What is cancer?
We're being graced with an official visit from Bob Weinberg this week. One of the things he wanted to do was make an opportunity to meet a group of PhD students and other junior scientists. This strikes me as an excellent instinct because these vastly famous people doing their tours of honour will always have the chance to meet the other famous and important scientists at the host institution, and they will usually have a chance to be paraded for the general public, but it's quite easy for them to miss the actual working researchers. So, I signed myself up to be on the waiting list if there were any spaces for post-docs after the opportunity had been offered to the PhD students, and there were some extra spaces, so I attended the meeting yesterday.
Someone asked me what exactly Weinberg is so famous for: basically he did the original work to prove that you can turn a normal cell into a cancer cell by blocking a couple of genes and injecting another couple, and what the minimum set of genes is. And he's by no means a one-hit wonder, he's been doing lots of exciting stuff in the quarter century since that landmark achievement. It's actually quite surprising that this was the first time I'd heard him speak; I'd have expected to run across him at conferences by now.
His talk was mixed; I could clearly see why he is so respected both as a scientist and as a communicator, and indeed I'm writing this post because I'm excited about what he had to say. But at the same time he didn't live up to his towering reputation. He talked down to the group quite badly; this may well be because his expectations of PhD students are based on the American system where a new PhD student has only had a science "major" and is still very much learning. So his talk was pitched at the wrong level for European PhD students who have completed an entire degree in their specialist subject, and are used to being treated as professional scientists albeit at an early stage of their training. He also admitted half way through the talk that he was sleep deprived and in a really horrible mood, and apologized for being unusually grumpy as a result (he was really thrown by some annoying computer problems at the beginning, when it took nearly 20 minutes to get the system set up to display his Powerpoint slides).
Both his grumpiness and his slight tendency to patronize caused me to behave rather abrasively, which I'm not particularly proud of. He started out by saying that he expected to be interrupted with lots of questions, so I took him at his word. But instead of asking questions which show how intelligent and engaged with his ideas I am, I found myself jumping on apparent flaws or omissions in his arguments and generally being a bit arsey. I doubt he was offended, but I also doubt I made a glowingly positive impression.
He started off by making what I thought was a really odd argument about cancer epidemiology. He showed some figures that point out that pretty much the only thing correlated with cancer incidence is access to screening and diagnosis. So ok, there's huge reporting bias in how we track the prevalence of types of cancer in populations or over time. I didn't find this as surprising or significant as he seemed think it was. I think the point was supposed to be that in spite of huge changes (usually alarming increases) in the reported incidence of various kinds of cancer, the (age-adjusted) death rates didn't really change between 1930 and 1990, or between different countries studied. So he postulated that one way to read these figures is that human intervention basically has no effect on cancer mortality; a certain proportion of people with a given type of cancer die no matter what anyone does, and a certain proportion survive because they were destined to survive anyway, (though they are likely to attribute it to some kind of faith healing or quackery).
OK, 1990 to the present there has actually been a measurable decline in mortality from breast cancer and a couple of types of leukaemia. So it's not all fatalism; medical advances are making a really profound difference here. He said that part of the decline in breast cancer mortality is explained by awareness of the risks of HRT so that it is no longer pushed at women as it was a generation ago. I wasn't convinced by that, because it is really only in the US that every middle aged woman took HRT, and then everyone stopped because of the breast cancer scare. According to Weinberg screening programmes and knowledge of some of the major genetic factors haven't made much difference, but he didn't really justify dismissing those factors. Breast cancer does also benefit from two of the only three new drugs that unquestionably outperform any therapy attempted since the 30s: tamoxifen and its friends, and the antibody-based therapy herceptin. The third unquestionably successful drug is Gleevec for a certain type of leukaemia. That gives really stunning results, like improving the 5 year survival rate from about 20% to about 95%, but it is only useful in one particular relatively rare type of leukaemia, so it doesn't register as a blip in overall population statistics.
If the glass is half-empty, it's depressing that humanity spent 60 years and unimaginable sums of money without making any measurable progress. If the glass is half-full, there have been three genuine breakthroughs in the past 15 years, so it could be that we're finally on the right track. (Also, no measurable progress might mean that the survival rate is improved from 5% to 10%, meaning thousands of people are alive who otherwise wouldn't be, or it might mean that patients get a year of decent quality of life rather than 6 months of misery, but of course would still count as mortality statistics.) FWIW, my old boss, who is not as famous as Weinberg but pretty famous, reckons that cancer will be a curable disease in our lifetime.
This stuff is more or less what all famous cancer researchers say. Some of you probably don't get as many chances to hear famous cancer researchers giving their spiel as I do, so I'm writing it here because I think it might be of interest. The really exciting bit was the second part of the talk though:
One of the most exciting results in cancer biology recently is that the only cells that are capable of giving rise to tumours are adult stem cells. This means that cells that normally don't grow don't suddenly turn rogue and start growing all over the place, as used to be believed (recently enough that I was taught this model at university in the late 90s). But in fact, cancer happens when cells that normally do grow, ie stem cells, start making tumours instead of healthy tissues.
If you generalize from this, you start to wonder how far cancer cells are really normal cells in the wrong situations, rather than total aberrations. Bear in mind that all cells in the body contain exactly the same genes, but use a subset of them to perform their correct functions. Cancer cells probably have, oh, half a dozen mutations, genetic changes. But that might mean they have six altered letters out of three billion which are identical to those of normal cells. How do such tiny changes alter the whole function of the body, even fatally in many cases? What if these altered cells aren't something entirely new, they're just switching to the wrong sort of program.
There are two circumstances where cells are "supposed" to grow rapidly and relatively independently. One is when the embryo is developing, when it has only a few months to grow from a single cell one tenth of a mm wide, to a baby-sized baby 50 cm long (there are very few tumours that grow that fast!). The other is when a person is injured, and needs to rapidly make new tissue to repair the damage. Weinberg suggested that both these situations are relevant in a tumour.
So, we can argue that a tumour acts like a wound site when there is no wound. It rapidly makes new blood vessels, which act to provide oxygen and nutrients to the centre of the tumour mass, but the blood vessels don't "know" that that is their "goal". The blood vessels start to grow because the body somehow "thinks" there is a wound there that needs to be repaired. The parts of the immune system which usually deal with wounds are all present at the sites of tumours; it was previously thought that this was a response to the presence of the "foreign" tumour, but in fact this doesn't make sense because the tumour isn't really foreign in the way that bacteria or other parasites are. So another way of looking at it is that the immune system, triggered inappropriately, actually causes the tumour. The immune cells are responding to a wound that isn't there, so they send out chemicals which signal the tumour cells to grow, as they would normally signal new tissue to develop and repair an actual wound.
Weinberg also pointed out that this may mean that surgery is a really problematic way of dealing with cancer. You cut out the tumour, which obviously does need to happen. But. It's impossible to eliminate absolutely every cell, and even a single stem cell left behind can regenerate the whole tumour, because that's what stem cells do. Even worse, surgery causes an actual wound, so all the immune system gubbins which is around will go into hyperdrive, making a really ideal environment for those stem cells to get going and grow like anything.
If this were the whole story, most cancers wouldn't be fatal. A tumour that does nothing except grow inexorably bigger is usually referred to as benign (this is a relative term, of course!) A malignant tumour is much more dangerous, for two reasons. Firstly, it actively invades the surrounding tissue, breaking down healthy tissue to make room for the tumour to grow. And secondly, pieces called metastases can break off and be carried round the body in the blood stream and lymph system, and cause new tumours all over the place. These metastatic tumours often can't be removed by surgery as there are too many of them, and it's often only a matter of time before they get into vital organs and cause a total system failure, otherwise known as death.
But there are some normal cells that are meant to invade the surrounding tissue, and meant to be able to move around the body and start growth at new sites. Namely, the cells of the early embryo. Weinberg's theory is that malignant cells turn on genes that are normally turned on at the moment when the blastocyst, the ball of frog-spawn like cells, starts to turn into an actual embryo with recognizable features. These genes help the cells to move around to position themselves in the right places to form specialized tissues, and also to invade other parts of the embryo and mother's uterus as necessary. So if these genes get turned on in an adult, you can get metastatic cells.
This feels like it could be a really productive novel way of looking at cancer. And I think it's cool!
Further reading:
1. Stem cells: the real culprits in cancer?. Rather impressive Scientific American article on cancer stem cells, aimed for a popular audience.
2. Reya et al, Stem cells, cancer, and cancer stem cells is a decent review of stem cells and cancer, if you have access to Nature and want to read something at a more advanced level than SciAm.
3. Campbell & Polyak, Breast Tumor Heterogeneity: Cancer Stem Cells or Clonal Evolution? is a less good review, also written by people who are skeptical of the cancer stem cells model, but has the advantage of being free.
4. Yang et al, Exploring a New Twist on Tumor Metastasis is a recent review by Weinberg himself of some of this connection between embryo development and metastasis.
Someone asked me what exactly Weinberg is so famous for: basically he did the original work to prove that you can turn a normal cell into a cancer cell by blocking a couple of genes and injecting another couple, and what the minimum set of genes is. And he's by no means a one-hit wonder, he's been doing lots of exciting stuff in the quarter century since that landmark achievement. It's actually quite surprising that this was the first time I'd heard him speak; I'd have expected to run across him at conferences by now.
His talk was mixed; I could clearly see why he is so respected both as a scientist and as a communicator, and indeed I'm writing this post because I'm excited about what he had to say. But at the same time he didn't live up to his towering reputation. He talked down to the group quite badly; this may well be because his expectations of PhD students are based on the American system where a new PhD student has only had a science "major" and is still very much learning. So his talk was pitched at the wrong level for European PhD students who have completed an entire degree in their specialist subject, and are used to being treated as professional scientists albeit at an early stage of their training. He also admitted half way through the talk that he was sleep deprived and in a really horrible mood, and apologized for being unusually grumpy as a result (he was really thrown by some annoying computer problems at the beginning, when it took nearly 20 minutes to get the system set up to display his Powerpoint slides).
Both his grumpiness and his slight tendency to patronize caused me to behave rather abrasively, which I'm not particularly proud of. He started out by saying that he expected to be interrupted with lots of questions, so I took him at his word. But instead of asking questions which show how intelligent and engaged with his ideas I am, I found myself jumping on apparent flaws or omissions in his arguments and generally being a bit arsey. I doubt he was offended, but I also doubt I made a glowingly positive impression.
He started off by making what I thought was a really odd argument about cancer epidemiology. He showed some figures that point out that pretty much the only thing correlated with cancer incidence is access to screening and diagnosis. So ok, there's huge reporting bias in how we track the prevalence of types of cancer in populations or over time. I didn't find this as surprising or significant as he seemed think it was. I think the point was supposed to be that in spite of huge changes (usually alarming increases) in the reported incidence of various kinds of cancer, the (age-adjusted) death rates didn't really change between 1930 and 1990, or between different countries studied. So he postulated that one way to read these figures is that human intervention basically has no effect on cancer mortality; a certain proportion of people with a given type of cancer die no matter what anyone does, and a certain proportion survive because they were destined to survive anyway, (though they are likely to attribute it to some kind of faith healing or quackery).
OK, 1990 to the present there has actually been a measurable decline in mortality from breast cancer and a couple of types of leukaemia. So it's not all fatalism; medical advances are making a really profound difference here. He said that part of the decline in breast cancer mortality is explained by awareness of the risks of HRT so that it is no longer pushed at women as it was a generation ago. I wasn't convinced by that, because it is really only in the US that every middle aged woman took HRT, and then everyone stopped because of the breast cancer scare. According to Weinberg screening programmes and knowledge of some of the major genetic factors haven't made much difference, but he didn't really justify dismissing those factors. Breast cancer does also benefit from two of the only three new drugs that unquestionably outperform any therapy attempted since the 30s: tamoxifen and its friends, and the antibody-based therapy herceptin. The third unquestionably successful drug is Gleevec for a certain type of leukaemia. That gives really stunning results, like improving the 5 year survival rate from about 20% to about 95%, but it is only useful in one particular relatively rare type of leukaemia, so it doesn't register as a blip in overall population statistics.
If the glass is half-empty, it's depressing that humanity spent 60 years and unimaginable sums of money without making any measurable progress. If the glass is half-full, there have been three genuine breakthroughs in the past 15 years, so it could be that we're finally on the right track. (Also, no measurable progress might mean that the survival rate is improved from 5% to 10%, meaning thousands of people are alive who otherwise wouldn't be, or it might mean that patients get a year of decent quality of life rather than 6 months of misery, but of course would still count as mortality statistics.) FWIW, my old boss, who is not as famous as Weinberg but pretty famous, reckons that cancer will be a curable disease in our lifetime.
This stuff is more or less what all famous cancer researchers say. Some of you probably don't get as many chances to hear famous cancer researchers giving their spiel as I do, so I'm writing it here because I think it might be of interest. The really exciting bit was the second part of the talk though:
One of the most exciting results in cancer biology recently is that the only cells that are capable of giving rise to tumours are adult stem cells. This means that cells that normally don't grow don't suddenly turn rogue and start growing all over the place, as used to be believed (recently enough that I was taught this model at university in the late 90s). But in fact, cancer happens when cells that normally do grow, ie stem cells, start making tumours instead of healthy tissues.
If you generalize from this, you start to wonder how far cancer cells are really normal cells in the wrong situations, rather than total aberrations. Bear in mind that all cells in the body contain exactly the same genes, but use a subset of them to perform their correct functions. Cancer cells probably have, oh, half a dozen mutations, genetic changes. But that might mean they have six altered letters out of three billion which are identical to those of normal cells. How do such tiny changes alter the whole function of the body, even fatally in many cases? What if these altered cells aren't something entirely new, they're just switching to the wrong sort of program.
There are two circumstances where cells are "supposed" to grow rapidly and relatively independently. One is when the embryo is developing, when it has only a few months to grow from a single cell one tenth of a mm wide, to a baby-sized baby 50 cm long (there are very few tumours that grow that fast!). The other is when a person is injured, and needs to rapidly make new tissue to repair the damage. Weinberg suggested that both these situations are relevant in a tumour.
So, we can argue that a tumour acts like a wound site when there is no wound. It rapidly makes new blood vessels, which act to provide oxygen and nutrients to the centre of the tumour mass, but the blood vessels don't "know" that that is their "goal". The blood vessels start to grow because the body somehow "thinks" there is a wound there that needs to be repaired. The parts of the immune system which usually deal with wounds are all present at the sites of tumours; it was previously thought that this was a response to the presence of the "foreign" tumour, but in fact this doesn't make sense because the tumour isn't really foreign in the way that bacteria or other parasites are. So another way of looking at it is that the immune system, triggered inappropriately, actually causes the tumour. The immune cells are responding to a wound that isn't there, so they send out chemicals which signal the tumour cells to grow, as they would normally signal new tissue to develop and repair an actual wound.
Weinberg also pointed out that this may mean that surgery is a really problematic way of dealing with cancer. You cut out the tumour, which obviously does need to happen. But. It's impossible to eliminate absolutely every cell, and even a single stem cell left behind can regenerate the whole tumour, because that's what stem cells do. Even worse, surgery causes an actual wound, so all the immune system gubbins which is around will go into hyperdrive, making a really ideal environment for those stem cells to get going and grow like anything.
If this were the whole story, most cancers wouldn't be fatal. A tumour that does nothing except grow inexorably bigger is usually referred to as benign (this is a relative term, of course!) A malignant tumour is much more dangerous, for two reasons. Firstly, it actively invades the surrounding tissue, breaking down healthy tissue to make room for the tumour to grow. And secondly, pieces called metastases can break off and be carried round the body in the blood stream and lymph system, and cause new tumours all over the place. These metastatic tumours often can't be removed by surgery as there are too many of them, and it's often only a matter of time before they get into vital organs and cause a total system failure, otherwise known as death.
But there are some normal cells that are meant to invade the surrounding tissue, and meant to be able to move around the body and start growth at new sites. Namely, the cells of the early embryo. Weinberg's theory is that malignant cells turn on genes that are normally turned on at the moment when the blastocyst, the ball of frog-spawn like cells, starts to turn into an actual embryo with recognizable features. These genes help the cells to move around to position themselves in the right places to form specialized tissues, and also to invade other parts of the embryo and mother's uterus as necessary. So if these genes get turned on in an adult, you can get metastatic cells.
This feels like it could be a really productive novel way of looking at cancer. And I think it's cool!
Further reading:
1. Stem cells: the real culprits in cancer?. Rather impressive Scientific American article on cancer stem cells, aimed for a popular audience.
2. Reya et al, Stem cells, cancer, and cancer stem cells is a decent review of stem cells and cancer, if you have access to Nature and want to read something at a more advanced level than SciAm.
3. Campbell & Polyak, Breast Tumor Heterogeneity: Cancer Stem Cells or Clonal Evolution? is a less good review, also written by people who are skeptical of the cancer stem cells model, but has the advantage of being free.
4. Yang et al, Exploring a New Twist on Tumor Metastasis is a recent review by Weinberg himself of some of this connection between embryo development and metastasis.