Monday, March 23, 2009

MicroRNAs and Mental Illness

ResearchBlogging.orgA paper published in Proceedings of the National Academy of Sciences caught my eye earlier this month. In this paper, Dr. Jannet Kocerha et al. describe the potential link between expression of a micro-RNA (miRNA) and schizophrenia. Because miRNAs are super interesting, and because I'm into understanding the molecular etiology of brain disorders, I thought it would be a good paper to blog about. Then it fell to the bottom of my Google Reader, as tends to happen when one is busy writing rotation proposals, teaching seventh grade science classes, and enjoying spring break. So, I apologize for my tardy reporting. That said, on to miRNAs!

Biochemistry fans will already be familiar with miRNAs, but it's worth explaining them as a preface to this paper because they were discovered relatively recently and tend to fly in the face of traditional biology rules. If you've taken a biology class, you probably learned about The Central Dogma: DNA makes RNA (through transcription) which makes protein (through translation). You may have also learned about the regulation of gene expression, which is a key concept in genetics. Regulated gene expression is what allows the different cells in your body to make different proteins and specialize in different functions, despite the fact that they all contain identical DNA. Depending on the cell, some genes will be turned on and others will be turned off. This allows each cell to focus on its unique function. There are several ways in which gene expression can be regulated, including transcription factors and epigenetic mechanisms. About 15 years ago, scientists discovered that certain RNA molecules can also regulate gene expression (Lee et al., 1993). RNAs that have this function are called "micro-RNAs," because they tend to be much smaller than messenger RNAs (mRNAs; the RNA molecules that carry the message used in protein translation).

miRNAs are freaky because they defy The Central Dogma. These non-coding RNAs (that is, RNAs that do not carry the sequence for translating a protein) arise from parts of the genome often characterized as "junk DNA" due to lack of association with any known protein product. It turns out that some of the "junk" is pretty important -- DNA that codes for miRNAs is essential for proper regulation of gene expression. If you mess with gene expression, all kinds of things can go wrong in a cell. If they regulate important genes, miRNAs can be critical for cellular function.

But how do miRNAs actually work? As I mentioned before, under normal circumstances, DNA makes RNA makes protein. In order for a gene to be expressed, it must be transcribed from DNA to mRNA. To make protein, the mRNA is then translated into a chain of amino acids. miRNAs work by disrupting the translation of mRNA into protein. These pieces of RNA form a sequence that is complementary to a small part of the target mRNA (in other words, where the mRNA has a C, the miRNA has a G; where the mRNA has an A, the miRNA has a U; etc.). When a complementary miRNA encounters its target, it will bind to the mRNA at the site of complementarity. This creates a double-stranded RNA, which prevents the cell's translation machinery from making protein out of the mRNA, and in some cases triggers RNA degradation.

So what does this mean? Well, it adds another possible layer of regulation for gene expression. Imagine a cell expresses Gene X. Some other process might control the expression of the complementary miRNA Y. So, in cases where Gene X is switched on but miRNA Y is switched off, Gene X leads to the production of Protein X. If miRNA Y does get switched on, though, Protein X doesn't get made -- even if Gene X is still being transcribed! All sorts of complicated feedback cycles can be created in this way, especially when certain miRNAs impact the production of other miRNAs, and on ad infinitum.

After that lengthy preface, I will now discuss the paper. Its main points and key criticisms can all be gleaned from the commentary by Dr. Joseph Coyle in the same issue of PNAS, but I will attempt to summarize the article in my own words. It's good practice for me!

Kocerha et al. were investigating a mouse model of schizophrenia in which a certain class of neurotransmitter receptor (the NMDA receptor) does not function properly. This model is based on work done in humans that has implicated decreased NMDA-receptor-mediated signaling in many brain disorders, including schizophrenia. For their paper, the researchers modeled NMDA hypofunction in several ways: they treated mice with a drug called dizocilpine, which is an NMDA receptor antagonist, and they examined mice with a mutation in a gene called Grin1, which is crucial for the formation of normal NMDA receptors. In either case, the mice exhibited "schizophrenia-like deficits" in motor behavior due to the experimental manipulations.

To identify which miRNAs might be associated with decreased NMDA receptor function, the scientists used a microarray. This technique allows us to measure whether any of a large number of RNAs are upregulated or downregulated by experimental conditions. In this case, microarray data showed that expression of a miRNA called miR-219 was significantly reduced by a single dizocilpine treatment. The researchers did not observe the same changes in mice chronically treated with dizocilpine, which is problematic because chronic treatment is considered by some to be a better model of schizophrenia than acute treatment. Still, the fact that dizocilpine had an effect on miR-219 expression made the miRNA worthy of further investigation. A similar effect to dizocilpine was seen in mice that possess a mutation in Grin1. These mice experience a 90% reduction in functional NMDA receptors that leads to locomotor behavior reminiscent of schizophrenic humans. The mutant mice were also found to have significantly decreased miR-219 in their brains.

To see whether the observed miRNA expression changes were indeed brought about specifically by the schizophrenia-like symptoms of dizocilpine treatment and Grin1 mutation, Dr. Kocerha and colleagues repeated their experiments after pre-treating the test subjects with haloperidol, an antipsychotic drug. They found that haloperidol effectively suppressed both the schizophrenia-like behaviors of the mice and the associated decrease in miR-219.

Thus convinced that miR-219 plays a role in NMDA receptor function, the researchers sought its mRNA target. Using bioinformatics, they looked for genes with RNA sequences complementary to the sequence of miR-219. One of the best candidates was CaMKIIγ, a crucial regulator of NMDA signaling. They tested miR-219's regulatory effect on CaMKIIγ by seeing whether miR-219 could inhibit the expression of the CaMKIIγ gene in cultured cells. The cells contained either normal CaMKIIγ (which has a sequence complementary to miR-219) or mutant CaMKIIγ (with mutations that made the sequence no longer complementary) connected to a reporter gene (luciferase). The reporter gene made it easy to quantify the amount of CaMKIIγ expressed in the cultured cells. These experiments showed that CaMKIIγ containing the normal miR-219 complementarity sequence was decreased by about 40% compared to CaMKIIγ that could not form double-stranded RNA with miR-219. Then, to test whether this effect was miR-219-dependent, the scientists used an antisense version of miR-219 to inhibit the reduction of CaMKIIγ. "Antisense" means the complementary sequence to the "sense" RNA that is normally found in the cell. In a way, they made a miRNA against the miRNA -- the antisense RNA binds to miR-219 perfectly, which prevents the miR-219 from binding to its complementary site on the CaMKIIγ mRNA. This antisense experiment restored a significant amount of CaMKIIγ expression, providing evidence that CaMKIIγ expression is indeed inhibited by miR-219. These results were further confirmed by measuring endogenous CaMKIIγ protein in cortical neuron cultures made to overexpress miR-219 and seeing a reduction in CaMKIIγ compared to controls.

Then came some really exciting stuff. After showing that miR-219 and CaMKIIγ interact in vitro, Kocerha et al. wanted to see if they could improve the schizophrenia-like symptoms of their mouse model by treating them with antisense miR-219. After infusing the brains of dizocilpine-treated mice with this miR-219 inhibitor (a tricky technique to master, because RNA degrades so easily), the team observed "markedly altered hyperlocomotion and stereotypy 30 min after dizocilpine administration ... when compared with mice receiving the LNA mismatch or saline." The mice treated with antisense miR-219 had less pronounced schizophrenia-like motor symptoms, although the effect of the drug was not completely eliminated. This seems to indicate that a significant portion of the behavioral effect of dizocilpine is mediated by miR-219.

Of course, we should take such experiments with a grain of salt, especially when animal models are being used to represent a complex psychiatric disorder like schizophrenia. Dr. Coyle points out one especially intriguing point in his commentary in PNAS: "It seems counterintuitive that reduction in miR-219 appears to be responsible for hyperactivity in the acute dizocilpine paradigm but reducing miR-219 levels with antisense infusion reverses dizocilpine-induced hyperactivity." Hopefully further research will be able to explain this apparent inconsistency. As for the broader implications of this study for schizophrenia patients, it's difficult to assess whether mice are experiencing hallucinations or mood disorders, so all of the results observed here are based on the motor symptoms of decreased NMDA receptor function (hypermobility, stereotyped repetitive movements). We already have drugs that can mimic the beneficial effect of miR-219 inhibition in these animal models, so at this point no one is advocating RNA interference as a potential treatment for psychiatric patients. This research does show a new mechanism by which schizophrenia symptoms and the drugs that relieve them may be interacting, however, which could lead to new avenues of drug development as we further elucidate the players in this biochemical pathway.

For those without subscriber access to PNAS, other reviews of the paper can be found at The Scripps Research Institute and Schizophrenia Research Forum.

References

Lee R.C., Feinbaum R.L., Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell 75(5): 843–54. (1993) DOI: 10.1016/0092-8674(93)90529-Y

Kocerha, J., Faghihi, M., Lopez-Toledano, M., Huang, J., Ramsey, A., Caron, M., Sales, N., Willoughby, D., Elmen, J., Hansen, H., Orum, H., Kauppinen, S., Kenny, P., & Wahlestedt, C. MicroRNA-219 modulates NMDA receptor-mediated neurobehavioral dysfunction Proceedings of the National Academy of Sciences 106(9): 3507-3512 (2009) DOI: 10.1073/pnas.0805854106

5 comments:

  1. I think "miRNAs are freaky because they defy The Central Dogma" is a bit misleading. The central dogma (which was only the dogma like 50 years ago) just talks about genetic information moving from nucleotides to coding proteins. If anything, reverse transcription/integration is a much freakier "violation" of the central dogma, since it implies that there is *heritable* information flow from RNA back to DNA. After all, no one is surprised that proteins regulate gene expression, so why not RNA? Are transcription factors violations of the central dogma?

    microRNAs, on the other hand, simply occupy the space of YANR (yet another non-coding RNA), which is the stomping grounds of things like snoRNAs, snRNAs, sRNAs (from bacteria), siRNAs, piwiRNAs, tRNAs, lincRNAs, and so on, which doesn't include the many sundry ways that mRNA folds can autoregulate gene expression. I think what's cool about microRNAs is how that they may regulate the expression of almost 30% of genes.

    ReplyDelete
  2. Eric, you make an excellent point about how non-coding RNAs fit into the central dogma (which of course is now known to have many exceptions). In my struggle to write an explanation of miRNAs that would illustrate their coolness to my non-scientist friends (and to avoid making this long post even longer), I was simplified a lot. But you're right, the gee-whiz excitement about non-coding RNAs has mostly worn off by now for biologists.

    ReplyDelete
  3. As a cognitive neuroscientist who is curious about these things, really appreciated your simple/clear explanation of this result and some genetics basics. Thanks very much :)

    ReplyDelete
  4. This is a great post, and why I'm going to subscribe to your blog. But, ahem, PUHLEASE! make your RSS full feed, instead of only the titles.

    ReplyDelete
  5. Hi caostaff,

    Glad you liked the post! I didn't realize that my RSS feed was truncated like that -- it should be publishing full length posts now.

    ReplyDelete