Mental Retardation Syndromes

I. Neurofibromatosis, Coffin-Lowry Syndrome, and the ras/ERK Cascade II. Angelman Syndrome

III. Fragile X Syndromes

A. Fragile X Mental Retardation Syndrome Type 1

B. Fragile X Mental Retardation Type 2

IV. Summary

As we discussed in the first chapter, our great capacity for learning and remembering plays an enormous role in forming our human potential. Moreover, our personal experiences, where we have learned and remembered specific items and events, define us as individuals. These truths are nowhere more evident than when we consider individuals with pronounced learning and memory deficits, present from birth. In this chapter, we consider human mental retardation syndromes and their underlying molecular basis. In some intellectually satisfying instances, we will actually be able to tie mechanisms for these disorders back into fundamental mechanisms for synaptic plasticity and learning that we have already discussed.1

1Parts of this chapter are adapted from Weeber and Sweatt (1).

There is another point that is important to make in the context of this chapter. Of all the various areas of cognitive neurobiology, the field of learning and memory has advanced the farthest into studies at the molecular level, based on a reductionist approach of using model systems simpler than the human. But how can one bridge the enormous distance from specific molecules to human cognition? Over the past few years, a number of groups, including research teams at Baylor College of Medicine, UCLA, Johns Hopkins University, and the University of Illinois to name a few specific examples, have undertaken to bridge this gap by studying naturally occurring human mental retardation syndromes. The philosophy of the approach is to use identified human genetic mutations that result in mental retardation and learning disorders as an entrée to beginning to understand the molecular basis of human cognition. As a practical matter, this translates into taking an identified human gene linked to a mental retardation syndrome and making knockout and transgenic mouse models. These models are then used to generate insights into the underlying molecular and cellular basis for the defect. The rationale is that this gives one insights into the analogous "knockout humans" and hence gives insights into the molecular neurobiology of human cognitive processing. Some specific examples of mental retardation syndromes where this approach has been applied are given in Table 1.

Even though this approach is at a very early stage, it is interesting that different studies of this sort have already begun to converge on two common signal transduc-tion cascades as being involved in human learning and memory: the ras/ERK cascade and its associated upstream regulators and downstream targets (see Figure 1) and the CaMKII system. In the first section of this chapter, I will describe exciting recent findings implicating dysfunction of the ras/ERK cascade in human mental retardation syndromes. I should emphasize that several of the ideas I will present in this section, where I present potential mechanistic links between various different human mental retardation syndromes, are at best educated guesses. However, I simply cannot resist beginning to synthesize a unified picture of critical molecular events in human learning. This extends perfectly the sorts of studies we have been discussing where learning was studied in rodent models.

In the second section of this chapter, I will discuss recent findings from my lab and Alcino Silva's lab implicating CaMKII in a form of human mental retardation. In the final section, I will discuss fragile X retardation syndromes, and we will see yet another instance of the basic studies of synaptic plasticity running head-on into studies of a human learning disorder. In this last section I will highlight work from Bill Greenough's lab, among others' that has begun to tie mechanisms of local dendritic protein synthesis in with molecular mechanisms of fragile X mental retardation type 1.

Before getting down to the serious business of this chapter, I want to relate a personal anecdote that illustrates the funny way that things sometime evolve in science. Alcino Silva and I are of the same scientific generation and, as such, have been competitors in a certain sense—we both set out as naive, optimistic young scientists to "solve memory," of course ideally before anyone else did. While I'm oversimplifying, Alcino basically staked his claim with CaMKII and I grabbed MAP kinases, each of us working on basic mechanisms of synaptic plasticity and memory that were described earlier in this book. Essentially as side projects, Alcino started working on neurofibromatosis mental retardation and my lab started working on Angelman Mental Retardation Syndrome (AS). In the first section of this chapter, I will describe work from Alcino's lab, highlighting the likely importance of MAP kinase in human memory, based on his studies of Neurofibromatosis. In the second section I am going to describe studies from my lab suggesting the importance of CaMKII in human memory, based on our studies of the Angelman Syndrome. In my mind, these are very satisfying examples of how following the clues that Nature gives us will always lead us to common ground.

I. NEUROFIBROMATOSIS, COFFIN-LOWRY SYNDROME, AND THE RAS/ERK CASCADE

Neurofibromatosis is an autosomal-dominant disease that exhibits a variety of clinical features, principally neurofibromas, or benign tumors of neural origin. Other characteristics can include skin discoloration (café au lait spots) and skeletal malformation. The gene that causes neurofibromatosis when mutated in humans is the neurofibromatosis type 1 oncogene, NF1. NF1 is distinct from the gene coding for a

TABLE 1 Mouse Models of Human Mental Retardation Syndromes

Human Mental Retardation Syndromes

Gene Product

Potential Targets

Mouse Model

Learning LTP Defects? Change?

References

Neurofibromatosis

Coffin-Lowry Syndrome

Angelman Syndrome

Fragile X mental retardation 1

Fragile X mental retardation 2

Rett Syndrome

Myotonic dystrophy

Down Syndrome (Trisomy 21)

Williams Syndrome

Neurofibromin 1 (NF1)

Ribosomal S6 kinase2 (rsk2)

Ubiquitin ligase (E6AP)

FMR1 protein (RNA binding proteins)

FMR2 protein (putative transcription factor)

Methyl-CpG binding protein 2 (MeCP2)

Dystrophin protein kinase (DMPK)

DS critical locus

DYRK1A (minibrain kinase homolog)

Superoxide dismutase (SOD)

WS critical locus: LIMK-1 Elastin Syntaxin 1A FKBP6 EIFH4

ras/ ERK, adenylyl cyclase, cytoskeleton CREB, ribosomal S6 protein p53 tumor suppressor protein, others?

Protein synthesis machinery, mRNA targeting, spine structure

Unknown— gene expression

Transcriptional repressors— regulation of unknown genes Na+ channels, Tau many others

Multiple genes including DYRK1A and SOD Unknown

Superoxide dependent processes—redox regulation of PKC, ras, transcription factors

Cytoskeleton, extracellular matrix, spine morphology

(strain dependent)

+ Morris and

Mervis (44)

second type of neurofibroma-related gene, NF2, which causes a different type of neurofibromatosis.

Heterozygous NF1 mutations result in human mental retardation in about 50% of cases. The heterogeneity of mutations in the NF1 gene is likely to contribute to its lack of complete penetrance for various phenotypes including the mental retardation pheno-type. Thus, while neurofibromatosis was initially identified and named for the neurofibroma tumor phenotype, the genetic mutation also causes mental retardation in humans (2, 3).

Adenylyl Cyclase Pathway

FIGURE 1 Signal transduction pathways involved in learning and memory. See text for dis-ussion and additional definitions. R1 = growth factor receptor tyrosine kinases; R2 = phospholipase C coupled receptors; R3 = adenylyl cyclase coupled receptors; R4= ligand-gated calcium channels. MAPs = Microtubulin associated proteins. CBP = CREB binding protein. CRE = cAMP response element. Figure reproduced from Weeber and Sweatt (1).

FIGURE 1 Signal transduction pathways involved in learning and memory. See text for dis-ussion and additional definitions. R1 = growth factor receptor tyrosine kinases; R2 = phospholipase C coupled receptors; R3 = adenylyl cyclase coupled receptors; R4= ligand-gated calcium channels. MAPs = Microtubulin associated proteins. CBP = CREB binding protein. CRE = cAMP response element. Figure reproduced from Weeber and Sweatt (1).

The product of the NF1 gene is neurofibromin, a multidomain molecule that has the capacity to regulate several intracellular processes, including the ERK MAP kinase cascade, adenylyl cyclase, and cytoskeletal assembly. In vivo human neurofibromin is expressed as the product of four different mRNA splice variants, and the type I and type II variants are abundantly expressed in the brain. Alternative splicing of exon 23a in the gene results in the type I and type II variants; type II neurofibromin includes the 23a exon product while type I does not. In a sophisticated study using knockout mouse technology, Alcino Silva and his colleagues based at UCLA identified the 23a exon product as being critical for learning (4). Mice lacking the 23a product exhibited learning problems but no apparent developmental abnormalities nor tumor predisposition, thus implicating the 23a-encoded protein domain as being involved in learning. The 23a-encoded domain of neurofibromin type I protein contributes to regulating the GAP (GTPase activating protein) domain of NF1, a domain that regulates interaction with the NF1 target ras (5, 6). As was described in Chapter 4, ras is a low-molecular-weight G protein coupled to downstream activation of the ERK cascade (see Figures 1 and 2).

In considering that the learning-associated exon of NF1 controlled its GAP activity, Alcino and his group proposed that loss of NF1 regulation of ras, specifically hyper-activation of ras, caused the learning disorder phenotype in Nf1 exon 23a-/-animals. However, because of the complexities of NF1 protein function, and indeed even the complexities of how the 23a exon

Nf1 Protein
FIGURE 2 Regulation of ras activity by GAPs and GEFs. See text for discussions. Figure reproduced from Weeber and Sweatt (1). Adapted from Sweatt (45).

product might itself regulate GAP activity in NF1, the conclusion was inferential. Thus, it was necessary to come up with an independent line of evidence that the NF-1 mutation-associated learning deficits were indeed due to mis-regulation of ras. In impressive follow-up studies published in Nature, Rui Costa in Alcino's lab directly tested their hypothesis that ras hyper-activation caused learning deficiencies in NF1-deficient animals. In this series of studies, they used the classical genetic approach of diminishing ras function through heterozygous ras gene deletion, as well as using a pharmacologic inhibitor of ras activity in vivo to probe for interactions of the NF1 gene product with the ras pathway (7, 39).

The mouse model that they used was an NF1 knockout mouse with a heterozygous deficiency. Costa et al. assessed learning using the Morris water maze paradigm, which as we have already discussed, measures hippocampus-dependent spatial learning. They observed that heterozygous NF1-deficient animals exhibited deficits in the Morris maze task, as assessed using a probe trial and by quantitating quadrant search time. Thus, as expected, heterozygous NF1-deficient mice mimicked aspects of the human learning defects associated with NF1 deficiency. Prior studies by another group (8) had shown directly that loss of NF1 led to aberrant activation of the ras/ERK cascade, specifically that NF1 heterozygous deficiency animals exhibited increased ras/ERK activity in non-neuronal cells. This observation directly suggested that alterations is ras activity occurred in these mice and, thus, could be causative of the learning phenotype, as Costa et al. had hypothesized based on their earlier studies with exon 23a-deficient mice.

To test this idea Costa et al. crossed NF1 heterozygous knockout animals with animals deficient in ras, reasoning appropriately that if hyperactive ras caused the NF1+/- learning phenotype, then genetically diminishing ras activity should rescue the animals to normal learning behavior. Costa et al. identified both N-ras and K-ras heterozygous deficiencies as rescuing the learning defect in NF1+/- mice (see Figure 3). These data strongly implicate ras, specifically N-ras and K-ras, as downstream effectors of NF1 in vivo. Moreover, these findings implicate hyperactivation of this pathway as the cause of learning deficiencies in NF1-deficient mammals, including, of course, the human.

In all studies using animal models constitutively deficient in a gene product, it

Mental Retardation Animals

FIGURE 3 Ras-dependent spatial learning in Nf1+/ animals. These data illustrate that learning deficits of Nf1+/- mice are ras-dependent. Results shown are from the hidden version of the water maze for the Nf1+/-/K-ras+/- population, n = 11 (wild type (WT), n = 24; Nf1+/-, n = 15; K-ras+/-, n = 15). (A) Latency to get to the platform over days. (B) Percent time spent in each quadrant during a probe trial. (C) Average proximity to the exact position where the platform was during training, compared with proximity to the opposite position in the pool. Panels D-F illustrate acquisition, percent time in quadrant, and proximity data for the Nf1+/-/N-ras+/-population (WT, n = 10; Nf1+/-, n = 10; N-ras+/-, n = 7; Nf1+/-/N-ras+/-, n = 9). Panels G-I illustrate data obtained with acute pharmacologic inhibition of ras by use of a farnesyl transferase inhibitor (FTI). Acquisition, percent time in quadrant, and proximity data for the different genotypes and treatments during the FTI rescue experiment (WT + FTI, n = 19; WTsaline, n = 18; Nf1+/-FTI, n = 18; Nf1+/-saline, n = 18). Quadrants are training quadrant (TQ), adjacent right, adjacent left and opposite quadrant (OP). Figure reproduced from Costa et al. (39).

FIGURE 3 Ras-dependent spatial learning in Nf1+/ animals. These data illustrate that learning deficits of Nf1+/- mice are ras-dependent. Results shown are from the hidden version of the water maze for the Nf1+/-/K-ras+/- population, n = 11 (wild type (WT), n = 24; Nf1+/-, n = 15; K-ras+/-, n = 15). (A) Latency to get to the platform over days. (B) Percent time spent in each quadrant during a probe trial. (C) Average proximity to the exact position where the platform was during training, compared with proximity to the opposite position in the pool. Panels D-F illustrate acquisition, percent time in quadrant, and proximity data for the Nf1+/-/N-ras+/-population (WT, n = 10; Nf1+/-, n = 10; N-ras+/-, n = 7; Nf1+/-/N-ras+/-, n = 9). Panels G-I illustrate data obtained with acute pharmacologic inhibition of ras by use of a farnesyl transferase inhibitor (FTI). Acquisition, percent time in quadrant, and proximity data for the different genotypes and treatments during the FTI rescue experiment (WT + FTI, n = 19; WTsaline, n = 18; Nf1+/-FTI, n = 18; Nf1+/-saline, n = 18). Quadrants are training quadrant (TQ), adjacent right, adjacent left and opposite quadrant (OP). Figure reproduced from Costa et al. (39).

is important to distinguish between the acute, ongoing need for the activity of the gene product versus a developmental necessity for the same protein. Alcino and his collaborators addressed this issue by acutely administering a farnesyl transferase inhibitor (which inhibits ras activity by disrupting its membrane association) and demonstrating that transient inhibition of ras in adult animals rescued the learning phenotype in NF1 heterozygous animals. Of course, farnesyl transferases act on other proteins besides ras, so a caveat to this experiment is the possibility of the inhibitor affecting other targets besides ras. However, their interpretation of a need for acute ras-dependent processes for adult learning and memory is also consistent with the wide variety of additional evidence we have already discussed implicating the ras/ERK cascade in learning.

Costa et al. next used their NF1+/-animal models to assess hippocampal long-term potentiation, using theta-burst

Mental Retardation Animals

FIGURE 4 Ras-dependent LTP deficits in Nf1+/- animals. In panels A through D; neurofibromin 1 heterozygous deficiency animals exhibit LTP deficits for a variety of LTP induction protocols. Panel d illustrates that K-ras heterozygous deficiency rescues the Nfl-associated LTP deficit, indicating that it results from an overactivation of ras. For each panel, percentage of baseline field EPSP is plotted over time. (A) Two-burst induction protocol (wild-type (WT), n = 5; Nf1+/-, n = 8). (B) Five-burst induction protocol (WT, n = 7; Nf1+/-, n = 8). (C) Ten-burst induction protocol (WT, n = 7; Nf1+/-, n = 11). (D) Two-burst induction protocol ras rescue experiment, where heterozygous deficiency of K-ras rescues the LTP in Nf1-deficient animals. Thus, LTP deficits in Nf1+/- mice are ras-dependent (WT, n = 13; Nf1+/-, n = 13; Nf1+/-/K-ras+/-, n = 8). Representative traces are shown from left to right for WT, Nf1+/-, and Nf1+/-/K-ras+/-. Horizontal bar, 10 ms; vertical bar, 1 mV. (E) Summary of the amount of LTP measured 40 minutes after induction under the different stimulation protocols. (F) LTP measured 40 minutes after induction in the Nf1+/-/K-ras+/- rescue experiment is comparable to wild-type. Data, figure, and legend reproduced from Costa et al. (39).

FIGURE 4 Ras-dependent LTP deficits in Nf1+/- animals. In panels A through D; neurofibromin 1 heterozygous deficiency animals exhibit LTP deficits for a variety of LTP induction protocols. Panel d illustrates that K-ras heterozygous deficiency rescues the Nfl-associated LTP deficit, indicating that it results from an overactivation of ras. For each panel, percentage of baseline field EPSP is plotted over time. (A) Two-burst induction protocol (wild-type (WT), n = 5; Nf1+/-, n = 8). (B) Five-burst induction protocol (WT, n = 7; Nf1+/-, n = 8). (C) Ten-burst induction protocol (WT, n = 7; Nf1+/-, n = 11). (D) Two-burst induction protocol ras rescue experiment, where heterozygous deficiency of K-ras rescues the LTP in Nf1-deficient animals. Thus, LTP deficits in Nf1+/- mice are ras-dependent (WT, n = 13; Nf1+/-, n = 13; Nf1+/-/K-ras+/-, n = 8). Representative traces are shown from left to right for WT, Nf1+/-, and Nf1+/-/K-ras+/-. Horizontal bar, 10 ms; vertical bar, 1 mV. (E) Summary of the amount of LTP measured 40 minutes after induction under the different stimulation protocols. (F) LTP measured 40 minutes after induction in the Nf1+/-/K-ras+/- rescue experiment is comparable to wild-type. Data, figure, and legend reproduced from Costa et al. (39).

stimulation. They found deficits in theta-burst LTP in NF1-deficient animals which, like the learning defects, were rescued by heterozygous deletion of ras (see Figure 4). One twist to the story is that Costa et al. found that alterations in GABAergic function are likely involved in the LTP phenotype, specifically GABAergic feed-forward inhibitory neurons in area CA1 that regulate cellular excitability and the likelihood of LTP induction under some conditions such as theta-burst stimulation. They observed that blocking this system rescued the effects of NF1 deficiency on

LTP, suggesting that NF-1 functioned to control this GABAergic system and, through this mechanism control the likelihood of LTP induction. They also directly observed enhanced GABAergic function in the NF1+/- animals in whole-cell recording of GABA inputs onto CA1 pyramidal neurons. Thus, the critical locus of NF1 function may be GABAergic interneurons. However, in additional, more recent studies, Alcino's group has also observed derangement of LTP in the pyramidal neurons of area CA1, using other types of LTP induction protocols that don't involve GABAergic interneurons. Thus, there is also an effect of NF1/ras in pyramidal neuron dendrites.

Having found that diminishing ras function led to a rescue of the phenotype, we are compelled to ask: what is the target of ras that leads to these derangements? As described in Chapter 4, ras regulates the MEK/ERK MAPK cascade in hippocampal neurons, and Ingram et al. (8) had previously reported elevated ERK activity in NF1-deficient animals. Thus, the most likely target is the raf/MEK/ERK signal transduction cascade. Moreover, other work by Alcino's research team has also nicely demonstrated an interaction of ras with the ERK cascade, using an approach they termed pharmacogenetics. They found that mice heterozygous for a null mutation of the K-ras gene (K-ras+/-) showed normal hippocampal ERK activation, LTP, and contextual conditioning in a conditioned place preference task. However, treatment with a low dose of MEK inhibitor, ineffective in wild-type controls, blocked MAPK activation, LTP, and contextual learning in K-ras+/- mutants. These data strongly indicated that K-Ras is upstream of MEK/ ERK signaling in hippocampus, and that acute activation of this pathway is involved in synaptic plasticity and memory.

Aside from direct ras/ERK cascade regulation another potential function of NF1 is regulating adenylyl cyclase; Tong et al. demonstrated that NF1 also regulates adenylyl cyclase in mammalian neurons (see reference 9 and Figure 1). They showed that neuropeptide and G-protein-stimulated adenylyl cyclase activity were reduced in mice completely deficient in NF1 activity. Even though the effects on adenylyl cyclase seen by Tong et al. were selective for animals with homozygous deletions of NF1, it is certainly worthwhile to consider that attenuation of subtle forms of regulation of adenylyl cyclase might play a role in learning deficits in heterozygous NF1-deficient animals as well. Of course, adenylyl cyclase also is upstream of the ERK MAP kinase cascade in hippocampal neurons (see Figure 1). Thus, NF1 may couple to erk via two pathways in the hippocampus, and both these effects may contribute to dysregulation of ERK in NF1-deficient mice and humans.

One of the targets of ERK in the hippocampus is CREB, and, as we have discussed, this molecule has been widely implicated in learning and memory in many species. ERK couples to CREB through the intervening kinase ribosomal S6 kinase (RSK2), which phosphorylates CREB at ser133 just like PKA. As I described in Chapter 7, in the mammalian hippocampus, it has been found that PKA cannot elicit CREB phosphorylation without going through ERK, thus ERK/RSK2 is an obligatory step in PKA regulation of CREB-mediated gene expression in mammalian hippocampal neurons (10, 11). The important implication of this in the present context is that rsk2 is the gene disrupted in human Coffin-Lowry Mental Retardation Syndrome (12, 13). Thus, the same pathway implicated in studies of Neurofibromatosis Mental Retardation, ERK/RSK2/CREB, has also been implicated as being involved in human learning and memory in an independent line of studies (see also Box 1).

Overall, while our current thinking is likely oversimplified, it is interesting that two different human mental retardation syndromes impinge upon the same signal transduction cascade, the ras/ERK/CREB cascade, which is coupled to regulation of gene expression in neurons. As we discussed earlier, it is noteworthy that this

Adult Dyslexia

Adult Dyslexia

This is a comprehensive guide covering the basics of dyslexia to a wide range of diagnostic procedures and tips to help you manage with your symptoms. These tips and tricks have been used on people with dyslexia of every varying degree and with great success. People just like yourself that suffer with adult dyslexia now feel more comfortable and relaxed in social and work situations.

Get My Free Ebook


Post a comment