Reported by Mervyn Monteiro

2000 August 2. Presenilins and Lewy body pathology. Lippa et al (173) used α-synuclein immunoreactivity to determine the frequency of Lewy body (LB) occurrence in familial Alzheimer's disease (FAD) cases linked to mutations in PS1. Interestingly, they found that approximately half of 14 afflicted brains contained LBs. Moreover, some of the LB-contain brains were from deceased patients whose clinical records were consistent with their having had Parkinsonism-like symptoms. In an earlier report, this group found LBs in approximately 60 percent of brains from patients with FAD-linked mutations in PS and APP genes. Fishel et al. (294) also reported that they had found LB by α-synuclein staining in six of 11 brains from individuals having FAD-linked mutations in PS2. Taken together, these results suggest that a significant fraction of FAD-linked mutations in PS1 and PS2 lead to LB pathology. Insight into what causes these inclusions may provide insight into the etiology of the disease.

Presenilins and the cell cycle

Mice disrupted in the murine presenilin 1 (PS1) gene die on day one after birth. However, these mutant mice survive to adulthood when the human PS1 transgene is expressed under the control of the Thy-1 promoter. Zheng et al. (599) reported, however, that these rescued mice develop extensive epidermal hyperplasia and neoplasia. The abnormalities appear to arise due to the restricted expression of the Thy-1 driven PS1 transgene to neurons and its failure, unlike the normal murine PS1 gene, to be expressed in skin. Zheng et al. provided evidence for the lack of expression of the human PS1 transgene in keratinocytes cultured from the rescued mice. In these cultures, β-catenin protein had a longer half-life and cyclin D1 expression was increased, compared to keratinocytes expressing endogenous murine PS1. Edward Koo (16) also provided supporting evidence that loss of PS1 expression leads to increased cyclin D1 expression. Since increased β-catenin and cyclin D1 are both associated with induction of cellular proliferation (both proteins are involved in the G1 to S phase transition of the cell cycle), these data suggest that presenilins may have important roles in the regulation of the cell cycle and in growth control.

The unfolded protein response

The unfolded protein response (UPR) is up-regulated during cellular stress and when misfolded proteins accumulate in the endoplasmic reticulum (ER). A number of presentations, including those by Katayama et al. (1176), De Strooper et al. (342), and several others at the meeting, suggested that the UPR is compromised in cells disrupted in the PS1 gene, and in cells expressing mutant FAD PS1 genes. Furthermore, Katayama and colleagues reported that the chaperone protein BiP was decreased in AD brains (Nature Cell Biol;1(8):479-85). De Strooper also reported at the meeting that they too had observed down-regulation of chaperone proteins, BiP and GRP94, in cell cultures prepared from PS1 and PS2 knockout mice.
Contrary to the these findings, Sato and colleagues (1186) reported that the UPR response was unaltered in PS1-deficient cells and in cells expressing FAD-linked PS1 mutations. In fact, Sato et al. reported that PS1-deficient cells, as well as neuroblastoma cell lines stably expressing several different FAD PS1 mutations, are able to efficiently induce transcriptional up-regulation of genes involved in the UPR. Thinakaran induced cellular stress with three different agents: tunicamycin, thapsigargin, and a Ca2+ ionophore. Each of these treatments caused induction of BiP (GRP78), GRP94, and CHOP mRNA to equivalent levels and with a similar profile in wild-type PS1 expressing cells, in PS1-/- cells, and in cells expressing FAD mutant PS1 proteins. Moreover, Thinakaran et al. did not find any noticeable difference in BiP protein levels in protein lysates of transgenic mice carrying FAD PS1 mutant genes or in human brains from FAD individuals.

The calcium connection

At least three different calcium-binding proteins are known to interact with presenilins: calsenilin, calmyrin and sorcin. Wasco (19) described studies of calsenilin, which has highest homology to the recoverin family of proteins. Calsenilin has two structural domains: a long N-terminal domain and a C-terminal domain containing four calcium-binding EF hand motifs and. Calsenilin binds to both full-length presenilin and C-terminal fragments of presenilins. Calsenilin protein is soluble when overexpressed in cells. However, when it is coexpressed with PS2, calsenilin fractionates predominantly with the membrane-bound presenilins. Membrane binding of calsenilin was not affected by changes in calcium levels, as treatment with EGTA did not cause any noticeable release of calsenilin from the membrane-bound pool. Calsenilin is cleaved during apoptosis. This cleavage is sensitive to zVAD treatment, indicating it is likely due to a caspase(s) action during apoptosis. Purified caspase 3 was found to cleave calsenilin at a site between the N-terminal and EF hand. The N-terminal domain of calsenilin was found to mediate binding to presenilin.

Lilliehook et al. (1172) reported on the physiology associated with overexpression of calsenilin in cells. In one approach, they used microarray analysis to examine changes in mRNAs profiles expressed in calsenilin-induced and non-induced cells. The identity of a number of mRNAs that were down regulated upon calsenilin expression was described. Two of these were phosphoinoside-3-kinase class 3 and inositol 1,4,5-trphosphate receptor (IP3R), type I. In immunoblot experiments they confirmed that IP3R was reduced in cells overexpressing calsenilin. They next examined how calsenilin overexpression affects calcium responses in H4 neuroglioma cells. Calsenilin expression led to reduce amplitude of calcium release induced by carbachol (agonist) treatment compared with non-expressing cells. Calsenilin expression also caused an elevation in amplitude of calcium release induced by thapsigargin treatment, suggesting that calsenilin increases the calcium content of the ER. Moreover, calsenilin overexpression enhanced apoptosis induced by both serum withdrawal and thapsigargin treatment. Taken together, the data suggest that calsenilin decreases IP3-mediated calcium signaling and increases calcium content of ER stores.
A common feature associated with FAD mutations in both presenilin and APP genes is the dysregulation of calcium signaling.

Consistent with this observation, Kim et al. (534 and 633) presented evidence that presenilins carrying FAD mutations have altered capacitance calcium entry (CCE). Similar CCE defects were recently reported by Leissring et al. (JCB 149:793-797). CCE is the influx of extracellular calcium as a compensatory mechanism to replenish calcium depleted from intracellular stores. Kim demonstrated that cultured neurons from PS1 knockout embryos had elevated CCE responses compared to neurons cultured from embryos expressing wild-type PS1. Similarly, cells stably expressing the PS1 D257A engineered mutation, which acts as a dominant negative mutation in terms of Aβ secretion, had elevated CCE responses. Kim also presented data demonstrating that FAD mutations in PS1 and PS2 genes cause reduced CCE.
To determine whether there is a molecular link between CCE and Aβ production, Kim et al. treated cells with SKF96365, a CCE inhibitor. As expected SKF96365 inhibited CCE responses but interestingly this treatment correlated with an increase in Aβ production. Moreover, Kim presented findings that overexpression of the putative Drosophila CCE channel, Trp6, causes a reduction in Aβ production. Finally, Kim showed that addition of Aβ to cultures does not affect CCE. Taken together these results suggest that FAD presenilin mutations cause reduced CCE responses in cells and that this defect is likely to be upstream of the effect on Aβ production. Since FAD PS mutations have increased agonist-induced IP3 regulated calcium release, the combined effects of these two alterations may result in low cellular stores of calcium which may be detrimental to cell survival.

Presenilins and APP cleavage

Yankner et al., (603) reported that cells cultured from mouse embryos disrupted in both PS1 and PS2 genes do not undergo any detectable proteolytic cleavage of APP that corresponds to c-secretase activity. Thus, these cells fail to produce any detectable Aβ fragments. Instead, cells lacking both presenilins produce an increase in two other C-terminal APP cleaved fragments, C83 and C99 fragments, which are derived from α- and β-secretase cleavage, respectively (see Nature Cell Biol 2:463-5). Similar results were reported at the meeting by De Strooper et al. (see Nature Cell Biol. 2:461-2). These results are consistent with the notion that absence of presenilins eliminates c-secretase cleavage of APP. However, Yankner reported that there was a noticeable change in the proportion of the two C-terminal fragments of cultures prepared from mouse embryos disrupted in one versus both presenilin genes. In single PS1 knockout mice the proportion of the C99 species represented approximately 15%-30% of the total C-terminal fragments. In contrast in PS1/PS2 double knockout mice the C99 fragment was reduced to only 5% of the total fraction. These results suggest that loss of presenilin expression not only affects c-secretase cleavage of APP but also β-secretase cleavage of APP.

Steiner, et al., (303) reported on mutation analysis of residue 384 in PS1. Residue 384 of wild-type PS1 is a glycine. However, a FAD PS1 missense mutation causing a glycine to alanine substitution has been identified. Haass showed that cells expressing the presenilin FAD G384A mutation produce the largest documented increase in Aβ42 production. In contrast, mutation of this glycine to a number of other residues produced relatively small changes in Aβ production. They reasoned that since residue 384 was important for c-secretase activity, and it is adjacent to the aspartate 385, which is one of two aspartates (the other being D257) proposed to be essential for presenilin-associated c-secretase activity, it could be part of a signature sequence conserved in other proteases. They, therefore, compared the sequence surrounding residue 384 to see whether there was homology to any other protein in databases. Not only did they find the surrounding sequence to be highly conserved in all presenilin proteins across species, but they also found homology to a sequence present in proteins belonging to the type 4-prepilin-peptidase family. The type 4-prepilin peptidases (TFPP) are membrane-bound proteases present in different bacterial species, and like presenilins, are thought to constitute a novel family of bi-lobed aspartate proteases (see JBC 275:1502-10). TFPP cleaves the bacterial prepilin leader sequence releasing pilin protein into the periplasm. The processed pilin is then assembled into pili, which has diverse functions including, toxin secretion, gene transfer, adhesion and infection. Haass indicated that presenilins and prepilin both possess a conserved motif (G/AXGDX where X is variable and GD is the sequence at residues 384 and 385, respectively, in PS1). The similarity in the protease cleavage functions of presenilins and TFPP and the functional relationship will surely be explored in future studies.

Iwatsubo et al (631) described mutagenesis experiments of the sequence, ALPALP, which is a highly conserved in the C-terminal region of all known presenilins. The C-terminal proline in the sequence is mutated in FAD. They hypothesized that the conserved sequence could be important in presenilin function and, therefore, proceeded to mutagenize both proline residues. They reported that when the N-terminal most proline is mutated to a leucine in PS2 it does not affect either Notch cleavage or Aβ production. In contrast, mutation of the C-terminal proline to a serine abolishes both Notch and cleavage as well as Aβ production. They propose that the C-terminal proline of the ALPALP sequence may be important in the folding and stability of presenilins.

Presenilin interactors

De Strooper et al. (602) used a yeast 2-hybrid interaction trap to identify telencephalin as a interactor of the C-terminal region of PS2. Telenecephalin is a member of the NCAM superfamily of molecules, and is expressed in hippocampus and dentate gyrus. Telencephalin is also known to interact with integrins and has been functionally implicated in promoting neurite outgrowth. It has also been implicated in LTP. Interestingly, in presenilin knockout mice, telencephalin was abnormally localized, suggesting that lack of presenilins affects either sorting or localization of telencephalin.

Livne-Bar described a genetic screen for PS-interacting genes in Drosophila (928). They used a P-element screen to identify modifiers of PSEN defects, including those able to rescue a PSEN lethal phenotype. About two dozen interacting genes were isolated by this screen. These genes fell into three broad classes: genes involved in the Notch signaling pathway, genes involved in programmed cell death, and cell cycle-regulated genes. One of the genes able to rescue PSEN null larvae was shibire. Shibire whose mammalian counterpart is dynamin, belongs to the GTPase superfamily. Dynamin and shibire have been shown to play crucial roles in receptor-mediated endocytosis, although there is some contention as to whether dynamin functions as a mechanochemical motor or as a regulatory molecule required to activate downstream effectors involved in coated vesicle formation.

Ken Kosik (634) provided evidence that disruption of the Drosophila presenilin gene produces neurogenic defects and accumulation of large ubiquitin-positive inclusions within cells. In these mutant Drosophila embryos, Armadillo (β-catenin) is also mislocalized. Kosik et al. used microarry analysis to determine changes in RNA expression in the brains of mice lacking or containing presenilin genes. Out of over one thousand transcripts analyzed, only a handful were upregulated. Substantially more genes were downregulated, some of which included frizzled receptor, CaM kinase, SOX13, a Rad23 homologue, Nibrin, and interleukin 7.

Comments

Make a Comment

To make a comment you must login or register.

Comments on this content

No Available Comments

References

No Available References

Further Reading

No Available Further Reading