kinase, and β-catenin signaling is therefore unaffected. This finding also illustrates the insensitivity of the Wnt pathway to the two GSK-3 isoforms. Indeed, there is no evidence that the Wnt pathway discriminates between GSK-3α or β in the pursuit of β-catenin regulation. This fact is not reflected in the Wnt literature, in which GSK-3β is given undue prominence. The bias perhaps derives from a difference in the efficacy of rescue of the zeste-white3 mutation in flies by GSK-3α and β in which the β isoform appeared more effective (29, 30). However, the amount of expression of the two proteins was not equalized in these experiments, and GSK-3α does substitute for the Drosophila kinase.

There are several outstanding questions regarding the regulation of GSK-3 by Wnts. First, the molecular mechanism by which Wnts inactivate GSK-3 is unclear. More precisely, the means by which Dishevelled represses GSK-3 activity has yet to be detailed. Addition of Dishevelled to a GSK-3-Axin-β-catenin complex in vitro reduces GSK-3 activity (80). A small protein termed GBP (GSK-3 binding protein; also known as FRAT, frequently rearranged in advanced T cell lymphomas) binds and inhibits GSK-3 in a mutually exclusive manner to Axin (81, 82). It is therefore possible that GBP may modulate GSK-3 association with Axin, although this would require a mechanism of GBP regulation by Wnts. GBP homologs have not been identified in the genomes of Drosophila or C. elegans, suggesting a species-specific function and excluding it from a canonical role in Wnt/Wg-mediated GSK-3 regulation. GSK-3 becomes phosphorylated in response to Wnt, which may contribute to its inactivation (31, 48). However, a direct role for this phosphorylation (as well as that of Dishevelled phosphorylation) has yet to be established: For example, it may be a consequential rather than causal modification. A second question is, how does Axin shield GSK-3 from other regulatory systems [such as the phosphatidylinositol 3′ kinase (PI3K) pathway (see below)]? Third, does the population of GSK-3 associate with Axin turnover, and is this important for regulation? That is, is there exchange of GSK-3, and, if so, does this contribute to the inactivation process? This is important because dissociation of GSK-3 from the Axin complex effectively blinds it to β-catenin, resulting in accumulation of this target protein. Fourth, is β-catenin the only molecule processed by the Axin-APC-GSK-3 complex? Given the high conservation of the components of this destruction factory, perhaps there are other proteins that bind to the β-catenin site on Axin and are similarly processed by GSK-3. Precedents for this idea include the F-box (a binding motif first noted in cyclin F) proteins of the SCF (Skp1-Cdc53-F-box)-like complexes that have limited specificity for target proteins.

Other Pathways to GSK-3

The PI3K pathway is often stimulated in response to mitogens and hormones (83). PI3 kinase-mediated production of 3′ phosphorylated phosphatidylinositols within the cytoplasmic leaflet of the membrane recruits a selection of pleckstrin homology (PH)-domain containing proteins, including several protein kinases such as protein kinase B (PKB, also known as Akt). The protein substrates of PKB have been avidly sought after the revelation that expression of activated PKB in cells (84-87) or in transgenic mice results in a prosurvival signal (88, 89). Targets for PKB include the proapoptotic protein Bad and members of the Forkhead family of transcription factors [reviewed in (90, 91)]. PKB also phosphorylates both isoforms of GSK-3 in mammals, as well as the Drosophila ortholog. Phosphorylation of GSK-3 by PKB at Ser21 (on the α isoform) or Ser9 (on the β isoform) causes inactivation and is the primary mechanism responsible for growth factor inhibition of this protein kinase (92-94) (Fig. 2). Other protein kinases also target these two serine residues, including p70 S6 kinase, Rsk1, and cAMP-dependent protein kinase (PKA) (95-97). Interestingly, activation of these pathways is not usually associated with stabilization of β-catenin. For example, although inhibition or genetic disruption of the GSK-3 ortholog in flies (Zeste-white3) results in a clearly defined cuticle phenotype because of stabilization of Armadillo, expression of activated alleles of PKB in flies does not result in this phenotype (98). In mammalian cells, insulin-mediated activation of PKB and its consequent phosphorylation and inactivation of GSK-3 are not associated with transcriptional activation of LEF-1 (99). Cells therefore discriminate in their responses to GSK-3 inhibition by the Wnt pathway versus other pathways (see Substrates below).

Fig. 2.

Anatomy of mammalian GSK-3α and GSK-3β. Serine and tyosine phosphorylation sites are indicated.

In addition to serine phosphorylation, GSK-3 is tyrosine phosphorylated (100) (Fig. 2). The site of this phosphorylation is within the T loop at a tyrosine analogous to that required for activation of the MAPK family of protein kinases. However, unlike the MAPKs, GSK-3 is generally predominantly tyrosine phosphorylated in resting cells. Whether this phosphorylation plays a regulatory role is uncertain, although its phosphorylation is required for GSK-3 function (100, 101). Neurotrophin withdrawal from PC12 cells and staurosporine treatment of SH-SY5Y neuronal cells increases the degree of tyrosine phosphorylation of GSK-3 (in this case, of Tyr216 in GSK-3β) (102). Induction of tyrosine phosphorylation correlates with increased activity, as well as cellular apoptosis. However, in most cells, GSK-3 is abundantly tyrosine phosphorylated and is highly active. The reason why the enzyme is inactive and nonphosphorylated in the neuronal cells is unclear. In Dictyostelium, GSK-3 activity is induced by cAMP receptors and is required for proper cell fate determination between stalk and spore cells (103, 104). A candidate tyrosine kinase termed Zak1 has been identified that tyrosine phosphorylates and activates slime mold GSK-3 (105). A mammalian cognate of ZAK1 remains to be found.

GSK-3 activity is thus tightly regulated in cells. Possible exceptions are the yeast orthologs. Although these proteins are phosphorylated on tyrosine and serine and threonine residues, regulation of their activity has yet to be shown (101, 106-108).

Biological Roles of GSK-3


GSK-3 was first identified as a key, negative regulator of glycogen synthase, the rate-limiting enzyme of glycogen synthesis [(109-112); reviewed in (113, 114)]. Of note, insulin-induced signals reduce the phosphorylation of the particular residues on glycogen synthase that are targeted by GSK-3. This occurs by a combination of events, including regulation of phosphatase activity and insulin-mediated inactivation of GSK-3, through insulin receptor substrate-1 (IRS-1), activation of PI3K, and consequent PKB phosphorylation of GSK-3.

Identification of protein substrates of GSK-3 has been confounded by several biochemical complexities. Unlike most protein kinases, GSK-3 does not phosphorylate many of its targets on the basis of simple primary sequence recognition. In the case of β-catenin, both GSK-3 and the substrate must be anchored to Axin: Simply mixing the purified kinase and β-catenin together does not result in phosphorylation (71). In many cases, a priming phosphorylation event by a distinct protein kinase is required that generates a recognition site for GSK-3 (115-118).

The complex requirements for GSK-3 substrate recognition may relate to the problem of signal discrimination. Solution of the three-dimensional structure of GSK-3β revealed an interaction between the NH2-terminal regulatory phosphorylation site and the catalytic domain (119, 120). When phosphorylated at Ser9, the NH2-terminal domain folds back and binds to Arg96, which is proximal to the substrate binding site. Biochemical analysis showed that this arginine is required for binding of the kinase to substrates that require a priming phosphorylation but does not impact binding of those that do not, such as β-catenin (121). Therefore, Ser9-phosphorylated GSK-3β would not be capable of binding (and hence phosphorylating) the phosphorylation-primed substrates but would still be capable of targeting those molecules that are independent of Arg96 binding (Fig. 3). The lack of effect of Wnt on GSK-3 substrates other than β-catenin is presumably due to the specificity introduced by Axin. Only the Axin-associated population of GSK-3 is regulatable by Wnts. Hence, GSK-3 activity can be selectively modified with respect to two different classes of targets: Axin-associated/Wnt-regulated targets and phosphorylation-primed/NH2-terminal phosphorylation-regulated targets (Fig. 3).

Fig. 3.

How signals discriminate between regulation of distinct GSK-3 substrates. Phosphorylation of GSK-3β at Ser9 selectively inhibits its phosphorylation of substrates that require prior phosphorylation (primed) compared with those that do not (such as β-catenin).

The search for physiological substrates of the kinase has been substantially aided by the finding that GSK-3 activity is inhibited by lithium (122-126). The effect is reasonably specific in that no other protein kinases are directly affected by this ion, but lithium is known to affect many other processes and enzymes, including inositol monophosphatase (122). As a consequence, attributing effects of lithium to inhibition of GSK-3 requires independent corroboration. That said, use of lithium as an inhibitor has increased the rate of identification of targets, such as Even-skipped (127), microtubule-associated protein-1B (128), and cyclin D (129), and has validated some previously characterized substrates, for example, c-Jun (130-133).

Although varied, several of the targets have similarities (Table 2). For example, phosphorylation of β-catenin by GSK targets it for degradation by the ubiquitin pathway (37). A similar effect is seen upon phosphorylation of cyclin D at Thr286 (129). Both of these proteins are associated with cellular proliferation. Mutation of the β-catenin GSK phosphorylation sites contributes to tumorigenesis (67-69, 134, 135), and expression of a Thr286 → A mutant of cyclin D results in cell transformation and tumor growth in nude mice (136). Several GSK-3 targets are transcription factors. These include members of the c-Jun family (130-132), C/EBPs (CCAAT enhancer binding proteins) (137), STATs (signal transducer and activators of transcription) in Dictyostelium (138), and NF-ATc (nuclear factor of activated T cells) (139-141). The effects of phosphorylation by GSK-3 tend to be inhibitory and include reduced affinity for DNA and promotion of nuclear export. Thus, inhibition of GSK-3 usually results in increased gene expression.

Fig. 4.

Involvement of GSK-3β in NF-κB activation. See text for details. TNFR-1, tumor necrosis factor receptor 1; TRAF2, tumor receptor-associated factor 2; TRADD, TNFR-1-associated signal transducer; RIP, receptor interacting protein.

Table 1.

Translator for the Wnt and Wingless signaling components in flies and mammals.

Table 2.

Substrates of GSK-3. The amino acids in bold are the direct targets of GSK-3, those underlined are the residues targeted by priming kinases (where applicable). Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; X, any amino acid; and Y, Tyr.

A role for GSK-3 in apoptotic signaling and inflammation was recently uncovered by genetic analysis. Generation of mice mutant for GSK-3β revealed a surprising transcriptional target for this kinase. Mice homozygous for the deleted allele die in utero around embryonic day 16. Sectioning of embryos at various developmental stages revealed nothing untoward except for extensive hepatocyte cell death (79). The phenotype is reminiscent of mice harboring defects in the NF-κB pathway, such as RelA knockout mice or IKK2/β mutant mice (142, 143). These mice die of hepatocyte cell death because of an imbalance in the antiapoptotic signaling induced by tumor necrosis factor-α (TNF-α). Like the RelA mutant mice, GSK-3β–/– hepatocytes were protected from apoptosis by injection of the mother with neutralizing antibodies against TNF-α. Mouse embryo fibroblasts derived from the GSK-3β-null animals were defective in TNF-induced activation of NF-κB and were particularly sensitive to cell death induced by TNF, an effect that could be mimicked by treatment with lithium. Although NF-κB activity was substantially reduced in the GSK-3β-null cells, IκB phosphorylation and nuclear translocation of NF-κB were apparently normal, indicating that the defect was independent of the IκB cytoplasmic sequestration pathway (144) (Fig. 4). This suggests that GSK-3β is required for nuclear functioning of NF-κB and implies that small molecule inhibitors of GSK-3 should be potent antiinflammatory agents. In support of this, dibromohymenaldisine was initially isolated on the basis of its inflammation suppressive properties (145-147) and was later shown to inhibit GSK-3 (148).

As mentioned previously, the Wnt pathway in these mice is not discernibly affected in the embryos, the lack of GSK-3β presumably being compensated by the remaining presence of GSK-3α. The insensitivity of the mice and isolated cells to NF-κB activation, however, suggests that this function of GSK-3β, at least, cannot be substituted by GSK-3α. The physical target of GSK-3β in the NF-κB pathway is presently unknown (Fig. 4).

Biological functions

The high degree of conservation of GSK-3 from yeast to mammals has facilitated extensive genetic analysis of this protein kinase. In Saccharomyces cerevisiae, four GSK-3 related genes have been identified: ScGSK-3 (also known as MDS1 and RIM11) (106, 107, 149), MRK1 (150), MCK1 (151, 152), and YOL128C (153). ScGSK-3 and MRK1 have 63% and 65% identity to the mammalian protein in the kinase domain, respectively, whereas MCK1 is substantially less related, at 52% identity. These four protein kinases are not essential for viability (154), but disruption of MCK1 or ScGSK-3 causes defects in the induction of meiotic genes, loss of chromosomes at low temperatures, and defective sporulation. Fission yeast contain at least one GSK-3-like gene implicated in cytokinesis (108). Several plant homologs have been identified, although their functions are relatively poorly understood (155, 156). Disruption of the single known GSK-3 homolog in Dictyostelium results in no effect on the growth within the amoebic life cycle but causes a cell fate switch upon engagement of the sporulation pathway (leading to excessive differentiation into prestalk and stalk cells) (103, 104). As mentioned above, mutations of the enzyme in Drosophila also result in cell fate switching, in this case mimicking ectopic expression of wingless (because Wg acts to inhibit the kinase activity) (27, 28). Overexpression of GSK-3 in Drosophila shortens the circadian periodicity of the fly, an effect traced to premature entry of the Period/Timeless transcriptional complex into the nucleus (157). GSK-3 phosphorylates Timeless and this accelerates nuclear translocation. Presumably, the phosphorylation rate of Timeless is usually constant and is a function of GSK-3 and phosphatase activity. The circadian rhythm effect is insensitive to the Wingless pathway (157). This suggests that regulation of GSK-3 by the PI3K pathway may influence cellular periodicity. Because this pathway also modulates cell size in flies, this may provide a mechanism for coupling these cellular determinants. The studies of GSK-3 relatives in simpler organisms have thus revealed specific roles for the enzyme in differentiation, transcription, circadian rhythms, and development.

Although less is understood about GSK-3 functions in mammals, attention has focused on the potential role of the kinase in brain pathophysiology. In particular, GSK-3 has been implicated in the generation of neurofibrilliary tangles associated with Alzheimer's disease. The tangles are formed from hyperphosphorylated tau (158, 159). GSK-3 is one of several serine kinases that can phosphorylate tau and exhibits specificity for most of the sites that are found phosphorylated in tau located in the fibrilliary tangles (so-called paired helical filament tau) (160-165). Moreover, transgenic overexpression of GSK-3β in brain results in increased tau phosphorylation (166, 167), as well as several hallmarks of Alzheimer's disease, including neurodegeneration, reactive astrocytosis, microgliosis, and the formation of apoptotic bodies. In intact cells, lithium reduces tau phosphorylation (123-125, 168). Several genes have been genetically associated with Alzheimer's disease, including the genes for amyloid precursor protein (the source of β-amyloid found in plaques) and presenilins. Two presenilin genes encode integral membrane proteins (169, 170). Presenilin 1 has been found in association with β-catenin [(171-175); reviewed in (176)]. The effects of the association of presenilins on β-catenin are controversial, with one report suggesting destabilization of β-catenin and induction of neuronal apoptosis (177), but other reports indicating no effect. Mutant presenilins exhibit a "gain-of-aberrant function," altering the trafficking of β-catenin, which may play a role in the etiology of the disease (178). However, wild-type presenilin 1 is not required for accumulation or nuclear transport of β-catenin, nor does it directly affect GSK-3 activity (178). Thus, the presenilins do not appear to be normal components of the Wnt/β-catenin pathway.

The inhibition of GSK-3 by lithium may have health implications, in that 0.1% of the North American population suffers from bipolar disorder. The mainstay of treatment is oral lithium at doses that generate serum concentrations in the range that is inhibitory for GSK-3 (122-126). Another mood-stabilizing drug, valproate, is also used in treating this condition and epilepsy. Valproic acid inhibits GSK-3 in vitro (179), supporting the idea that modulation of GSK-3 activity may have therapeutic benefit in behavioral conditions, in addition to benefits for treatment of Alzheimer's disease.

Recently, a number of potent and selective GSK-3 inhibitors have been developed, including dibromohymenaldisine (147, 148), 3-anilino-4-arylmaleimides (180), and two drugs poetically developed by Glaxo SmithKline (GSK) termed SB-415286 and SB-216763 (181). These latter molecules are neuroprotective in primary neurons induced to die by inhibition of the PI3K pathway (182). These findings suggest that suppression of GSK-3 by PKB plays an important role in survival signaling, consistent with proapoptotic effects caused by transgenic overexpression of GSK-3 (167, 183). Small molecule inhibitors of GSK-3 may therefore have several therapeutic uses, including the treatment of neurodegenerative disease, bipolar disorder, and inflammatory disease. There are some possible clouds to this silver lining. In the heart, GSK-3 appears to suppress cardiac hypertrophy (184, 185). Chronic suppression of this kinase may therefore lead to cardiac complications. Further, the GSK-3 blockers increase cellular β-catenin levels, which may increase the propensity for developing certain cancers. Clearly, the clinical utility of these drugs awaits animal and human trials.


GSK-3 keeps turning up in the most unpredictable places. Since its discovery in 1980 and its cloning in 1990, study of this unassuming enzyme has resulted in a sequence of surprising findings, many more of which are sure to follow. It is a difficult gene product to pigeonhole, because it touches many aspects of biology from development to cancer. As its secrets have been revealed, there are some underlying methods to the madness. By acting as a cellular brake in some ancient organism, perhaps the protein simply accrued similar responsibilities because cells can presumably make do with just one broad-based inhibitor. Along the way, though, it has been coopted by other pathways, further diversifying its functionality. As therapeutics to this fascinating molecule enter clinical testing, it is difficult to predict how safe it will be to suppress its action. GSK-3 has a lot of cellular dependents and it certainly would not be advisable to disturb the hornet's nest.

Perhaps the name is not so bad. A functionally correct name would be too long and clumsy to remember. It also serves as a reminder of the pioneering role of glycogen metabolism in understanding signal transduction. And, if the name is good enough for a multinational pharmaceutical company, perhaps it is okay for an esoteric kinase.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 111.
  112. 112.
  113. 113.
  114. 114.
  115. 115.
  116. 116.
  117. 117.
  118. 118.
  119. 119.
  120. 120.
  121. 121.
  122. 122.
  123. 123.
  124. 124.
  125. 125.
  126. 126.
  127. 127.
  128. 128.
  129. 129.
  130. 130.
  131. 131.
  132. 132.
  133. 133.
  134. 134.
  135. 135.
  136. 136.
  137. 137.
  138. 138.
  139. 139.
  140. 140.
  141. 141.
  142. 142.
  143. 143.
  144. 144.
  145. 145.
  146. 146.
  147. 147.
  148. 148.
  149. 149.
  150. 150.
  151. 151.
  152. 152.
  153. 153.
  154. 154.
  155. 155.
  156. 156.
  157. 157.
  158. 158.
  159. 159.
  160. 160.
  161. 161.
  162. 162.
  163. 163.
  164. 164.
  165. 165.
  166. 166.
  167. 167.
  168. 168.
  169. 169.
  170. 170.
  171. 171.
  172. 172.
  173. 173.
  174. 174.
  175. 175.
  176. 176.
  177. 177.
  178. 178.
  179. 179.
  180. 180.
  181. 181.
  182. 182.
  183. 183.
  184. 184.
  185. 185.
  186. 186.
View Abstract

Navigate This Article