Posted 23 August 2006

E-mail discussion Printable version

Return to Discussion Text

Stefan Lichtenthaler led this live discussion on 17 December 2003.

Participants: Stefan Lichtenthaler, University of Munich; Gabrielle Strobel, Alzheimer Research Forum; Tobias Hartmann, University of Heidelberg; Dave Teplow, Brigham and Women's Hospital; Lit-Fui Lau, Pfizer, Inc.; Michael Irizarry, Massachusetts General Hospital; Gil Westmeyer, Ludwig-Maximilians University, Munich; Yong Shen, Sun Health Institute; Mike McDonald, Vanderbilt University; Larry Nault, Matrix Integrated Linkage; Terry Reisine, ALSP; Huaxi Xu, Burnham Institute, San Diego; Dane Liston.

Note: The transcript has been edited for clarity and accuracy.

Stefan Lichtenthaler
Hi, Dave, good to see you.

Dave Teplow
Hi, Stefan, looking forward to your illuminating insights.

Gabrielle Strobel
Hi, Susanne and Bryce, and welcome. Hello, Lit-Fui, great to have you! And welcome, Michael! All, Michael reported at New Orleans that BACE activity goes up with age in all areas tested of mouse and human brain. Correct me if I'm wrong, Michael.

Perhaps we could start by Stefan stating briefly what he thinks BACE's physiological function may turn out to be, and how this ties in with the immune system? Perhaps not all present here today have had a chance to read your latest paper yet. Everyone, while Stefan is typing in his reply, just a reminder that this is an open forum where you are welcome to enter your questions and thoughts freely.

Stefan Lichtenthaler
At this point, the function of BACE is still very speculative. It may well start the degradation of membrane proteins, which would fit well with its cellular localization in the endosomes.

Gabrielle Strobel
Stefan, degradation or activation?

Stefan Lichtenthaler
On the other hand, BACE may play a modulatory role for the function of the ADAM proteases. For example, APP secreted by α-secretase seems to have a neurotrophic function, whereas the BACE-cleaved APP does not seem to have this function. Thus, the decision of whether APP is cleaved by α- or β-secretase may have different functional outcomes.

Gabrielle Strobel
What drives this decision?

Stefan Lichtenthaler
Similarly, for our new BACE substrate, PSGL1, we observed an ADAM-cleavage as well as a BACE-cleavage, although in that context, it still remains unclear whether this leads to different functions of the secreted protein.

Gabrielle Strobel
Everyone, lest I grill Stefan, please feel free to chime in. This is informal.

Dave Teplow
How much of the BACE cleavage is occurring at the cell surface and how much after internalization?

Stefan Lichtenthaler
Under circumstances where BACE is not overexpressed, most BACE cleavage should occur after internalization and not at the cell membrane.

Dave Teplow
If so, then how much of BACE's function is related to regulatory cleavage of the APP ectodomain?

Stefan Lichtenthaler
At the cell surface, the membrane protein has a chance to undergo cleavage by an ADAM protease. But upon internalization (perhaps because the protein is not needed anymore at the cell surface) it can undergo cleavage by BACE.

Tobias Hartmann
Is there any BACE cleavage on the cell surface, given a pH close to 7?

Stefan Lichtenthaler
Tobias, Dennis Selkoe recently had a JBC paper showing that BACE cleavage may occur at the cell surface if general endocytosis is inhibited. He used a dominant-negative dynamin mutant. Dave, in your opinion, where does BACE cleave?

[Editor's note: See Chyung and Selkoe, 2003.]

Tobias Hartmann
In your estimate, what percentage of intracellularly BACE cleaved AβPP will become secreted, rather than internally digested. My observation, when I compare p3 and Aβ levels with AαPP and AβPP levels, is that most AβPP never gets secreted.

Dave Teplow
Teleologically, why would the cell require a BACE-like activity in endosomes if these organelles continue to mature into lysosomes, in which other, less specific endoproteases can function?

Tobias Hartmann
Dave, that's the point I want to get at; maybe most AβPP is never designed to reach the outside, while AαPP has a function—at least in neurite sprouting, which would be an important clue for other substrates.

Stefan Lichtenthaler
Tobias, would you say that the AβPP is further degraded in the endosomes or lysosomes?

Tobias Hartmann
I very much guess so; it has to end up somewhere. Since it disappears, lysosomes are one option that is tempting.

Dave Teplow
Addressing Tobias's question, is anyone aware of x-ray crystallography studies of the C-terminus of the APP ectodomain? These might reveal how β- or α-cleaved ectodomains could function differently.

Stefan Lichtenthaler
I am not aware of a C-terminal structure.

Tobias Hartmann
X-ray, not that I know of, but the N-terminus of Aβ—the difference between BACE and α cleavage—is very flexible with no specific structure. Whatever this might be, the possibility has to be considered that AαPP and AβPP (or Aβ/p3) act together as uneven signaling molecules, possibly working in concert with γ-CTF nuclear signaling. Anything known about NTF/CTF (nuclear) signaling known from other substrate candidates?

Stefan Lichtenthaler
BACE in the endosomes could contribute to the degradation of a membrane protein. But, in contrast to a lysosomal protease, it would allow the CTF to move to the nucleus.

Dave Teplow
This is a very interesting comment, Stefan. Maybe I'm looking at the wrong end, and differences in physiological function reside at the N-terminus of the CTF!

Gabrielle Strobel
How about a function in synaptic transmission? Colin Dingwall and others reported last month on a set of corresponding BACE1 transgenics, one expressing BACE1 only in neurons and the other a knockout. It was shown that transgenics exhibited a bold, exploratory behavior and show elevated 5-hydroxytryptamine turnover, while the knockouts were timid and less exploratory. Both lines were viable, fertile, normal morbidity. This suggests an unexpected role for BACE1 in neurotransmission, perhaps through changes in APP processing and Aβ levels, the authors write. What do you all think?

[Editor's note: See Harrison et al., 2003.]

Stefan Lichtenthaler
Gabrielle, this brings us to the question of additional BACE substrates. I am convinced that we have overlooked some phenotypic changes in BACE knockout mice, since we haven't looked in the right way or, in other words, not asked the right question yet. However, it is clear that BACE cannot have an essential function in development.

Lit-Fui Lau
Has anybody looked at the BACE KO under stressed conditions?

Stefan Lichtenthaler
Lit-Fui, I don't know any study where the knockouts were stressed.

Stefan Lichtenthaler
How about this? BACE does not have one preferred substrate, because then we could expect a more prominent phenotype in the BACE knockout mice.

Tobias Hartmann
Last comment, because I have to leave, very much in favor of your suggestion, Stefan.

Gil Westmeyer
Okay—I see that the discussion is going into one specific direction—but does no one believe in the physiological function of Aβ? You cannot have this without BACE! Why don't we discuss the hypothesis put forward by Stefan whether BACE is involved in immunologic processes. From the discussion so far, one could think that we don't know anything—but that is not the case, i.e., sialyltransferase PSGL1.

Mike McDonald
What do we know about BACE and sialyltransferase?

Stefan Lichtenthaler
The sialyltransferase has been shown by the Hashimoto lab to be a substrate for BACE.

[Editor's note: See Kitazume et al., 2001; Kitazume et al., 2003.]

Yong Shen
Did they find it in AD brains?

Stefan Lichtenthaler
We have shown that PSGL1 is a substrate for BACE. Both, PSGL1 and the sialyltransferase are proteins with an immunological function. That led us to speculate that a BACE knockout may have an immunological phenotype.

[Editor's note: See Lichtenthaler et al., 2003.]

Gabrielle Strobel
Perhaps I should add that a group from Merck and the University of Illinois, Chicago reported in New Orleans that sialyltransferase ST6Gal1, in their in-vitro experiments with a BACE inhibitor, did not appear to be a BACE substrate.

Michael Irizarry
Stefan, BACE appears to cut APP at both Asparagine-1 and Glutamine-11 of Aβ. Is it known if PSGL1, or ST6Gal1, or the sodium channel have more than one BACE cleavage site? What do other people in the discussion think of the significance of the Glu11 site of cleavage of APP by BACE?

Stefan Lichtenthaler
Michael, for PSGL1 and ST6Gal1, only one cleavage site has been identified and there is no evidence for a second one. For the sodium channel, I haven't seen the data yet.

Gabrielle Strobel
All, for clarification: the sodium channel some are referring to here regards Dora Kovacs' finding, reported at the Neuroscience meeting, that a voltage-gated sodium channel subunit is cleaved by BACE. Correct me if I am wrong. (This would fit with the Dingwall paper, too, would it not?)

Michael Irizarry
Gabrielle, you are correct; the BACE substrate reported by Dora Kovacs is the sodium channel subunit.

Dave Teplow
Stefan, this "immunological phenotype" idea is interesting, and as an ex-immunologist, I can say that studies of the effects of BACE KOs early in life may reveal nothing because the threshold of immunological deficits may not occur until late in life.

Gabrielle Strobel
Let's talk a little more about the mouse models, if we can. γ-secretase knockouts were embryonic lethal, telling us it's indispensable; BACE knockouts are viable, telling us it's dispensable. Those are crude first answers (and, of course, we know much more about γ-secretase by now). My questions: What would be the ideal mouse models to find out BACE's biological function? This is the time for wish lists, after all. Anyone?

Yong Shen
Gabrielle, maybe conditional transgenic and mutant APP transgenic.

Gabrielle Strobel
Yong, the Dingwall mice are conditional in that they express human BACE 1 driven by the CAMK promoter, only in neurons.

Gil Westmeyer
Yong, it seems probable that APP is not the physiological substrate for BACE, so APP transgenic may not help. I think it is just a question of characterizing the existing mice.

Dave Teplow
Michael, one reasonable interpretation of our sequencing data is that the alternative cleavage sites are used as a result of a regulatory phenomenon mediated by the interaction of APP and BACE with other proteins modulating their function.

Stefan Lichtenthaler
Dave, any idea about such other proteins?

Dave Teplow
Stefan, I wish I did, but that would simply be "hand waving."

Terry Reisine
Stefan, what are the publications showing that BACE has poor activity in cutting wild-type APP?

Stefan Lichtenthaler
Terry, for example, the Gruninger-Leitch paper in JBC. They used in-vitro experiments with all different kinds of synthetic peptides.

Terry Reisine
Stefan, if BACE is very poor at cutting wild-type APP, then what is the enzyme involved in cutting APP in most patients with AD?

Gil Westmeyer
Terry, I would answer that BACE is still the best candidate; it is just that AD is a chronic disease in which accumulative effects do count!

Terry Reisine
Gil, it might be that BACE is the only target accepted and the one most studied, but not necessarily the best one.

Gabrielle Strobel
Everyone, it strikes me as vaguely contradictive that BACE cleavage of APP is strictly required for Aβ generation. Aβ is made by most cells of the body throughout life, yet BACE gets sorted to different cellular compartments and cleaves APP poorly. How can that be?

Stefan Lichtenthaler
What about this? BACE cleaves a large number of different proteins, but only to a small extent. Like APP and PSGL1, these proteins may mainly be cleaved by ADAM-proteases. This could be the reason that an ADAM knockout is severe, whereas a BACE knockout only shows subtle effects.

Terry Reisine
Stefan, if what you say is correct, then making drugs to block BACE could cause many problems, especially if such drugs need to be used over long periods of time.

Gabrielle Strobel
Very interesting, Stefan. Then what happens in AD? Maybe AD-related changes (in translation, sorting, localization, recycling, perhaps) would make an even safer target than BACE itself? Or is it simply the increased activity, as Mike has found?

Michael Irizarry
Gabrielle, sorting, localization, recycling are such broad processes that attempting to target these therapeutically would likely affect many other cellular processes. The hope would be that selective BACE inhibitors would only affect β-secretase processing of APP, and minimally affect other BACE-related cleavages, perhaps by selectively affecting BACE in brain, or BACE in neurons or glia.

Stefan Lichtenthaler
Terry, sure, that needs to be considered. However, it is clear that under normal, e.g., non-stressed conditions, the BACE knockout seems to be phenotypically nearly normal. In addition, a BACE-directed therapy would not need to inhibit BACE completely and yet have an Aβ-lowering effect over the longer time.

Dave Teplow
Stefan et al., I can't help it. I need to bring biophysics in here. The comment of Stefan's is excellent, because it points out the fact that small changes in Aβ concentration can result in huge differences in clinical status. This is the result of the strong concentration dependence of the assembly reaction of Aβ.

Gabrielle Strobel
Dave, would this imply that a mild inhibition of BACE would be sufficient, and that there is a greater "therapeutic index or window" with BACE than with γ-secretase?

Dave Teplow
Gabrielle, I think your suggested implication is reasonable.

Gil Westmeyer
Terry, tell me which one you mean—why not study the usual suspect, especially since there is no overt phenotype? As Stefan has just pointed out, small therapeutic alteration might suffice.

Stefan Lichtenthaler
Terry, just to add some more. I still think that BACE is an excellent drug target, but the identification of additional BACE substrates or additional BACE-caused phenotypic changes may allow us to see what we should look at in a patient when testing BACE inhibitors.

Gil Westmeyer
Just to add: A lot of side effects can be controlled. That is the everyday life of clinicians. It could be a lot worse!

Dane Liston
Does anyone have a guess as to how much reduction in Aβ levels will be needed to have therapeutic effect (reduce deposition). This should apply to β- and γ-secretase.

Dave Teplow
Dane, I think your question assumes that "deposits" are a key factor in disease pathogenesis. I would dispute this.

Terry Reisine
Stefan, you are correct in assuming that effects on other proteins being cleaved by BACE over a long time don't cause major problems. Also, as Dane says, there is little information linking the molecular endpoint, reduction of Aβ production, with clinical improvement.

Dane Liston
Dave, agreed, but they may provide a quantifiable marker of activity/effect.

Dave Teplow
Dane, yes, and this would be very valuable diagnostically and prognostically.

Stefan Lichtenthaler
Terry, what is your opinion on BACE being the β-secretase? What other enzyme would you have in mind?

Terry Reisine
Stefan, there is evidence of cysteine proteases that are able to cleave wild-type APP better than Swedish mutant APP. Whether or not they are better candidates than BACE is still not established, because the in-vivo work has not been done.

Stefan Lichtenthaler
Terry, I agree. In that recent paper they should have used BACE knockout mice to validate their hypothesis.

Gabrielle Strobel
To pick up Terry's point about Aβ reduction and clinical improvement, did you all see the data on BACE knockout/PD-APP mice from the Elan team, which had no amyloid deposition or Aβ, but still had memory deficits!?

Stefan Lichtenthaler
Gabrielle, concerning the PD-APP mice crossed to the BACE knockout: If they have memory deficits although there is no BACE cleavage, would you assume, that the missing sAβPP is causing the memory deficit?

Dave Teplow
Gabrielle, this is related to the earlier question about CTF function in controlling neuronal physiology (and potentially, viability).

Gabrielle Strobel
Dave, can you elaborate a bit? You mean a CTF from α cleavage?

Dave Teplow
Gabrielle, precisely, or from other regulated cleavage events.

Gabrielle Strobel
Okay, thanks.

Gabrielle Strobel
Stefan, I don't know; I just think this is really intriguing. Your suggestion is one possibility, and might suggest that BACE's normal role is to release a biologically active sAβPP, as you posited. Or perhaps it is just a result of the overexpression of mutant foreign APP in a mouse, a sort of artifact.

Stefan Lichtenthaler
Gabrielle, that's why we need to see the data. The study you referred to was presented as a poster at SFN, right?

Gabrielle Strobel
Stefan, yes. And other BACE knockout/Tg2576 crosses reported improvement of the memory deficit, so it is not clear, I should add….

[Editor's note: The first of these papers has since appeared in print, see ARF related news story.]

Huaxi Xu
Gabrielle's point is a good one. People should do a better control experiment (e.g., overexpress other proteins).

Gabrielle Strobel
As we are nearing the end of the hour, let me feed in a question from the intro text that I believe we have not addressed yet: In AD, what factors "divert" BACE1 from its physiological substrates and toward more APP cleavage? Stefan, anyone?

Gil Westmeyer
Maybe we can include in our discussion which effect cholesterol levels might have on the activity of BACE. Of course, sporadic cases should be the focus of our interest! What about combining the knowledge we have about risk factors and the secretases?

Gabrielle Strobel
Please do—another topic that has not gotten its due so far. More, anyone…?

Stefan Lichtenthaler
Gil, would you suggest to use statins to target both γ- and β-secretase?

Dave Teplow
Gil, this [cholesterol] could be one of the "proteins" I implied could exert a controlling effect on BACE cleavage (or that of other secretases).

Gabrielle Strobel
Brain cholesterol is not necessarily increased in AD, Gil. Wrong?

Gil Westmeyer
Dave, I was just pointing to the fact that all known players in AD can be found in lipid rafts, and that APP processing might be correlated with this "compartment." I do not know any specific study, however, about cholesterol in the brain.

Dave Teplow
Gil, an excellent point!

Gabrielle Strobel
We have reached the end of the hour. Everyone is welcome and encouraged to continue on as long as they like. But before people start dropping out, I want to thank Stefan for his excellent background text and insight here today, and all of you for coming and contributing. We will circulate a transcript and hopefully have much more data to discuss in a follow-up chat in a year or so. The chat software remains open.

Gil Westmeyer
Thanks, everyone—and don't get too old!

Larry Nault
Thanks, everyone and happy holidays.

Stefan Lichtenthaler
Thanks very much, Gabrielle, for organizing the chat. In a year's time, we will likely have more substrates and can better judge the biological function of BACE.

Gabrielle Strobel
Happy holidays, and schoene Weihnachten to the German contingent.