Live Discussion: β Amyloid Degradation: The Forgotten Half of Alzheimer's Disease
Return to Discussion Text
Live discussion held 12 September 2002 with Wesley Farris and Malcolm Leissring, Brigham and Women's Hospital and Harvard Medical School.
Participants: Chastity Whitaker, Chris Eckman, Claudia Almeida, Douglas Feinstein, Craig Atwood, Dan Raper, Detlef Schmicker, Elizabeth Eckman, Gabrielle Strobel, Keith Crutcher, KS, Lou Hersh, Malcolm Leissring, Matthew LaVoie, Nilufer Ertekin-Taner, Paul Shapiro, Reisuke, Rina Yamin, Stefan Mansourian, Steve Estus, Sylvain Lesne, Tony Turner, Wes Farris.
Note: The transcript has been edited for clarity and accuracy.
Gabrielle Strobel: Hello everyone and welcome all to today's chat. I am Gabrielle Strobel, managing editor of the Alzheimer Research Forum, and will be moderating today.
Malcolm Leissring and Wes Farris: OK. I thought we would start by discussing what genetic evidence there might be for the involvement of decreased β-amyloid degradation as a cause of AD.
Chris Eckman: Nilufer, this question may be best for you.
Nilufer Ertekin-Taner: The main lines of genetic evidence, in my opinion come from the mouse/rat knock-out/mutant studies showing elevations of brain and plasma amyloid β levels in mice that lack the NEP and PLAU genes, and in rats that have mutations in IDE. In addition, our group has shown in three different series that variants in PLAU are associated with risk for LOAD and that some of these variants are associated with elevated plasma Ab levels in LOAD families. I can continue but would like to give others the chance to reply.
Gabrielle Strobel: It looks as though two separate Ab-degrading enzymes, IDE and PLAU, could be under the linkage peak on chromosome 10. How likely is that?
KS: Personally, I do not think this is a problem. Although improbable, it is not impossible to imagine a clustering of activities that are related.
Malcolm Leissring and Wes Farris: According to Tony Brookes, there may be at least two, possibly three, with alpha-T-catenin being the third. I should have said according to Brookes and Goate, and Tanzi and....
Claudia Almeida: What is PLAU?
Nilufer Ertekin-Taner: PLAU=gene name for urokinase-type plasminogen activator (uPA). There is also evidence from Tanzi and Brookes laboratories about significant associations with variants in IDE and AD.
Chris Eckman: The knockout studies show elevations [of Ab] clearly in brain for ECE, NEP and IDE [knockouts] and in plasma for PLAU [knockouts]. To me this is considerable.
Malcolm Leissring and Wes Farris: Has anyone looked at plasminogen KOs?
Gabrielle Strobel: I vaguely recall a talk by Luc Buee in Stockholm on tPA knockout mice (raised for a different purpose) that had thioflavin-positive amyloid plaques in brain. Did anyone else hear this?
Nilufer Ertekin-Taner: Yes this work was presented at Hot Topics, I believe.
Sylvain Lesne: Definitively yes. I am one of the co-workers of that story.
Gabrielle Strobel: Sylvain, tell us more, please.
Steve Estus: I have heard this and been puzzled. We have preliminary data that tPA XO/Hsiao mice have less Ab, if anything. Still concerned about confounds.
Malcolm Leissring: Hmmmm…. Different genetic backgrounds??
Sylvain Lesne: Well, we have now extra data confirming the role of the tPA/plasmin axis in Ab degradation.
Nilufer Ertekin-Taner: What is the age of your tPA knockouts when they start showing amyloid deposits Sylvain?
Sylvain Lesne: Between 12-14 months old.
Sylvain Lesne: We have preliminary data about plasmin KO mice and we will soon have APP-tPA KO mice at the same age.
Nilufer Ertekin-Taner: Could measurements at different age points possibly explain differences between Steve's and Sylvain's results?
Chris Eckman: Nilufer: You and Steve Younkin have looked at knockouts in this system already. It might be worthwhile to state the data again for the room.
Craig: I may have missed this but are the tPA knockouts on a wild type or APP-transgenic background?
Steve Estus: The tPA knockout mice have a B6 background. Most of the mice in the initial run were F2s, which we ran out to 11 months old to quantify Ab in collaboration with Steve Younkin.
Craig: Thanks Steve, so it is mouse Ab that is depositing.
Sylvain Lesne: As I mentioned to Dennis Selkoe last AD meeting, three mice had clearly at least 30-40 Ab "plaques." Did your mice come from Peter Carmeliet's lab?
Stefan Mansourian: Have you been able to confirm that these are plaques by immunohistochemistry? It is sometimes possible to get spurious "plaques" by thioflavin-T staining...
Sylvain Lesne: Absolutely. I performed IHC with anti-Ab42 specific antiserum.
Stefan Mansourian: Were you able to repeat this with any other Ab42 antiserum, or with Ab40 antiserum?
Wes Farris: Any thoughts on why this would be the only mouse model with deposition of Ab?
Nilufer Ertekin-Taner: In Steve Younkin's laboratory measurements of plasma and brain Ab were done in PLAU knock-out mice. We found significant elevations in plasma Ab in knockout mice compared to wild type. This increase was more enhanced when the mice were aged. We did not see significant elevations in brain amyloid levels.
Sylvain Lesne: We did not check whether tPA KOs were eliciting enhanced plasma Ab levels.
Craig: Did you try mouse versus human Ab-specific antibodies?
Sylvain Lesne: This is currently under investigation in the lab. But these deposits are also immunoreactive to a Ab1-10 directed antiserum, so we are also looking this way.
Wes Farris: Any thoughts on why this mouse model would show deposits of endogenous Ab, unlike the presenilin [model]?
Gabrielle Strobel: Nilufer, that is interesting. Malcolm and Wes, how about plasma Ab and also about insulin levels in IDE knockout mice?
Wes Farris: The elevation of both plasma A-β and insulin are being actively looked into in the IDE KO mice. Preliminary evidence suggests that insulin is markedly elevated.
Gabrielle Strobel: Malcolm and Wes, you wrote about rats with IDE mutations. Do they have AD-like behavior or learning deficits? Or pathology? Or, assuming there are more Ab protofibrils around, LTP impairment, as in the Walsh et al. experiments?
Wes Farris: We have not looked for any of the phenotypes you mentioned yet--just A-b levels.
Craig: Wes, would high levels of insulin compete for IDE with Ab aggregation?
Wes Farris: Craig, I don't know that A-b aggregation is induced by IDE.
Malcolm Leissring: IDE does not induce Ab aggregation. That original finding was later found to be flawed.
Wes Farris: Lou, have you looked at increases in any other IDE substrates in your IDE KO mice?
Malcolm Leissring: So we have one group (Lesne's) that has found deposits and increased A-b in tPA KOs, but two others that have found no differences (Saido--see his comment), or the opposite (Estus').
Sylvain Lesne: I spoke with Professor Saido last July and we were surprised to have such different results.
Craig: What about degradation [of Ab by IDE]?
Douglas Feinstein: Is it known how IDE function is normally regulated? E.g. by phosphorylations?
Lou Hersh: There is no evidence I'm aware of for any regulation of IDE.
Malcolm Leissring: Doug, there are some reports that calcium and ATP may modulate IDE, but the evidence is weak and old...
Craig: So if there are high levels of insulin competing for IDE this would not prevent Ab degradation? Or vice versa?
Wes Farris: One hypothesis that we have is that hyperinsulinemia may cause increased levels of A-β by competing with IDE's degradation of A-b.
Lou Hersh: The problem with insulin and IDE is, where is the insulin and where is the "functional" IDE?
Douglas Feinstein: Is IDE neuronal?
Malcolm Leissring: Functional IDE has been shown to be on the cell surface in numerous studies.
Lou Hersh: Douglas-I don't think IDE has been looked at real carefully in the brain.
Stefan Mansourian: Douglas: Yes, it has been found in cultured neurons and PC12 cells (Vekrellis, 2000) and by IHC in human and mouse neurons.
Lou Hersh: Stefan: I'm not convinced cultured cells reflect the in vivo situation, particularly PC12 cells.
Stefan Mansourian: But we and others (Conrad Talbot, for example) have seen IDE in neurons by immunohistochemistry...
Douglas Feinstein: Lou, Stefan: IDE localization should be done in mice/rats; possibly developmentally, possibly in the mut[ant] APP transgenics.
Malcolm Leissring: Several zinc-metalloproteases that lack signal peptides are found on cell-membranes. Examples are IDE, NRDc, neurolysin, and thimet oligopeptidase, and perhaps several others…If it is a problem for IDE, it is a problem for all these zinc metalloproteases.
Lou Hersh: Malcolm, I seem to recall that the levels of secreted IDE are very low.
Sylvain Lesne: We see a 2-3-fold increase of Ab accumulation in primary cultured neurons from tPA-/- mice (compared to WT). The effect was reversed by adding tPA in tPA deficient neurons.
Steven Estus: Sylvain, that's amazing. Do you have to add exogenous plasminogen to see this effect?
Sylvain Lesne: No, we did not add exogenous plasminogen into the culture media of our cultured tPA-/- neurons. We also confirmed the requirement of plasmin to let this occur by blocking the effect with a2AP. alpha2-antiplasmin(a2AP).
Steven Estus: Sylvain, perhaps you and I are the only ones on this particular thread, but have you had a chance to look at cellular Ab in this paradigm?
Sylvain Lesne: This is a point I want now to look at.
Steven Estus: Sylvain, our data are not necessarily contradictory, in that we are looking at the tPA/Hsiao mice. Have you looked at APP transgenic animals?
Sylvain Lesne: We use the PD-APP mice crossed with our tPA-/- or plasmin-/- mice.
Steven Estus: Sylvain, I misunderstood. Has all of your work been in the PD-APPs?
Sylvain Lesne: Not at all, we used single transgenic animals and we are now using bigenic mice to test whether the lack of tPA would potentiate/exacerbate Ab deposition.
Sylvain Lesne: Professor Estus, did you observe any changes in Ab loads in your bigenic mice (tPA-Tg2576)as compared to same-age Tg2576 mice?
Steven Estus: We observed decreased Ab burden as quantified by ELISA.
Claudia Almeida: Where is Ab accumulating, extracellularly?
Gabrielle Strobel: Claudia, primarily extracellularly but that is in flux somewhat. Increasingly, intracellular Ab seems to be detected inside neurons. Gunnar Gouras had a presentation in Stockholm about age-related increases of intracellular Ab42 in AD-vulnerable neurons. What does this mean for our topic today. Anyone?
Douglas Feinstein: Gabrielle..how do they know they [are] AD vulnerable if they are still there? Perhaps those are the healthier neurons.
Gabrielle Strobel: Douglas, they were neurons from typically affected brain areas, I believe.
Chris Eckman: I agree with Lou. I think it is telling that several different genes linked to Ab degradation all elevate Ab in the knockout animals. These include ECE (1 and 2) , NEP, IDE and PLAU. Is this not the ultimate test? Honestly, I think that there are likely to be regional as well as subcellular/extracellular differences regarding which enzyme is involved, but the knockouts are quite telling.
Wes Farris: Lou, do you believe that IDE is on the cell membrane? Where do you think it sees Ab?
Lou Hersh: Wes, intracellularly. That's where it appears to see insulin.
Malcolm Leissring: I think the case for secreted IDE is more difficult, but the case for cell-surface associated IDE is very clear. It is on the surface of neurons, that's for sure.
Wes Farris: Lou, is it in in membranous vesicles or the cytosol?
Claudia Almeida: Lou, what is the data that IDE sees insulin intracellularly?
Malcolm Leissring: IDE antibodies. IDE antibodies were injected and reported to affect insulin degradation.
Douglas Feinstein: Lou, Wes, so IDE trafficking could play a role in Ab removal?
Wes Farris: Doug, IDE clearly plays a role in A-b degradation in KO mice, if that is what you mean by trafficking. If this is not what you mean, please clarify. Thanks.
Douglas Feinstein: Wes, trafficking, as movement from intracellular organelles to membrane surface. I would think membrane-IDE might be more efficacious to degrade extracellular Ab.
Wes Farris: Douglas, I agree with you about membranous IDE being more relevant, but I would say the jury is still out over whether IDE is on the membrane. Our lab has shown IDE on the surface of neurons, but we as yet don't have a good mechanism as to how it gets there (no transmembrane domain).
Douglas Feinstein: Wes thanks.. I was just looking at an older Selkoe paper concluding IDE mostly works extracellularly.
Gabrielle Strobel: How about glial cells? Do we know what their contribution is to the secretion of Ab-degrading enzymes?
Wes Farris: I don't know of any specific studies of glial degradation, but they clearly could play a role.
Chris Eckman: Very interesting question Gabrielle. The honest answer is we do not have any idea collectively.
Malcolm Leissring: IDE is expressed in and secreted from mouse microglial cells. It seems they need to be primed first with LPS. Intriguingly, activation of purinergic receptors on microglia causes the release of leaderless cytosolic proteins like IL-1β (and IDE) through a mechanism called microvesicle shedding. Perhaps this is one mechanism by which IDE is secreted.
Lou Hersh: Doug. If I remember from Roth's paper, the amount of cell surface IDE he found was vanishingly small.
Stefan Mansourian: Chris: Are ECE-1 and -2 located primarily in neurons or glia?
Nilufer Ertekin-Taner: UPAR (UPA receptor) is found on microglia and its expression was shown to be increased upon treatment with Ab.
Chris Eckman: ECE-1 and 2 are ubiquitous, with ECE-2 mostly in neurons but certainly they could play a role there (in glia?).
Steven Estus: Regarding expression of the plasmin members, plasmin/tPA/PLAU are expressed in neurons. tPA, PAI-1, PAi-2 are also expressed in microglia.
Malcolm Leissring: Chris, doesn't ECE-1 suffer the same problem of not having a signal peptide?
Lou Hersh: In terms of IDE trafficking we know that some is targeted to the peroxisomes through a C-terminal signal.
Malcolm Leissring: Lou, yes, you make a good point. IDE is known to be trafficked across membranes--definitely into peroxisomes, so why not into other compartments or onto the cell surface??
Douglas Feinstein: Malcolm, hence, increasing IDE levels should be 'protective'?
Malcolm Leissring: Yes, that's the theory at least.
Chris Eckman: ECE is clearly a transmembrane protein and ECE catalyzed Ab degradation is optimal at a slightly acidic pH, indicating intracellular to me.
Malcolm Leissring: Chris, Which ECE?
Lou Hersh: Malcolm-I don't think the peroxisomal targeting sequence gets proteins to the cell surface.
Malcolm Leissring: IDE also gets to mitochondria.
Lou Hersh: Malcolm, where's the evidence for that?
Malcolm Leissring: Lou, I haven't published it yet, but we have found that the initial 41 amino acids of IDE encodes a mitochondrial targeting sequence; EGFP fusions containing this sequence are targeted to mitochondria very efficiently.
Gabrielle Strobel: Are any of those proteases good drug targets, given how promiscuous they seem to be and that we would want to increase their function?
Chris Eckman: Very good question Gabrielle. The real answer is we do not know. Clearly overexpression of some of these has been tried and the mice seem OK. Remember though, that all of these can certainly also do other things. The real tests still need to be done. Can overexpression of any enzyme (physiological or not) result in decreases in Ab that are therapeutically useful? I know several of us have this target in our sights and I am certain we will be hearing more in the future.
Nilufer Ertekin-Taner: Gabrielle, your point about potential drug targets is interesting. If we could detect the people who develop AD because they have "defective" forms of these enzymes, then one could think about compensating for their defect, as opposed to overexpressing them. Ultimately, genetic tests that could develop as a result of the current research could help us identify such individuals, who would be candidates for treatment.
Douglas Feinstein: Nilufer... the possibility is that the proteases are not defective per se, but either their subcellular localization, modifications, etc.. so genetics wouldn't necessarily work.
Nilufer Ertekin-Taner: Douglas. What I meant by defective is not simply underexpression but includes all of the potential problems with folding trafficking etc. you are referring to. If we can find the genetic mutations leading to these problems, then we have a test for identifying individuals at risk.
Douglas Feinstein: Nilufer...I absolutely agree
Gabrielle Strobel: Are there drugs to induce the overexpression of proteases?
Malcolm Leissring:We're working on it.
Malcolm Leissring: On the therapeutic side, what does everyone think of the reports that NEP is upregulated by injection of Ab?
Chris Eckman: The data look decent, but as with everything, a confirmation would be good.
Lou Hersh: I agree with Chris. This is surprising and needs to be replicated. We haven't seen NEP upregulated in Ab mice.
Malcolm Leissring: There is another report that says that injection of vehicle alone causes the same effect.
Chris Eckman: The ascorbic acid paper?
Wes Farris: Yes, Chris, that is the paper.
Malcolm Leissring: Yeah, I find it confusing that Ab injection causes tangles in one model and prevents plaques in another.
Chris Eckman: Many people have tried Ab injections over the years to create AD models. The reality is that it never worked, presumably because it was cleared, by what I do not know. Remember there are a wealth of other ways to remove peptides in addition to direct catabolism.
Malcolm Leissring: Good point, Chris.
Gabrielle Strobel: We are nearing the end of the hour. Please continue chatting for as long as you like but before people begin dropping out I want to thank everyone for coming and for this fun discussion. Also, those who have not corresponded with Nico before, please send me your email addresses to firstname.lastname@example.org so I can send you the edited transcript for your review. And please send me citations you may have mentioned. Thanks everyone!
Malcolm Leissring: I'd like to thank everyone for coming to the chat.
Wes Farris: Nilufer, what is your latest thinking about the chromosome 10q linkage?
Nilufer Ertekin-Taner: Wes. Evidence from multiple laboratories indicates the existence of a locus (loci) on chromosome 10 that is linked with risk for AD. Our evidence suggests that this locus/loci act through Ab. Multiple genes reside under the peaks, and we need to sort them out by both genetic and functional studies. I do think it possible that multiple genes exist on chromosome 10q that yield these signals.
Gabrielle Strobel: Is this cluster of three possible LOAD risk factor genes, then, why this peak is so large? Are other genetic risk factors for LOAD more scattered throughout the genome and therefore harder to find?
Nilufer Ertekin-Taner: Gabrielle. Multiple factors could cause the linkage peak to be large or narrow (study size, recombination information, heterogeneity). The existence of multiple genes in a region will not necessarily lead to narrower peaks, quite the contrary.
Wes Farris: Nilufer, what are your best candidate genes (that are public) to date?
Nilufer Ertekin-Taner: Wes, we presented our data on two genes at WAC, PLAU and VR22. Variations in the former are associated with LOAD in our case-control series and AB in the families but do not account for the linkage signal. Variations in the latter account for a substantial proportion of our linkage signal, however.
Douglas Feinstein: This may be a repeat question…. But has anyone examined if inflammatory stimuli decrease IDE, or other putative degrading enzymes?
Wes Farris: Douglas, LPS increases secretion of IDE, but I don't think that inflammatory stimuli decrease IDE, nor have I heard that this has been studied.
Malcolm Leissring: Doug, to my knowledge I have not seen much work looking at the IDE/inflammation connection, but the possibility that cytokines and IDE might be secreted by similar mechanisms from microglia makes it an interesting possibility.
Douglas Feinstein: I was thinking along the lines of the a cyclic phenomenon in which low amounts of amyloid--> inflammatory --> decrease IDE etc.
Paul Shapiro: Are there any decent commercially available antibodies for IDE?
Wes Farris: Paul, not yet, but I heard that Richard Roth may make his monoclonal commercially available.
Gabrielle Strobel: How about MMPs? Takaomi Saido mentioned them in his second comment posted earlier today. There are MMP inhibitors that, I believe, failed in anti-angiogenesis trials. Would it make sense to try them in AD models?
Malcolm Leissring: That would seem to cause AD, Gabrielle!!
Gabrielle Strobel: Yikes, of course.
Malcolm Leissring: Perhaps we should look into that, though!
Wes Farris: Nilufer, unfortunately I missed the international conference this year. What is VR22? I heard you may have found some allelic association to IDE in some families--is this correct?
Nilufer Ertekin-Taner: Wes, VR22 is a novel a-t-catenin. We do not see an effect for the IDE polymorphisms we tested in the families.
Gabrielle Strobel: Wes, check out the news coverage of Nilufer's excellent talk on the ARF website; it has a short summary.
Gabrielle Strobel: That leads me to repeat my question: Are there drugs to induce expression of these proteases? At that Jamaica meeting, Mike O'Neal talked about a compound made for another purpose that turned out, unexpectedly, to induce BDNF expression...
Malcolm Leissring: I'm afraid that the bench beckons me...I must bid you all adieu!! Signing off... thank you everyone for coming!!
Chris Eckman: Thanks for inviting me to this chat. I think it is great to finally see people really interested in Ab removal. I honestly believe that this will result in the identification of additional individuals at risk and may someday result in new therapeutics.
Nilufer Ertekin-Taner: I need to get going, too. Thanks everybody.
Gabrielle Strobel: Bye Nilufer, thanks for coming.
Wes Farris: Thanks, Nilufer. Thanks to everyone for participating. Bye!
Douglas Feinstein: Nice discussion, bye.
Detlef Schmicker: Thanks everybody, Detlef.
Gabrielle Strobel: Good bye everyone, and join us next time. We will post a transcript of this discussion.
Steven Estus: Thanks for organizing this chat. I have to run now, but it has been informative. If anyone wants to talk more later, please feel free to write.
Gabrielle Strobel: Thanks Steve, I am sure Sylvain will...