6 December 2008. For the first time, scientists from two labs have established co-cultures of astrocytes and stem cell-derived human motor neurons, and used the systems to probe the mechanism of motor neuron death in amyotrophic lateral sclerosis as well as search for possible treatments. In back-to-back papers published December 4 in Cell Stem Cell, groups led by Fred Gage at The Salk Institute in La Jolla, California, and Kevin Eggan at Harvard University, show that cultured human motor neurons are killed by rodent or human astrocytes expressing a mutation that causes ALS, just as cultured rodent motor neurons are. The results provide further confirmation that glia, not just the motor neurons themselves, contribute to non-cell autonomous motor neuron death in ALS. The cultured cells also offer a fast way to screen for treatments, and could even be developed in patient-specific fashion to identify personalized therapies for neurodegenerative diseases.
“The work is part of a growing effort to use human embryonic stem cells to model human disease,” Gage said. Animal models take time to breed and grow up, and may not accurately mimic human disease. Cells differentiated from embryonic stem cells are closer to the cells in adult humans, and are certainly a faster route for drug screening than whole organisms, making them a valuable addition to animal models. Similar in vitro systems might provide models for diseases such as Alzheimer’s, Parkinson’s, or spinal muscular atrophy, the scientists said.
The two research groups, working independently, used similar means to the same end. Both labs studied human stem cell-derived motor neurons cultured with glial feeder cells expressing mutant human superoxide dismutase 1 (SOD1-G37R or SOD1-G93A), which is responsible for 10 to 20 percent of familial ALS cases. The toxicity of mSOD1 glia to wild-type motor neurons has been shown before in animal models (Boillée et al., 2006; Yamanaka et al., 2008) and in vitro with rodent cells (Nagai et al., 2007 and see ARF related news story). Gage, Eggan, and colleagues proved the same is true in human cells. “They have done something very important, I think,” said Serge Przedborski of Columbia University in New York, who was not involved with either study. “Even if it’s incremental, it’s still a significant increment.”
The researchers used similar protocols to generate motor neurons. Both groups started by allowing suspensions of embryonic stem cells to form embryoid bodies. They then used retinoic acid, sonic hedgehog or a sonic hedgehog agonist, growth factors and other factors to induce differentiation along the motor neuron lineage. A feeder layer of astrocytes completed the cultures.
Paolo Di Giorgio, first author on the Eggan paper and now located at The Salk Institute, plated the stem cell-derived neurons on astrocytes isolated from the cortex of nontransgenic, wild-type human SOD1-, or SOD1-G93A-expressing mice. Carol Marchetto, first author on the Gage paper, used a different astrocyte source—cells from human fetal brain transfected with a lentiviral vector expressing SOD1-G37R or wild-type SOD1. Przedborski noted that the ideal feeder layer would be astrocytes isolated from someone who had ALS—“That would have been beautiful,” he said—but admitted that such cells would be difficult to obtain.
Both groups used a GFP marker, controlled by the motor neuron-specific promoter Hb9, to confirm differentiation into the right kind of cells. Marchetto used a lentiviral vector to transduce the differentiated cells, and Di Giorgio used electroporation on the embryonic cells, to insert the Hb9::GFP marker. In both cases five to 10 percent of cells showed Hb9 activity and other motor neuron markers. “We are very content that we’ve actually made a motor neuron through our in vitro differentiation,” said Gabriella Boulting, an author on the Eggan lab paper.
Marchetto took motor neuron confirmation a step further and showed that when co-cultured with muscle cells, the differentiated motor neurons formed rudimentary neuromuscular junctions. Using patch-clamp analysis, she found that Hb9::GFP-positive cells were electrophysiologically active. “It’s not only a cell that has the right costume or attire, it’s also a cell that can recapitulate the function of a motor neuron,” Przedborski said.
The Salk scientists found that Hb9::GFP-positive motor neuron numbers dropped by half after four weeks of co-culture with human astrocyte feeder cells harboring a SOD1-G37R mutation, compared to untransfected astrocytes or those transfected with wild-type SOD1. Similarly, the Harvard scientists found that motor neuron numbers dropped by more than half after 10 days growing on SOD1-G93A rodent astrocytes, and were reduced even further after 20 days. The findings confirm that mSOD1 astrocytes are sufficient to kill wild-type motor neurons derived from human embryonic stem cells. Di Giorgio also found that media conditioned by mSOD1 astrocytes was poisonous to motor neurons, suggesting a soluble factor, or factors, mediates the toxicity.
The researchers then used their in vitro systems to probe the mechanism for astrocyte-mediated toxicity. In both cases, the mSOD1-containing astrocytes expressed markers for inflammation. Di Giorgio and colleagues used oligonucleotide arrays to identify 53 genes that were upregulated in the mSOD1 glia compared to nontransgenic cells. Of those, 13 have a known role in inflammation or immunity.
The Harvard group then zeroed in on the prostaglandin D2 receptor, which was upregulated more than 14-fold in mSOD1 glia. To test if signaling downstream of the glial prostaglandin D2 receptor might mediate toxicity to motor neurons, Di Giorgio treated nontransgenic astrocytes and motor neurons with prostaglandin D2 (PGD2); this killed approximately 80 percent of the motor neurons. They then used an inhibitor of the prostaglandin D2 receptor, MK 0524, to partially protect cultured motor neurons from PGD2, saving about a third of the cells. Though these results suggest that PGD2 is toxic to motor neurons, they do not prove that the prostaglandin is toxic via astrocytes, or that toxicity is specific for motor neurons as happens in ALS. It is possible that PGD2 is simply toxic to all neurons, Przedborski noted.
Marchetto found that mSOD1 astrocytes expressed glial fibrillary acidic protein (GFAP), a marker for astrocyte activation, as well as inflammatory factors such as inducible nitric oxide synthase (iNOS), the neuroendocrine secretory protein chromagranin A (CHGA), and the NADPH oxidase NOX2. NOX2 produces reactive oxygen species (ROS), so Marchetto then assessed whether antioxidants might ameliorate astrocyte ROS production or relieve astrocyte toxicity to motor neurons. The NOX2 inhibitor apocynin, as well as the antioxidants α-lipoic acid and the flavonoid epicatechin, decreased the numbers of astrocytes producing ROS in a monoculture, while apocynin also rescued motor neurons co-cultured with mSOD1 astrocytes. Apocynin has been shown to increase lifespan in a mouse model of ALS (Harraz et al., 2008).
The in vitro system is a convenient model because compounds can initially be screened for attenuation of astrocyte ROS production or inflammatory marker expression without adding the motor neurons that are time-consuming to produce. Promising candidates can then be further tested in co-culture with stem cell-derived motor neurons and eventually in mSOD1 mice.
Since induced pluripotent cells can be coaxed to form motor neurons (Dimos et al., 2008 and see ARF related news story), Gage envisions eventually culturing patient-specific motor neurons to identify the treatments that will be most effective. These in vitro systems, then, have the potential not only to discover drugs for general use, but also to develop personalized medicines.—Amber Dance.
Marchetto MCN, Muotri AR, Mu Y, Smith AM, Cezar GG, Gage FH. Non-cell-autonomous effect of human SOD1-G37R astrocytes on motor neurons derived from human embryonic stem cells. Cell Stem Cell 2008 December 4;(3):649-657. Abstract
Di Giorgio FP, Boulting GL, Bobrowicz S, Eggan KC. Human embryonic stem cell-derived motor neurons are sensitive to the toxic effect of glial cells carrying an ALS-causing mutation. Cell Stem Cell 2008 December 4;(3):637-648. Abstract