Feature: Collapsing breast cancer’s niche

This feature appeared in the July/August 2011 issue of Australian Life Scientist. To subscribe to the magazine, go here.

This story really only started about six years ago when Sydney pathologist and researcher Sandra O’Toole was doing her PhD at The Johns Hopkins Hospital in the U.S., working on developmental pathways in cancer.

The lab she joined was famous for its work on hedgehog signalling and cancer, and O’Toole was looking specifically at the role of hedgehog in pancreatic cancer. However, on moving back to Australia her interest shifted more to breast cancer.

“There was very little known about hedgehog pathways in breast tumours – only a handful of papers – so I decided to focus my studies in that area,” says O’Toole.

Hedgehog signalling is a key regulator of animal development, with hedgehog being the extracellular ligand that sits at the pathway start. Hedgehog proteins and pathways are highly conserved from flies to human. Indeed, hedgehog signalling was first discovered in fruit fly (Drosophila) embryos where it helps to set up the body plan: what goes where and next to what.

The pathway is named for the somewhat spiky projections on fly larvae produced when hedgehog is mutated. Hedgehog signalling remains important throughout development and also in the adult, but when the pathway malfunctions, diseases such as cancer may result.

O’Toole started out by looking at a cohort of breast cancer patients collected through St Vincent’s Hospital and the Garvan Institute of Medical Research in Sydney. This cohort comprises almost 300 patients, for whom there is long-term follow-up data and clinical histories, as well as tissue samples from the original breast cancer surgeries.

O’Toole and colleagues first measured the levels of various hedgehog and hedgehog-related proteins in the patient sample to find any association with clinicopathological features such as length of survival, metastases and so on.

“We found that the hedgehog level was associated with poor outcome and a higher risk of having metastatic disease. In particular, there was a very strong association with basal-like breast cancer, a very aggressive form of the disease,” she says.

“Patients with this type of breast cancer do very poorly and one reason for that is the lack of any targeted therapies in these cases.” O’Toole is referring to the fact that most breast cancers are estrogen or progesterone receptor-positive and the cancer growth may be limited by drugs like tamoxifen that block the action of these hormones.

---PB---

More recently, cancers positive for HER2 (human epidermal growth factor receptor 2) have also been identified and these are similarly treated using Herceptin. Such specifically-targeted therapies for breast cancer are making quite a large impact on survival, according to O’Toole, particularly with the aggressive HEP2-positive tumours.

“What is left are these so-called ‘triple-negative’ breast cancers, which tend to occur in younger patients and constitute 15-20 per cent of all breast cancer cases.” Of these, about half to two thirds have basal-like cancer and there is an intense interest right now is finding new medications that can treat this kind of breast cancer.

About this time, new Garvan Institute group leader, Alex Swarbrick, entered the plot, fresh from the lab of Nobel laureate J. Michael Bishop in San Francisco. “Alex had been doing some fantastic mouse transplantation studies in the U.S. looking at oncogenes in cancer, and he continued with them back here in Sydney,” says O’Toole.

“So, we got together to ponder whether we could manipulate the mouse model to decide if the findings from my human samples were just observational or did they really have a functional impact. In other words: is hedgehog signalling important in breast cancer, and if so, what might the mechanism look like?”

The project appealed to Swarbrick because of the strong clinical data already in place and the great potential to go from that point. “In my world most people start with their cell lines, then go to mouse models and eventually try to validate their results in human tissues,” he says.

“But Sandra had already done a huge amount of clinical analysis in breast cancer patients and so the question for me was then: can we look for interesting roles of hedgehog signalling in the etiology of the disease?”

Mice showing the way

After a bit of preliminary ‘fiddling about’ (a technical term), O’Toole and Swarbrick settled on a mouse model that develops a malignancy quite similar to basal-like breast cancer in humans, with similar pre-malignant to fully invasive stages.

It was developed using Swarbrick’s retroviral expression system. “The human patients showed quantitative stepwise increases in hedgehog expression the closer you got to invasive cancer, compared to normal tissues, so the higher the level of hedgehog, the more aggressive the cancer.

Similarly, hedgehog expression in the mouse model seemed to correlate with tumour aggressiveness, poor differentiation and a bad outcome, and strongly with the basal-like phenotype,” says Swarbrick.

Surprisingly, when the researchers then isolated cancer cells from the hedgehog-overexpressing mice and grew them on their own in culture to develop a cell line, they acted just like control cells: no apparent gain-of-function in terms of growth rate or increased migration or invasion or anything.

---PB---

This was a bit of a worry, to say the least. Much to everyone’s relief, all was revealed when those cells were transplanted into immune-deficient mice, so back into the in vivo context of mouse tissue where they could chat happily with all the other cells.

“The transplanted mice developed tumours that were much larger than control tumours, they were more locally invasive, and were more associated with invasion of the lymphatic system – an additional step to metastasis,” says O’Toole. Voila, a major phenotype, and a strongly invasive cancer.

The team then did the reverse experiment of blocking hedgehog expression using an antibody to see just how much these nasty tumours actually need hedgehog? Stopping the hedgehog in this way resulted in smaller tumours and, interestingly, it also changed the pattern and rate of metastatic spread. All findings were definitely painting hedgehog as the bad guy.

Looking at the transplanted mice in detail revealed even more, as Swarbrick explains. “Tumours formed in the mice carrying our cell line showed high levels of hedgehog protein in the epithelial cancer cells, as expected. However, the pathway was clearly also turned on in the surrounding rim of ‘normal’ stromal cells.

These fibroblasts just lit up when stained for hedgehog pathway markers. Thus the surrounding stroma was providing a niche or friendly environment that somehow supported the aggressive growth of the tumour.”

To further prove the point, when this cosy arrangement between high hedgehog-expressing breast cancer cells and the surrounding stroma was disrupted by otherwise occupying hedgehog (using the blocking antibody), the mouse tumours withered and stopped spreading.

According to Swarbrick, although it has been shown before that breast cancer cells can create their own survival niche by communicating their needs to the healthy cells that surround them, “we had basically found a way to block the metastatic capacity of the cancer cells. It was really quite dramatic.”

So, the team’s predominant evidence supported a paracrine type of signalling (one cell type affecting the function of a nearby different cell type), both in the clinical samples and in the mouse models.

“Scientifically, this is one of the really interesting things to come out of the work: how the cancer cells communicate with the adjacent normal stromal cells. Hopefully by blocking that paracrine conversation in breast cancer via the hedgehog pathway, we could stop the really bad effects,” says O’Toole. “We don’t yet know what the stromal cells are feeding back, and that is something we are keenly pursuing.”

To Swarbrick, the whole scenario just reeks of a cancer stem cell niche, really nasty looking cancer cells, shabbily dressed and very aggressive. “Our hypothesis is that the niche set up with the stromal cells provides a stem cell-like environment for these poor-outcome cancers.”

---PB---

Synergy

These findings, which were published recently in Cancer Research, represent a tremendous effort by a lot of people, explains O’Toole. “And it demonstrates the strength in linking clinical researchers like myself with basic researchers like Alex.

He brings that angle of really thinking about what is happening in the cells signalling-wise, while I am more focused on what is happening in the tissue and cancers – a great combo. “We were also lucky enough to have a lot of input from Professor Neil Watkins from Monash Institute of Medical Research in Melbourne. He was one of the first to show the importance of hedgehog in lung cancer, and I met him while I was working in the U.S.. He has given us some fantastic intellectual input about the project direction.”

Of course, there was also significant clinical input into this type of study, including from medical oncologist, Catriona McNeil, and other local pathologists such as Associate Professor Ewan Millar, O’Toole adds.

“In fact, Catriona is working to set up a clinical trial to look at hedgehog-blocking drugs in triple-negative breast cancer cases, using existing agents that work a little further down the hedgehog pathway (small-molecule inhibitors of Smoothened).

“Some of these drugs have already gone through Phase I and II trials and this is a big advantage – they don’t kill people and indeed are quite well tolerated. Meanwhile we are trying to develop a more specific drug that may block the initial conversation between the normal and cancer cells because our data suggests that this might be more effective, but that will be a few years off.”

As always, a stack of interesting findings means a stack more things to find out about. However, O’Toole and Swarbrick have a pretty clear idea of how to proceed, at least for the near future. “Understanding the nature of the reverse signal coming back from the stroma to the hedgehog-overexpressing cancer cells is going to be a key point in understanding this whole system,” says Swarbrick.

“It also potentially throws up another therapeutic target if it turns out that the tumour-promoting effects of hedgehog depend on the stromal cells. We predict that the stromal cells would be less able to resist treatment because they are less plastic genetically and therefore far less able to just mutate a gene and get around the drug effect.”

Current experiments to address this goal involve taking tumours from the mouse models and fractionating the constituent cells into stromal and epithelial for transcriptional profiling. “This will help us start to understand what gene expression changes are going on in each of the respective cell types involved, what changes are likely to have the greatest impact and what is the kind of dance synergy between the two.

“For example, we can look for any expression changes in the stroma that are happening in response to hedgehog activation in the epithelium or, conversely, look for changes in the epithelium in response to those changes in the stroma.”

Both Swarbrick and O’Toole are also very keen to see their study go from human to mouse and the lab, and then back to humans. “Hopefully with this multidisciplinary team we are well placed to move it along relatively quickly. And on the drug development aspect, we just have to keep agitating to make it happen.”

---PB---

Hedgehog and cancer

People twigged to the possible role of hedgehog in cancer quite a long time ago, with figures of up to 30 per cent of all human cancers involving the hedgehog pathway. The first discoveries were made in syndromes like Gorlin’s syndrome where people were predisposed to cancers such as medulloblastoma and basal cell carcinoma.

What is interesting about those diseases is that they are generally driven by a mutation in the pathway itself that confers constitutive activation so the pathway acts in a cell-autonomous manner, meaning that the cell with the mutation is the one in which hedgehog is hyperactive.

In more recent years evidence is emerging that other kinds of adult sporadic cancers like lung, prostate and now breast, show not only this cell-autonomous activity, but you also see cases in which there is a paracrine type of signalling going on in which one cell type affects the function of a neighbouring different cell type.

It gets particularly interesting because there are many developmental systems where paracrine hedgehog signalling is very important. So it seems like these cancers are somehow actually reactivating that developmental paracrine signalling instead of instituting some kind of crazy, new gain-of-function cancer mutation. Indeed, this it is a bit of an open question at present – in some cancers it is clearer while in others the mechanism at play is far less clear.

This idea of hedgehog signalling being involved in cancer through a defunct embryonic development mechanism that becomes reactivated in cancer has opened the door for new therapeutic possibilities. Given that such a mechanism might not be required in most normal adult tissues, inhibiting the hedgehog pathway may offer a selective way of attacking just the tumour cells.

This feature appeared in the July/August 2011 issue of Australian Life Scientist. To subscribe to the magazine, go here.