The smallest entity of life is the single cell, which exists not only as single cell organisms, but as evolution proceeds, as members of a bigger and more complex living organism. During the progression of life, an organism encounters many experiences, and encodes these experiences as memories or knowledge.

While it is clear that in larger organisms that this occurs in the brain or equivalent higher order centers, a similar memory process also occurs in single cells within the nucleus- an equivalent higher order control center in which the cell’s genome is stored. Rather than using neural circuits, single cells encode memory in their nucleus by chemical modifications (or methylation) in the cell’s genome. Distinct patterns of methylation results in differential regulation of gene expression, which in turn pertains to the cell’s identity and/or lineage. These DNA methylation patterns are often coined “epigenetic modifications”, and that cellular memory is often referred to as “epigenetic” memory.

With the emergence of induced pluripotent stem cells (iPS cells) technology, where somatic cells are reprogrammed into pluripotent stem cells using an armory of molecular approaches (including nuclear transfer and more recently Yamanaka’s direct reprogramming with transcription factors), scientists soon realize that the cell’s epigenetic memory (or methylation marks) could influence how well somatic cells could be successfully reprogrammed. In the recent study by the Nature paper Kim et al (2010), it appears that many iPS cells retain methylation marks in their genome that dictate their lineage of origin.

Not only does this result in expression of genes specific to their lineage of origin, but also in the suppression of genes pertaining to other lineages, and thus impinge on the pluripotency of iPS cells.

As scientists unveil the complex network of enzymes that control epigenetic patterns in the cell’s genome, they now have at their disposal molecular tools- including demethylating agents- to manipulate the cell’s epigenetic memory. In Kim’s article, a proposed strategy to improve the pluripotency (and ES properties) of iPS cells is to use demethylating agents to wipe out these annoying lineage specific epigenetic marks or “troublesome memories” that may impede cell reprogramming. In a sense, the authors are suggesting that by inducing cellular amnesia, we can force somatic cells to assume the pure identity and phenotype of pluripotent embryonic stem cells, and to differentiate successfully into cells of other lineages.

Quite a neat trick! But the demethylation only generates cells with a blank slate with no epigenetic memory. According to the data presented in this article, it appears that in order to sustain a pluripotent state, pluripotent cells must have epigenetic marks to silence all lineage specific genes. Similarly, to assume the identity of distant cell lineages via differentiation, wouldn’t the cells need to establish another set of epigenetic marks to turn off their lineage of origin? Well then, how are these epigenetic marks established? According to Kim’s article, these epigenetic marks can be established with forced expression of transcription factors driving either pluripotency or differentiation into distant lineages. Moreover, specific cell conditions can drive the establishment of epigenetic marks crucial for attaining pluripotency or the identity of distant lineages.

Overall, the Kim article suggests that transcription factors and passaging in specific culture conditions might be producing key epigenetic marks, in lieu of the epigenetic memory of the cell’s lineage of origin, to reprogram iPS cells. But exactly how this can happen still remains a mystery.

As I perused Kim’s article and the News piece in the September 16 issue of Nature, I began thinking of what might occur in cancer stem cells. According to Peter Dirks’ first article on the discovery of brain tumor stem cells in 2004 (Singh et al., 2004), brain tumor stem cells are multipotent cells that are capable of differentiation into neurons and glial cells (astrocytes and oligodendrocytes). But unlike the “normal” neural stem cells that gives rise either to cells of neuronal or glia (not both), brain tumor stem cells can differentiate into both lineages simultaneously, giving rise to differentiated neuronal and glial “cancer” progeny that make up the jumbled and variable array of cells populating the tumor.

A bit déjà vu, but this seems to be what the demethylating agent in Kim’s article is actually doing to render iPS cells more capable of assuming lineages distant from the cell’s lineage of origin. Could cancer be really cellular amnesia- where cancer cells “forget” their lineage of origin, and hence begin to behave erratically? Moreover, cancer stem cells (or their stemness phenotype) are a temporary epigenetic phenomenon as reported by Roesch et al (2010). To that end, it is also possible that these “forgetful” cancer cells may be aberrantly establishing epigenetic marks (such as partial stem cell epigenetic program) to continue growing into a tumor, and to survive by evading stressful environmental conditions such as the host’s immune response or anticancer treatments. 

It seems that Kim’s article has opened an interesting world into the cell’s epigenetic memory, and their implications not only in iPS research, but in my opinion also in cancer research. As the molecular secret underlying the cell’s epigenetic memory gradually unfolds, molecular scalpels will soon become available to understand the workings of the cell’s epigenetic memory and how epigenetic amnesia and inception could influence cellular behavior. Perhaps similar to Dom Kobb (Leonardo DiCaprio) in "Inception", the successful iPS reprogramming or cancer treatments may lie in the cellular “inception” of the appropriate epigenetic memories.


Kim et al (2010). Nature. 467(7313):285-290.

Dirks et al (2004). Nature. 432(7015):396-401.

Roesch et al (2010). Cell. 141, 583–594.