Live Discussion: New Theories of Presenilin Function

Chris Weihl led this live discussion on 12 December 2000. Readers are invited to submit additional comments by using our Comments form at the bottom of the page.

View Transcript of Live Discussion — Posted 30 August 2006

Background Text

Introduction
Among the numerous publications describing presenilin's putative role in membrane trafficking, apoptosis, development, signal transduction and the amyloid cascade, two sets of papers have been published looking at novel pathogenic mechanisms for the elusive 8 transmembrane domain protein associated with familial Alzheimer's disease. One set of papers deals with the role that presenilin may have in the coordination of endoplasmic reticulum resident chaperone proteins and their response to cellular stress (Katayama et al., 1999; Niwa et al., 1999). The second set of papers looks at the role of presenilin in synaptic plasticity using neurophysiological paradigms associated with learning (Barrow et al., 2000; Parent et al., 1999; Zaman et al., 2000).

Presenilin and the Unfolded Protein Response (UPR)
Ever since the discovery of presenilin-1 and its homologue presenilin-2, researchers have attempted to correlate its role with the phenotypic changes seen in Alzheimer's patients. Several lines of evidence suggest that PS plays an important role in amyloid production and its subsequent deposition. Other investigators have demonstrated that PS1 mutations associated with FAD increase a cells susceptibility to apoptosis and neuronal cell death. Two recent studies attempt to link these two phenomenon by investigating the role of PS in a novel cellular signaling pathway associated with the unfolded protein response (UPR) (Katayama et al., 1999; Niwa et al., 1999).

What is the UPR?
In brief, when a cell's endoplasmic reticulum (the place where proteins are synthesized) is overloaded with misfolded proteins, as in times of stress, the UPR allows for an increase in the transcription of resident ER chaperone proteins such as GRP78/Bip. The upregulation of these chaperones aid in the maintenance of protein tertiary structure, diminishing the load of detrimental "junk" ER misfolded protein. This pathway, while not essential for the cell's everyday function, is invaluable under times of cellular stress and instrumental in cellular survival (for review see Welihinda et al., 1999).

What Are the Molecular Mechanisms Involved in the UPR
The UPR was initially characterized in yeast, but several homologous pathways have now been described in mammalian cells. In yeast the pathway involves two proteins, an ER transmembrane protein (Ire1) and a nuclear transcription factor (Hac1). Hac1 recognizes UPR elements on promoters for GRP78/Bip and other molecular chaperones. In yeast, the ER lumenal domain of Ire1 senses the level of free GRP78/Bip. When free GRP78/Bip is low as in times of stress (presumably because GRP78/Bip is complexed with misfolded protein in the ER lumen), Ire1 dimerizes, autophosphorylates and becomes an RNA endonuclease. The Ire1 endonuclease then splices a constitutively produced but inactive Hac1 mRNA into an actively translated mRNA. Once the newly spliced Hac1 mRNA is translated, the Hac1 protein then enters the nucleus and activates specific UPR elements on molecular chaperone genes. The upregulation of ER molecular chaperones such as GRP78/Bip is the final result of the UPR.

In mammals this process is less well characterized. While hIre1· and hIre1, maintain RNA endonuclease function in vitro, they do not appear to function this way in vivo. Current data suggests that after hIre1· and hIre1, autophosphorylation, they are cleaved and transported to the nucleus where they can help in the trans-activation of Hac1 homologues. This process again results in an upregulation of ER molecular chaperones.

Does Mutant PS1 Perturb the UPR in Cells?
Katayama and colleagues initially looked at the susceptibility of mutant PS1 stably expressing neuroblastoma cells to "ER stressors." These stressors included tunicamycin (which prevents protein glycosylation) and the calcium ionophore, A23187 (which depletes ER stores of calcium). As expected the mutant PS1 expressing cells were more susceptible to these stresses as suggested by earlier investigators (Guo et al., 1996; Guo et al., 1999b).

In order to characterize the cellular response to these stressors in mutant PS1-expressing neuroblastoma cells, Katayama looked at the mRNA expression levels of a specific ER resident molecular chaperone, GRP78/Bip, that is known to be upregulated with the application of these ER stressors. While basal levels of GRP78/Bip were unaltered amongst the expressed transgenes, 6 hours after the application of tunicamycin a 50-30% decrease in GRP78 expression was seen in mutant PS1 expressing cell lines. This data was further reproduced in transiently expressing HEK293 cells and knock-in mice expressing mutant PS1.

To determine whether this decrease was due to a defect in the cells' unfolded protein response, Katayama performed gel shift assays with lysates from PS1-WT or PS1 mutant expressing cells. They demonstrated that there was a decrease in the activation of the UPR promoter element on GRP78/Bip. Niwa and colleagues performed a similar set of experiments using fibroblasts from PS1 knockout mice. The level of GRP78/Bip mRNA was decreased ~40% in the PS1 knockout cells 7 hours after the ER stressor tunicamycin was added. Katayama also demonstrated a significant decrease in the levels of GRP78/Bip in the brains of sporadic and familial AD patients when compared with unaffected control patients.

What Role Does PS1 Play in the Unfolded Protein Response Pathway?
Katayama demonstrates the colocalization of PS1 with Ire1 and co-immunoprecipitates full-length wild type and mutant PS1 with overexpressed Ire1 in vivo. These data suggest a putative interaction between PS1 and the most upstream element of the UPR pathway, Ire1. Furthermore, Katayama show that the phosphorylation and presumable activation of Ire1 is diminished in PS1-mutant expressing cells. Katayama and colleagues speculate that PS1 may serve as a molecular tether between Ire1 oligomers and phosphatases associated with Ire1 regulation. This hypothesis is reminiscent of other proposed mechanism for PS1. Takashima et al. suggested that PS1 binds GSK3-beta as well as its substrates, tau and beta-catenin. By tethering a kinase with its substrate, PS1 modulates the phosphorylation of tau and the phosphorylation and subsequent degradation of beta-catenin (Takashima et al., 1998; Zhang et al., 1998).

Niwa further extends these studies by demonstrating that PS1 knockout fibroblasts have a decrease in their ability to cleave and subsequently transport the C-terminus of hIre1 to the nucleus from the ER when compared with PS1-WT expressing controls. They speculate that PS1 regulates the activity of, or serves as, the gamma-secretase responsible for the cleavage of hIre1 since its cleavage site is within the ER membrane. This hypothesis is similar to the proposed role of PS1 as the gamma-secretase involved in the processing of APP, and Notch (Haass and Mandelkow, 1999).

How Do Changes in the UPR Pathway Result in AD and Amyloid Production?
It is intriguing to speculate that the neuronal cell loss seen in FAD patients may be due to an impaired UPR. Perhaps neurons are more sensitive to specific environmental insults (or ER stressors) that may precipitate the neuronal cell loss seen in FAD patients expressing mutant PS1. Furthermore, several studies have demonstrated an increase in multi-ubiquitinated protein inclusions in AD patient brain tissue suggesting an overload of misfolded proteins (Alves-Rodrigues et al., 1998). It is interesting that overexpression of GRP78/Bip rescues PS1-mutant expressing cells from the ER stressors tunicamycin and calcium ionophore, A23187 (Katayama et al., 1999). A similar study overexpressed HSP70, a cytosolic molecular chaperone, and rescued cells from cell death induced by mutations in Cu/Zn superoxide dismutase-1 (SOD-1) that are associated with familial ALS patients (Bruening et al., 1999). Moreover this same study demonstrated a decrease in HSP70 chaperoning function in transgenic mice expressing mutant SOD-1 (Bruening et al., 1999).

Another study found that APP transiently associated with the GRP78/Bip as it moved through the secretory pathway. Overexpression of GRP78/Bip decreased the amyloidogenic phenotype of the APP Swedish mutation by lowering the ratio of A-beta1-42/1-40 (Yang et al., 1998). The authors speculated that an increase in APP's association with GRP78/Bip decreased the likelihood of gamma-secretase cleavage. Moreover, a decrease in the cells UPR would result in a decrease in GRP78/Bip and may allow for APP to be conformationally more susceptible to gamma-secretase (Yang et al., 1998).

A new paper by Sato and colleagues casts doubt on the putative role of PS and its mutations in the ER stress response. This paper, in the December 2000 Nature Cell Biology, uses similar strategies as the antecedent investigations by Katayama and Niwa (see above) to explore the role of PS and FAD associated mutant PS in the unfolded protein response (UPR). Contrary to these previous studies, Sato finds no difference in the UPR between PS1-WT, PS1 knockout or FAD mutant PS1 expressing cells, mice or FAD patients. This addendum will attempt to highlight the major differences between the studies placing special emphasis on the results and experimental design.

Sato and colleagues comprehensively address the role of PS1 in the UPR. However, contrary to Niwa et al., Sato finds no difference in the UPR using PS1 knockout fibroblasts, as well as, PS1/PS2 knockout cells after measuring the levels of GRP78/Bip and CHOP mRNA and protein levels following tunicamycin treatment. Moreover, Sato also finds no difference in the activity of IRE1 in PS1 deficient cells. However, Niwa's study addressed the translocation of IRE1 from the cytosol to the nucleus, whereas Sato's study used immunoblots to assess the phosphorylation state of the IRE1 protein. While this difference seems trivial, Niwa and colleagues speculate that PS1 has a direct role in the cleavage and hence transport of the IRE1 protein, not the phosphorylation state. This difference in technique and concomitant result is reminiscent of studies investigating the translocation vs. the stability of beta-catenin in similar cell lines (see previous panel discussion). However, in support of Sato, the most downstream event in the UPR pathway, GRP78/Bip mRNA levels, is tested and no difference was found between the treatment groups.

Using several different cell lines and transgenic mice that express FAD associated mutant PS1, Sato and colleagues again challenge the previous studies of Katayama. Sato demonstrates no difference in the levels of GRP78/Bip mRNA/protein levels or the phosphorylation state of IRE1 following ER stress in mutant expressing cell lines. However, Katayama assessed the activity of IRE1 differently in the initial study. They used gel shift assay to assess the functional activation of the UPR in stimulated cells. It remains to be determined if subtle changes in experimental design may confound the distinct differences in these two papers. Sato also contradicts the initial study of Katayama and finds no difference in basal GRP78/Bip protein levels from FAD patient and transgenic mouse tissue.

Finally, a recent paper by Sato, Imaizumi, et al. demonstrates that a splice variant of PS2, which is enriched in sporadic AD brains, can associate with IRE1 and plays a direct role in the UPR by downregulating GRP78/Bip expression . This study further unifies the hypothesis that PS increases the susceptibility of FAD patient brains to specific stresses that modulate the ER stress response.

The role of the UPR in Alzheimer's disease and amyloid production is intriguing. The papers by Katayama and Niwa propose a novel mechanism of action for PS1 and its mutations that unify the current phenotypes of enhanced cell death and A-beta production.

Presenilin and Synaptic Plasticity
The pathology seen in PS1 mutant expressing transgenic mice has been disappointing. Although the mice do demonstrate an increase in the ratio of A-beta1-42/1-40, they do not develop appreciable amyloid plaques even when co-expressed with human APP. This is in contrast to the APP mutant expressing transgenic mice, which do demonstrate amyloid plaque deposition (presumably because of the high level of mutant APP expression). In an effort to investigate the role of PS1 on more subtle phenotypes such as synaptic plasticity, three independent groups investigated the neurophysiologic properties of neurons from FAD transgenic mice (Barrow et al., 2000; Parent et al., 1999; Zaman et al., 2000).

The role of PS1 at the synapse has not been fully investigated. While most studies demonstrate that PS1 and its fragments are localized to nuclear, ER and golgi membranes, some groups have shown that PS1 is present on synaptic vesicles and at pre- and post-synaptic terminals (Beher et al., 1999; Efthimiopoulos et al., 1998; Georgakopoulos et al., 1999; Lah et al., 1997). The role of PS1 in the development of neuronal pathways has also not been appropriately addressed by investigators. PS1 knockout mice have significant perturbations in embryonic pattern formation (Shen et al., 1997). Moreover, one study using double transgenic mice expressing both mutant PS1 and mutant APP demonstrated a reorganization of the synaptic terminals of the basal forebrain suggesting that mutations associated with FAD altered the anatomical structure of the developing brain (Wong et al., 1999). These studies suggest that PS1 may alter synaptic plasticity in PS1 mutant transgenic mice by directly participating in synaptic transmission or by altering the brains neuronal architecture.

The simplest form of learning and memory occurs between two neurons, one pre-synaptic and another post-synaptic. Synaptic stimulation using the correct frequency, amplitude and duration can create a lasting response in the post-synaptic neuron. Neuronal plasticity involves multiple mechanisms including membrane potential, receptor density, calcium release and gene transcription. More complex mechanisms of learning have been described that involve several pre- and post-synaptic neurons. The most well characterized neurophysiologic mechanism of higher learning is long-term potentiation (LTP) in the CA1 and CA3 regions of the mammalian hippocampus. In this paradigm, tetanic stimulation of post-synaptic neurons by the pre-synaptic neuron can produce a prolonged potentiation of the post-synaptic neurons response to future stimulation that can last for several hours. Neurophysiologists presume that alterations in these simple learning paradigms might result in higher learning deficits in animals.

Does Mutant PS1 Alter Synaptic Function?
In order to address whether mutations in PS1 alter the neurophysiology of mammalian brains, three independent groups explored the synaptic plasticity in FAD transgenic mice (Barrow et al., 2000; Parent et al., 1999; Zaman et al., 2000). Each group studied well-characterized neurophysiologic paradigms using hippocampal brain slices from unique mutant PS1 expressing mice (PS1A246E, PS1deltaE9, PS1M146L, PS1M146V).

Parent and colleagues measured several neurophysiologic parameters in their experiments using field excitatory postsynaptic potential (fEPSP) at the Schaffer collateral-CA1 synapse in hippocampal slices. They found no difference in the basal synaptic transmission including maximum fEPSP slope, maximum fEPSP amplitude or the basal synaptic strength. However upon high frequency stimulation used to elicit long-term potentiation (LTP), mutant PS1 expressing animals had a larger initial amplitude that was more persistent than PS1-WT and non-transgenic control littermates.

Barrow and colleagues examined similar parameters using intracellular recordings from CA3 pyramidal neurons. Their data demonstrated that following a train of 10 action potentials there was significant increase in the amplitudes of the after-hyperpolarizations. Moreover they also found that mutant PS1 expressing animals had a larger amplitude and more persistent response to LTP induction when compared with controls. Barrow and colleagues speculated that this may be due to a change in the release of ER stores of intracellular calcium as previously demonstrated (Guo et al., 1996). Using PS1-mutant expressing hippocampal pyramidal cell neurons they showed an increase in the rise and rate of intracellular calcium release following neuronal depolarization. Their data demonstrates that mutant PS1 may alter the intracellular ER calcium stores and hence increase its release upon depolarization. Increased ER calcium release may contribute to the enhanced neuronal plasticity seen in PS1 mutant transgenic brain slices.

Zaman and colleagues confirmed the prior two studies. They found enhanced and elevated LTP in PS1-mutant expressing brain slices at CA1 pyramidal neurons. In addition, they propose that the increase in synaptic plasticity due to enhanced calcium release may alter GABAA inhibitory input at the CA1 hippocampal neuron. Zaman used pharmacologic manipulation to either inhibit or enhance GABAA inhibitory transmission. Normally, in non-transgenic mice brain slices, when GABAA is inhibited by picrotoxin, LTP is enhanced and when GABAA is potentiated with a benzodiazepine, LTP is decreased. However in PS1 mutant transgenic mice brain slices, GABAA inhibition produced no effect and GABAA potentiation restored LTP to wild-type controls. This finding suggested that GABAA inhibition was upregulated in PS1 mutant expressing mice to compensate for the enhanced synaptic excitatory activity.

How Do Changes in Synaptic Plasticity Contribute to AD?
On the simplest level it is easy to speculate that alterations in LTP, either increased or decreased, could lead to alterations of learning and memory that are associated with the progression of AD in patients. Another possible scenario proposed by Zaman and colleagues postulates that the constant increase in intracellular calcium during neuronal stimulation may burden the neurons causing them to die or improperly function. Support for this hypothesis is found in a paper by Mattson and colleagues describing an increased sensitivity of FAD transgenic mice to glutamate mediated excitotoxicity (Guo et al., 1999a). Finally, as shown by Zaman, the neuronal architecture may be altered so as to compensate for the changes associated with mutant PS1 expression. Zaman proposes that pharmacologic agents aimed at decreasing the synaptic activity, such as benzodiazapines, may be protective in AD. This thought is intriguing in light of a clinical study that showed a decrease in the incidence of AD in patients chronically using benzodiazipines for sleep (Fastbom et al., 1998).

Conclusion
It is important for researchers to continue to search for mechanisms by which mutations in PS1 and PS2 cause familial Alzheimer's Disease. An increase in apoptosis or changes in amyloid production are only phenotypes. While therapies can be aimed at rescuing these phenotypes, it is also prudent to explore therapies targeted at the underlying mechanisms related to these phenotypic changes. The aforementioned papers suggesting roles for PS1 in the unfolded protein response or synaptic transmission shed new light onto potential roles for PS in AD.

Questions for Discussion and Future Investigation

• Does mutant PS1 increase FAD patient brains' susceptibility to ER stresses, even though this may not occur through the UPR pathway?
• Can experimental design account for the differences between the reports (on UPR)?
• How does a gain of function mutation in PS1 result in the same phenotype as PS1 knockouts?
• Can one unified global function of PS be attributed to its effects on amyloid production, apoptosis, signal transduction, synaptic transmission and development?
• Assuming that PS is the gamma secretase, how will future treatments such as gamma secretase inhibitors affect PS's function on other cellular pathways?
• Describe future treatments that may be aimed at correcting the defect in these newly identified roles of PS?
• What directions are you currently pursuing in regards to your initial observations?
• Do your results agree with the other sets of investigator's papers? Please describe.
• Is Alzheimer's Disease research too narrowly limited to the amyloid hypothesis? Do you have other hypotheses related to your own work?

References
Alves-Rodrigues, A., L. Gregori, and M.E. Figueiredo-Pereira. 1998. Ubiquitin, cellular inclusions and their role in neurodegeneration. Trends Neurosci. 1998 Dec;21(12):516-20. Abstract.

Barrow, P.A., R.M. Empson, S.J. Gladwell, C.M. Anderson, R. Killick, X. Yu, J.G. Jefferys, and K. Duff. 2000. Functional phenotype in transgenic mice expressing mutant human presenilin-1. Neurobiol Dis. 2000 Apr; 7:119-26. Abstract.

Beher, D., C. Elle, J. Underwood, J.B. Davis, R. Ward, E. Karran, C.L. Masters, K. Beyreuther, and G. Multhaup. 1999. Proteolytic fragments of Alzheimer's disease-associated presenilin 1 are present in synaptic organelles and growth cone membranes of rat brain. J Neurochem. 199 Apr; 72:1564-73. Abstract.

Bruening, W., J. Roy, B. Giasson, D.A. Figlewicz, W.E. Mushynski, and H.D. Durham. 1999. Up-regulation of protein chaperones preserves viability of cells expressing toxic Cu/Zn-superoxide dismutase mutants associated with amyotrophic lateral sclerosis. J Neurochem. 1999 Feb;72:693-9. Abstract.

Efthimiopoulos, S., E. Floor, A. Georgakopoulos, J. Shioi, W. Cui, S. Yasothornsrikul, V.Y. Hook, T. Wisniewski, L. Buee, and N.K. Robakis. 1998. Enrichment of presenilin 1 peptides in neuronal large dense-core and somatodendritic clathrin-coated vesicles. J Neurochem. 1998 Dec; 71:2365-72. Abstract.

Fastbom, J., Y. Forsell, and B. Winblad. 1998. Benzodiazepines may have protective effects against Alzheimer disease. Alzheimer Dis Assoc Disord. 12:14-7. No abstract available.

Georgakopoulos, A., P. Marambaud, S. Efthimiopoulos, J. Shioi, W. Cui, H.C. Li, M. Schutte, R. Gordon, G.R. Holstein, G. Martinelli, P. Mehta, V.L. Friedrich, Jr., and N.K. Robakis. 1999. Presenilin-1 forms complexes with the cadherin/catenin cell-cell adhesion system and is recruited to intercellular and synaptic contacts. Mol Cell.1999 Dec; 4:893-902. Abstract.

Guo, Q., K. Furukawa, B.L. Sopher, D.G. Pham, J. Xie, N. Robinson, G.M. Martin, and M.P. Mattson. Alzheimer's PS-1 mutation perturbs calcium homeostasis and sensitizes PC12 cells to death induced by amyloid beta-peptide. Neuroreport. 1996 Dec 20; 8:379-83. Abstract.

Guo, Q., W. Fu, B.L. Sopher, M.W. Miller, C.B. Ware, G.M. Martin, and M.P. Mattson. 1999a. Increased vulnerability of hippocampal neurons to excitotoxic necrosis in presenilin-1 mutant knock-in mice. Nat Med. 1999 Jan;5:101-6. Abstract.

Guo, Q., L. Sebastian, B.L. Sopher, M.W. Miller, C.B. Ware, G.M. Martin, and M.P. Mattson. Increased vulnerability of hippocampal neurons from presenilin-1 mutant knock-in mice to amyloid beta-peptide toxicity: central roles of superoxide production and caspase activation. J Neurochem. 1999 Mar; 72:1019-29. Abstract.

Haass, C., and E. Mandelkow. Proteolysis by presenilins and the renaissance of tau. Trends Cell Biol. 1999 Jun; 9:241-4. Abstract.

Katayama, T., K. Imaizumi, N. Sato, K. Miyoshi, T. Kudo, J. Hitomi, T. Morihara, T. Yoneda, F. Gomi, Y. Mori, Y. Nakano, J. Takeda, T. Tsuda, Y. Itoyama, O. Murayama, A. Takashima, P. St George-Hyslop, M. Takeda, and M. Tohyama. Presenilin-1 mutations downregulate the signalling pathway of the unfolded-protein response. Nat Cell Biol. 1999 Dec;1:479-485. Abstract.

Lah, J.J., C.J. Heilman, N.R. Nash, H.D. Rees, H. Yi, S.E. Counts, and A.I. Levey. Light and electron microscopic localization of presenilin-1 in primate brain. J Neurosci. 1997 Mar; 17:1971-80. Abstract.

Niwa, M., C. Sidrauski, R.J. Kaufman, and P. Walter. A role for presenilin-1 in nuclear accumulation of Ire1 fragments and induction of the mammalian unfolded protein response. Cell. 1999 Dec 23; 99:691-702. Abstract.

Parent, A., D.J. Linden, S.S. Sisodia, and D.R. Borchelt. Synaptic transmission and hippocampal long-term potentiation in transgenic mice expressing FAD-linked presenilin 1. 1999 Feb; Neurobiol Dis. 6:56-62. Abstract.

Sato, N., Imaizumi, K., Manabe, T., Taniguchi, M., Hitomi, Y., Takagi, T., Kudo, T., Tsuda, T., Itoyama, Y., Makifuchi, T., Fraser, P., St. George-Hyslop, P., and Tohyama, M. (2000). Increased production of b-amyloid and vulnerability to ER stress by an aberrant spliced form of presenilin-2. J Biol Chem. 2000 Oct 12 . Abstract.

Sato, N., Urano, F., Leem, J.-Y., Kim, S.-H., Li, M., Donoviel, D., Bernstein, A., Lee, A., Ron, D., Veselits, M., Sisodia, S., and Thinakaran, G. (2000). Upregulation of BiP and CHOP by the unfolded-protein response is independent of presenilin expression. Nature Cell Biol Dec 2000. 2 (12) 863-870. Abstract.

Shen, J., R.T. Bronson, D.F. Chen, W. Xia, D.J. Selkoe, and S. Tonegawa. Skeletal and CNS defects in Presenilin-1-deficient mice. Cell. 1997 May 16; 89:629-39. Abstract.

Takashima, A., M. Murayama, O. Murayama, T. Kohno, T. Honda, K. Yasutake, N. Nihonmatsu, M. Mercken, H. Yamaguchi, S. Sugihara, and B. Wolozin. Presenilin 1 associates with glycogen synthase kinase-3beta and its substrate tau. Proc Natl Acad Sci U S A. 1998 Aug 4; 95:9637-41. Abstract.

Welihinda, A.A., W. Tirasophon, and R.J. Kaufman. 1999. The cellular response to protein misfolding in the endoplasmic reticulum. Gene Expr. 7:293-300. Abstract.

Wong, T.P., T. Debeir, K. Duff, and A.C. Cuello. 1999. Reorganization of cholinergic terminals in the cerebral cortex and hippocampus in transgenic mice carrying mutated presenilin-1 and amyloid precursor protein transgenes. J Neurosci. 19:2706-16. Abstract.

Yang, Y., R.S. Turner, and J.R. Gaut. 1998. The chaperone BiP/GRP78 binds to amyloid precursor protein and decreases Abeta40 and Abeta42 secretion. J Biol Chem. 273:25552-5. Abstract.

Zaman, S.H., A. Parent, A. Laskey, M.K. Lee, D.R. Borchelt, S.S. Sisodia, and R. Malinow. 2000. Enhanced synaptic potentiation in transgenic mice expressing presenilin 1 familial Alzheimer's disease mutation is normalized with a benzodiazepine. Neurobiol Dis. 7:54-63. Abstract.

Zhang, Z., H. Hartmann, V.M. Do, D. Abramowski, C. Sturchler-Pierrat, M. Staufenbiel, B. Sommer, M. van de Wetering, H. Clevers, P. Saftig, B. De Strooper, X. He, and B.A. Yankner. 1998. Destabilization of beta-catenin by mutations in presenilin-1 potentiates neuronal apoptosis. Nature. 395:698-702. Abstract.

Participants: Chris Weihl, Gopal Thinakaran, Kudo, Kazunori Imaizumi, Nikolaos Robakis, Weiming Xia, June Kinoshita

Note: Transcript has been edited for clarity and accuracy.

Cweihl: Welcome to the panel discussion on PS1 and new theories of its functions. Let's begin by discussing the role of PS1 in the unfolded protein response (UPR).

kudo: Dr.Imaizumi and Dr.Thinakaran, firstly we should show what is the problem between our paper and Sato's paper (Nature Cell Biol, Dec 2000).

Thinakaran: Sure.

Cweihl: My first question is general. Do mutations in PS cause an increase sensitivity to ER stressors? Dr. Thinakaran?

Thinakaran: Good point. I haven't directly tested the sensitivity issue in my lab. However, a recent paper by Robert Siman casts serious doubts.

Cweihl: Please explain? Dr. Imaizumi and do you care to comment as well?

Imaizumi: We observed that FAD-PS1 mutation do sensitize cells to ER stressors.

Thinakaran: Siman and colleagues used neurons from their knock-in mice to address whether there is increased cell death, caspase cleavage, TUNNEL labeling etc with glutamate stress and also typical ER stress. Their paper just got published in J. Neuroscience. Also, he presented his data at the Neuroscience meeting.

Imaizumi: I also saw his data.

Cweihl: Dr. Imaizumi what are the key differences between your studies and the Thinakaran paper? Dr. Thinakaran, I'd like to ask you the same question.

Thinakaran: I will wait until Dr. Imaizumi has a chance to address.

Imaizumi: I think [they are] differences of treatment of stressors, dose and time course.

Cweihl: Please explain the differences.

Imaizumi: [Dr. Thinakaran's] data show that middle dose of tunicamycin affects sensitivity to ER stress, but higher or lower dose did not affect it. I personally thought that the doses were comparable and the time points the same as well.

Thinakaran: I am not quite sure what Dr. Imaizumi refers to as low/mid/high doses. We only used one dose and not a dose-response study. In fact, we tried hard to keep the dose and time of treatment as close to their Nature Cell Biology paper as possible.

Cweihl: Dr. Imaizumi. what happens to the UPR at high vs. low vs. mid dose?

Imaizumi: At high dose, the differences of ER stress response are difficult to detect. High dose causes cell death.

Thinakaran: But your Nature Cell Biology paper did report clear differences at the same dose we tried. This I can't explain. Dr. Imaizumi, if low dose shows the difference between wt and mutant PS1, and only high dose causes cell death (presumably more in mutant cells), how can we explain this difference in cell death based on "differential sensitivity"?

Imaizumi: At low dose, as cell death does not occur, the UPR is activated in normal cells.

kudo: I would like to add to Dr. Imaizumi's comment. Dr. Thinakaran checked relatively late time-points but we checked early points after adding stress.

Thinakaran: Dr. Kudo, not true. Our time points 5 h was shorter than the 6 h treatment reported in the Nature Cell Biology paper.

Cweihl: To me the differences between the papers were related to cell lines and experimental analysis. I thought that the differences between the two papers was reminiscent of the initial studies between cell lines looking at beta catenin stability vs. translocation of beta catenin to the nucleus. Please comment Dr. Thinakaran?

Thinakaran: Unlike catenin papers, both Imaizumi's study and our study (data from David Ron's lab and my own) examined "activation" of IRE. There are many ways of looking at activation. But, ultimately the function is the coordinate induction of UPR genes.

Imaizumi: I detected the disturbed activation of PERK and IRE1.

Cweihl: What is the best way to look at activation of IRE? Translocation, phosphorylation or ultimately mRNA levels?

Imaizumi: Phosphorylation.

Thinakaran: In my opinion, gene expression. It is the readout. Phosphorylation is like writing a check. Gene expression is cashing it.

Imaizumi: PS1 mutant affects the phosphorylation of IRE1 within 30 min. after treatment with tunicamycin.

Cweihl: I agree with Dr. Thinakaran. However you look at protein levels and mRNA levels, the protein levels are the sensor to activate the UPR. Is this a confounding problem in your experimental design?

Thinakaran: Again, I refer to Katayama et al. paper, where the difference was seen clearly at later time points. Is there a problem in reproducing that data?

Nikos enters

Imaizumi: It is reproducible data.

Nikos: Hi from Robakis lab. We had some problems logging in

Thinakaran: Dr. Imaizumi, then I fail to understand why you insist on the early time points.

Imaizumi: PERK and IRE1 are activated very early after ER stress.

guest10 enters

Imaizumi: I think this activation may be critical for induction of GRP78 mRNA.

Thinakaran: As far as I can tell (as an outsider to the ER stress field), no one has shown a functional role for "delay" of 10-15 min in activation of IRE or PERK. If I remember the data correctly, after 45 minutes of treatment, there is NO difference in the activation of either kinase.

Cweihl: Dr. Imaizumi, do your results agree with the Niwa paper in Cell?

guest10: This is Weiming Xia from Center for Neurologic Diseases. I forgot my password. Early onset?

Imaizumi: No, we could not detect any differences of UPR in PS1 knockout cells.

Cweihl: So there is agreement between your and Thinakaran's paper in regard to KO lines?

Imaizumi: Yes

Thinakaran: Has anyone talked with Dr. Walter about this?

Imaizumi: I have not talked with him.

Cweihl: So can we attribute the differences between the papers to experimental design? Dr. Thinakaran?

Thinakaran: I don't think experimental design is the problem. Our study followed Dr. Walter's experiment with regards to BiP Northern - concentration, time course.

Cweihl: Dr. Robakis or Dr. Xia do you care to comment on the differences between these two papers?

Nikos: (Robakis) Not at this point because we don't have experimental data to support one or another.

Imaizumi: We could not detect the cleaved fragments of IRE1.

Thinakaran: Oh oh...Someone really ought to sort this issue with Dr. Peter Walter. The AD field is citing his Cell paper whenever there is a chance to say PS1 regulates this cleavage, not detected by any other lab. Maybe other lab's haven't done their study as carefully as Walter's lab.

Cweihl: Let us shift discussion to the other topic.: PS1 and synaptic transmission. Do we have a consensus that PS1 mutations cause a perturbations in synaptic transmission?

Nikos: (Robakis) We think the action is at the cell surface including synaptic contacts where PS1 interacts with cadherins and participates in the cleavage of APP and Notch1.

Cweihl: Please explain, Dr. Robakis? Does PS1 exist at these synaptic contacts, Dr. Robakis?

Nikos: Yes it does. We have shown that in our paper in Molecular Cell. PS1 concentrates at synaptic and cell-cell contact sites.

Cweihl: Dr. Thinakaran can you comment on Dr. Robakis points?

Thinakaran: Dr. Robakis, has anyone addressed whether mutations in cadherins lead to altered APP or Notch metabolism?

Weiming Xia: I have not done any experiments to favor one side. My guess is that experimental methods used in two papers lead to the difference.

Nikos: PS1 regulates cell-cell adhesion and we believe that it also participates in synaptic structure. Dr Thinakaran, we have evidence that PS1 binds directly to cadherins and we have mapped the site (paper under review).

Thinakaran: What baffles me is that the levels of PS in cells are not nearly 1/10 of the levels of cadherins or other cell junction structural proteins.

Cweihl: Several lines of evidence suggest that the role of PS in synaptic output is at the level of the ER and calcium release or reuptake. Can this be the case as well as changes at the synapses?

Thinakaran: Synaptic transmission can be altered by proteins that are not localized in the synapse. There are several [lines of] evidence for this.

Nikos: We don't say that all of the PS1 molecules bind to cadherins. A portion of total PS1 is bound to cadherins and only at cell-cell contact sites.

Cweihl: Dr. Thinakaran, then could a trafficking issue be involved that alters the synaptic sites?

Thinakaran: Definitely.

Nikos: If calcium is not there and therefore cell-cell contacts are not forming, the cadherin-PS1 complex falls apart (see our paper in Mol. Cell).

Cweihl: Please explain and do you have any evidence for the trafficking synaptic proteins in particular? Dr. Thinakaran?

Nikos: In the absence of cell-cell contacts PS1 concentrates in the ER-Golgi (see the above paper).

Thinakaran: Is anyone studying second messenger cascade as a possible level of regulation by PS? Studies are under way in Sam Sisodia's and our lab to address the trafficking of proteins.

Nikos: We have already have shown that PS1 is found in vesicles and we have proposed in 1997 (Efthimiopoulos et al.) that PS1 trafficks to the cell-cell surface.

Cweihl: Dr. Robakis please expand on how PS1 association with cadherins alters synaptic transmission and in particular multiple tetanic stimulation paradigms that favor calcium levels?

Nikos: Again the data we have is that PS1 binds directly to E-cadherin and stabilizes the cadherin-catenin complex (paper under review). We don't have data that PS1 mutation affects synaptic transmission.

Thinakaran: The effect on synapse doesn't t necessarily happen in adult. It may be predetermined during development - for example a change in the number of spines and synaptic contacts during development.

Cweihl: Dr. Thinakaran, is there evidence for alterations developmentally in synaptic architecture in FAD patients or mice?

Thinakaran: The story on synaptic transmission is only starting to unfold. I am just throwing new ideas into the discussion.

June: Gopal touches on an issue that has interested me. Namely, how might effects of FAD PS mutations during development contribute to AD pathogenesis in carriers of those mutations?

Thinakaran: June, actually I am being the voice for my wife Dr. Angele Parent, who first published the LTP study on PS mice.

Cweihl: Is there evidence that changes in NOTCH can alter synaptic architecture?

Thinakaran: PS1 must be such a busy protein to be able to bind to cadherins, catenins and about 100 other proteins. Is there any cell-type selectivity of interactors?

Cweihl: Dr. Thinakaran, perhaps these bindings are all transient.

Nikos: (Robakis) Yes it seems that PS1 transiently binds to all these proteins.

Thinakaran: The lateral movement on the plane of the bilayer for a polytopic protein is unlikely to allow "transient" binding to a number of proteins localized throughout the cell - ER, cell surface etc.

Nikos: (Robakis) We have also evidence that PS1 expressing neurons are more resistant to degeneration than neurons with lower levels of PS1 (Giannakopoulos et al. American J. Pathology 1997).

Cweihl: Dr. Robakis does this mean that PS1 is protective and FAD mutations result in loss of function?

Nikos: (Robakis) Our data is on sporadic AD and suggests that PS1 somehow helps neuronal survival. June: Gopal, I'm wondering whether effects of mutant PS1 on Notch processing, for example, might alter neuronal populations or architecture in such a way as to reduce "cognitive reserve."

Thinakaran: June, I think the jury is still out on the issue whether PS mutations influence Notch processing. In our lab it has been inconsistent.

Cweihl: I would like to finally ask one broad question that can be commented upon by all present. Can we identify one unified function for PS1 and its mutations that address all of its multiple roles in synaptic transmission, ER stress, A-beta production, signaling etc.? And how might therapies that are targeted at PS1 disrupt its many putative roles?

Nikos: (Robakis) With regards to cadherin we have experimental evidence that the PS1-cadherin interaction is stronger than the PS1-catenin interaction with respect to detergent extraction

Weiming Xia: I would love to say it is the cleavage by PS1.

Thinakaran: Trafficking

Taisuke: Intra/juxtamembranous cleavage by presenilins.

Imaizumi: Trafficking and ER stress.

Cweihl: So would gamma secretase inhibitors have detrimental effects to synaptic transmission or notch cleavage or IRE1 cleavage?

Thinakaran: Perhaps we can change Presenilin to Polyfunctionalin PF1 and PF2.

Nikos: (Robakis) we cannot identify one unified function for PS1 yet that can can explain its role in Alzheimer's

Thinakaran: Chris, I think that treatments that target PS1 haven't been characterized well in terms of effect on other proteins.

Weiming Xia: DeStrooper and Wolfe showed that gamma secretase inhibitor blocked the Notch cleavage.

Cweihl: Is that a bad thing?

Thinakaran: Maybe the companies are looking at selective inhibitors. Most of the inhibitors are based on gamma-site of APP, so it surprises me that they all also block Notch.

Cweihl: Well it is nearing the hour and I would love final comments by Drs. Thinakaran , Imaizumi, Robakis and Xia.

Weiming Xia: It is not true that [gamma-secretase inhibitors] ALL ALSO block Notch. At least with different IC50.

Thinakaran: I think that cleavage and trafficking may be related functions of PS1. ER degradation in general is a likely area that PS may have a role - this will overlap with all the interactions and functions that we have been discussing.

Nikos: (Robakis) We believe that PS1 participates in the processing and presumably modulates the function of many type 1 receptors. We already know some, including APP, Notch1 and cadherins.

Thinakaran: Is cadherin "processed" by PS?

Nikos: (Robakis) We don't know yet. We are currently studying that.

Weiming Xia: It is interesting to identify additional "co-factors"/substrates involved in the high molecular weight complexes.

Cweihl: Thank you all for participating in this interesting discussion. Any other questions can be posted to the web site and will be directed to the investigator. Thank you again.

Gopal: I'd like to thank June and Chris for arranging this, and all the participants for an interesting discussion. Imaizumi: Thank you for interesting discussion.

Weiming Xia: Thank you, and have a wonderful holiday.

Nikos: Dr. Georgakopoulos represented the Robakis lab. Thank you. Weiming Xia: Gopal, watch out while you are driving....

June: Oh yes, I heard about the snow storm in Chicago. How are things there?

Cweihl: I'm at home.

June: Hurrah for the Internet!

Thinakaran: Me too.

Cweihl: ... and don't plan to leave.

June: By the way, Christian Haass wasn't able to participate today due to technical problems... Good bye for now!

Weiming Xia: My best regards to all of you. See you at the next live discussion.

END

Additional Comments
Kazunori Imaizumi: I would like to comment on differences of data between our and Dr.Thinakaran's labs. Previously, we provided the cells that we used in Nature Cell Biology paper (1,479-485,1999) to Dr.Thinakaran. Using these cells, they examined the induction of GRP78/BiP protein levels ( I think that they should examine the levels of BiP mRNA). They informed me that no significant differences of BiP induction were observed. Although the same cells were used in both labs, results were not consistent with each other. The causes may be due to the differences of experimental conditions (dose, cell types, medium change before stimulation) or sampling RNA or proteins. In these items, I think medium change before treatment with agents is most important because we could not obtain constant data if it was not performed. As we reported previously, the decrease of BiP mRNA induction in cells expressing FAD PS1 mutations is at 30 % compared with the controls. To detect the subtle defects in the UPR, the cells are carefully dealt with under the same conditions.