Once thought to be a simple go-between, RNA is recently attracting much more attention as an important regulator of gene expression in its own right. At the Society for Neuroscience annual meeting, 15-19 November in Washington, DC, Rosa Rademakers of the Mayo Clinic in Jacksonville, Florida, presented evidence for the involvement of a microRNA (miRNA) in frontotemporal dementia, which is the cause of 10 to 20 percent of early onset dementia cases. The research is published in the December 1 Human Molecular Genetics.

MicroRNAs can degrade or repress translation of their targets, and also upregulate mRNA expression in some cases. The short sequences, 21 to 23 nucleotides, have been linked to neurodegenerative diseases such as Alzheimer’s (see ARF related news story and ARF news story); for review, see Bushati and Cohen, 2008), and Rademakers predicts that miRNA dysfunction will be implicated in many more conditions.

Rademakers’ research is the first evidence for miRNA involvement in frontotemporal dementia. She and colleagues found that the human miRNA miR-659 binds to the 3’-untranslated region of the progranulin mRNA, blocking translation. People with a progranulin mutation that enhances this binding can have significantly lowered progranulin levels, leading to frontotemporal dementia, Rademakers said, making the mutation a major risk factor for disease.

Scientists had previously linked frontotemporal lobe degeneration with ubiquitin- and TDP-43-positive inclusions (FTLD-U) to mutations in the progranulin gene (PGRN) (Baker et al., 2006, Cruts et al., 2006, and see ARF related news story). (Another form of frontotemporal dementia exhibits tau-positive inclusions.) All of the more than 60 PGRN mutations previously associated with FTLD-U are dominant, loss-of-function mutations that cause premature termination of the PGRN mRNA (Gass et al., 2006). These mutations cut progranulin protein levels by approximately 50 percent in heterozygotes, which is sufficient to cause disease (Van Damme et al., 2008).

Progranulin’s role in the nervous system is unclear. It acts as an anti-inflammatory in its full-length form but, when proteolytically cleaved to form granulins, is pro-inflammatory. Progranulin is present in neurons and microglia. In the periphery, progranulin is involved in wound healing (for review, see He and Bateman, 2003).

Rademakers and colleagues discovered the link between miRNA-659 and progranulin when they sequenced PGRN genes from 378 FTLD patients. Among 339 patients with no PGRN coding mutation, they found a polymorphism, rs5848, in the 3’UTR that did not exhibit Hardy-Weinberg equilibrium: more of the patients (55) were TT homozygous at this position than would be predicted by chance. TT homozygotes made up 16.2 percent of the FTLD patients, compared to 9.3 percent of control subjects. Among 59 FTLD-U patients without other FTLD-linked mutations, the rs5848 TT genotype frequency was 25.4 percent. All TT subjects showed the same neuropathology, with neuronal cytoplasmic inclusions and short, thin neurites in the cortex. The similar pathology suggests a common cause of disease among these patients. While the coding-sequence PGRN mutations are dominant, heterozygotes for rs5848 TT did not have increased susceptibility to FTLD-U, suggesting the mutation is recessive.

The location of the mutation, outside the coding sequence, suggests that it affects PGRN expression rather than function. On Western blots, extracts from rs4858 TT patients had progranulin levels reduced by approximately 30 percent compared to CC homozygous control extracts. However, quantitative RT-PCR found no difference in the PGRN mRNA levels between CC and TT individuals. “That was the first time we thought maybe microRNAs are involved,” Rademakers said. Using computer analysis, Rademakers predicted that a single miRNA, miR-659, could bind to the site of the mutation. In silico, changing the cytosine to a thymine altered the PGRN mRNA-miRNA binding, shifting miR-659’s position and allowing it to bind the transcript more tightly, with three additional nucleotide pairings. After confirming that miR-659 is expressed in human brain tissue, including the frontal and temporal neocortex, Rademakers hypothesized that miR-659 binds to PGRN mRNA, blocking translation, and that the mutant mRNA binds miR-659 more tightly. This knockdown of PGRN, in turn, could cause FTLD.

To test this hypothesis, the scientists turned to cell experiments. When they transfected human M17 neuroblastoma cells with miR-659, they expressed less progranulin, while cells transfected with a nonspecific control miRNA had normal PGRN levels. To further quantify miR-659’s activity, Rademakers and colleagues engineered a luciferase reporter with the PGRN 3’-UTR. In mouse N2A neuroblastoma cells (used to avoid any effects from endogenous human miR-659), miR-659 decreased luciferase expression in a dose-dependent manner, further suggesting miR-659 acts on the GRN 3’-UTR to knock down expression.

Rademakers posits that the lower one’s progranulin levels, the higher the risk of developing FTLD-U. A single T allele appears not to lower PGRN translation below the acceptable threshold, but two mutations can repress enough GRN translation to cause disease.

“The data do look convincing,” said Alison Goate, a geneticist at Washington University in St. Louis. “They provide pretty good evidence that there is an increased risk of FTLD-U associated with people who are homozygous for this variant.” However, she noted that the patient numbers were small—only 59 subjects in the second genetic analysis, and only 14 for the GRN expression studies. “In some ways, they were pretty lucky they found it,” Goate said.

MicroRNAs appear poised to be major players in neurodegenerative disorders, and Rademakers said her research is more evidence for that hypothesis. However, she recognizes that other scientists may take some convincing. “People are very skeptical of these novel mechanisms that become popular,” she said. Goate also noted that TDP-43 is implicated in several neurodegenerative diseases, and that perhaps GRN has a role in promoting TDP-43-based diseases other than FTLD-U. In the future, Rademakers intends to analyze sequence variation and expression levels of miRNAs to further characterize their role in FTLD.—Amber Dance


  1. The manuscript by Rademakers and colleagues provides evidence that increased binding of miR-659 to the 3’UTR of the GRN gene could underlie an important risk for TDP-43-positive frontotemporal dementia (FTLD-U). These data bring strong clinical support for the role of microRNAs in neurodegenerative disorders in humans. These results are consistent with a loss of function of the GRN gene in the disease, further linking gene dosage effects in neurodegenerative disorders (as seen, e.g., with APP in Alzheimer disease and SNCA in Parkinson disease).

    I think Amber Dance did a fantastic job reviewing the highlights of this paper. I would like to discuss additional issues with regard to certain technical and mechanistic aspects of these findings, which could be taken into account when interpreting the data.

    First, miR-659, located on chromosome 22 in humans, seems to be relatively very weakly expressed in adult brain (with cycle threshold [Ct] values of approximately 32 as measured by qRT-PCR). Therefore, whether endogenous miR-659 levels are sufficient to regulate GRN levels in vivo remains speculative. Mechanistically, one must envisage that regulation of GRN mRNA by miR-659 occurs in a cell-autonomous fashion. One possibility, not shown here, is that miR-659 is expressed in specific cell types, such as the granular cell layer of the cerebellum where GRN protein is decreased (it should be noted that the qRT-PCR for miR-659 was performed on whole tissues). In my opinion, this would strongly strengthen the biological significance of the proposed mode of regulation.

    Here, the authors use basic, but widely accepted in vitro systems to validate their hypothesis. First, artificial overexpression of miR-659 (at a concentration of 12 nM) in human M17 neuroblastoma cells leads to decreased expression of endogenous GRN protein levels (note that inverse experiments using antisense oligonucleotides to block endogenous miR-659 was not performed, possibly due to the extremely low levels of this microRNA in these cells). Whether GRN mRNA levels are affected in these conditions is not shown. Then, additional studies were conducted in mouse Neuro2A cells using luciferase-based constructs containing the GRN 3’UTR. In these latter experiments, functional effects on GRN expression are seen with the mutant TT construct at concentrations starting at 5 pM of exogenous miR-659. Again from a mechanistic point of view, it would be interesting to see whether the “increased” binding (i.e., increased sequence complementarity) of miR-659 to the mutant TT allele causes an siRNA effect (thus degradation of mRNA). It should be noted, however, that, in affected patients, GRN mRNA (from total tissue sections) is not affected.

    Interestingly, the predicted target site (more particularly the “seed” sequence) for miR-659 in the GRN 3’UTR is only conserved in humans, and is not found in other mammals including mouse and dog (e.g., see www.targetscan.org). Similarly, miR-659 is, at least for now, only found in humans. Interestingly, the GRN 3’UTR is quite short (approximately 300 bp in length). In comparison, the BACE1 and APP 3’UTRs, which equally have functional microRNA target sites, are approximately 4,000 bp and 2,000 bp in length, respectively.

    Overall, these findings provide novel and important clues into the development of FTLD-U. In addition, this study contributes to the potential role of microRNA pathways in the development of neurodegenerative disorders in human. I agree that relatively few patients were analyzed here to make definitive conclusions with regard to the biological relevance of these findings.

Make a Comment

To make a comment you must login or register.


News Citations

  1. BACE in Alzheimer’s—Does MicroRNA Control Translation?
  2. Number 107: MicroRNA Gets to First BACE in AD Brain
  3. Birds of a Feather…Mutations in Tau Gene Neighbor Progranulin Cause FTD

Paper Citations

  1. . Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature. 2006 Aug 24;442(7105):916-9. PubMed.
  2. . Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature. 2006 Aug 24;442(7105):920-4. PubMed.
  3. . Mutations in progranulin are a major cause of ubiquitin-positive frontotemporal lobar degeneration. Hum Mol Genet. 2006 Oct 15;15(20):2988-3001. PubMed.
  4. . Progranulin (granulin-epithelin precursor, PC-cell-derived growth factor, acrogranin) mediates tissue repair and tumorigenesis. J Mol Med (Berl). 2003 Oct;81(10):600-12. PubMed.

Further Reading

Primary Papers

  1. . Common variation in the miR-659 binding-site of GRN is a major risk factor for TDP43-positive frontotemporal dementia. Hum Mol Genet. 2008 Dec 1;17(23):3631-42. PubMed.