health & science

Why does cancer treatment work only for some people?

Why does chemotherapy cure some people’s cancers, yet have almost no effect on others? And why do so many promising new cancer drugs fail in clinical trials? The answer to both questions, says Professor Bill Wilson, co-director of the Auckland Cancer Society Research Centre, is that every cancer is its own individual “freak of nature” – even a relapsed patient has a different disease than what they suffered the first time.  

A small team of researchers in Auckland is looking for genes in cancer cells that reveal how sensitive the individual cancer will be to chemotherapy and other treatments. They’re hoping that, one day, oncologists will be able to tell people whether the expense or awful side effects of treatment will be worth it, not just on average, but for them.

Gene-hunting techniques can also be used to identify which cancers will succumb to expensive new cancer drugs, hopefully raising the drugs’ success rates and persuading Pharmac to fund them.

Eloise Gibson spoke to Bill Wilson about the search for genes that might determine someone's survival.

He says the team is getting faster and more accurate results from using the gene-editing tool CRISPR.

Newsroom: So about a year ago I interviewed Francis Hunter from your research team, and I’ve been thinking ever since about something he said: that we don’t really know why some people respond so well to chemotherapy and other people don’t respond at all. He said you were trying to find out using CRISPR. But it took me a whole year to get back to him about it and now he’s gone to (American drug giant) Johnson & Johnson! Thanks for stepping in to talk to me instead. How did this work get started? Was Francis the instigator?

Bill Wilson: Yes. Francis Hunter is a very talented young man and we hope to entice him back one day. During his PhD - I was his supervisor - we looked at using an earlier (pre-CRISPR) technology to find out what genes are important to cancer’s sensitivity to one of the drugs we were developing. That earlier technology is not as specific as CRISPR. We got some important results and he (Francis) really led that area. He spent some time in Toronto during his PhD, and it was during that time that CRISPR really burst on the scene as a better way of doing this.

NR: So CRISPR's much faster?

BW: It’s a much more powerful way of doing the same thing, trying to interrogate the whole genome at once, and so it was quite transformative for us. We were the first people in this part of the world to use that approach.

NR: So, as I understand it, the situation for an oncologist right now is that chemo may or may not work for a particular person sitting in front of them. There’s a percentage chance it will, depending what cancer you have, but there’s also a huge unknown variation, even among people who have seemingly similar cancers. But it’s not well understood why chemo works so well for some people and not others. Is that right?

BW: That’s exactly the problem. We know there’s a big variation in response but we still don’t really lift the lid on exactly what’s gone wrong in individual tumours, in most cases. That is starting to change with the so-called molecularly-targeted therapies, which inhibit the (gene) drivers responsible for the (particular) cancer. So, most famously that’s the drug known as Gleevec.

Ed: Gleevec was hailed as a wonder drug almost two decades ago because of its amazingly high success rate for treating chronic myeloid leukaemia, a rare form of cancer that affects certain types of white blood cells. It worked startlingly well because researchers found exactly what, genetically speaking, was driving the exact form of cancer and made a drug to stop it.

BW: Gleevec was the first case where this was possible, in chronic myeloid leukeamia, by using an inhibitor against the protein that is altered in those cancer cells, that drive the growth of the cancer. We’ve been trying to do the same thing in other human cancers ever since, with varying success. We’ve been reasonably successful with lung cancers now, there's quite a high proportion of lung cancers where we can find their driver mutation and use an inhibitor against it. But that’s not done in conventional chemotherapy.

NR: Chemo is more of a one-size-fits-all treatment.

BW: Yes.

NR: What the range of success rates for chemo? Does it work for say, half of people? You mentioned earlier that chemo is still used in combination even with some of the cutting-edge immunotherapy drugs, so it’s success rates obviously still matter, even though we're getting these exciting new treatments now.

BW: A good example is the recent FDA (U.S. Food and Drug Administration) approval of Keytruda (a new immunotherapy cancer drug) for head and neck cancer. On average, when Keytruda is combined with chemotherapy, it results in something like a two month improvement in overall survival. So it’s not huge. Some patients derive absolutely no benefit, the tumour just continues on as if nothing has happened. Some patients might gain up to another two years. This makes a huge difference, for example for Pharmac, if they are looking at the health economics of whether to fund this expensive medicine.

That’s for the use of Keytruda with chemotherapy, but we know that the exact same variation applies with chemotherapy alone. In fact, it can make the difference between an effective treatment and absolutely no advantage whatsoever.

NR: So, for some proportion of patients, chemo will fix the problem and others will get no benefit for what I imagine are quite distressing and awful side effects?

BW: That’s one issue, and another is that if you select the wrong treatment – the treatment that’s not matched to the individual patient’s tumour – then you’re losing time. You’re losing the opportunity to try a treatment that may be more effective.

This is a changing culture in oncology, accepting that every tumour is an individual freak of nature.

Although we tend to bundle them into groups such as lung cancer and head and neck cancer, increasingly we are starting to look at them for their individual characteristics and match treatment accordingly.

We’re only at the beginning of the process. It’s going to take decades, I think, to give full effect to that vision, but that’s the path we’re now on. It happened first with the molecularly targeted agents (such as Gleevec) because if the target (gene mutation) is there, versus when it's not there, the difference is day and night. The cancer cell might be 1000 times more sensitive to Drug X than a cell that doesn’t have it.

With conventional chemotherapy, it’s more like a tenfold range. So it’s not quite as dramatic, but a tenfold difference in dose to achieve the same outcome, clinically, is absolutely huge. So it’s something we really do need to understand and it’s been a neglected problem.

NR: One thing I’ve heard about chemo is that there is a high proportion of cancers for which, if the first round of chemo doesn’t work, it’s kind of all over for that person. Chemo has the highest chance of working on the first round. So there was a feeling that oncologists should hit the first round very hard.

But now researchers are getting very interested in the immune system and they are realising that in fact, over-dosing is dangerous in terms of the patient’s own immune system's ability to get in and help kill the cancer. If that’s right, you wouldn’t want to overdo chemo -- not just for the quality of life of the patient, but also because it may work worse if you do. Does that make sense, and is it right?

BW: It makes complete sense. And that’s part of the transition towards more nuanced individualised cancer treatment. The right dose for the right patient at the right time. So it’s about individualising the dose and the choice of treatment. And it changes over time, because with almost every form of cancer treatment, over time, resistance emerges. That maybe be less so with these new immune approaches, but with molecularly-targeted drugs and chemotherapy, the emergence of resistance is inevitable, so it’s a matter of asking the question again when the patient relapses and again (later), as we understand the new properties of this tumour that’s presenting itself.

NR: Everyone reading this will probably have heard about these new immune therapies and they will have caught on to the fact that researchers are very excited about them. [Ed: Immunotherapy drugs work by increasing the person's own immune response and sometimes stopping the cancer from using its normal tricks to hide from the immune system.]

With all this excitement about these new therapies can you explain why we still need chemo?

BW: Not all patients respond to immune therapies and we don’t fully understand the reasons for that, there are many. Some tumours just don’t look foreign enough to the immune system…so they are essentially invisible to the immune system. And the environment within tumours is not conducive to the efficient operation of immune systems, the micro-environment tends to be hypoxic (lacking oxygen) and acidic, with high concentrations of lactic acid. Immune cells tend to shut down in these environments.

Immune targeting is a hugely important advance, but, ultimately, it will take its place beside the other tools that we have and we need to learn when it is appropriate and when it’s not and how to overcome the barriers I just alluded to.

NR: When I spoke to Francis, he explained that it takes a few gene mutations to give a person cancer, but there are also all these other genes that don’t determine whether you get cancer, but might be involved in whether or not you survive it.

It sounds like your work is about trying to look at those genes – the ones that determine whether treatment works. And you're trying to personalise chemotherapy as much as possible. Are you looking at the genetic characteristics of cancer itself, or is it the person’s genome outside of the cancer?

BW: We’re looking at the cancer genome.  But not just what they call the driver mutations – the two, three, four or five mutations that are required to turn a normal cell into a cancer cell. It’s the rest of the genome, that doesn’t necessarily drive the cancer but does determine its sensitivity to treatment. Our to find out which genes are active in each individual tumour and what the implications are.

(Because) the vast majority of genes in cancer cells are normal genes, the same as the ones found in the normal human genome. A few will have mutated, but most have exactly the same DNA sequence as in normal cells. However, some of these normal genes are turned on in cancers (as in, they are busier than they were in the normal cell they evolved from), and some are turned off. And the pattern differs between different cancers, just as it does between different normal tissues.

NR: So there are a couple of genes you’ve found already that you think might play a role in determining whether chemo works for someone, can you talk me through them?

BW: We’ve used these CRISPR techniques to mutate every single known human gene in big populations of cancer cells and then we ask which cells grow throughout the treatment, [so] which ones have become resistant to chemo.

And then we find out which genes have been inactivated in cells that are resistant to chemo drugs, like DNA cross-linkers.

Ed: In cancer treatment, DNA cross-linking agents kill cancer cells by damaging their DNA and stopping them from dividing.

BW: They (cross-linkers) are the very earliest class of chemo drugs ever developed and they still play an important role in cancer treatment. We have found two genes in particular that appear to have a major in role in the sensitivity of cancer cells to these (chemo) agents. One is a gene called Schlafen 11. The German word means “sleep”. And the protein made by this gene monitors the DNA replication process, as the DNA unzips and copies itself. If the replication process runs into trouble, Schlafen 11 puts the cell to sleep permanently.

It’s an executioner.

Because it’s just too dangerous if DNA replication goes awry in a normal cell, it’s too dangerous for that cell to continue so that cell is executed.

NR: So it’s like a benign suicide bomber?

BW: Exactly, in the interests of the overall organism, it’s safer not to let cells continue in that state. We’ve found Schlafen 11 has a critical role in also putting cells to sleep when they have DNA crosslinks, so if they have high expression of Schlafen 11, they are particularly sensitive to these DNA cross-linking agents. Another group of researchers has found the same thing.

We are using drugs that release these DNA cross-linkers selectively in the tumour, and exploiting the lack of oxygen in tumours, to target delivery. But that’s only one of several genes we’ve identified.

We don’t believe the activity of one gene alone is going to be the whole answer, it’s going to require understanding the activity of whole sets of genes that determine the overall sensitivity of the cell to anticancer agents. It’s a complex problem.

NR: So I’ve missed something here – does increased expression of this German-sleep-suicide-bomber, as I’ve decided to call it, is that helpful to the effect of chemotherapy or unhelpful? As a patient, you would want increased expression of this gene, is that right?

BW: Exactly. You want high Schlafen 11 in your cancer.

NR: So this is an example of how – if this was to hold up in various clinical trials – you would hope that an oncologist could sequence someone’s cancer and say ‘You have high expression of this gene, therefore we are very hopeful of the effectiveness of first-generation, old-style chemo for you, and it's worth putting yourself through the pain of going through it”?

BW: Yep. And Francis probably mentioned some work he’s been doing in another gene…one that he’s found that is highly expressed in may human cancers relative to normal tissues, and it appears to confer a better outcome to DNA cross-linking chemo in ovarian cancer patients. So that gives you a flavour of the kind of direction we are heading in.

NR: So how many people are working on this and what are the next steps?

BW: This has become quite a major area of activity for us in the Auckland Cancer Society Research Centre, so at the moment there’s a group of about six people, not all of them full-time, focused on the use of these tools.

Traditionally we’ve been a drug discovery group, but a few years ago we realised that we’ve had the experience of taking drugs into clinical development that have failed because we didn’t really, fully understand how they were working. We didn’t fully understand how to identify the patients that would be appropriate.

So my group kind of took a step backwards and sideways, around 2014, to invest heavily in the tools to identify biomarkers (showing whether treatments would work) and co-develop the biomarkers with the new anti-cancer agents. We are doing that now in a couple of projects, where we are developing new anti-cancer drugs and also using these CRISPR techniques at the same time to explore the bio-markers we’ll need to match them to the appropriate patients.

NR: So these techniques are equally applicable to working out the effectiveness of new drugs, as well as older chemotherapy treatments. But could you also use them as a layer of due diligence, before you go to a very expensive clinical trial of a new treatment? And maybe one day an oncologist could even use them on a person before recommending a treatment?

BW: Absolutely. It’s my hope that we are going to see quite a major change in cancer medicine over the next two or three decades. At the moment we have this horrible problem that all the drugs being approved by the FDA come with obscene price tags. But they will slowly come off patent. So the range of these very effective new therapies that will become available at reasonable cost is going to expand, and it’s going to be a different environment for our children by the time they are in the high cancer risk age bracket.

And I think the funding investment will start to move from the medicines themselves to the diagnostics, to really in-depth understanding of, what is this cancer and what are the opportunities for treatment?

NR: The economics should improve, not just from drugs coming off patent, but also from drugs being used for the right people, shouldn't they? There are so many tragic ‘Pharmac won’t fund my drugs’ stories in the news, and Pharmac always seems to say ‘Well, this drug costs X Zillion dollars and has a 20 percent chance of giving someone an extra three months of life, and we’re not funding it because something else would save a lot more lives.’ But the calculus would look a lot better if you could say, 'Yes it costs a zillion dollars, but there’s an 80 percent chance that for this person it will work, and you don’t need to fund it for those other people who won’t benefit.'

BW: That’s right, and we have a major project looking at exactly at that area, again initiated by Francis, and led by Dr Barbara Lipert, about a target antibody for a form of breast cancer treatment that Pharmac does not currently fund. The public would know it by the name Kadcyla. It’s highly effective in some patients and it’s not effective in others. We’re using CRISPR to identify the determinants, so we can identify the patients who will respond. Pharmac approval is not yet forthcoming, but we think this will make a major difference.

And that will change the health economics equation radically for that (treatment).

Can you help our journalists uncover the facts?

Newsroom is committed to giving our journalists the time they need to uncover, investigate, and fact-check tough stories. Reader donations are critical to buying our team the time they need to produce high-quality independent journalism.

If you can help us, please donate today.


Newsroom does not allow comments directly on this website. We invite all readers who wish to discuss a story or leave a comment to visit us on Twitter or Facebook. We also welcome your news tips and feedback via email: Thank you.

With thanks to our partners