Introduction

As Tolstoy wrote in Anna Karenina, “Happy families are all alike; every unhappy family is unhappy in its own way.” The same might be said of the brain. Normal brains look and behave similarly, but those affected by different neurodegenerative diseases are unique, since each disease targets a select subset of cells. In Parkinson disease, for example, dopaminergic neurons in the substantia nigra suffer, while their neighbors in the tegmentum remain relatively unscathed. In amyotrophic lateral sclerosis, lower motor neurons bear the brunt of disease, while upper motor neurons are less likely to be affected. This selective cell vulnerability presents a puzzle, particularly in the case of disease caused by inherited mutations: If the mutation is present in every cell in the body, what leads some neurons to resist its effects while others degenerate?

Analyzing differences between affected and unaffected cells might provide the answer, but separating them has not always been easy. Increasingly, researchers have taken advantage of modern technology, such as laser capture microscopy, to isolate individual cells and compare their gene expression patterns. The mRNAs expressed in each cell type may hint at their vulnerabilities. Microarrays and other screening tools, researchers hope, will help determine why some cells falter, and others survive particular disease onslaughts.

How are these types of analyses advancing our understanding of selective cell vulnerability? A Webinar led on 14 September 2010 by Eva Hedlund of the Karolinska Institute in Stockholm, Sweden, explored the value of microdissection and microarray analysis in studying neurodegenerative diseases. Hedlund discussed her latest results on ALS, and Chee-Yeun Chung of MIT’s Whitehead Institute shared her data on selective cell vulnerability in Parkinson disease. Rickard Sandberg, also from the Karolinska Institute, presented a new technique—RNA deep sequencing—that allows him to discover not only which mRNAs are present in tissues, cell lines, or single cells, but which splice forms they represent. Joining these presenters for a panel discussion were Stanislav Karsten of the University of California in Los Angeles, and Stephen Ginsberg of the Nathan Kline Institute in Orangeburg, New York.
  

Webinar Recordings

Given some technical difficulties experienced during the first presentation, we had re-recorded Dr. Eva Hedlund’s talk, which is available here below. The following talk, that of Chee-Yeun Chung, starts at minute 30, while that of Rickard Sandberg, begins at minute 48.

Slide Presentations

  • View a larger version of Eva Hedlund's slides.

  • View a larger version of Chee-Yeun Chung's slides.

  • View a larger version of Rickard Sandberg's slides.

Background

Background Text
By Amber Dance

In general, neurodegenerative diseases target specific cells, leaving many of their closest neighbors relatively unscathed. What makes some cells so susceptible to disease while others are tough enough to withstand the same insults? Using laser capture microdissection and microarrays, researchers are starting to answer that question by comparing gene expression profiles among cells.

Eva Hedlund, of the Karolinska Institute in Stockholm, Sweden, recently reported on selective cell vulnerability in motor neuron diseases (Hedlund et al., 2010). She and her colleagues noted that although motor neurons degenerate in ALS, spinal muscular atrophy (SMA), and spinobulbar muscular atrophy (SBMA), each disease hits specific cell populations. All three conditions sicken ventral motor neurons, while only ALS and SBMA affect lower cranial nerves. Upper cranial nerves are usually spared.

To determine the differences among these three neural populations, Hedlund and colleagues used laser capture microdissection (LCM) to isolate motor neurons from different areas—midbrain cranial nerves, brain stem, and cervical spinal cord—of normal rats. They then used microarray analysis to look for differences in gene expression between single cells. The data indicate that IGF-1 and IGF-II, which can be neuroprotective, are highly expressed in upper cranial neurons—perhaps explaining how they resist degeneration. On the other hand, members of the ubiquitin-based proteolysis system are more strongly expressed in spinal motor neurons. Ubiquitin-mediated proteolysis has been implicated in motor neuron disease.

Chee-Yeun Chung of MIT’s Whitehead Institute has adopted a similar LCM strategy to study Parkinson disease (Chung et al., 2005). Chung and colleagues found that in mice, vulnerable A9 dopaminergic neurons, compared to A10s, express more of the pro-apoptotic genes caspase-7 and Bcl2-like 11, perhaps explaining why A9 neurons are more susceptible to neurodegeneration. More recently, the researchers found that in mice, primates, and humans, the transcription factor orthodenticle homeobox 2 is highly expressed in A10 neurons compared to A9 (Chung et al., 2010), raising the possibility that this protein might be neuroprotective. Further, they found that overexpressing this transcription factor in cultured neurons protected the cells from the toxin MPP+, which induces Parkinson disease in rodents. The work has also led them to examine the protein phosphatase inhibitor G-substrate and the ras-related protein RAB3B as protective factors in A10 neurons (Chung et al., 2009; Chung et al., 2007).

One shortcoming in standard microarray analysis is that while it can determine which genes are transcribed when and where, it does not always inform on alternative splicing patterns. To study transcriptomes at this level of detail, Rickard Sandberg, also at the Karolinska Institute, applies a new technique: deep RNA sequencing, also known as RNA-seq (reviewed in Wang et al., 2009; Ramsköld et al., 2009). For RNA-seq, researchers fragment and reverse transcribe whole RNA samples from tissues, cell lines, or single cells, then sequence all the cDNAs. Common splice forms will show up many times in the sequencing results, while rarer forms will not appear as often, allowing scientists to quantify splice form levels (Wang et al., 2008). RNA-seq is more sensitive than microarrays.

Alzforum is pleased to have Hedlund and her co-panelists share their data and expertise during this Webinar. As always, we welcome your comments both prior to and during the event.

Comments

Make a Comment

To make a comment you must login or register.

Comments on this content

  1. I note with interest the forthcoming Webinar on this topic. I would like to comment that Yan Cao, at Affymetrix, and I did some informal work a few years ago on the number of neurons that were required in order to provide sufficient starting material for then current Affymetrix arrays. The number was in the range of 500-1,000, using two rounds of amplification. Newer arrays may be more sensitive, but still require significant numbers of cells.

    My colleagues at The Translational Genomics Research Institute (TGen), especially Winnie Liang, have used more recent Affymetrix arrays with Applied Biosystems Arcturus Laser Capture Microscopy (which we found leaves appreciable material behind) and still required significant numbers of cells with amplification.

    In 1998, we published a paper (Chow et al., 1998) using immunohistochemically identified single neurons with two rounds of modified Eberwine amplification and cDNA arrays with radioactive reporting. The conclusion I would draw is that "true" single-cell array studies are currently possible with two rounds of amplification, cDNA arrays, and radioactivity.

  2. The theme of selective cell vulnerability in neurodegenerative (but also in other amyloid) diseases is of the utmost importance to better understand the molecular basis of the pathological aggregate-cell interaction and its consequences. Most data so far reported highlight the importance of a number of biochemical features providing the cell the ability to resist the toxic effects of aggregates. Besides general protective mechanisms, including the complement of molecular chaperones and the efficiency of the ubiquitin-proteasome system, many other factors may contribute to the efficiency with which cells face the toxic insult of amyloids. Three have emerged as particularly important: the efficiency of the mechanisms exploited by the cell to control Ca2+, to control ROS levels, and the biochemical and biophysical properties of the cell membranes (notably the plasma membrane) (1). It is well established that amyloids, particularly the early oligomeric forms, interact with cell membranes, modifying their biophysical features and permeability. In most cases, this results in derangement of intracellular free Ca2+ levels, inducing a condition of oxidative stress. In this regard, cells with efficient Ca2+ pumps and/or high antioxidant capacity appear more resistant. It has also been repeatedly reported that amyloid interactions with the cell membrane, in most cases the earliest event in amyloid toxicity, occur at specific sites, notably lipid rafts, and depend on the membrane lipid content; in fact, amyloid/membrane interactions appear to be favored by the presence of anionic sites, such as those provided by the ganglioside GM1, and disfavored by the membrane-stiffening effect of cholesterol. GM1 also appears to favor protein/peptide misfolding and aggregation (2,3). However the complex interplay between the content of the two lipids is not fully understood, even though it is known that different cells, or the same cells in different functional states, display different amounts of these and other lipids.

    <p>More recently, amyloid polymorphisms and microenvironmental conditions in tissue have also emerged as important. Actually, in vitro, different destabilizing conditions may result in differently misfolded forms of the same protein/peptide, giving rise to distinct aggregation pathways populated by structurally different intermediates with variable toxicities. This may also hold true in tissue. A recent report showed that the huntingtin exon-1 gene product, with an expanded polyglutamine, can, under different conditions, aggregate in vitro into fibrils displaying different flexibilities and toxicities. Apparently, htt amyloids extracted from different regions of the brains of HD mice had comparable conformations and toxicities (4).

    </p><p>Finally, we have recently reported that the structural features of amyloids are important for their ability to interact with, and to permeabilize, the cell membrane, and, accordingly, for their toxicity (5). These data, together with recent findings that amyloid fibrils can undergo fragmentation in the presence of shearing forces or lipid membranes (6,7), and that such behavior is likely to occur even in tissue (8), support the idea that amyloid cytotoxicity in vivo, in addition to specific biochemical features of a cell type (including the levels of antioxidant defenses and Ca2+-ATPase activities, the extent of chaperone complement, and the efficiency of the machineries aimed at protein turnover), can also depend on local environmental features. This view is further strengthened by our recent findings showing that cytotoxicity cannot be considered a general property of early amyloids (5). Rather, it appears to be a property that emerges from a complex interplay between oligomer conformational and biophysical properties and the biochemical features of the cell membrane resulting from its lipid content.

    </p><p>This work is still at the beginning, and many other players involved in amyloid toxicity modulation in tissue are expected to emerge as the research focuses on different aspects.

    References:

    . Insights into the molecular basis of the differing susceptibility of varying cell types to the toxicity of amyloid aggregates. J Cell Sci. 2005 Aug 1;118(Pt 15):3459-70. PubMed.

    . Formation of amyloids by Abeta-(1-42) on NGF-differentiated PC12 cells: roles of gangliosides and cholesterol. J Mol Biol. 2007 Aug 24;371(4):924-33. PubMed.

    . Formation of toxic Abeta(1-40) fibrils on GM1 ganglioside-containing membranes mimicking lipid rafts: polymorphisms in Abeta(1-40) fibrils. J Mol Biol. 2008 Oct 17;382(4):1066-74. PubMed.

    . Distinct conformations of in vitro and in vivo amyloids of huntingtin-exon1 show different cytotoxicity. Proc Natl Acad Sci U S A. 2009 Jun 16;106(24):9679-84. PubMed.

    . A causative link between the structure of aberrant protein oligomers and their toxicity. Nat Chem Biol. 2010 Feb;6(2):140-7. PubMed.

    . Fibril fragmentation enhances amyloid cytotoxicity. J Biol Chem. 2009 Dec 4;284(49):34272-82. PubMed.

    . Lipids revert inert Abeta amyloid fibrils to neurotoxic protofibrils that affect learning in mice. EMBO J. 2008 Jan 9;27(1):224-33. PubMed.

    . Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synapse loss near senile plaques. Proc Natl Acad Sci U S A. 2009 Mar 10;106(10):4012-7. PubMed.

References

Paper Citations

  1. . Cell type-specific gene expression of midbrain dopaminergic neurons reveals molecules involved in their vulnerability and protection. Hum Mol Genet. 2005 Jul 1;14(13):1709-25. PubMed.
  2. . The transcription factor orthodenticle homeobox 2 influences axonal projections and vulnerability of midbrain dopaminergic neurons. Brain. 2010 Jul;133(Pt 7):2022-31. PubMed.
  3. . Functional enhancement and protection of dopaminergic terminals by RAB3B overexpression. Proc Natl Acad Sci U S A. 2009 Dec 29;106(52):22474-9. PubMed.
  4. . An endogenous serine/threonine protein phosphatase inhibitor, G-substrate, reduces vulnerability in models of Parkinson's disease. J Neurosci. 2007 Aug 1;27(31):8314-23. PubMed.
  5. . RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet. 2009 Jan;10(1):57-63. PubMed.
  6. . An abundance of ubiquitously expressed genes revealed by tissue transcriptome sequence data. PLoS Comput Biol. 2009 Dec;5(12):e1000598. PubMed.
  7. . Alternative isoform regulation in human tissue transcriptomes. Nature. 2008 Nov 27;456(7221):470-6. PubMed.

External Citations

  1. Hedlund et al., 2010

Further Reading

Papers

  1. . Neurotoxicity induces cleavage of p35 to p25 by calpain. Nature. 2000 May 18;405(6784):360-4. PubMed.