Many had trouble reaching the brain. Or they got in but couldn’t stay, ousted in short order by P-glycoprotein. At long last, drug developers have overcome these and other hurdles, and well over a decade of effort developing β-secretase (BACE1) inhibitors is starting to pay off. At the Alzheimer’s Association International Conference held 14-19 July 2012 in Vancouver, Canada, three companies reported promising Phase 1 data on brain-penetrant compounds that block BACE1 activity. The protease cuts amyloid precursor protein (APP) to kick off production of the Aβ peptides believed to be a molecular culprit of Alzheimer’s disease. “Even though it’s been very difficult to develop BACE inhibitor drugs, they are now coming along. They are in clinical trials. They can get into the brain, and they inhibit BACE1 and reduce Aβ levels dramatically in the CSF. I’m very excited about the possibilities,” said Robert Vassar of Northwestern University Medical School in Chicago, whose lab cloned and characterized BACE1 (Vassar et al., 1999).




Crystal structure of BACE1, a leading target in AD drug development research. Image courtesy of Wikimedia

Eli Lilly and Co., Indianapolis, Indiana, appeared to set the pace for the newest batch of BACE1 inhibitors. At last year’s AD/PD International Conference in Barcelona, Spain, and later in a Journal of Neuroscience paper, the company reported that it finally had in hand an oral compound (LY2811376) with nice drug properties that got into the brain and reduced Aβ in healthy volunteers (see ARF related conference story and May et al., 2011). Unfortunately, that molecule never reached Phase 2. Rat studies linked the Lilly compound with retinal pigment epithelial defects—perhaps the only consolation being the absence of these problems in BACE knockout mice, which indicated they were an off-target effect and that blocking BACE1 could still conceivably work as an Alzheimer’s intervention.

In Vancouver, Lilly scientists reported preclinical and Phase 1 data on a new BACE1 inhibitor, as did Eisai and Merck. In addition, though not presenting at AAIC, Roche has an oral BACE1 inhibitor (RG7129) in Phase 1 (see Santarelli Q&A), and High Point Pharmaceuticals, North Carolina, is recruiting people with mild cognitive impairment (MCI) or mild AD for a 28-day course of its BACE1 inhibitor (HPP854) in a Phase 1 safety study. The Philadelphia-based biotech company Vitae Pharmaceuticals is partnering with Boehringer Ingelheim to develop oral BACE compounds that lower brain Aβ in AD mouse models. Other companies including AstraZeneca, CoMentis, and Takeda Pharmaceutical Company are also developing BACE inhibitors (see ARF related news story), but did not present at AAIC this year. Genentech has an entirely different approach—bispecific antibodies with one arm targeting BACE and the other recognizing transferrin receptor to boost brain penetrance (see ARF related news story on Atwal et al., 2011, and Yu et al., 2011). Below is a summary of the data Lilly, Eisai, and Merck presented at AAIC.

Lilly’s Patrick May reported that the compound LY2886721 showed good selectivity (i.e., did not inhibit other aspartyl proteases such as cathepsin D, pepsin, and renin) and reduced Aβ in a dose-dependent manner in HEK293Swe cells and in primary neurons from PDAPP mice. Despite a short half-life of three hours in mice, a 3-30 mg/kg dose lowered brain Aβ by 20 to 65 percent, relative to vehicle-treated groups, and the effect lasted up to nine hours after dosing. In beagle dogs, where the molecule has a longer half-life, a 0.5 mg/kg dose halved cerebrospinal fluid (CSF) Aβ in nine hours, and plasma Aβ levels were still down after 24 hours. Before moving into CSF biomarker studies, the Lilly team conducted an initial single ascending-dose study evaluating safety, tolerability, and of plasma Aβ pharmacodynamics following 1-35 mg oral doses of LY2886721 to healthy adults.

In Phase 1 studies, people took the compound by mouth, then had CSF samples collected at baseline and every one to four hours for 36 hours (single-dose study), or 24 hours after the final dose (multiple-dose study). Scientists measured CSF levels of APP’s non-amyloidogenic (sAPPα, C83) and amyloidogenic (sAPPβ, C99) cleavage products. The single-dose trial enrolled healthy adults for LY2886721 (10 or 35 mg), or placebo, and participants in the multiple-dose study got 5, 15, or 35 mg of the inhibitor, or placebo, once daily for two weeks. The compound lowered CSF Aβ40, Aβ42, and sAPPβ concentrations, while increasing sAPPα, consistent with what is expected for BACE1 inhibition, Lilly’s Brian Willis reported. On a poster by Ferenc Martenyi and colleagues, the inhibitor showed good pharmacodynamic (PD) and pharmacokinetic (PK) properties, and was safe and well tolerated up to six weeks from final dosing in the 68 subjects who completed the Phase 1 studies. The compound has entered a six-month Phase 2 trial of the two higher doses in early AD patients whose disease is ascertained with a positive brain amyloid PET scan by florbetapir, May said in his talk.

Eisai is also developing an oral BACE inhibitor (E2609), and the main news so far is much the same. According to company scientists, it is highly potent and selective, crosses the blood-brain barrier, and seems safe and well tolerated in healthy people thus far. It did well enough in early clinical studies to progress to Phase 2. Eisai’s single-dose Phase 1 trial assessed safety and plasma readouts only; its multiple ascending-dose trial included CSF measures as well. In the single-dose study, young healthy volunteers received 5, 10, 25, 50, 100, 200, 400, or 800 mg of inhibitor, or placebo, while an elderly cohort got a single 50 mg dose. The scientists collected blood samples before dosing and at various times up to 144 hours afterward. Plasma Aβ1-x levels dropped in a dose-dependent manner, with 52 percent clearance for 5 mg doses and 92 percent at 800 mg. At higher doses (200-800 mg), plasma Aβ1-x remained about 40 percent below baseline even at six days, when most of the compound was washed out, Eisai’s Robert Lai reported. All doses of E2609 had acceptable tolerability with no severe adverse events, and the safety profile was similar between young and old participants. In preclinical studies, the compound lowered Aβ levels in brain, CSF, and plasma of rats and guinea pigs, and in CSF and plasma of nonhuman primates, as shown on posters by Tatsuto Fukushima and Fiona Lucas.

Bruce Albala presented preliminary findings from Eisai’s multiple ascending-dose trial, where healthy adults ages 50-85 took the BACE inhibitor by mouth once a day for 14 days and had blood samples taken before, during, and after dosing to assess pharmacokinetics (PK) and pharmacodynamics (PD). They also had spinal taps two days before dosing and 12 hours after the final dose on day 14. Albala presented data on the 25, 50, 100, and 200 mg cohorts; the group that got the highest (400 mg) dose is still being analyzed. According to Albala, safety and tolerability looked good, with headache as the most common side effect. E2609 reaches steady state fairly quickly (five to seven days). Rising doses yield linear increases in CSF levels, indicating the compound gets into the brain and levels do not plateau, Albala said. Importantly, the inhibitor seems to be reducing amyloid production, as CSF Aβ1-x levels decreased 46 percent in the 25 mg cohort and dipped steadily further with rising doses, up to 80 percent in people taking 200 mg. CSF Aβ1-40, Aβx-40, Aβ1-42, and Aβx-42 levels all dropped as well. “These preliminary results suggest E2609 is a potent BACE1 inhibitor that results in consistent, sustained reduction of human Aβ1-x,” Albala said. Also this May, Eisai started dipping its toes into AD by recruiting MCI patients who have a confirmed biomarker for amyloid-β into a single-dose, single-center Phase 1 study.

For its part, Merck presented a slew of posters detailing preclinical and Phase 1 studies of its BACE1 inhibitors. Its lead compound (MK-8931) held its own in Phase 1, leaving some to wonder why it was not included in the talk session featuring Lilly’s and Eisai’s compounds. Merck enrolled a total of 88 healthy young adults—40 for a rising single-dose study in Belgium, 48 for a rising multiple-dose study at a U.S. site—taking blood and CSF samples at baseline, and every two hours from final dosing to 36 hours after. Indeed, “it is fairly intensive for the subjects, and we are very grateful they volunteer. We have to use a much larger needle than used for standard lumbar punctures,” Merck’s Mark Forman told Alzforum, noting that the use of larger needles increases the frequency of side effects such as headache and back pain. His poster, and another by Jack Tseng, claimed the compound has good PK and PD properties, showing dose-proportional and consistent responses among subjects at doses ranging from 2.5 to 550 mg. Maria Michener presented a cisterna magna-ported rhesus monkey model that enables more detailed PK/PD assessments because researchers can do serial CSF and plasma sampling over the course of weeks or months (see ARF ICAD story and ARF SfN story for more on this primate model).

In Phase 1, a single dose (100 or 500 mg) of the Merck inhibitor sent CSF Aβ40 and Aβ42 down more than 90 percent. In the multiple-dose study, CSF Aβ dropped 50 and 80 percent with 10 and 40 mg doses, respectively, while higher doses achieved more than 90 percent Aβ suppression. Furthermore, the compound’s 20-hour half-life makes it ideal for once-daily dosing, Forman said. A poster by K. Christopher Min showed the compound behaving similarly in a PK/PD study of healthy Japanese subjects.

According to a mathematical model developed by Huub Jan Kleijn, Julie Stone, and other company scientists, more than 90 percent of subjects can expect their CSF Aβ levels to be halved with a 12 mg dose of the Merck inhibitor. Doses around 30-40 mg should suppress Aβ by 75 percent, the model predicts. Population modeling on a poster by Lei Ma suggests that age, race, and body weight should have modest effects on MK-8931’s PK properties. Data from this analysis will help the scientists select doses for future Phase 2 trials. The company has done a small pharmacodynamics study in 30 AD patients to see if levels of Aβ reduction predicted by the model hold up in real life. Participants received 12 or 40 mg of the compound, or a higher dose (60 mg) for safety purposes. Dosing for this study was finished in June, and additional trials in AD patients are planned for later this year, Forman told Alzforum.

As further proof of principle, Merck scientist Lynn Hyde presented two posters showing that a preclinical BACE inhibitor (MBI-3) curbs amyloid buildup in the highly aggressive TgCRND8 mouse model of AD. The compound worked in preventive and therapeutic paradigms—that is, 11-week treatment of eight-week-old mice without plaques, or 18-week treatment of six-month-old mice with plaques. Treated TgCRND8 mice did not develop the phenotypes seen in BACE1 knockout mice, i.e., reduced prepulse inhibition, peripheral nerve hypomyelination, and poor navigation in the Morris water maze (ARF related news story on Savonenko et al., 2008). This suggests to the Merck scientists that these might be primarily developmental issues. Some of these data were presented at a Keystone symposium earlier this year (see ARF related conference story).

This movement on the BACE inhibition front has come with serendipitous timing. Days before the AAIC got underway, scientists reported in Nature the discovery of a protective mutation (A673T) near APP’s BACE cleavage site, which reduced Aβ production by 40 percent in human cell lines and lowered sporadic AD risk (ARF related news story on Jonsson et al., 2012). This mutation affects the same APP site as a different mutation that increases Aβ production and causes familial AD (A673V). “That is almost incontrovertible evidence that Aβ plays a critical, early role in AD,” Vassar told Alzforum. “You increase it, people get AD. You decrease it, people are protected. Genetically, it goes both ways.”

Still, BACE1 inhibitors have a long road ahead, as the field awaits data on clinical efficacy and possible side effects. On the former, some scientists fear the compounds may do little for people already showing symptoms. “If [BACE inhibitors] are tested in patients with MCI or early AD, this may be too late to show clinical efficacy,” noted Stefan Lichtenthaler of the German Center for Neurodegenerative Diseases, Munich, Germany, in an e-mail to Alzforum. “My guess is that these compounds would do better as a preventive drug, not as a therapeutic. But let’s see.” Pfizer announced last week that one of its highly anticipated Phase 3 trials of intravenous bapineuzumab showed no benefit in ApoE4 carriers with mild to moderate AD (see ARF related news story).

Weihong Song of the University of British Columbia, Canada, stressed the need to look for potential side effects. “No matter how well your compound can inhibit BACE … does it inhibit other things? That is a major question,” he said in an Alzforum phone interview after co-chairing the BACE therapeutics session with Lichtenthaler at AAIC. Recent work by Lichtenthaler and colleagues, and by Bart de Strooper’s group at the University of Leuven, Belgium, reveals additional BACE1 substrates. The work suggests that the secretase could play a role in forming connections in the brain (ARF related news story on Kuhn et al., 2012 and Zhou et al., 2012). This fits with research by Vassar and colleagues showing axon guidance defects in BACE1 knockout mice (Rajapaksha et al., 2011). “There are some mild wiring problems in the mouse olfactory system, and we are concerned these may be more general phenomena. I don’t know how serious it will be in humans,” Vassar said. “It’s not a red light for BACE inhibitors. It’s just a note of caution as we move forward with these compounds in clinical trials.”—Esther Landhuis.


  1. What could possibly go wrong with the use of BACE inhibitors for the prevention or treatment of AD?

    The latest flurry of results from pharmaceutical companies showing that BACE inhibitors reduce the levels of amyloid-β in the brain has led to a renewed wave of optimism that this approach may lead to an effective treatment for AD, possibly even a cure. This optimism has been amplified by the recent report of a mutation in the APP gene, A673T, that reduces BACE activity and protects against AD (Jonsson et al., 2012).

    We have recently suggested that there is a plausible explanation for the effect of the APP A673T mutation on AD rates that goes beyond the conventional wisdom of “less Aβ protects, more Aβ harms” involving a fine-tuning effect on Aβ levels by the A673T mutation that improves the adaptive response of APP to AD-causing brain stress (Castello and Soriano, 2012). In other words, there is likely a range of Aβ that optimizes the adaptive response of APP to brain stress, and the A673T mutation helps to maintain Aβ levels within that range.

    It is very unlikely that any laboratory-led BACE manipulation may result in a better therapeutic outcome than what this mutation provides. Indeed, a reduction in Aβ levels through BACE inhibition that is too dramatic would compromise Aβ physiological function and have a catastrophic impact on neuronal function, even if there were no fears of potential BACE side effects, which is also an unlikely scenario (see, e.g., note of caution by Weihong Song above).

    Overall, everything we know very strongly suggests that inhibiting BACE activity in the brain to treat AD may not be the most sensible way forward.


    . A mutation in APP protects against Alzheimer's disease and age-related cognitive decline. Nature. 2012 Aug 2;488(7409):96-9. PubMed.

    . Rational heterodoxy: Cholesterol reformation of the amyloid doctrine. Ageing Res Rev. 2012 Jul 6; PubMed.

  2. BACE1 inhibition is an important potential theurapeutic arm of the β amyloid clearance phenomenon in Alzheimer's which needs to be realized clinically.

    However, another BACE1 inhibitor, minocycline, which is also neuroprotective and is already an approved therapeutic agent, is now undergoing trials in cognitively normal individuals and patients with mild cognitive impairment (MCI) or Alzheimer's disease (AD) at Huntington Medical Research Institute. Patients and controls will undergo clinical screening, neuropsychological tests, blood and urine analyses, quantitative magnetic resonance imaging (MRI), and 1H and 13C magnetic resonance spectroscopy (MRS). Each individual will receive minocycline oral administration for four weeks initially, after which MRI, MRS, and neuropsych results will be recorded. If no adverse side effects occur, subjects will continue minocycline administration for an additional five months.

    A study by Ferretti demonstrated recently that minocycline corrects early, pre-plaque neuroinflammation and inhibits BACE1 in a transgenic model of Alzheimer's disease-like amyloid pathology (Ferretti et al., 2012).

    There are a small group of Alzforum members who are interested in the emerging role of pathogens in Alzheimer's disease, and feel that β amyloid clearance problems may be the result of pathogenic inflammation (spirochetal) with and without viruses.

    Chronic spirochetal infection can cause slowly progressive dementia, cortical atrophy, and amyloid deposition in the atrophic form of general paresis. There is a significant association between Alzheimer's disease (AD) and various types of spirochetes (including the periodontal pathogen Treponemas and Borrelia burgdorferi), and other pathogens such as Chlamydophyla pneumoniae and herpes simplex virus type 1 (Miklossy, 2011). Miklossy’s lab at the University Medical School (CHUV), Lausanne, Switzerland, exposed mammalian glia and neuronal cells in vitro to Borrelia burgdorferi spirochetes and bacterial lipopolysaccharides. Morphological changes analogous to amyloid deposits were observed at two to eight weeks' exposure. Increased levels of β amyloid precursor protein and hyperphosphorylated tau were detected by Western blot (Miklossy et al., 2004).

    Seven out of 10 brains from the Harvard McLean Brain Bank were positive for Borrelia DNA. Alan MacDonald demonstrated this, and feels “Borrelia burgdorferi infection is the root cause of at least 70 percent of Alzheimer's disease, based on the detection of positive in-situ DNA hybridization results in the cytoplasmic granulovacuolar bodies of hippocampal neurons (with no positive signals detected in the nucleus) for flagellin B DNA sequences of Borrelia burgdorferi" (MacDonald, 2007).

    A randomized controlled trial of doxycycline and rifampin for patients with Alzheimer’s disease demonstrated that cognitive decline was statistically improved in the treatment arm over placebo (Loeb et al., 2004).

    Minocycline protected basal forebrain cholinergic neurons from murine-p75-saporin immunotoxic lesioning in animal models (Hunter et al., 2004). Minocycline attenuates neuronal cell death and improves cognitive impairment in Alzheimer’s disease models (Choi et al., 2007). Minocycline does not affect amyloid-β phagocytosis by human microglia cells. Minocycline attenuates the release of TNF-α by human microglia upon exposure to Aβ, SAP, and C1q (Familian et al., 2007).

    Moderate magnetic field therapy (0.5 Tesla) in 15 Alzheimer’s patients was demonstrated by this author in 2006 on the hypothesis that the outer protein of the Borrelia burgdorferi bacteria is strongly electron-negatively charged and will be repelled by the negative pole of a 0.5 Tesla electromagnet below and a positive pole above the patient's head. Cognition improved. However, this improvement was gradually lost from several weeks to six months in these patients in an open-label, IRB-approved pilot study. All patients had moderate to severe Alzheimer's disease (Nichols et al., 2006).

    The mechanism may also be related to crosstalk between SMF and IL-6, as well as the upregulation or downregulation of over 2,600 genes (Wang et al., 2009).

    See also:

    Nichols et al. Medical Hypothesis; elctromagnetic field therapy using Magnetic Molecular Energising (MME) and Antibiotic Therapy: A Pilot Study. 31/01/2006. E pub; European Biology and Bioelectromagnetics.


    . Minocycline corrects early, pre-plaque neuroinflammation and inhibits BACE-1 in a transgenic model of Alzheimer's disease-like amyloid pathology. J Neuroinflammation. 2012;9:62. PubMed.

    . Emerging roles of pathogens in Alzheimer disease. Expert Rev Mol Med. 2011;13:e30. PubMed.

    . Borrelia burgdorferi persists in the brain in chronic lyme neuroborreliosis and may be associated with Alzheimer disease. J Alzheimers Dis. 2004 Dec;6(6):639-49; discussion 673-81. PubMed.

    . Alzheimer's neuroborreliosis with trans-synaptic spread of infection and neurofibrillary tangles derived from intraneuronal spirochetes. Med Hypotheses. 2007;68(4):822-5. PubMed.

    . A randomized, controlled trial of doxycycline and rifampin for patients with Alzheimer's disease. J Am Geriatr Soc. 2004 Mar;52(3):381-7. PubMed.

    . Minocycline protects basal forebrain cholinergic neurons from mu p75-saporin immunotoxic lesioning. Eur J Neurosci. 2004 Jun;19(12):3305-16. PubMed.

    . Minocycline attenuates neuronal cell death and improves cognitive impairment in Alzheimer's disease models. Neuropsychopharmacology. 2007 Nov;32(11):2393-404. PubMed.

    . Minocycline does not affect amyloid beta phagocytosis by human microglial cells. Neurosci Lett. 2007 Apr 6;416(1):87-91. PubMed.

    . Moderate strength (0.23-0.28 T) static magnetic fields (SMF) modulate signaling and differentiation in human embryonic cells. BMC Genomics. 2009;10:356. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Barcelona: Out of Left Field—Hit to The Eye Kills BACE Inhibitor
  2. Q&A With Roche’s CNS Leader Luca Santarelli
  3. Getting to First BACE: BACE1 Inhibition Takes a Step Forward
  4. Smuggling Antibodies to BACE Across the Blood-Brain Barrier
  5. Madrid: γ-secretase Dimers, A New Model, A Drug in Clinic
  6. San Diego: Merck Reports BACE Inhibition in Primates
  7. Down to BACE-ics—Old Mouse a New Model for Schizophrenia?
  8. Keystone: Therapies Around ApoE—Has Their Time Come?
  9. Protective APP Mutation Found—Supports Amyloid Hypothesis
  10. No Pony in There: Bapi Fails Mild to Moderate ApoE4 Carriers
  11. BACE Secrets: Newly Identified Substrates May Regulate Plasticity

Paper Citations

  1. . Beta-secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE. Science. 1999 Oct 22;286(5440):735-41. PubMed.
  2. . Robust central reduction of amyloid-β in humans with an orally available, non-peptidic β-secretase inhibitor. J Neurosci. 2011 Nov 16;31(46):16507-16. PubMed.
  3. . A therapeutic antibody targeting BACE1 inhibits amyloid-β production in vivo. Sci Transl Med. 2011 May 25;3(84):84ra43. PubMed.
  4. . Boosting brain uptake of a therapeutic antibody by reducing its affinity for a transcytosis target. Sci Transl Med. 2011 May 25;3(84):84ra44. PubMed.
  5. . Alteration of BACE1-dependent NRG1/ErbB4 signaling and schizophrenia-like phenotypes in BACE1-null mice. Proc Natl Acad Sci U S A. 2008 Apr 8;105(14):5585-90. PubMed.
  6. . A mutation in APP protects against Alzheimer's disease and age-related cognitive decline. Nature. 2012 Aug 2;488(7409):96-9. PubMed.
  7. . Secretome protein enrichment identifies physiological BACE1 protease substrates in neurons. EMBO J. 2012 Jul 18;31(14):3157-68. PubMed.
  8. . The neural cell adhesion molecules L1 and CHL1 are cleaved by BACE1 protease in vivo. J Biol Chem. 2012 Jul 27;287(31):25927-40. PubMed.
  9. . The Alzheimer's β-secretase enzyme BACE1 is required for accurate axon guidance of olfactory sensory neurons and normal glomerulus formation in the olfactory bulb. Mol Neurodegener. 2011;6:88. PubMed.

Other Citations

  1. PDAPP mice

External Citations

  1. Phase 1 safety study
  2. single-dose trial
  3. multiple-dose study
  4. Phase 2 trial
  5. single-dose Phase 1 trial
  6. multiple ascending-dose trial
  7. Phase 1 study
  8. pharmacodynamics study

Further Reading


  1. . The phosphatidylinositol 3-kinase inhibitor wortmannin alters the metabolism of the Alzheimer's amyloid precursor protein. J Neurochem. 1999 Dec;73(6):2316-20. PubMed.
  2. . Robust central reduction of amyloid-β in humans with an orally available, non-peptidic β-secretase inhibitor. J Neurosci. 2011 Nov 16;31(46):16507-16. PubMed.
  3. . Alteration of BACE1-dependent NRG1/ErbB4 signaling and schizophrenia-like phenotypes in BACE1-null mice. Proc Natl Acad Sci U S A. 2008 Apr 8;105(14):5585-90. PubMed.
  4. . Secretome protein enrichment identifies physiological BACE1 protease substrates in neurons. EMBO J. 2012 Jul 18;31(14):3157-68. PubMed.
  5. . The neural cell adhesion molecules L1 and CHL1 are cleaved by BACE1 protease in vivo. J Biol Chem. 2012 Jul 27;287(31):25927-40. PubMed.