The thing that is so intriguing about cancer stem cells (CSCs), from a cancer researcher’s perspective, is their powerful potential as therapeutic targets. While CSCs can create or regenerate tumours (causing relapse in some patients despite our best efforts to prevent it – more on this in my previous post on CSCs), they also present a way to completely eliminate cancer.
It seems pretty obvious that if we had some way to kill off the CSCs in a tumour, radiation and chemotherapy could be more successful in curing patients. So naturally, cancer researchers are drawn to exploring the potential of this idea in the lab.
However, as scientists studying CSCs with the goal of progressing cancer treatment discoveries, we need a model of the disease to use in the lab (for a number of reasons, we can’t experiment directly on actual patients). And herein lays the conundrum: what do we use as a model? In choosing what to use, there are several things that need to be considered. (1) Is it reflective of the cancer process and tumours in patients? (2) Will there be sufficient resources to perform proper optimization? (3) Will the findings be reproducible and reliable?
There are a number of different kinds of experimental strategies cancer biologists can use for their research, including computer modelling (in silico), cells in a dish (in vitro), and animals (in vivo). While we could have a lengthy discussion about all of these types – and everything in between – I’d like to focus on our in vitro – cell culture – options for studying human CSCs.
In my experience, there are two general sources of cells for in vitro work: (1) patient tumour samples, from which we grow cells in a dish; and (2) cell lines (long-living cells, originally grown from patient tumours). Wait – what? There’s a difference? Yes and no.
No, because technically they are the both cells grown in a dish, and come from patient tumours (see Figure). Yes, because of the nature by which they have been grown in culture – and this makes a big difference, for reasons we don’t fully understand.
When I say ‘patient samples,’ I am referring to cells isolated from a tumour/tissue samples within hours of it being taken from the patient, and grown in a dish thereafter for a short period of time (few months or less), during which, all experiments are performed. Alternatively, when I say ‘cell line’, I am referring to cells that were isolated from a patient tumour/tissue, and have been grown and distributed (by scientists and distributors like American Type Tissue Culture) for many years. Some of these cell lines – such as Henrietta Lacks’ HeLa cells – have been around for decades.
So with these options, which do you choose? Each has its pros and cons, and I (along with all CSC researchers) have had to consider all of them (see Table).
Cell lines are great. They can grow seemingly forever, so you never run out; if you don’t need them for a while, you can freeze them, and thaw later; they are super convenient to buy, and you can readily have a cell line of interest. BUT, cell lines still only come from one tumour – and if we know anything about cancer, it’s that no two tumours are the same – so you almost always need more than one. Most importantly, they’ve been in culture for so long, they are often not what they were inside the patient – so are we really modelling the disease anymore? If you think about it, a dish in an incubator with media ( cell food) could select for entirely different characteristics than the environment inside the patient: there is reduced competition for food, cells no longer need to evade the immune system, and researchers themselves manipulate the media components. So what survives? The cells best suited to growing in a dish (think natural selection), which are hardly the same things with which we started,(albeit, sometimes this is intentional).
Obviously patient samples must be the way to go, right? They come right from the patient, and are used right away, so they are super-relevant to the original tumour. Except, you need a consenting cancer patient to donate their tumour. And the cells need to not die before you use them. And you need enough cells to test things multiple times, never mind optimize. Forget about them surviving a freeze-thaw cycle… you see where I’m going with this (if not, check out this post about my experiences with patient samples and research).
So there are a couple of solutions (that I know of) to this dilemma.
First, sometimes you can get lucky with a patient sample, and it spontaneously just keeps growing. The trick here is to keep it ‘minimally cultured’ – meaning, you don’t keep growing the same batch of cells continuously, but freeze many vials of cells from the early days of culture, to use much later. Why is this important? Well, it keeps them distinct from cell lines, for one. And, some have found long-term culture to do some nasty/weird things (e.g. normal tissues can acquire cancerous traits; standard culture conditions can promote genetic changes in cancer cells).
Option two: do all your work with cell lines AND with patient samples. This way, you have a consistent source of cells through all experiments, and can divide your tiny number of patient sample cells among one or two experiments each. And hopefully, the cell line results match up with those from the patient samples – a process referred to as validation.
Oftentimes, it comes down to what the researcher is specifically investigating. For example, if I want to start classifying breast CSCs by their genetic profiles, I would likely use patient samples, since I need to account for the wide range of differences among breast cancer patients, and isolating DNA from patient samples is typically easier than growing the cells. Or, say I want to test a large number of drugs to find one that kills CSCs. For this, I would likely choose a few cell lines, so I can run my tests as much as I want to determine the best few compounds, then maybe test those in a few ‘minimally-cultured’ patient cell lines.
Realistically, we never choose just one model system. We tend to use a combination of in vitro models to answer research questions. And the most dedicated of us will go beyond cells in a dish, to try get as close as possible to mimicking the human tumour. Never doubt the significance of a new model system for CSC biology – it just might be the one to push the frontier of cancer research.
Please note that much of this is based on my own experiences and thoughts (except where research is cited), and information should be taken with a grain of salt. Are you a CSC researcher? Tell me your opinion/strategy on Twitter: @Sara_M_Nolte!
Rosland G.V., Svendsen A., Torsvik A., Sobala E., McCormack E., Immervoll H., Mysliwietz J., Tonn J.C., Goldbrunner R. & Lonning P.E. & (2009). Long-term Cultures of Bone Marrow-Derived Human Mesenchymal Stem Cells Frequently Undergo Spontaneous Malignant Transformation, Cancer Research, 69 (13) 5331-5339. DOI: 10.1158/0008-5472.CAN-08-4630
Li Y., Héroux P. & Kyrychenko I. (2012). Metabolic restriction of cancer cells in vitro causes karyotype contraction—an indicator of cancer promotion?, Tumor Biology, 33 (1) 195-205. DOI: 10.1007/s13277-011-0262-6
Sara M. Nolte
Latest posts by Sara M. Nolte (see all)
- Crazy for CRISPR! - December 5, 2016
- Made in Canada solutions for the scarcity of stem cell donors? - August 8, 2016
- From bench to bedside: clinical trials and cancer drug approval in Canada - September 28, 2015