After attending a particularly fascinating seminar given by Diana Bianchi, cancer researcher Peter Geck returned to his office at Tufts University School of Medicine to design an unorthodox experiment. Geck had access to hundreds of breast tissue biopsy samples removed from women over the last several decades, and he decided to design PCR primers to look for male cells in those female samples.
Bianchi, also a researcher at Tufts University, studies fetal DNA circulating in mothers during pregnancy and has developed several prenatal tests currently in use. While a talk on this topic seems unlikely to inspire a cancer researcher’s next study, the ideas proved to be a siren song too powerful for Geck to resist.
“When I graduated from medical school, we learned that the placenta was a complete barrier, that nothing moved in and out except oxygen and food. But that is completely wrong. Both ways, there is an exchange of cells,” Geck said. “These cells actually invade another person.”
Geck’s research focuses on breast and prostate cancer. Pregnancies, particularly those that occur at a young age, are known to protect against developing breast cancer. Geck wondered whether invading fetal cells might play that protective role. His first step was to look for fetal cells in the breasts of women with and without cancer—in this case, male fetal cells since the Y chromosome presented an attractive marker for easy identification.
“When the picture came together, it was a surprise. There is a huge number of normal healthy females who carry male cells in the breast,” said Geck. More than half of the samples they tested contained male cells. This number is even more remarkable when you consider the fact that these tissues were taken from a random sampling of women, including an unknown proportion who were never pregnant with sons.
Not long ago, Geck’s results might have been met with skepticism. Biology courses taught that animals contained the same genome in each cell, albeit with minor rare mutations or changes. Chimeric animals—those harboring more than one population of cells due to a mutation during embryogenesis—were considered unusual and valuable for cell lineage studies. However, several recent studies have demonstrated chimerism in different human organs. What this means for human health is largely unexplored.
Geck’s study, which was published in the International Journal of Cancer (1), gave the first indication of a role for non-tumor chimerism in cancer. “Most of the women who get these microchimeric infiltrations look like they are really and significantly protected from breast cancer,” said Geck. His data indicate that male cell integration increased the odds of remaining cancer-free by 375%.
While Geck speculates that this infiltration of stem cells during pregnancy may explain in part why women generally live longer and stay healthier to older ages than men, cancer cells carrying Y chromosomes tell a darker story.
Only 21% of cancer samples tested carried male cells, but there were far more male cells in tumors than in healthy samples. So in contrast to protecting against cancer, in greater numbers, fetal cells may in fact promote tumor growth.
According to Geck, much effort in cancer research is currently devoted to identifying cancer stem cells. One possibility is that when too many fetal cells infiltrate the breast, they actually cause cancer. On the other hand, they may have nothing to do with causation at all, and as Geck noted, are simply carried along with cancer progression.
Cancer itself is known for heterogeneity and somatic variation between cells, where entire chromosome arms may be translocated, and large insertions, deletions, and aneuploidy are common. Increases in microchimerism have been correlated with autoimmune disease and neurodegenerative disorders. Not to mention that some serious somatic mutations that are lethal when transmitted through the germline manifest as disease only in mosaic form, such as McCune-Albright syndrome, Proteus syndrome, or trisomies for chromosomes 8, 9, or 14.
The infiltration of fetal cells accounts for many occurrences of microchimerism, but there are other means to generate chimerism. Mothers can transfer maternal cells to the fetus, or even transfer cells from previous pregnancies to younger siblings in the womb. But aside from cell transfer, microchimerism results from mutations during development, such as chromosomal abnormalities, copy number variations, small insertions and deletions, or point mutations, leading to lines of cells carrying slight variations in their genomes.
Naturally occurring somatic mutations are certainly of interest to researchers, but they are more challenging to study than invading cells. While PCR identifies known sequences, such as the presence of a Y chromosome within a female sample, when the mutation is unknown, researchers must turn to discovery approaches such as next-generation DNA sequencing. Currently, these studies are most often conducted using bulk samples, where minor variations occurring in only a small fraction of cells get lost amongst the overwhelming number of reads from the majority of cells. So to really study native somatic mutations, researchers need to use single cell sequencing techniques.
Gilad Evrony is an MD/PhD student in the laboratory of Christopher Walsh at Boston Children’s Hospital who recently developed a single neuron sequencing approach to look for microchimerism in the brain. “The cause of a significant fraction of neurologic diseases remains unknown, including epilepsy, autism, cortical malformations, and schizophrenia. These are diseases in which some individuals have a known genetic mutation, but many others remain unexplained. We suspected that somatic mutations could explain some cases of these major diseases, an idea that until now has been largely unexplored,” Evrony said.
Pursuing this idea, Evrony combined flow cytometry for single cell sorting and optimized an established multiple displacement amplification method, a PCR-free method for modifying small amounts of DNA such as minute quantities isolated from single cells, to produce enough starting material for sequencing. Most importantly, he conducted extensive quality control experiments on the performance of single cell capture and amplification for sequencing.
Using this approach, Evrony and his colleagues reported the identification of somatic L1 transposon insertions in human caudate and cortical neurons in the journal Cell (2). While these insertions were rare, “the most exciting and surprising aspect of our work has been that the somatic mutations we discovered gave us a unique window into proliferation and migration patterns of clones or lineages of cells in the human brain. This is the first time we have been able to trace lineages of cells in the normal human brain,” explained Evrony. “We hope our paper introducing single neuron sequencing is the first step towards a whole suite of methods that will be developed to explore this.”
Geck and Evrony both pointed out the importance of using the correct starting material when studying microchimerism—a mutation in a small portion of cells in the brain may not show up in sequences from the blood, for example. So although new sequencing technologies readily provide sequence data, obtaining the right starting material, especially when single cells are needed, can be complicated. At the other end of these studies, working through a maze of sequencing data also presents significant challenges.
“[Structural variation] is considered the hardest type of variation to study, but that doesn’t mean that it’s not important,” said Mark Gerstein, a professor of biomedical informatics at Yale University. “In fact, it affects a lot of base pairs in the genome both in the normal germ line genome and also in the cancer genome.”
Gerstein’s lab has created numerous bioinformatics tools for identifying structural variation in large sequencing datasets, including the CNVnator, which he recently used to show that copy number mosaicism in iPS cells already exists in parental skin fibroblasts rather than arising during the process of inducing pluripotency in those fibroblasts (3).
CNVnator reviews sequencing reads mapped on a reference genome to identify those areas where few reads are present, which can represent deletions, or areas with far more reads than expected, indicating duplications. While this approach works well for the immense numbers of short reads created using Illumina sequencing technology, Gerstein is attentive to evolving technologies as well.
“We are constantly thinking about how that technology can evolve to get longer reads so we can look at more complex events.” A promising new technology is the single-molecule PacBio platform, where very long reads “would be really transformative in terms of structural variation because you could much more directly read these variations out from the reads and there would be less computing; it would be a less speculative endeavor.”
In addition to improving sequencing technology, new computational approaches will add significantly to researchers’ knowledge of the extent of microchimerism and structural variation, enabling further studies into the health implications of this phenomenon and even the mechanisms that give rise to it. “Maybe some of the mechanisms that give rise to natural somatic variation give rise to cancer variation or amplify it in the process,” Gerstein suggested.
Although at the moment there are more unknowns than knowns in this nascent field, “it is not an epiphenomenon that these cells get through and we happen to find them.” Geck said. “I think there is a big future, but we are just at the beginning now.”
1. Dhimolea et al. 2013. High male chimerism in the female breast shows quantitative links with cancer. Int J Cancer 133(4): 835-842.
2. Evrony et al. 2012. Single-neuron sequencing analysis of L1 retrotransposition and osmatic mutation in the human brain. Cell 151: 483-496.
3. Abyzov et al. 2012. Somatic copy number mosaicism in human skin revelaed by induced pluripotent stem cells. Nature 492: 438-445.