In the March 17 Science Signaling, researchers led by Yong-Keun Jung at Seoul National University, South Korea, reported a new modulator of γ-secretase that appears to preferentially affect cleavage of amyloid precursor protein. Stress-associated ER protein 1 (SERP1) promotes γ-secretase activity and Aβ generation, the authors claim. Intriguingly, cells activate SERP1 under conditions of cellular stress such as high glucose levels. In diabetic mice, knockdown of SERP1 lowered Aβ. “These data provide insight into a mechanism for selectively inhibiting γ-secretase to modulate Aβ production,” Jung wrote to Alzforum.

  • ER protein SERP1 boosts γ-secretase activity and Aβ production.
  • Conversely, it lowers Notch cleavage.
  • SERP1 expression rises in diabetics, helping explain their risk for AD.

Lucía Chávez-Gutiérrez at VIB-KU Leuven Center for Brain and Disease Research, Belgium, said the findings may have relevance for Alzheimer’s disease. “The paper … highlights a potential pathogenic role for SERP1-mediated upregulation of Aβ production in diabetes and in Alzheimer’s disease,” she wrote to Alzforum (full comment below). Scientists were more skeptical about whether the findings offer a therapeutic target (see comments by Weiming Xia and Grace Stutzmann below).

Researchers have attempted to find physiological modulators of the γ-secretase complex before, proposing candidates such as β-arrestin 2 and G-protein coupled receptor 3 (Gpr3) (Dec 2012 news; Oct 2015 news). One high-profile putative modulator, γ-secretase activating protein (GSAP), did not replicate in subsequent studies (Sep 2010 news; Jan 2014 news). 

Act on APP Only. ER stress prompts more SERP1, which stabilizes the γ-secretase complex in lipid rafts, leading to more APP cleavage and Aβ. [Courtesy of Jung et al., Science Signaling/AAAS.]

To find other physiological modulators of γ-secretase, Jung and colleagues previously ran a gain-of-function screen in HEK 293 cells, looking for genes that affected the enzyme’s activity (Han et al., 2014). This turned up SERP1, a single-pass transmembrane protein of 64 amino acids that stimulated γ-secretase. SERP1 sits in the ER membrane and is activated by ER stress (Yamaguchi et al., 1999; Lee et al., 2003). 

In the new study, joint first authors Sunmin Jung and Junho Hyun delved into mechanism. They confirmed that, in two different cultured cell lines expressing mutant APP, overexpressing SERP1 bumped up secreted Aβ by as much as 70 percent, while knocking down SERP1 suppressed Aβ. Crucially, however, SERP1 had milder and opposite effects on Notch. SERP1 overexpression lessened Notch cleavage by 30 percent, and knockdown had no effect. SERP1 appears to be a negative regulator of Notch cleavage, the authors noted.

Further experiments elucidated what lay behind these observations. The effect hinged on the makeup of the γ-secretase complex. The mature complex comprises four subunits: presenilin, nicastrin, presenilin enhancer 2, and anterior pharynx-defective phenotype 1. APH1 itself comes in three isoforms, A, B, and C. SERP1 stabilized the binding of nicastrin to the APH1A isoform, thus boosting the amount of mature γ-secretase. Notably, APH1A isoform was previously implicated in pumping up amyloidogenic processing of APP (May 2009 news). 

Co-immunoprecipitation showed that SERP1 directly interacted with both APH1 and nicastrin. SERP1 also increased the amount of γ-secretase found in lipid rafts, where APP gets cleaved (see image above). Notch is cleaved elsewhere, suggesting one mechanism for substrate selectivity (Vetrivel et al., 2005). 

Chávez-Gutiérrez considers these interaction data convincing. Weiming Xia at Boston University agreed. “They did an excellent job in dissecting the molecular architecture of SERP1 and APH1A/Nicastrin,” he wrote to Alzforum (full comment below).

How would ER stress alter the picture? In cultures, the authors stressed cells by inhibiting the ER calcium channel SERCA. This increased SERP1 expression and Aβ production. When the authors knocked down SERP1 before applying the channel blocker, Aβ levels did not change, confirming that SERP1 was crucial to this response. The stress response was mediated by the ER stress sensor inositol-requiring enzyme 1. Blocking IREP1 prevented the increase in SERP1.

Because high blood sugar causes ER stress, the authors examined a mouse model of diabetes induced by injecting the pancreatic β cell poison streptozotocin. These mice had increased SERP1 expression and more γ-secretase complex in the brain. When the authors injected 3xTg mice with streptozotocin, Aβ production bumped up 20 percent. As in cell culture, the effect on Aβ could be prevented by knocking down SERP1.

Chávez-Gutiérrez found this telling. “The tripartite relationship between glucose-triggered ER stress, SERP1 levels and Aβ production may explain the link between diabetes and AD,” she wrote.

Is there evidence that SERP1 contributes to AD? The authors examined postmortem AD and control brains. In AD brain, SERP1 was up about 10-fold in hippocampal tissue, and fivefold in parietal lobe.

Because of SERP1’s selective effect on APP cleavage, the authors believe it could be a therapeutic target. They will look for inhibitors that disrupt the interaction between SERP1 and γ-secretase, Jung wrote to Alzforum. On that point, Xia cautioned that SERP1 knockout mice have poor glucose tolerance and increased ER stress. He believes more might be gained by revisiting previously identified γ-secretase modulators, evaluating their effects in animal and human studies using newly developed plasma biomarkers of amyloid, tau, and neurodegeneration. Likewise, Stutzman noted that it may be premature to begin thinking of harnessing this pathway for a therapeutic strategy, in part because SERP1 affected Aβ40 and Aβ42 in equal measure.—Madolyn Bowman Rogers


  1. Jung et al. have discovered SERP1, a facilitator for the assembly of the γ-secretase complex and the generation of amyloid β protein (Aβ). They did an excellent job in dissecting the molecular architecture of SRRP1-APH1A/Nicastrin, and provided detailed in vitro and in vivo evidence to support its role in γ-secretase cleavage of APP and Notch. The question is, is SERP1 a target for AD therapeutic development?

    Unfortunately, the SERP1 knockout mice carry too many problems. As the authors point out, the SERP1 KO mice have impaired glucose tolerance. Pancreatic islets from SERP1 KO mice showed a significant delay of insulin synthesis/secretion after glucose challenge (Hori et al., 2006). Knocking out SERP1 enhances ER stress and suppresses protein translation, which could contribute to growth retardation and increased mortality. Jung's group discovered this potential target but faces a daunting challenge in validating this target for AD drug development.

    There are a number of high-profile γ-secretase modulating factors, like SERP1, that have been discovered and explored as AD therapeutic targets in the past 15 years. They include TMP21, G protein-coupled receptor 3, and β-arrestin 1. The lack of progress in therapeutic development based on these modulators could be due to the failure of several clinical trials of γ-secretase inhibitors. At the time of stopping the clinical trial of the γ-secretase inhibitor semagacestat nine years ago, non-selective inhibition of Notch cleavage by the γ-secretase was considered as the primary reason for the failure. In this regard, targeting SERP1 would avoid this problem and reduce Aβ production without affecting Notch signaling.

    With recent advances in plasma biomarker discoveries, the AT(N) system (Amyloid deposition, Tau pathology, and Neurodegeneration) could be efficiently and economically utilized to screen preclinical AD subjects for trial enrollment. Since targeting secretases (β- and γ-secretases) acts upstream of AD pathological change, manipulating modulators like SERP1 will readily illustrate changes of AT(N) biomarkers in animal and human studies. It is time to revisit these modulators and determine their efficacies in the context of AT(N) biomarkers, with the ultimate goal of preventing neurodegeneration.


    . Deletion of SERP1/RAMP4, a component of the endoplasmic reticulum (ER) translocation sites, leads to ER stress. Mol Cell Biol. 2006 Jun;26(11):4257-67. PubMed.

  2. These findings about SERP1 are interesting from a cell biology/protein handling perspective, and they introduce novel ways of thinking about how cellular metabolism and stress can influence protein handling and signaling cascades. The authors present a compelling mechanism by which SERP1 function is altered to preferentially drive γ-secretase cleavage of APP and increase Aβ peptides under ER stress conditions.

    That said, I think it's premature to begin thinking of harnessing this pathway for a therapeutic strategy. Reasons include the apparently equal effects on Aβ40 and 42 peptides, or at least their suppression by shSERP1, which is inconsistent with the preferential increase of the more pathogenic Aβ42 species in AD and the increase in Aβ42:40 ratios in the 3xTg mice. Other than Figure 6 (panel 6b and c), there is no analysis of Aβ40 or 42 levels, so it is unclear how to link this to amyloidopathy.  Also, reducing Aβ40 may introduce undesirable effects as this peptide is associated with physiological functions as well. Thus, reducing SERP1 levels or function lacks the required specificity to target specific Aβ peptides associated with proteinopathy. 

    On a broader scale, it remains increasingly unclear how, or if, reducing amyloid levels in the brain may improve cognitive function or overall outcomes in AD patients. Based on the long list of failures in this domain, this variation of a theme doesn't seem sufficiently different to offer convincing hope for this approach.

  3. This paper presents SERP1 as a direct modulator of γ-secretase activity under ER stress, and highlights a potential pathogenic role for the SERP1-mediated upregulation of Aβ production in diabetes and in Alzheimer’s disease (AD).

    The authors convincingly demonstrate that the modulatory effect is mediated by a direct interaction of SERP1 with the NCT-APH subcomplex that increases its levels, as well as the steady-state levels of the γ-secretase complex. SERP1 not only enhances γ-secretase complex levels, but also localizes its activity to lipid rafts. Thus, this effect has further implications for the γ-secretase processing of substrates, as it favors the cleavage of substrates co-localizing with the protease in rafts (APP) over substrates with a non-lipid raft localization (e.g., Notch).

    Since ER stress regulates SERP1expression levels, the authors investigated whether ER stress has an effect on γ-secretase activity and consequently Aβ production. Interestingly, the report shows that under hyperglycemia-triggered ER stress, increased SERP1 levels result in elevated γ-secretase steady-state levels and higher Aβ production. The tripartite relationship between glucose-triggered ER stress, SERP1 levels, and  Aβ production may explain the link between diabetes and AD.

    Intriguingly, the authors found increased abundance of SERP1 (~9.6-fold) in postmortem hippocampal AD brain tissue, relative to age-matched, non-AD samples. Of note, Hata et al. recently reported increased γ-secretase activity in detergent-resistant fractions (lipid rafts) prepared from postmortem AD brain samples, relative to controls.

    Collectively, these observations support the contribution of γ-secretase activity/increased Aβ production to sporadic AD.


    . Enhanced amyloid-β generation by γ-secretase complex in DRM microdomains with reduced cholesterol levels. Hum Mol Genet. 2020 Feb 1;29(3):382-393. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Could β-Arrestin Provide New Way to Halt Aβ Accumulation?
  2. G-Protein Receptor Knockout Rescues Several Models of Alzheimer’s
  3. There’s a GSAP for That: Novel APP Partner a New Therapeutic Target?
  4. GSAP Revisited: Does It Really Play a Role in Processing Aβ?
  5. Double Paper Alert—Keystone Presentations Now in Press

Research Models Citations

  1. 3xTg

Paper Citations

  1. . OCIAD2 activates γ-secretase to enhance amyloid β production by interacting with nicastrin. Cell Mol Life Sci. 2014 Jul;71(13):2561-76. Epub 2013 Nov 24 PubMed.
  2. . Stress-associated endoplasmic reticulum protein 1 (SERP1)/Ribosome-associated membrane protein 4 (RAMP4) stabilizes membrane proteins during stress and facilitates subsequent glycosylation. J Cell Biol. 1999 Dec 13;147(6):1195-204. PubMed.
  3. . XBP-1 regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response. Mol Cell Biol. 2003 Nov;23(21):7448-59. PubMed.
  4. . Spatial segregation of gamma-secretase and substrates in distinct membrane domains. J Biol Chem. 2005 Jul 8;280(27):25892-900. PubMed.

Further Reading

Primary Papers

  1. . SERP1 is an assembly regulator of γ-secretase in metabolic stress conditions. Sci Signal. 2020 Mar 17;13(623) PubMed.