Fully mature liver cells from laboratory mice have been transformed directly into functional neurons by researchers at the Stanford University School of Medicine. The switch was accomplished with the introduction of just three genes and did not require the cells to first enter a pluripotent state. It is the first time that cells have been shown to leapfrog from one fundamentally different tissue type to another.
“These liver cells unambiguously cross tissue-type boundaries to become fully functional neural cells,” said Marius Wernig, MD, PhD assistant professor of pathology and a member of Stanford’s Institute for Stem Cell Biology and Regenerative Medicine. “Even more surprising, these cells also simultaneously silence their liver-gene expression profile. They are not hybrids; they are completely switching their identities.”
The cells make the change without first becoming a pluripotent type of stem cell — a step long thought to be required for cells to acquire new identities.
The researchers used a technique developed by Stanford bioengineer Stephen Quake, PhD, to analyze the gene expression profiles of individual hepatocytes (liver cells) and fibroblasts to show that both types of transformed cells not only begin looking and acting like true neurons, they also decisively shut down nearly all gene expression associated with their former, very different identities.
“This is fascinating,” said Wernig. “We can imagine ways that the three introduced factors could stimulate neural gene expression, but how do they also down-regulate two completely unrelated donor networks — those of skin and liver cells?”
Understanding how this down-regulation works will help scientists and clinicians determine whether these so-called transdifferentiated cells can be used to learn more about diseases or even be safely used in human therapy. It would not be good, for example, if newly derived neurons began to again express skin or liver proteins. It also may help researchers understand the process of development, during which cells commit to certain fates while also turning off other potential pathways.
Wernig and Marro began investigating whether hepatocytes could transform into neurons because the fibroblasts they first transformed into neurons in 2010 are a notoriously messy groups of cells. Fibroblasts can be found in almost any organ in the body and contain mixtures of cell types. This made it extremely difficult to identify a cell-of-origin for the resulting neurons and to figure out exactly how big of a developmental leap the cells were making.
In contrast, hepatocytes are fairly homogenous and well-defined. Developmentally speaking, they are also worlds away from neurons: Hepatocytes arise from one of three classes of embryonic tissue called the endoderm; neurons from the ectoderm. The remaining tissue, the mesoderm, is, for the most part, sandwiched between the two. To put it simply: Your innards mostly arise from endoderm, your nervous system and the outer layer of your skin from ectoderm, and your connective tissue and muscles from mesoderm. Transforming endodermal cells into ectodermal cells is a testament to the power of the transdifferentiation technique.
To accomplish the transformation of the hepatocytes, the researchers used a virus to introduce the same three genes that they used for the fibroblasts: Brn2, Ascl1 and Myt1l. As with the fibroblasts, the hepatocytes began to exhibit neuronal characteristics within two weeks, and express neuronal genes within three weeks. Simultaneously, the cells began to suppress the expression of liver-specific genes.
Marro and Wernig used a sophisticated cell-labeling technique to confirm that the new neurons had indeed arisen from the former liver cells, and Fluidigm dynamic polyermerase chain reaction assays to analyze gene expression patterns of individual neuronal cells. They found that even “true” neurons express low levels of liver genes in the form of transcriptional noise. However, the newly differentiated neurons did express marginally higher levels of the same genes.
“Although the donor gene program is dramatically shut down, there are some remnants of their former life, like a kind of a memory,” said Wernig. “But the vast majority of expressed genes demonstrate a clear dominance of the neuronal transcription program.” Furthermore, the fact that the newly derived neurons generate electrical signals and form junctions with other neurons, and that they exhibit no residual liver function, indicates that this memory has no functional relevance, according to Wernig.
Direct lineage reprogramming is possible across different defined germ layers
Terminally differentiated endodermal cells can be directly converted to neurons
Reprogramming factors induce silencing of the donor-specific transcriptional program
Induced neuronal cells retain a small but detectable epigenetic memory
Several recent studies have showed that mouse and human fibroblasts can be directly reprogrammed into induced neuronal (iN) cells, bypassing a pluripotent intermediate state. However, fibroblasts represent heterogeneous mesenchymal progenitor cells that potentially contain neural crest lineages, and the cell of origin remained undefined. This raises the fundamental question of whether lineage reprogramming is possible between cell types derived from different germ layers. Here, we demonstrate that terminally differentiated hepatocytes can be directly converted into functional iN cells. Importantly, single-cell and genome-wide expression analyses showed that fibroblast- and hepatocyte-derived iN cells not only induced a neuronal transcriptional program, but also silenced their donor transcriptome. The remaining donor signature decreased over time and could not support functional hepatocyte properties. Thus, the reprogramming factors lead to a binary lineage switch decision rather than an induction of hybrid phenotypes, but iN cells retain a small but detectable epigenetic memory of their donor cells.