Bioinformatics Identifies Eight Master Regulators of ALS Neurodegeneration

July 2, 2015

During amyotrophic lateral sclerosis, motor neurons can fall victim to toxic secretions from the astrocytes that are supposed to support them. As scientists puzzle over the identity of those killer molecules, Serge Przedborski and colleagues at Columbia University in New York are working on how motor neurons respond to them. Using an unbiased, computational approach, they identified eight candidate transcription factors that seem to be involved. As they report in the July 2 Cell Reports online, one of the molecules surprised them—NF-κB, a pro-inflammatory transcription factor usually thought to protect cells. It seems to have a different role in the context of ALS, Przedborski said.

When Przedborski and other scientists isolate wild-type motor neurons and expose them to astrocytes containing the ALS gene SOD1, or even just media from those astrocytes, the motor neurons die by necroptosis, a form of programmed cell death related to inflammation (see Apr 2007 news). The same holds true for neurons exposed to astrocytes from people with sporadic ALS, and neurons in live mice that model the diseases (see Feb 2014 news; Oct 2011 news). How do astrocyte toxins initiate necroptosis in neurons? Przedborski had examined every molecular pathway he could think of without finding the answer. He turned to colleague Andrea Califano, co-senior author on the paper, for help. Califano developed a bioinformatics technique to identify transcriptional factors that serve as “master regulators” of biological processes (Lefebvre et al., 2010). The Master Regulator Inference algorithm (MARINa) starts with mRNA profiles and known data about which factors switch on which genes, and works backward to infer the transcription factors likeliest to cause an observed mRNA expression pattern.

Interactome interactions. Researchers profiled mRNAs and ran bioinformatics analyses to identify master regulators of neurotoxicity , which they then verified in cell culture. [Image courtesy of Cell Reports, Ikiz et al.]

In Przedborski’s lab, first author Burcin Ikiz developed a model system to generate those mRNA profiles. She cultured spinal motor neurons from embryonic mice, and exposed them to mSOD1 astrocyte-conditioned media (ACM) for different lengths of time. Ikiz was looking for the “point of no return,” when the neurons were committed to necroptosis even if she removed the astrocyte-tainted media. The neurons hit this point at three days; after that, half of them would be dead within the week. Ikiz compared the transcriptomes of those three-day neurons to those of neurons exposed to innocuous ACM from nontransgenic mice or those toting wild-type SOD1. She found 620 genes up- or downregulated by exposure to mSOD1 astrocyte media.

Ikiz then worked with joint first author Mariano Alvarez, from Califano’s group. Alvarez fed those mRNA profiles, plus an interactome from mouse brain that described the positive and negative targets of transcription factors, into MARINa. While Califano said this technique has a solid track record in cancer (Carro et al., 2010; Aytes et al., 2014), he was not certain it would be effective for neurobiology. “Much to Serge’s surprise, it actually worked,” quipped Califano. “We went from knowing very little about the effectors … to having a pretty precise repertoire.” Alvarez came up with 23 transcription factors that could explain the mRNA profiles Ikiz observed.

To test the role of those transcription factors in necroptosis, Ikiz silenced each individually (for one, she was unable to obtain effective silencing RNAs). For eight factors, silencing improved the survival of motor neurons exposed to the tainted ACM, but not motor neuron survival in general, suggesting they regulated a specific response to killer astrocytes.

Nf-κB stood out among the eight, because its activation already had been linked to inflammation and neurodegeneration in ALS (see Dec 2011 news; Frakes et al., 2014; Akizuki et al., 2013). To explore NF-κB’s role further, Ikiz performed a number of experiments. She observed that when neurons were exposed to mSOD1 ACM, NF-κB subunits were likelier to enter the nucleus, and bound more strongly to DNA. When she treated neurons with wedelolactone or BMS-345541, drugs that block NF-κB activity, the cells were protected from the toxic astrocytes. The same was true when she transduced cells with a dominant-negative version of IκB, which inhibits NF-κB—the cells withstood mSOD1 ACM.

Those results suggest that NF-κB and the other candidate master regulators are activated in mouse neurons exposed to the secretions of astrocytes with mSOD1. But what about people, in whom mSOD1 accounts for only a fraction of ALS cases? Ikiz repeated some of her experiments using motor neurons derived from human embryonic stem cells, and astrocytes from people who had sporadic ALS. Media from these astrocytes killed the neurons, but as with the mouse cultures, wedelolactone protected them.

Przedborski would not have predicted that NF-κB drives necroptotis. “We were very surprised,” he said. “If you ask most people who know about NF-κB, they will immediately tell you that NF-κB plays a survival role. Here, if anything, we would argue that it could drive the death of motor neurons.” He noted that much prior research on necroptosis was performed in immune cells, not neurons. Though NF-κB may be protective in lymphocytes, it could promote cell death in the context of motor neurons and ALS, he suggested.

Other researchers who did not participate in the study were willing to accept that possibility. “Califano’s method is pretty robust and reliable,” said Bin Zhang of the Icahn School of Medicine at Mount Sinai in New York. “With the systems-biologic approach, you always find something unexpected.” Junying Yuan of Harvard Medical School thought Ikiz’s findings fit in with NF-κB’s biology. The transcription factor is known to promote inflammation, and inflammatory cytokines can lead to necroptosis, she commented.

Both Yuan and Zhang noted that Ikiz and Mariano’s findings raise further questions, and both said they would like to see human data backing up these candidates as master regulators. “We do not know [the] upstream activators of these transcriptional factors, nor the downstream mediators that lead to necroptosis,” Yuan commented. Plus, Zhang added, researchers have not worked out how all these regulators would work together to drive programmed cell death. Przedborski’s group has already started to analyze a couple of the master regulators, he said.

As for Califano, his methods are already offering clues to other kinds of neurological disease, including Alzheimer’s. With collaborators at Columbia, he has applied MARINa and other algorithms to identify genetic variants linked with AD (Chen et al., 2014) and transcription factors that might control disease progression (Aubry et al., 2015). In the latter study, Califano and co-senior author Michael Shelanski came up with 38 possible master regulators at work in AD brains. Among their findings, they implicated p53 and dysregulation of acetylation in the disease.—Amber Dance

To view commentaries, primary articles and linked stories, go to the original posting on here.

Copyright © 1996–2015 Biomedical Research Forum, LLC. All Rights Reserved.