Schizophrenia Research Forum - A Catalyst for Creative Thinking

PAK Inhibitors Restore Spines in DISC1 Mouse Model

April 18, 2014. Experimental cancer drugs called PAK inhibitors can promote dendritic spines in a genetic mouse model of psychiatric disorders, reports a study published April 3 in the Proceedings of the National Academy of Sciences. Led by Akira Sawa at Johns Hopkins University, Baltimore, Maryland, the study focuses on fixing the loss of dendritic spines, the recipients of excitatory inputs from other neurons—a deficit reported in schizophrenia and also found in mice lacking normal amounts of protein encoded by disrupted in schizophrenia 1 (DISC1). The researchers developed compounds that inhibit p21-activated kinases (PAKs), and these protected or restored spines in the mice, as well as rescued deficits in sensory gating similar to those seen in schizophrenia.

The study offers up a new candidate drug development strategy for schizophrenia. PAKs catalyze changes to cell structure and morphology, including the transformations that turn a normal cell into a cancerous one. Inhibiting PAKs might be a way to stave off the loss of dendritic spines (and their resident synapses) reported in postmortem schizophrenia brain (Glantz and Lewis, 2000). This idea has recently gotten a boost from a genetic study that implicated copy number variants containing one member of the PAK family, PAK7, in schizophrenia (see SRF related news report).

But the new study came to PAKs through a different line of inquiry—DISC1, the gene disrupted in a Scottish family enriched for mental illness. In 2010, Sawa and colleagues reported that rat neurons low in DISC1 did not respond normally to the spine-building electrical activity through the N-methyl-D-aspartate (NMDA)-type of glutamate channels: Instead of showing a spurt of spine growth, the dendrites had only sparsely scattered spines (see SRF related news report). The researchers determined that this was because kalirin-7 (Kal-7), normally tethered to DISC1, was freed in DISC1 knockdown neurons to spur spine loss via Kal-7’s interaction with Rac1, a GTPase that in turn activates PAKs. The researchers reasoned that blocking the action of PAKs could block the program of excessive pruning.

Protect and rescue
To watch the dynamic changes unfold, first author Akiko Hayashi-Takagi and colleagues confirmed their previous finding with time-lapse photography. Upon NMDA receptor activation of in vitro rat cortical neurons, spines swelled in control neurons, but shrank away in neurons transfected with DISC1 interfering RNA (RNAi).

Treatment with one of three PAK inhibitors developed by the team, however, staved off the shrinkage. When the PAK inhibitors were added to the dish an hour prior to NMDA receptor activation, no significant differences emerged in spine size between control neurons and neurons with DISC1 knockdown 20 minutes later. Two of three PAK inhibitors also prevented the loss in spine density in the DISC1 knockdown neurons.

The PAK inhibitors could also protect neurons from the ill effects of a prolonged DISC1 knockdown. Though DISC1 loss initially leads to more spines, this trend reverses with time, leading to sparser and smaller spines. Adding the PAK inhibitors at the same time as the RNAi for DISC1 resulted in spine density and size no different from neurons with control RNAi seven days later. This effect was also dose dependent and did not seem to alter normal spines in controls.

The researchers also found some evidence that the PAK inhibitors could resuscitate shrunken spines: Adding PAK inhibitors after five days of DISC1 knockdown—when the spines are presumably withered—brought about an increase in spine size, but not density, three days later.

In vivo
The researchers then tested their PAK inhibitors in living mice. First, they orchestrated a DISC1 knockdown in cortical neurons of the prefrontal cortex by injecting DISC1 RNAi in utero. At postnatal day 35—mouse "adolescence"—the researchers imaged dendritic spines through a small hole in the skull by using two-photon microscopy to get a baseline. Mice with DISC1 knockdown showed about half the spine density that controls did.

However, daily injections of one of the PAK inhibitors, called FRAX486, for 25 days erased this difference. In a behavioral test of prepulse inhibition—in which the startle response to a sound is attenuated by a preceding sound—the DISC1 knockdown mice showed a strong startle—a deficit also observed in schizophrenia. But the DISC1 knockdown mice treated with FRAX486 showed less of a deficit, though not at control levels.

That PAK inhibitors could sculpt neurons and behavior alike is encouraging and provides some hope to a field running short on therapeutic ideas. But whether PAK inhibitors are viable treatments for people with schizophrenia awaits much additional research.—Michele Solis.

Hayashi-Takagi A, Araki Y, Nakamura M, Vollrath B, Duron SG, Yan Z, Kasai H, Huganir RL, Campbell DA, Sawa A. PAKs inhibitors ameliorate schizophrenia-associated dendritic spine deterioration in vitro and in vivo during late adolescence. Proc Natl Acad Sci U S A. 2014 Apr 3. Abstract

Comments on News and Primary Papers

Primary Papers: PAKs inhibitors ameliorate schizophrenia-associated dendritic spine deterioration in vitro and in vivo during late adolescence.

Comment by:  Albert H. C. Wong
Submitted 16 April 2014
Posted 16 April 2014

These are very interesting and novel results using knowledge of DISC1 and dendritic spine biology to test new drugs targeting a pathway implicated in schizophrenia. It is encouraging to see successful, mechanism-based approaches to discover new therapeutic interventions for schizophrenia. Although there is a long road to clinical translation, efforts such as these offer hope that drugs with new targets will emerge. The preventative potential of the PAK inhibitors is also very important, since all current treatments are symptomatic and do not change the course of illness.

View all comments by Albert H. C. Wong

Primary Papers: PAKs inhibitors ameliorate schizophrenia-associated dendritic spine deterioration in vitro and in vivo during late adolescence.

Comment by:  Amy Ramsey
Submitted 23 April 2014
Posted 23 April 2014

New Research Identifies PAK1 as a Promising Target for Schizophrenia Treatment
Schizophrenia has been described as a disease of the synapse, in part because postmortem brain studies have uncovered a loss of dendritic spines, the physical structures of glutamate synapses (Glantz and Lewis, 2000; Sweet et al., 2009; Garey et al., 1998). Interestingly, the same deficits in dendritic spines are seen in a number of pharmacological and genetic animal models of the disorder (Lee et al., 2011; Ramsey et al., 2011; Elsworth et al., 2011; Chen et al., 2008). Despite these intriguing observations, it has been difficult to move beyond correlations to show that spine deficits cause schizophrenia symptoms, and that reversing these deficits will reverse symptoms.

Even though altered spine density is observed in other psychiatric diseases besides schizophrenia, it could be a useful endophenotype with predictive validity for novel schizophrenia therapies. There is now a large body of literature demonstrating that changes in spine density are correlated with changes in neuron physiology and behavior (Moser et al., 1994; Leuner et al., 2003; Yuste and Bonhoeffer, 2001; Sanders et al., 2012). From this literature it is clear that RhoGTPase signaling is a common final pathway to modulate spine dynamics by rearranging the actin cytoskeleton (Oh et al., 2010; Xie et al., 2007; Zhou et al., 2009; Zhang et al., 2005; van Galen and Ramakers, 2005; Tolias et al., 2005; Soderling et al., 2007; Asrar et al., 2009; Cahill et al., 2009; and see figure below). What has been lacking until recently is a pharmacological tool to directly manipulate spine number by targeting RhoGTPase signaling. Although there are a number of psychoactive drugs that modulate spine density, their mechanism of action generally affects neurotransmitter receptor function and has only indirect effects on the molecular machinery that controls spine architecture.

For these reasons, a recent study by Hayashi-Takagi et al. (Hayashi-Takagi et al., 2014) represents a significant breakthrough for both basic neuroscience research and future schizophrenia pharmacotherapy. Their study, published in the Proceedings of the National Academy of Sciences USA, demonstrates that deficits in spine density and behavior are improved by drugs that directly target the RhoGTPase pathway. Three PAK1 inhibitors were studied for their ability to either prevent spine loss or reverse spine loss caused by knockdown of DISC1 in primary neurons. The compounds used in the study, FRAX120, FRAX355, and FRAX486, were identified by Afraxis in a screen for drugs that selectively inhibit PAK1 with limited effects on related kinases PAK2-4. One of the compounds was previously shown to improve phenotypes in a mouse model of Fragile-X syndrome (Dolan et al., 2013). However, the current study describes additional PAK1 inhibitors and is the first to test their potential for schizophrenia treatment.

The groundwork for this study was an earlier paper by the authors, which showed that DISC1 is a negative regulator of Rac1 signaling, and that DISC1 deficiencies lead to an overactivation of Rac1 and its downstream effector PAK1 (Hayashi-Takagi et al., 2010). They also showed that constitutive activation of Rac1 led to spine shrinkage and spine loss, providing a molecular mechanism to explain why DISC1 deficiency causes spine deficits.

The current study provides further support of this mechanism linking DISC1 to Rac1 signaling, since PAK1 inhibition prevented spine loss caused by DISC1 knockdown. The PAK1 inhibitors were most effective to prevent spine loss, but also had some ability to reverse pre-existing spine loss when DISC1 knockdown occurred before PAK1 inhibition. This is an important observation to consider if these drugs proceed to clinical studies, since they may be most efficacious with early intervention. The study also showed that the prevention of spine loss was accompanied by behavioral improvements in mice with a knockdown of DISC1.

This work focused on the molecular pathway linking DISC1 to PAK1, but the Rac1-PAK1 pathway is largely conserved in all cell types as the mechanism to orchestrate cytoskeletal rearrangements (Hall, 1998; Etienne-Manneville and Hall, 2002). Although it would seem that inhibiting this pathway would be toxic, it may not, in fact, be problematic since PAK1 inhibitors are showing promise as anti-cancer medicines (Yi et al., 2010).

A more pressing question is whether PAK1 inhibitors could be effective for other forms of schizophrenia that do not involve DISC1 dysfunction. The authors have convincingly demonstrated the link between DISC1 and the Rac1 pathway, but these studies haven’t been done for the many other susceptibility genes that are implicated in schizophrenia. There is room for optimism, however, since PAK1 is part of the final common pathway for cytoskeletal rearrangements. Perhaps in the several animal models with spine deficits, we will see the same improvements at the cellular and behavioral level by using PAK1 inhibitors.

The shape and number of dendritic spines is regulated by the RhoGTPase signaling cascades, which work in concert to orchestrate the polymerization and de-polymerization of actin into branched filaments. In neurons, there are three RhoGTPases that have been studied most extensively: RhoA, Rac1, and Cdc42 (colored green). RhoGTPase activity is modulated by the activation of membrane receptors such as receptor tyrosine kinases, G protein-coupled receptors, or ionotropic receptors such as the NMDA receptor pictured above. Guanine nucleotide exchange factors (GEFs) such as Tiam1 and Kalirin promote the activation of Rac1, while DISC1 has been shown to negatively regulate Rac1 by sequestering Kalirin. When DISC1 is reduced, Kalirin is no longer sequestered, and Rac1 is activated. Once activated, Rac1 promotes the activity of its downstream effector PAK1, a kinase that in turn phosphorylates LIMK1. While transient Rac1 activation leads to increased spine size through actin rearrangements, prolonged Rac1 activation has the opposite effect of causing spine retraction and elimination. The recent paper by Hayashi-Takagi et al. demonstrates that PAK1 inhibitors can prevent or reverse spine deficits caused by DISC1 knockdown. Figure adapted from: van Galen and Ramakers (2005) Prog Brain Res 147: 295-317.


Glantz LA, Lewis DA. Decreased dendritic spine density on prefrontal cortical pyramidal neurons in schizophrenia. Arch Gen Psychiatry . 2000 Jan ; 57(1):65-73. Abstract

Sweet RA, Henteleff RA, Zhang W, Sampson AR, Lewis DA. Reduced dendritic spine density in auditory cortex of subjects with schizophrenia. Neuropsychopharmacology . 2009 Jan ; 34(2):374-89. Abstract

Garey LJ, Ong WY, Patel TS, Kanani M, Davis A, Mortimer AM, Barnes TR, Hirsch SR. Reduced dendritic spine density on cerebral cortical pyramidal neurons in schizophrenia. J Neurol Neurosurg Psychiatry . 1998 Oct ; 65(4):446-53. Abstract

Lee FH, Fadel MP, Preston-Maher K, Cordes SP, Clapcote SJ, Price DJ, Roder JC, Wong AH. Disc1 point mutations in mice affect development of the cerebral cortex. J Neurosci . 2011 Mar 2 ; 31(9):3197-206. Abstract

Ramsey AJ, Milenkovic M, Oliveira AF, Escobedo-Lozoya Y, Seshadri S, Salahpour A, Sawa A, Yasuda R, Caron MG. Impaired NMDA receptor transmission alters striatal synapses and DISC1 protein in an age-dependent manner. Proc Natl Acad Sci U S A . 2011 Apr 5 ; 108(14):5795-800. Abstract

Elsworth JD, Morrow BA, Hajszan T, Leranth C, Roth RH. Phencyclidine-induced loss of asymmetric spine synapses in rodent prefrontal cortex is reversed by acute and chronic treatment with olanzapine. Neuropsychopharmacology . 2011 Sep ; 36(10):2054-61. Abstract

Chen YJ, Johnson MA, Lieberman MD, Goodchild RE, Schobel S, Lewandowski N, Rosoklija G, Liu RC, Gingrich JA, Small S, Moore H, Dwork AJ, Talmage DA, Role LW. Type III neuregulin-1 is required for normal sensorimotor gating, memory-related behaviors, and corticostriatal circuit components. J Neurosci . 2008 Jul 2 ; 28(27):6872-83. Abstract

Moser MB, Trommald M, Andersen P. An increase in dendritic spine density on hippocampal CA1 pyramidal cells following spatial learning in adult rats suggests the formation of new synapses. Proc Natl Acad Sci U S A . 1994 Dec 20 ; 91(26):12673-5. Abstract

Leuner B, Falduto J, Shors TJ. Associative memory formation increases the observation of dendritic spines in the hippocampus. J Neurosci . 2003 Jan 15 ; 23(2):659-65. Abstract

Yuste R, Bonhoeffer T. Morphological changes in dendritic spines associated with long-term synaptic plasticity. Annu Rev Neurosci . 2001 ; 24():1071-89. Abstract

Sanders J, Cowansage K, Baumgärtel K, Mayford M. Elimination of dendritic spines with long-term memory is specific to active circuits. J Neurosci . 2012 Sep 5 ; 32(36):12570-8. Abstract

Oh D, Han S, Seo J, Lee JR, Choi J, Groffen J, Kim K, Cho YS, Choi HS, Shin H, Woo J, Won H, Park SK, Kim SY, Jo J, Whitcomb DJ, Cho K, Kim H, Bae YC, Heisterkamp N, Choi SY, Kim E. Regulation of synaptic Rac1 activity, long-term potentiation maintenance, and learning and memory by BCR and ABR Rac GTPase-activating proteins. J Neurosci . 2010 Oct 20 ; 30(42):14134-44. Abstract

Xie Z, Srivastava DP, Photowala H, Kai L, Cahill ME, Woolfrey KM, Shum CY, Surmeier DJ, Penzes P. Kalirin-7 controls activity-dependent structural and functional plasticity of dendritic spines. Neuron . 2007 Nov 21 ; 56(4):640-56. Abstract

Zhou Z, Meng Y, Asrar S, Todorovski Z, Jia Z. A critical role of Rho-kinase ROCK2 in the regulation of spine and synaptic function. Neuropharmacology . 2009 Jan ; 56(1):81-9. Abstract

Zhang H, Webb DJ, Asmussen H, Niu S, Horwitz AF. A GIT1/PIX/Rac/PAK signaling module regulates spine morphogenesis and synapse formation through MLC. J Neurosci . 2005 Mar 30 ; 25(13):3379-88. Abstract

van Galen EJ, Ramakers GJ. Rho proteins, mental retardation and the neurobiological basis of intelligence. Prog Brain Res . 2005 ; 147():295-317. Abstract

Tolias KF, Bikoff JB, Burette A, Paradis S, Harrar D, Tavazoie S, Weinberg RJ, Greenberg ME. The Rac1-GEF Tiam1 couples the NMDA receptor to the activity-dependent development of dendritic arbors and spines. Neuron . 2005 Feb 17 ; 45(4):525-38. Abstract

Soderling SH, Guire ES, Kaech S, White J, Zhang F, Schutz K, Langeberg LK, Banker G, Raber J, Scott JD. A WAVE-1 and WRP signaling complex regulates spine density, synaptic plasticity, and memory. J Neurosci . 2007 Jan 10 ; 27(2):355-65. Abstract

Asrar S, Meng Y, Zhou Z, Todorovski Z, Huang WW, Jia Z. Regulation of hippocampal long-term potentiation by p21-activated protein kinase 1 (PAK1). Neuropharmacology . 2009 Jan ; 56(1):73-80. Abstract

Cahill ME, Xie Z, Day M, Photowala H, Barbolina MV, Miller CA, Weiss C, Radulovic J, Sweatt JD, Disterhoft JF, Surmeier DJ, Penzes P. Kalirin regulates cortical spine morphogenesis and disease-related behavioral phenotypes. Proc Natl Acad Sci U S A . 2009 Aug 4 ; 106(31):13058-63. Abstract

Hayashi-Takagi A, Araki Y, Nakamura M, Vollrath B, Duron SG, Yan Z, Kasai H, Huganir RL, Campbell DA, Sawa A. PAKs inhibitors ameliorate schizophrenia-associated dendritic spine deterioration in vitro and in vivo during late adolescence. Proc Natl Acad Sci U S A . 2014 Apr 3. Abstract

Dolan BM, Duron SG, Campbell DA, Vollrath B, Shankaranarayana Rao BS, Ko HY, Lin GG, Govindarajan A, Choi SY, Tonegawa S. Rescue of fragile X syndrome phenotypes in Fmr1 KO mice by the small-molecule PAK inhibitor FRAX486. Proc Natl Acad Sci U S A . 2013 Apr 2 ; 110(14):5671-6. Abstract

Hayashi-Takagi A, Takaki M, Graziane N, Seshadri S, Murdoch H, Dunlop AJ, Makino Y, Seshadri AJ, Ishizuka K, Srivastava DP, Xie Z, Baraban JM, Houslay MD, Tomoda T, Brandon NJ, Kamiya A, Yan Z, Penzes P, Sawa A. Disrupted-in-Schizophrenia 1 (DISC1) regulates spines of the glutamate synapse via Rac1. Nat Neurosci . 2010 Mar ; 13(3):327-32. Abstract

Hall A. Rho GTPases and the actin cytoskeleton. Science . 1998 Jan 23 ; 279(5350):509-14. Abstract

Etienne-Manneville S, Hall A. Rho GTPases in cell biology. Nature . 2002 Dec 12 ; 420(6916):629-35. Abstract

Yi C, Maksimoska J, Marmorstein R, Kissil JL. Development of small-molecule inhibitors of the group I p21-activated kinases, emerging therapeutic targets in cancer. Biochem Pharmacol . 2010 Sep 1 ; 80(5):683-9. Abstract

View all comments by Amy Ramsey

Comments on Related News

Related News: DISC1 and SNAP23 Emerge In NMDA Receptor Signaling

Comment by:  Jacqueline Rose
Submitted 2 March 2010
Posted 2 March 2010
  I recommend the Primary Papers

The newly published paper by Katherine Roche and Paul Roche reports SNAP-23 expression in neuron dendrites and examines the possible role of this neuronal SNAP-23 protein. To this point, SNAP-23 has traditionally been discussed in reference to vesicle trafficking in epithelial cells (see Rodriguez-Boulan et al., 2005 for review), so it is of interest to determine the function of SNAP-23 in neurons. Suh et al. report that surface NMDA receptor expression and NMDA-mediated currents are inhibited following SNAP-23 knockdown. Further, SNAP-23 knockdown results in a specific decrease in NR2B subunit insertion; previously, the NR2B subunit has been reported to preferentially localize to recycling endosomes compared to NR2A (Lavezzari et al., 2004). Given these findings, it is reasonable to conclude that SNAP-23 may be involved in maintaining NMDA receptor surface expression possibly by binding to NMDA-specific recycling endosomes.

Interestingly, there is recent evidence that PKC-induced NMDA receptor insertion is mediated by another neuronal SNARE protein; postsynaptic SNAP-25 (Lau et al., 2010). It is possible that activity-induced NMDA receptor trafficking is mediated by SNAP-25, while baseline maintenance of NMDA receptor levels relies on SNAP-23. Other evidence to suggest a strictly regulatory role for SNAP-23 in neuronal NMDA insertion is the finding that activity-dependent receptor insertion from early endosomes has previously been reported to be restricted to AMPA-type glutamate receptors (Park et al., 2004). However, it is possible that activity-induced insertion of AMPA receptors occurs via a distinct endosome pool than NMDA receptors; AMPA and NMDA receptor trafficking has been reported to proceed by distinct vesicle trafficking pathways (Jeyifous et al., 2009).

Although SNAP-23 may not be involved in activity-dependent early endosome receptor trafficking, it is possible that SNAP-23 operates in other pathways linked to activity-induced NMDA receptor trafficking. For instance, SNAP-23 may be the SNARE protein by which lipid raft shuttling of NMDA receptors occurs. SNAP-23 has been found to preferentially associate with lipid rafts over SNAP-25 in PC12 cells (Salaün et al., 2005). As well, NMDA receptors have been found to associate with lipid raft associated proteins flotilin-1 and -2 in neurons (Swanwick et al., 2009). Lipid raft trafficking of NMDA receptors to post-synaptic densities has been reported to follow global ischemia (Besshoh et al., 2005), and the possibility remains that under certain circumstances, NMDA trafficking occurs by lipid raft association to SNAP-23.

Taken together, the discovery of post-synaptic SNARE proteins offers several avenues of research to determine their roles and functions in glutamatergic synapse organization. Further, investigating disruption of synaptic receptor organization presents several possibilities for potential etiologies of disorders linked to compromised glutamate signaling like schizophrenia.


Besshoh, S., Bawa, D., Teves, L., Wallace, M.C. and Gurd, J.W. (2005). Increased phosphorylation and redistribution of NMDA receptors between synaptic lipid rafts and post-synaptic densities following transient global ischemia in the rat brain. Journal of Neurochemistry, 93: 186-194. Abstract

Jeyifous, O., Waites, C.L., Specht, C.G., Fujisawa, S., Schubert, M., Lin, E.I., Marshall, J., Aoki, C., de Silva, T., Montgomery, J.M., Garner, C.C. and Green, W.N. (2009). SAP97 and CASK mediate sorting of NMDA receptors through a previously unknown secretory pathway. Nature Neuroscience, 12: 1011-1019. Abstract

Lau, C.G., Takayasu, Y., Rodenas-Ruano, A., Paternain, A.V., Lerma, J., Bennet, M.V.L. and Zukin, R.S. (2010). SNAP-25 is a target of protein kinase C phosphorylation critical to NMDA receptor trafficking. Journal of Neuroscience, 30: 242-254. Abstract

Lavezzari, G., McCallum, J., Dewey, C.M. and Roche, K.W. (2004). Subunit-specific regulation of NMDA receptor endocytosis. Journal of Neuroscience, 24: 6383-6391. Abstract

Park, M., Penick, E.C., Edward, J.G., Kauer, J.A. and Ehlers, M.D. (2004). Recycling endosomes supply AMPA receptors for LTP. Science, 305: 1972-1975. Abstract

Rodriguez-Boulan, E., Kreitzer, G. and Müsch, A. (2005) Organization of vesicular trafficking in epithelia. Nature Reviews: Molecular Cell Biology, 6: 233-247. Abstract

Salaün, C., Gould, G.W. and Chamberlain, L.H. (2005). The SNARE proteins SNAP-25 and SNAP-23 display different affinities for lipid rafts in PC12 cells. Journal of Biological Chemistry, 280: 1236-1240. Abstract

Suh, Y.H., Terashima, A., Petralia, R.S., Wenthold, R.J., Isaac, J.T.R., Roche, K.W. and Roche, P.A. (2010). A neuronal role for SNAP-23 in postsynaptic glutamate receptor trafficking. Nat Neurosci. 2010 Mar;13(3):338-43. Abstract

Swanwick, C.C., Shapiro, M.E., Chang, Y.Z. and Wenthold, R.J. (2009). NMDA receptors interact with flotillin-1 and -2, lipid raft-associated proteins. FEBS Letters, 583: 1226-1230. Abstract

View all comments by Jacqueline Rose