The γ-secretase complex continues to tantalize researchers seeking to treat Alzheimer’s disease. Since this enzyme finishes the job of snipping APP into Aβ, interfering with it could lower amyloid and perhaps delay disease onset. Even as γ-secretase inhibitors have faced setbacks due to side effects (see, e.g., ARF related conference story), γ-secretase modulators (GSMs) have gained luster as a potentially safer approach. But exactly how GSMs work and even to what protein they bind remain burning questions, and it appears the answers may be different for different classes of GSMs. In the September 6 Journal of Biological Chemistry, researchers led by Harald Steiner and Christian Haass at Ludwig-Maximilians University, Munich, Germany, and Karlheinz Baumann at F. Hoffmann-La Roche Ltd., Basel, Switzerland, report that their second-generation GSMs bind directly to the N-terminal end of presenilin, the enzymatically active protein in the γ-secretase complex. However, only some GSMs from other groups show the same behavior in their assay. The authors first presented this data last July at the 2011 Alzheimer’s Association International Conference in Paris, France. It follows similar reports at AD/PD 2011 in Barcelona, Spain, from several groups who also showed that their second-generation GSMs directly bind presenilin (see ARF related news story).

The findings contrast with the sometimes-confusing literature on first-generation GSMs. These molecules belong to the class of non-steroidal anti-inflammatory drugs (NSAIDs) and include such compounds as sulindac sulfide and flurbiprofen. Some researchers have reported that NSAID-based GSMs lower Aβ by acting on the substrate, APP, rather than on the enzyme complex (see ARF related news story on Kukar et al., 2008 and ARF related news story on Richter et al., 2010). However, a later study questioned this finding (see Beel et al., 2009), and other papers suggest these GSMs may act directly on presenilin (see ARF related news story on Lleó et al., 2004; Eriksen et al., 2003; Takahashi et al., 2003; and Beher et al., 2004).

Second-generation, higher-potency GSMs, under development by several groups, are structurally distinct from NSAID-based compounds, and may have different targets. For example, TorreyPines Therapeutics, Inc., which has since folded, reports that its second-generation compounds primarily bound to the Pen2 protein component of the γ-secretase complex, and to a lesser extent to presenilin-1 (see ARF related news story on Kounnas et al., 2010).

Steiner and colleagues investigated the binding specificity of their second-generation GSMs, RO-02, RO-03, and RO-57. These are bridged aromatic aminopyrimidine-derived compounds that are potent in the low nanomolar range. To identify the molecular target, first author Amelie Ebke created a labeling compound by adding photoactivatable and biotin groups to the end of RO-57. The modified molecule, dubbed RO-57-BpB, was able to lower Aβ in cell culture, albeit with lower affinity than unmodified RO-57. When Ebke and colleagues irradiated the cell culture with UV light, the photosensitive end of RO-57-BpB bound covalently to the target molecule. The authors then pulled down the target using streptavidin, which binds biotin. Importantly, high amounts of RO-57, RO-02, and RO-03 competed with RO-57-BpB and reduced labeling, demonstrating that the modified molecule was binding its normal target. Over several experiments, the authors found that RO-57-BpB pulled down the N-terminal fragment of presenilin-2, and less strongly the N-terminal fragment of presenilin-1, but did not capture any other γ-secretase subunits or APP.

One of the advantages of this system is that it looks at binding in native cellular conditions, rather than in dissociated membranes, Baumann told ARF. Another plus is that the competitive assay confers specificity, he added. Other scientists praised the paper’s methodology. “The biochemistry in this paper is outstanding,” noted Joanna Jankowsky at Baylor College of Medicine, Houston, Texas.

The authors also tested several other classes of second-generation GSMs. They found that two GSMs under development at the pharma company Eisai (E2012 and ES2), which are structurally similar to RO-02, efficiently competed against the labeling molecule, but only one of the two TorreyPines compounds did. A second-generation GSM developed by Merck, GSM-1, competed well for presenilin-2 labeling but only weakly affected presenilin-1, suggesting the binding site differs between the two presenilins. In addition, Ebke and colleagues found that a few first-generation GSMs, including sulindac sulfide and fenofibrate, showed some binding to presenilin. In contrast, various γ-secretase inhibitors did not compete with RO-57-BpB, indicating GSMs do not bind at the enzyme’s catalytic site. This suggests that GSMs act allosterically, i.e., by altering the enzyme’s shape rather than directly interfering with the catalytic site.

The differential results highlight the idea that not all GSMs act in the same way. “Ultimately, I think the truth will be that there will be flavors of GSMs that interact with substrate, substrate and presenilin/γ-secretase, and combinations of these with lipid membrane interactions,” Todd Golde at the University of Florida, Gainesville, wrote to ARF (see full comment below).

The next step is to pin down the exact amino acid residues that second-generation GSMs are binding to, Baumann told ARF. This knowledge would enable researchers to develop molecular models of GSM action and ultimately use that information to design more potent drugs. Jankowsky agreed, telling ARF, “The more we know about where different modulators bind, the more we will understand about the mechanism by which they act. We then can go in and do rational drug design, so that we end up with something that is the third-generation γ-secretase modulator.”—Madolyn Bowman Rogers


  1. My take is that this is fundamentally sound but that some GSMs clearly interact with substrate. Ultimately, I think the truth will be that there will be flavors of GSMs that interact with substrate, substrate and presenilin/γ-secretase, and combinations of these with lipid membrane interactions. Indeed, competition studies show only partial competition.

    View all comments by Todd E. Golde
  2. This is an intriguing paper suggesting that highly potent, second-generation GSMs directly target the N-terminal fragment of presenilin (PS), the catalytic engine of the γ-secretase complex. After the possibility was raised that NSAID-based GSMs target the substrate, several studies regarding the molecular action of GSMs have been reported. However, the specificity of these compounds was questioned, as the first-generation GSMs showed very low potency (over 100 microM for IC50). Moreover, these GSMs affected the structure of PS revealed by FLIM (Lleó et al., 2004).

    Here, Ebke et al. in collaboration with Roche used a novel photoaffinity probe with excellent potency and clearly showed that PS is a direct target. This is consistent with the study by TorreyPines (Kounnas et al., 2010) that immobilized GSMs bound with the PS1-NTF and Pen-2. Moreover, we also reported that GSM-1, which is an acidic phenylpiperidine GSM with excellent potency, directly targets to PS NTF (Ohki et al., SfN abstract, 2010).

    Intriguingly, these compounds showed different modulation properties against γ-cleavage; Roche/TorreyPines GSMs increased and decreased Aβ37/38 and Aβ40/42, respectively. In contrast, GSM-1 only affected Aβ38 and 42 generating activity. However, both types of GSM directly target PS1-NTF. Notably, these compounds share the binding site within the PS1-NTF; this manuscript and Ohki et al., unpublished observation). Collectively, these data suggest that the structural change in PS-NTF is critical to the modulation of the γ-secretase activity.


    . Nonsteroidal anti-inflammatory drugs lower Abeta42 and change presenilin 1 conformation. Nat Med. 2004 Oct;10(10):1065-6. PubMed.

    . Modulation of gamma-secretase reduces beta-amyloid deposition in a transgenic mouse model of Alzheimer's disease. Neuron. 2010 Sep 9;67(5):769-80. PubMed.

    View all comments by Taisuke Tomita
  3. Comment by Sascha Weggen, Thorsten Jumpertz, Andreas Rennhack, Bruno Bulic
    Despite substantial progress in the chemical and preclinical development of γ-secretase modulators (GSMs), the molecular mechanism and target of GSMs have remained controversial. The first GSMs were discovered in the class of non-steroidal anti-inflammatory drugs (NSAIDs) and displayed only low potency in the micromolar range. Subsequently, GSMs with nanomolar potency and favorable pharmacological properties have been reported in two major structural classes: carboxylic acids with structural similarities to NSAIDs (acidic GSMs), and compounds based on bridged aromatics that do not resemble NSAIDs and lack a carboxylic acid group (non-acidic GSMs). Earlier binding studies with a photo-probe based on the low-potency NSAID GSM flurbiprofen suggested targeting of the substrate APP (Kukar et al., 2008). In contrast, affinity purification studies with a potent non-acidic GSM identified a GSM binding site in the γ-secretase subunit PEN2 (Kounnas et al., 2010). However, these last experiments were performed under detergent conditions that do not support γ-secretase activity, and it was not investigated whether the affinity ligand preserved its GSMs activity after immobilization. Importantly, a number of prior observations had strongly suggested a primary binding site of GSMs within presenilin (PS), the catalytic subunit of γ-secretase (Zettl et al., 2010).

    Ebke et al. have now synthesized a photo-probe based on a series of potent non-acidic aminopyrimidine GSMs. This photo-probe retained GSM activity with an IC50 of approximately 0.5 μM in a cell-free assay using partially purified γ-secretase. Photo-affinity labeling studies with membrane fractions from HEK293 cells showed that this photo-probe bound the N-terminal fragment of PS. Binding was not observed after omission of UV irradiation or after co-incubation with an excess of a parental compound, demonstrating specificity of the labeling. None of the remaining γ-secretase subunits or the substrate APP was labeled by the photo-probe. In competition studies with other, structurally different non-acidic GSMs, some, but not all, compounds reduced binding of the photo-probe to PS, hinting at some variability in the GSM binding site.

    As pointed out by the authors, competition studies are not always straightforward to interpret, and might be affected by physicochemical properties of the compounds, e.g., solubility. In competition studies with a single potent acidic GSM, reduced binding of the photo-probe to PS2 but not PS1 was observed. These ambiguous results do not resolve the issue of whether acidic and non-acidic GSMs target the same binding site within PS. In conference proceedings, Taisuke Tomita and Takeshi Iwatsubo have previously reported binding of an acidic GSM photo-probe to the N-terminal fragment of PS. Using photo-affinity labeling in living cells, we have also confirmed binding of potent GSMs to PS.

    Taken together, these studies clearly establish PS as the molecular target of second-generation GSMs. Whether these findings can be extended to GSMs in the class of NSAIDs remains unclear. Ebke et al. have performed competition studies with the NSAID GSM sulindac sulfide, and have observed strongly reduced binding of the non-acidic GSM photo-probe indicative of overlapping binding sites. As discussed by the authors, these results are surprising, given the at least 100-fold lower potency of sulindac sulfide as compared to the photo-probe. In addition, competition experiments are generally not able to distinguish between direct and allosteric competition for a specific binding site. Finally, the specificity of sulindac sulfide, first known as a COX-binder, is questionable. The indene scaffold in sulindac sulfide appears as a frequent hitter in various high-throughput screenings, suggesting the possibility of multiple binding sites within the γ-secretase complex. Therefore, it remains possible that sulindac sulfide has a primary binding site within the substrate APP and reduces binding of the photo-probe to PS through allosteric interaction. However, because of the superior properties of the photo-probes based on second-generation GSMs, the overall structural similarities between NSAID GSMs and second-generation acidic GSMs, the prior evidence from cell biological studies indicating a primary binding site of NSAID GSMs within PS, and the new evidence from competition experiments by Ebke et al., we regard this possibility as less likely. In any case, while NSAIDs have been tremendously useful to establish the fundamentals of γ-secretase modulation, it appears that it is time to move on and concentrate on the molecular mechanism of potent and clinically relevant second-generation GSMs.


    . Modulation of gamma-secretase reduces beta-amyloid deposition in a transgenic mouse model of Alzheimer's disease. Neuron. 2010 Sep 9;67(5):769-80. PubMed.

    . Substrate-targeting gamma-secretase modulators. Nature. 2008 Jun 12;453(7197):925-9. PubMed.

    . Exploring the chemical space of gamma-secretase modulators. Trends Pharmacol Sci. 2010 Sep;31(9):402-10. PubMed.

    View all comments by Bruno Bulic
  4. Amelie Ebke and colleagues' new report on the binding of γ-secretase modulators (GSMs) directly to presenilin is a big step in the understanding of how these small molecules can modulate the enzyme's activity.

    This study definitively identifies the N-terminal fragment of presenilin (PS-NTF) as the binding site of GSMs, confirming previous studies that had identified γ-secretase more generically as the binding site of those compounds. As for the previous reports of carboxylic acid GSMs (sulindac sulfide, flurbiprofen) binding to APP, recent studies seem to confirm this interaction (Sagi et al., 2011; Kukar et al., 2011) but the report by Ebke and colleagues also shows that sulindac sulfide can compete binding of the second-generation non-acid GSM photoprobe on PS-NTF. This apparent discrepancy might actually be the result of binding promiscuity of sulindac sulfide and could potentially account for its lower potency.

    Interestingly, Taisuke Tomita and Takeshi Iwatsubo from the University of Tokyo, using a similar approach (as reported at the last AD/PD meeting in Barcelona) also identified PS-NTF as the binding site of a GSM. Their probe was derived from the carboxylic acid GSM-1, and they were able to precisely identify the hydrophobic region of transmembrane domain 1 of PS1 as the binding site. It is thus very interesting that the authors saw only a weak displacement of their probe with GSM-1 on PS1. This suggests that different chemical scaffolds might bind in different regions of PS1. Another interesting observation is the difference between PS1 (weak displacement) and PS2 (strong displacement) when competing with GSM-1. It suggests that compounds bind differently to both presenilins but doesn't tell us if it is the result of a difference in conformation or protein sequence. This might have an implication in the development of new GSMs.

    Another very interesting observation from the present report is the major difference observed between the two TorreyPines compounds in their ability to displace the probe, where TP1 competes at about 50 percent and TP2 doesn't compete at all. Both compounds have the same in-vitro potency and differ structurally only at one atom position where TP1 contains a pyridyl group and TP2 a phenyl group. Given their almost identical structures, it would be surprising that they bound in a different region. It is more likely that the kinetics of binding might be different. This is something to keep in mind for future studies.

    The authors conclude by suggesting studies that will aim at the refinement of the binding site. This will be extremely valuable in light of the GSM-1 binding data, the lack of overlap with new generation GSMs on PS1, and the differential binding to PS1 and PS2. It might also be relevant to test compounds that show different binding properties to PS1 and PS2 to see if this affects their potency on Aβ modulation.


    . Substrate sequence influences γ-secretase modulator activity, role of the transmembrane domain of the amyloid precursor protein. J Biol Chem. 2011 Nov 18;286(46):39794-803. PubMed.

    View all comments by Jean-Francois Blain

Make a Comment

To make a comment you must login or register.


News Citations

  1. Paris: Semagacestat Autopsy and Other News of Trial Tribulations
  2. Barcelona: Allosteric γ Modulation Moves Toward Clinic
  3. Surprise! Some γ-Secretase Modulators Work by Targeting APP
  4. γ-Secretase Modulators—More Substance for Substrate Theory
  5. FRETting Pays off—NSAIDs Target Presenilins, Reduce Aβ42
  6. New γ-Secretase Modulators Reduce Aβ42, Avoid Notch

Paper Citations

  1. . Substrate-targeting gamma-secretase modulators. Nature. 2008 Jun 12;453(7197):925-9. PubMed.
  2. . Amyloid beta 42 peptide (Abeta42)-lowering compounds directly bind to Abeta and interfere with amyloid precursor protein (APP) transmembrane dimerization. Proc Natl Acad Sci U S A. 2010 Aug 17;107(33):14597-602. PubMed.
  3. . Nonspecificity of binding of gamma-secretase modulators to the amyloid precursor protein. Biochemistry. 2009 Dec 22;48(50):11837-9. PubMed.
  4. . Nonsteroidal anti-inflammatory drugs lower Abeta42 and change presenilin 1 conformation. Nat Med. 2004 Oct;10(10):1065-6. PubMed.
  5. . NSAIDs and enantiomers of flurbiprofen target gamma-secretase and lower Abeta 42 in vivo. J Clin Invest. 2003 Aug;112(3):440-9. PubMed.
  6. . Sulindac sulfide is a noncompetitive gamma-secretase inhibitor that preferentially reduces Abeta 42 generation. J Biol Chem. 2003 May 16;278(20):18664-70. PubMed.
  7. . Selected non-steroidal anti-inflammatory drugs and their derivatives target gamma-secretase at a novel site. Evidence for an allosteric mechanism. J Biol Chem. 2004 Oct 15;279(42):43419-26. Epub 2004 Aug 10 PubMed.
  8. . Modulation of gamma-secretase reduces beta-amyloid deposition in a transgenic mouse model of Alzheimer's disease. Neuron. 2010 Sep 9;67(5):769-80. PubMed.

Further Reading

Primary Papers

  1. . Novel γ-secretase enzyme modulators directly target presenilin protein. J Biol Chem. 2011 Oct 28;286(43):37181-6. PubMed.