A batch of stories this week spans current approaches to biomarkers being developed for early detection of Alzheimer disease and as surrogate markers for clinical trials of therapeutic candidates. The goal is tests that will discriminate earlier, more easily and more clearly those people with Alzheimer disease or at risk of developing it, and predict clinical progression. The realization that AD starts wrecking the brain years before symptoms of dementia appear makes early detection the critical lever to pry open the window of opportunity for using new treatments that are now appearing on the horizon.

Four papers stand out this week. One, from researchers of the Alzheimer’s Disease Cooperative Study Group, describes a relatively straightforward method for measuring hippocampal atrophy to predict how AD will progress. Another, the work of Anne Fagan and colleagues at Washington University in St. Louis, Missouri, looks at the predictive value of CSF levels of amyloid-β (Aβ) peptides and tau proteins in normally aging people. In the third report, Elaine Peskind and colleagues at the University of Washington in Seattle use those same CSF biomarkers to assess the effects of statins on the brain in a small group of normal healthy people. And lastly, a look to the future, with an expression profiling approach to a possible blood test for Parkinson disease, coming from Clemens Scherzer and colleagues at Harvard Medical School in Boston.

Imaging MCI
Validating biomarkers of early AD by studying people with mild cognitive impairment (MCI) is a circular proposition. In many (but not all) people, this condition is an early sign of AD, but that diagnosis is uncertain until later, when the person progresses to more overt dementia (or not). So validating biomarkers in MCI requires prospective studies where groups are followed for years.

In this regard, a study set up starting in 1999 by the Alzheimer’s Disease Cooperative Study (ADCS) Group to test the effects of donepezil or vitamin E on the progression of AD has been a great help. The trial was disheartening in that it found no positive drug effect for vitamin E and little for donepezil (Petersen et al., 2005), but the investigators did put together a large cohort of well-characterized subjects with MCI that is still being followed. Rates of conversion to AD have been carefully established, with the goal of correlating progression to biological measures that might eventually serve as predictive markers.

One such measure is hippocampal atrophy. Loss of hippocampal volume is one of the earliest signs of ongoing AD. This brain area shrinks detectably before clinically measurable dementia appears, and people with it have an increased chance of progressing from MCI to AD (Jack et al., 1999). But calculating volumes from MRI scans is complicated, and limited to a few locations with the expertise.

In the new work, first author Charles DeCarli at the University of California at Davis, and colleagues, describe a method for estimating hippocampal atrophy from MRI images that is simpler than the standard analysis. In a report published in the January Archives of Neurology, they show that rating atrophy on a qualitative scale predicts progression from MCI to AD in a prospective study of 190 people who were part of the ADCS.

The subjects all got an MRI scan at the beginning of the study, and from those images the researchers measured multiple hippocampal dimensions and combined them into a rating on a scale of 0 (no atrophy) to 4 (severe atrophy). Ratings were relatively easy to perform and agreed well between independent raters. In the prospective part of the study, about one-third of the patients (66) progressed to dementia in the 3-year follow-up period; on average, they had higher ratings than those who did not progress. Among those with the highest scores (>2), 75 percent progressed from MCI to AD. The researchers found that a score of greater than 2 was associated with a more than twofold increased risk of developing dementia within 3 years.

If they set 2.0 as the cutoff value, the sensitivity of the method to detect incipient AD compared favorably to more quantitative measures, but was far easier to do.

“These data show that the use of a relatively simple, clinically applicable MTA [medial temporal atrophy] rating scale significantly increases the likelihood of identifying individuals with amnestic MCI who are destined to progress to dementia within 3 years above standard clinical evaluation, which includes MMSE testing,” the authors write. “Given the relatively high prevalence of MCI in the general population, use of this method as part of routine clinical evaluation may help identify individuals who might benefit from increased surveillance and future treatment.” But, they concede, the study involved a very selective cohort and needs to be tested in the general population.

Warning: AD Ahead
A shrinking hippocampus and signs of MCI signal destruction of brain cells that may be irreversible, so one goal of biomarker research is to move diagnosis forward to even earlier phases. One approach has been to look for protein markers in cerebrospinal fluid, including the neurotoxic Aβ and phospho-tau peptides, to get a window into ongoing brain pathology. A new paper from Washington University researchers Anne Fagan, David Holtzman, and colleagues advances this technique by showing that CSF levels of Aβ42 and tau proteins, in particular an elevated total tau/Aβ42 or p-tau181/Aβ42 ratio, mark cognitively normal individuals who are heading for dementia.

Previous studies by this group and others have shown that in patients with AD, CSF concentrations of Aβ40 and Aβ42 peptides are lower than normal, while total tau and phospho-tau levels are increased (see ARF related news story). This new study, published online January 8 in the Archives of Neurology, extends those findings by showing that people in even the earliest stages of AD (very mild to mild cognitive symptoms) already show changes in these biomarkers that resemble those seen in people with AD.

Moreover, in a prospective study, the researchers show elevation of total tau/Aβ42 or p-tau181/Aβ42 in cognitively normal people was associated with an increased risk of conversion to dementia during a follow-up period, which averaged 3-4 years. Of 61 people who started the study in the cognitively normal group, 13 went on to develop measurable impairment. The subjects in the top 15 percent of values for tau/Aβ42 or p-tau181/Aβ42 had a 4-5 times higher risk of conversion during the follow-up period. The lowest 15 percent had a lower rate. The results support the idea that CSF levels of plaque and tangle-derived peptides are potentially useful antecedent markers of future dementia, the authors conclude.

The study also strengthened the relationship between CSF Aβ42 concentrations and brain amyloid load as measured by PET imaging with Pittsburgh compound B. A previous study by the same group, in a small number of people, showed that lower CSF Aβ42 correlated with higher amyloid in the brain, suggesting that brain amyloid may represent an Aβ sink (see ARF related news story). The present paper contains results from PET imaging of 50 people, and the relationship holds up, indicating that low CSF Aβ42 levels can be a useful surrogate for brain amyloid load.

Watching Statins at Work
One application of CSF and imaging markers is to clinical trials. There they can help with patient selection and also serve as potential surrogate markers of the clinical effects of the drug at hand. A study in the January Journal of Alzheimer’s Disease shows just such a use. Robert Riekse, working with Elaine Peskind and coworkers at the University of Washington and the VA Puget Sound Healthcare System in Seattle, measured CSF Aβ40 and 42, tau and phospho-tau, as well as soluble amyloid precursor protein (APP) α and β fragments and F2 isoprostanes to assess the effects of statin treatment on Alzheimer-related proteins in the brain.

In a small trial of 23 healthy adults between 34 and 87 years old, the researchers treated 10 people for 14 weeks with simvastatin, which penetrates the brain well, and 13 with pravastatin, which stays outside the CNS. They then assessed CSF marker levels. The researchers found no change in Aβ or any amyloid marker, but did see a decrease in p-tau181 in all subjects treated with simvastatin. In the pravastatin group, half the subjects showed declines in p-tau181, while half showed increases. There were no changes in any other markers in either group.

The differences in the effects of the two statins may have been due to differences in brain penetration. Interestingly, the decline in phospho-tau in some patients on the impermeant pravastatin correlated with their CSF serum albumin ratio, a measure of blood-brain barrier integrity, suggesting that in those patients with lower p-tau, pravastatin might have had access to the brain. The results point to CNS penetration as a critical determinant of whether a statin will affect markers, and support the idea that differences in accessibility could cause some of the variation seen with these agents in epidemiological and clinical trials for AD.

The changes in p-tau after simvastatin treatment were small, and in this group of healthy people, they were not consistent with all previous studies done in AD patients. Nonetheless, the results support a role for cholesterol metabolism in the formation of neurofibrillary tangles, and suggest that CNS-permeant statins might affect AD pathology.

As Simple as a Blood Test?
Finally, there is a new paper from Boston researchers who combed through gene expression profiles in blood cells to find tell-tale signs of neurodegeneration in early Parkinson disease (PD). First author Clemens Scherzer presented some of this work at the Society for Neuroscience Meeting in 2005 (see ARF related news story), and the paper appeared this week in PNAS online.

The course of PD resembles AD in that the target neurons degenerate for years before symptoms appear, early disease is hard to identify, and there is no test for risk of future disease. The same rationale, then, drives the search for biomarkers, a search that has some researchers looking outside the brain at more easily accessible tissues.

To do that, Scherzer and colleagues assembled blood samples from 105 people, half of them with early PD, and the rest either healthy controls or people with other neurodegenerative diseases. From computer analysis of the gene expression profiles of each, they built a set of eight marker genes whose expression highly correlated with early PD. By measuring expression levels in any person, they could then calculate an individual PD risk score. Application of this risk marker to a separate population of 39 people predicted PD better than current risk factors.

Similar approaches looking at gene expression in peripheral tissue have been undertaken for AD (see ARF related news story and meeting coverage on Tony Wyss-Coray’s work). In fact, a study by Scherzer with James Lah at Emory University in Atlanta was the first to find low expression of the apolipoprotein E receptor LR11/SorLa gene in blood cells from AD patients (Scherzer et al., 2004). That study provided insight into the AD disease process, as the researchers showed subsequently that the protein is a modifier of APP processing, and an imminent genetics study identifies the SorLa gene as a new LOAD risk gene.

In the PD study, a closer consideration of the eight marker genes showed they do not appear to be functionally related in any one pathway or process. However, each one is expressed in brain, and three have been linked to PD or neurodegeneration previously. In addition to those eight, the investigators looked at an additional 22 genes that showed significant changes in PD vs. control to present some new insights into potential disease processes. Some of these genes could represent molecular surrogates of the disease, the authors write, but more work will be needed to confirm that.—Pat McCaffrey


  1. Statins are one of the major “miracle” drugs developed by the pharmaceutical industry. They lower cholesterol (LDL) levels effectively, yet have few side effects [1,2]. They have additional appeal because they appear to exhibit beneficial actions toward other diseases such as osteoporosis, stroke, and inflammatory disorders. The reason for this appears to arise from the ability of statins to inhibit palmitoylation, which inhibits a variety of signal transduction pathways. Inhibition of signaling by rac and rho is particularly important because these pathways are thought to mediate inflammation. Statins also appear to modulate eNOS and neuroprotective proteins such as Bcl-2 [3-5].

    The hypothesis that statins might be beneficial in therapy of Alzheimer disease is appealing because these medications have so few side effects. Whether statins are actually beneficial for treating AD remains in question, though. Many epidemiological studies suggested that statins might protect against AD, and these are supported by two small prospective clinical trials. But some large prospective clinical trials failed to observe a benefit associated with statins. Is this because statins don’t help in AD or are there other issues? One important question is whether all statins are equal. Newer statins, such as atorvastatin and rosuvastatin are more potent than the earlier statins, such as lovastatin. Simvastatin appears to be intermediate in potency, being sufficiently powerful to inhibit inflammation under some conditions, but not as potent as the newest statins [6,7]. The older statins, including lovastatin and simvastatin, are more lipophilic and penetrate the blood-brain barrier better than the newer hydrophilic statins such as atorvastatin or rosuvastatin, as well as pravastatin (which is actually “older”) [8,9]. This raises the possibility that lipophilic statins might exhibit actions that differ from those of hydrophilic statins. This could account for some of the confusion present in the clinical literature as it relates to dementia. Two large prospective trials examining whether simvastatin or atorvastatin might benefit patients with AD are ongoing, which should clarify this issue.

    In the meantime, another debate in the field questions the mechanism of statins’ action in the brain. Multiple studies demonstrate that statins can lower Aβ secretion in cells, depending on the dose [10,11]. This presents the appealing hypothesis that statins might be able to delay the incidence or progression of AD by reducing Aβ production. However, most studies have failed to observe Aβ lowering in humans [12].

    The current study by Riekse and colleagues addresses both the issue of comparative statin efficacy and putative mechanism of action. They compare the effects of pravastatin (which does not penetrate the BBB well) and simvastatin (which does penetrate the BBB) using an elegant prospective in vivo human trial examining CSF. Neither statin reduced Aβ, but simvastatin did reduce phospho-tau levels.

    The interpretation of these data is nuanced. The absence of any change in Aβ levels suggests that statins don’t inhibit Aβ secretion in the brain. However, the dose of simvastatin (20 and 40 mg/day) was only moderate. Data from the cardiovascular literature indicate that one must use 80 mg/day of simvastatin to inhibit the HMG CoA reductase pathway sufficiently to reduce inflammation. However, the group did examine and observe a decrease in C-reactive protein, which suggests that the inflammatory pathway was inhibited. However, I think it is wise to interpret this study conservatively with respect to Aβ. The data are interesting because they show that simvastatin treatment induces a statistically significant reduction in phospho-tau. This suggests that simvastatin might be capable of exhibiting neuroprotective actions, which is consistent with data from Gibson Wood’s group showing that statins modulate growth factors and neuroprotective proteins [5]. The absence of any effect of pravastatin on phospho-tau highlights the potential importance of lipophilicity in statin action.

    One final point to note is that the volunteers were all nondemented and many subjects were quite young (with the pravastatin group being younger than the simvastatin group). This means that the changes in phospho-tau were not necessarily related to the degenerative processes associated with AD.

    The main impact of this study is that it highlights potential differences in action among the statins and illuminates possible actions of statins in the brain.


    . The cellular biochemistry of cholesterol and statins: insights into the pathophysiology and therapy of Alzheimer's disease. CNS Drug Rev. 2004 Summer;10(2):127-46. PubMed.

    . Re-assessing the relationship between cholesterol, statins and Alzheimer's disease. Acta Neurol Scand Suppl. 2006;185:63-70. PubMed.

    . Effects of the 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors, atorvastatin and simvastatin, on the expression of endothelin-1 and endothelial nitric oxide synthase in vascular endothelial cells. J Clin Invest. 1998 Jun 15;101(12):2711-9. PubMed.

    . 3-Hydroxy-3-methylglutaryl coenzyme a reductase and isoprenylation inhibitors induce apoptosis of vascular smooth muscle cells in culture. Circ Res. 1998 Sep 7;83(5):490-500. PubMed.

    . Chronic administration of statins alters multiple gene expression patterns in mouse cerebral cortex. J Pharmacol Exp Ther. 2005 Feb;312(2):786-93. PubMed.

    . Comparative effects of simvastatin and atorvastatin in hypercholesterolemic patients with characteristics of metabolic syndrome. Clin Ther. 2003 Jun;25(6):1670-86. PubMed.

    . Comparison of efficacy and safety of atorvastatin and simvastatin in patients with dyslipidemia with and without coronary heart disease. Am J Cardiol. 2002 Mar 15;89(6):667-71. PubMed.

    . Statins of different brain penetrability differentially affect CSF PLTP activity. Dement Geriatr Cogn Disord. 2006;22(5-6):392-8. PubMed.

    . In vivo and in vitro blood-brain barrier transport of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors. Pharm Res. 1994 Feb;11(2):305-11. PubMed.

    . Simvastatin strongly reduces levels of Alzheimer's disease beta -amyloid peptides Abeta 42 and Abeta 40 in vitro and in vivo. Proc Natl Acad Sci U S A. 2001 May 8;98(10):5856-61. PubMed.

    . Statins cause intracellular accumulation of amyloid precursor protein, beta-secretase-cleaved fragments, and amyloid beta-peptide via an isoprenoid-dependent mechanism. J Biol Chem. 2005 May 13;280(19):18755-70. PubMed.

    . Plasma levels of beta-amyloid(1-40), beta-amyloid(1-42), and total beta-amyloid remain unaffected in adult patients with hypercholesterolemia after treatment with statins. Arch Neurol. 2004 Mar;61(3):333-7. PubMed.

    View all comments by Benjamin Wolozin
  2. This is an interesting and useful paper. It confirms in a large cohort
    of MCI subjects that hippocampal atrophy is an important marker of
    risk of progression to AD. More importantly, however, the authors show
    that qualitative assessment using an easily applied visual rating
    scale can be used to identify those at high risk of progression to
    clinical AD—this is of more direct relevance to current clinical
    practice than complicated manual (or other) measurements of
    hippocampal volume. The implication is that MCI patients who do not
    have an MRI or, having had an MRI, do not have their hippocampal
    atrophy adequately evaluated, have not had their risk of progression to AD properly assessed.

    View all comments by Nick Fox
  3. This study advances efforts to define individuals with MCI at greatest risk for AD, and it will be important in follow-up studies that define the sensititivy, specificity, positive and negative predictive value of MTA measures for distinguishing who will/will not progress to AD or other dementing disorders.

    View all comments by John Trojanowski
  4. This study adds further to our ability to exploit chemical and imaging biomarkers for the diagnosis of early AD, while emphasizing the need to identify other biomarkers that may predict who will progress to develop Abeta and tau pathologies prio to the onset of these AD pathologies.

    View all comments by John Trojanowski
  5. It is always an interesting prospect that gene expression patterns in blood may have utility as molecular markers, in this case for Parkinson disease. Using rigorous statistical methods, Clemens Scherzer and colleagues identified a battery of genes whose expression patterns in blood differ in individuals with PD from those without PD. Conceptually, identification of a biomarker phenotype in peripheral blood is clinically very attractive (for identification of PD or any disease or condition), and this study certainly advances the field in this regard and provides the impetus for further study of these identified candidate markers.

    It remains unclear, however, what the gene profile is really measuring—this is the “state-versus-trait” issue that plagues most, if not all, biomarker studies. The real question, of course, is whether this or any set of molecular markers can identify individuals with “preclinical” disease. Similar to AD, PD pathology develops over many years, perhaps decades, with clinical manifestations becoming apparent only after significant and substantial neuronal cell death has taken place (estimated to be about 70 percent of vulnerable dopaminergic neurons in PD). For biomarkers to have practical utility, they must be able to identify individuals in the preclinical phases of the disease, before such dramatic neuronal cell death has taken place, in order for putative treatments to have the best chance to preserve normal function in affected individuals. It will be very interesting to see if the molecular signatures identified in this study are able to identify individuals with PD pathology, but prior to the onset of symptoms.

    View all comments by Anne Fagan
  6. This study presents an intriguing preliminary finding supporting the involvement of lipophilic statins in AD tau pathophysiology. Statins are known to modulate enzymes involved in tau phosphorylation either directly or indirectly through inflammatory or Rho-mediated pathways, and the present results are consistent with this literature.

    However, the authors’ assertion that the primary use of statins is in the delay of neurodegeneration rather than in the effective treatment of the disease may be premature, given that the definitive clinical trials designed to address this very question have yet to read out. Nevertheless, the paper adds to a growing body of literature supporting the pleiotropic effects of statins and potential benefit in AD beyond modulation of cholesterol-specific pathways.

    View all comments by Holly D. Soares

Make a Comment

To make a comment you must login or register.


News Citations

  1. Diagnosis of AD—Does Spinal Fluid Hold the Key?
  2. Brain Imaging Speaks Volumes about AD and the Aβ Sink
  3. SfN: Early Steps toward Parkinson Disease Blood Test
  4. Looking outside the Brain for Early Signs of AD
  5. Translational Biomarkers in Alzheimer Disease Research, Part 5

Paper Citations

  1. . Vitamin E and donepezil for the treatment of mild cognitive impairment. N Engl J Med. 2005 Jun 9;352(23):2379-88. Epub 2005 Apr 13 PubMed.
  2. . Prediction of AD with MRI-based hippocampal volume in mild cognitive impairment. Neurology. 1999 Apr 22;52(7):1397-403. PubMed.
  3. . Loss of apolipoprotein E receptor LR11 in Alzheimer disease. Arch Neurol. 2004 Aug;61(8):1200-5. PubMed.

Further Reading

Primary Papers

  1. . Cerebrospinal fluid tau/beta-amyloid(42) ratio as a prediction of cognitive decline in nondemented older adults. Arch Neurol. 2007 Mar;64(3):343-9. Epub 2007 Jan 8 PubMed.
  2. . Effect of statins on Alzheimer's disease biomarkers in cerebrospinal fluid. J Alzheimers Dis. 2006 Dec;10(4):399-406. PubMed.
  3. . Molecular markers of early Parkinson's disease based on gene expression in blood. Proc Natl Acad Sci U S A. 2007 Jan 16;104(3):955-60. PubMed.
  4. . Qualitative estimates of medial temporal atrophy as a predictor of progression from mild cognitive impairment to dementia. Arch Neurol. 2007 Jan;64(1):108-15. PubMed.