Many researchers believe that slowing the buildup of amyloid in the brain holds the key to delaying or preventing Alzheimer’s disease. Scientists previously reported that selective serotonin reuptake inhibitors (SSRIs)—a class of commonly used antidepressants—suppressed Aβ and lowered amyloid plaque load in mice. Now, a follow-up report claims that the drugs inhibit Aβ production in people, as well. Yvette Sheline, formerly at Washington University in St. Louis and now at the University of Pennsylvania, Philadelphia, and John Cirrito at WashU, published the human data in the May 14 Science Translational Medicine.

A single dose of the SSRI citalopram, marketed under the brand name Celexa, cut Aβ production by a third in the cerebrospinal fluid (CSF) of healthy volunteers. Future studies may reveal whether chronic dosing produces the same effect, and if this would lessen amyloid buildup in the brain, Sheline said. If there is a sustained benefit, then SSRIs may join immunotherapy and BACE inhibitors as another possible route to curb amyloid pathology. Many SSRIs have been approved by the FDA and they have acceptable safety profiles.

Other researchers called the findings promising, while noting that the road to a potential therapy remains long. “It seems clear that citalopram decreases Aβ production and aggregation in both mice and humans. The big question is, can that translate into a clinical preventative or therapeutic effect?” asked William Klunk at the University of Pittsburgh. 

SSRIs block the reuptake of serotonin by neurons, making this neurotransmitter more available in the synaptic cleft and pumping up its signaling. Several groups have shown that serotonin dampens Aβ production in vitro, and also lightens plaque load and preserves cognition in 3xTg AD model mice (see Robert et al., 2001Arjona et al., 2002Nelson et al., 2007). In previous studies, Sheline and colleagues found that a single dose of any of several SSRIs cut Aβ levels in the interstitial fluid (ISF) of young APP/PS1 mice by a quarter, and four months of treatment halved plaque load. 

The authors wondered if people taking SSRIs might be similarly protected from amyloid buildup. To investigate this, the researchers recruited 186 cognitively normal volunteers with an average age of 69. Amyloid PET scanning revealed that participants who had used any antidepressants in the last five years had significantly less amyloid in their brains than those who had not taken the drugs (see Cirrito et al., 2011). This retrospective study could not establish if SSRI use suppressed amyloid production, however.

To explore a direct connection, Sheline and colleagues used stable isotope labeling kinetics (SILK), a technique developed by Randall Bateman at WashU to measure production of new proteins in the brain. Researchers intravenously administer 13C-labeled leucine to track newly generated Aβ, then collect CSF samples at regular intervals (see Apr 2009 news story). This allows them to determine rates of Aβ production and clearance. Sheline recruited 23 healthy adult volunteers below the age of 50. SILK revealed that two 30 mg doses of citalopram two hours apart suppressed Aβ production by 37 percent over the next 45 hours, without affecting clearance. The authors arrived at the dosage based on dose-response studies in mice. This amount is on the high end of what people take for depression, the authors note. 

What might citalopram do long term? For this question, human data is not available, but Sheline and colleagues looked at the effects of chronic citalopram administration in 6-month-old APP/PS1 mice, which actively deposit amyloid. The authors followed the fate of individual plaques in live animals using high-power, two-photon microscopy. Four weeks of citalopram arrested the growth of existing plaques and prevented new ones from forming. It did not clear deposits.

Stopping Plaque in Its Tracks:

In mice treated with placebo, amyloid plaques grow and new deposits appear (top panels). In mice taking citalopram, plaques stay static (bottom panels). [Image courtesy of Science Translational Medicine/AAAS.]

Commentators praised the methodology of the study, but cautioned that many questions about the therapeutic potential of SSRIs remain. Mark Mattson at the National Institute on Aging, Bethesda, Maryland, noted that the current study lacks behavioral data in mice, leaving open the question of clinical efficacy. Richard Wurtman at the Massachusetts Institute of Technology in Cambridge suggested that before advancing to large-scale human trials, it would be worth mining large epidemiological datasets to see if antidepressant use correlates with a lower risk of AD. Paul Aisen at the University of California, San Diego, wrote to Alzforum, “There is not yet strong evidence that citalopram reduces brain amyloid in humans; the effect on CSF Aβ42 is intriguing, but the relationship between CSF amyloid peptides and brain amyloid is complex.”

Nonetheless, researchers agreed that the data merit further human studies. “I think trials in people with mild cognitive impairment or who have biomarker evidence of AD should be seriously considered,” Constantine Lyketsos at Johns Hopkins University, Baltimore, told Alzforum. Citalopram may not be the best SSRI to use, however. A recent trial for agitation in AD found that the drug causes some adverse effects, including accelerated cognitive decline and abnormal heart rhythms, said Lon Schneider at the University of Southern California, Los Angeles (see Feb 2014 news story). The Food and Drug Administration now recommends that people take no more than 40 mg/day of citalopram. It further specifies that people over 60 should take no more than 20 mg/day, because the drug builds up in the blood of older adults and can increase the risk for heart arrhythmias.

For this reason, Sheline plans to use a different SSRI in future studies. All drugs in this class have similar effects on Aβ in mice, she told Alzforum. She will investigate the effect of chronic dosing by treating cognitively normal volunteers between the ages of 65 and 85 with an SSRI for two weeks, comparing baseline and final levels of Aβ in spinal fluid. If Aβ stays low, the next step would be a larger study using PET amyloid imaging to look for any changes in brain amyloid load. These studies might eventually pave the way for a longer clinical trial in cognitively healthy older adults at risk for AD, Sheline said.—Madolyn Bowman Rogers. 


  1. This study was well-conducted and is elegant because it shows consistent effects of the drug citalopram in transgenic mice and in humans. Regarding the potential use of citalopram in Alzheimer’s disease patients, Mark Mattson correctly points out that positive behavioral data are lacking for citalopram in transgenic mice, and Lon Schneider correctly notes that a recent study found that citalopram accelerated cognitive decline in AD patients (Porsteinsson et al., 2014).

    I would like to add that this is the third drug, after semagacestat (Doody et al., 2013) and avagacestat (Coric et al., 2012), for which a lowering effect on CSF Aβ production in humans is associated with detrimental cognitive effects in AD patients. The detrimental effect of citalopram on cognition in AD patients described in the JAMA paper is not trivial. In only nine weeks, the drug at 30 mg/day significantly (p = 0.03) lowered, compared with placebo, MMSE by 1.05 points (95 percent confidence interval 0.13-1.97 points). This effect size corresponds to at least six months of accelerated cognitive decline.

    It will be interesting to see in the future if other drugs that inhibit the production of Aβ in CSF in humans will also cause detrimental cognitive effects in AD patients.


    . Effect of citalopram on agitation in Alzheimer disease: the CitAD randomized clinical trial. JAMA. 2014 Feb 19;311(7):682-91. PubMed.

    . A phase 3 trial of semagacestat for treatment of Alzheimer's disease. N Engl J Med. 2013 Jul 25;369(4):341-50. PubMed.

    . Safety and Tolerability of the γ-Secretase Inhibitor Avagacestat in a Phase 2 Study of Mild to Moderate Alzheimer Disease. Arch Neurol. 2012 Aug 13;:1-12. PubMed.

  2. If we believe both the Porsteinsson results regarding cognition and the new data on SSRIs lowering Aβ, this may—together with the results of the γ-secretase inhibitor trials—indicate that lowering Aβ production as such is not only useless but may even be harmful for people with cognitive impairment. In any case, I would strongly support Richard Wurtman's suggestion to use existing clinical data (the various large epidemiology data bases should provide sufficient material) before starting any new trials based on this hypothesis.

  3. This elegant study is another contribution to the demonstration that activation of the serotonergic neurotransmission could have a beneficial effect in the context of AD. Such data should lead us to explore which serotonin receptor subtypes mediate such action. Among the 14 different receptors that respond to serotonin, all but 5-HT3 receptors are G protein-coupled receptors (GPCRs). More than 30 percent of the currently marketed drugs target GPCRs, and some already target serotonin receptors. We have recently demonstrated that chronic administration of a 5-HT4 receptor agonist is able to slow down amyloid pathology conjointly with cerebral inflammation and also to prevent cognitive deficits in 5XFAD mice (Giannoni et al., 2013).

    I would like to point out that activation of 5-HT4 receptors (and probably of other 5-HT receptors subtypes) is efficient in slowing down the progression of the pathology by activation of the non-amyloidogenic pathway and the release of the neuroprotective sAPPα fragment. Thus, chronic treatments using 5-HT4 agonists should be conducted in early stages of the disease, when the amyloid load is not too high. Then, they could delay amyloid accumulation. Translation in the clinic is far away, and we need safe drugs that will not increase the QT interval. That means back to the screening of molecules …


    . Early administration of RS 67333, a specific 5-HT4 receptor agonist, prevents amyloidogenesis and behavioral deficits in the 5XFAD mouse model of Alzheimer's disease. Front Aging Neurosci. 2013 Dec 24;5:96. PubMed.

    View all comments by Sylvie Claeysen

Make a Comment

To make a comment you must login or register.


Research Models Citations

  1. 3xTg
  2. APPSwe/PSEN1dE9

News Citations

  1. Studies Reveal New Hope, Old Problems With AD Biomarkers
  2. Citalopram Calms Agitation in Alzheimer’s, but Carries Risks

Paper Citations

  1. . The human serotonin 5-HT4 receptor regulates secretion of non-amyloidogenic precursor protein. J Biol Chem. 2001 Nov 30;276(48):44881-8. PubMed.
  2. . Effect of a 5-HT(2C) serotonin agonist, dexnorfenfluramine, on amyloid precursor protein metabolism in guinea pigs. Brain Res. 2002 Sep 27;951(1):135-40. PubMed.
  3. . Prophylactic treatment with paroxetine ameliorates behavioral deficits and retards the development of amyloid and tau pathologies in 3xTgAD mice. Exp Neurol. 2007 May;205(1):166-76. PubMed.
  4. . Serotonin signaling is associated with lower amyloid-β levels and plaques in transgenic mice and humans. Proc Natl Acad Sci U S A. 2011 Sep 6;108(36):14968-73. PubMed.

External Citations

  1. recommends

Further Reading

Primary Papers

  1. . An antidepressant decreases CSF Aβ production in healthy individuals and in transgenic AD mice. Sci Transl Med. 2014 May 14;6(236):236re4. PubMed.