What is cancer? - Livre d'Or








Miscellaneous. Eclectic. Random. Perhaps markedly literate, or at least suffering from the compulsion to read any text that presents itself, including cereal boxes. * Blogroll * Strange words * More links * Bookies * Microblog * Recent comments * Humans only * Second degree * By topic * Cool posts * Writing * New post

Tags

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *



livredor
What is cancer?
Friday, 01 February 2008 at 10:19 am
Tags:

Previous Entry Next Entry


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.


Whereaboooots: KI, Flemingsberg, Sweden
Moooood: impressedinspired
Tuuuuune: Vienna Teng: Lullabye for a stormy night
Discussion: 15 contributions | Contribute something
Tags:

Previous Entry Next Entry




Contribute something
View all comments chronologically



rho: default
From:rho
Date:February 1st, 2008 12:48 pm (UTC)
1 hours after journal entry, 01:48 pm (rho's time)
(Link)
Thank you for posting this! I think that it's cool as well.
(Reply to this comment) (Thread)
cartesiandaemon: default
From:cartesiandaemon
Date:February 1st, 2008 12:52 pm (UTC)
1 hours after journal entry
(Link)
That was actually really interesting. So, oversimplifying, a cancer cell is apparently a cell with a mutation that makes it think it's at a wound site? As opposed to any number of related potential phenomena? For instance[1], it's not a normal cell being fed wrong signals, nor a foetal cell, nor typically a cell with some other similar mutation?

Though do these cases occur also? Is a benign tumour composed of different sorts of cells, or encapsulated so it can't spread, or both?

And a cancer has enough awareness to grow genuine tissue (eg. with blood vessels, etc) but not to grow whatever sort of organ is appropriate to that part of the body.

OK, that's just thinking aloud. But its good to be up to date, both with cancer, and you.
(Reply to this comment) (Thread)
livredor: p53
From:livredor
Date:February 1st, 2008 01:35 pm (UTC)
2 hours after journal entry, 02:35 pm (livredor's time)

(Link)
Ooh, I like it when people ask intelligent questions about my essays! Oversimplifying is probably not your friend when it comes to understanding cancer, but I will try to answer without too many complicating details and exceptions and it-dependses.

a cancer cell is apparently a cell with a mutation that makes it think it's at a wound site?
In some cases it may be, yes. One of the good drugs I mentioned, herceptin, probably works because of this. In about 30% of cases of breast cancer, the cells are hyper-sensitive to signals telling them to grow, because they have lots of sensors whereas normal cells only have a few. Herceptin blocks off all those sensors and that seems to get rid of the cancers pretty impressively. So in a normal breast, cells will only grow if the breast is actually injured and needs to be repaired. In a herceptin-sensitive breast tumour, the tiniest little injury, say a bruise, causes the hyper-sensitive cells to "think" there is a wound and start growing manically in order to repair the non-existent wound. According to this theory, giving the patient the drug is equivalent to telling these cancer cells that the wound is healed now, so they can stop trying to repair it.

it's not a normal cell being fed wrong signals, nor a foetal cell, nor typically a cell with some other similar mutation?
Seems to be a combination. This is all a bit speculative, but one possible scenario is that immune system is over-enthusiastic about sending signals that there is a wound there (instead of the cell being over-enthusiastic about receiving them). We don't actually know of any specific cases where it has been proved that a defective immune system is the root cause as this model would suggest.

Asking whether a cancer cell is actually a foetal cell is a bit of a semantic question; it's clearly an adult cell, because it comes from an adult. It may switch on some of the genes that should normally only be switched on in a foetus. According to Weinberg's model, this kind of cell isn't the one that starts the tumour in the first place, but it may be the one that turns a benign tumour malignant. The cells that do this are usually the ones at the edge of the tumour, so this second and more dangerous mis-programming could also be a result of receiving the wrong signals from the surrounding immune components.

Some other mutations: they have to be there also, because normal cells have fail-safe mechanisms which makes them commit suicide if they get wrong signals. The ones which become tumours lose those fail-safe mechanisms and just grow anyway.

Is a benign tumour composed of different sorts of cells, or encapsulated so it can't spread, or both?
Mostly the former I think. In particular, it contains about 2% of cells that keep dividing to make new cells (cancer stem cells). But these cells can't break away from the tumour mass, can't break down other tissues to invade them, can't survive for very long in the bloodstream, and can't grow in distant sites which don't have the same environment that they're used to. If they also turn on the foetal genes, they gain the ability to do all those things.

And a cancer has enough awareness to grow genuine tissue (eg. with blood vessels, etc) but not to grow whatever sort of organ is appropriate to that part of the body.
Awareness is a funny term to use, but if I take it metaphorically I can see what you mean. Cancer cells can be spookily good at making at least a rough approximation of normal tissue; most tumours contain a mixture of different cell types which are related to the ancestral normal tissue. They just make way too much of it, or make it inappropriately.

There are pancreatic tumours which make insulin, for example. But if bits of them break off and lodge themselves in the lungs, having a bunch of miniature pancreases in your lungs is very bad for you, even if they are quite effective pancreases! The weirdest tumours are teratomas; they come from germ cells, and they make little lumps of mixed-up baby, like with teeth and hair and sometimes even tiny limbs, but all joined together anyhow and not in the proper shape of a baby.
(Reply to this comment) (Up thread) (Parent) (Thread)
cartesiandaemon: default
From:cartesiandaemon
Date:February 1st, 2008 02:01 pm (UTC)
2 hours after journal entry
(Link)
Ooh, I like it when people ask intelligent questions about my essays!

Oh, good. I was going to decry, but then decided a good description would be "intelligent, ignorant, questions" :)

Oversimplifying is probably not your friend when it comes to understanding cancer

ROFL. Well, no. Admittedly, taking ten years out of my life to study treatments for cancer is a hell of a lot more worthy than most things that crop up on my friends page that I would have to study to really participate in. (Including, certainly, the physics of obscure and inconsistent imaginary worlds, but also perhaps things like the laws governing american indian reservations -- someone needs to be on the ball, but I probably can't really help myself :)).

However, for what its worth, it feels like a useful level of knowledge from this end. Of course, not complete, but knowing the *sorts* of things that might happen is good.

One of the good drugs I mentioned, herceptin, probably works because of this.

Ah! That was going to be my next question :)

So that's a good treatment for those cases. Presumably one of the side-effect being (slightly or dangerously?) reduced response to actual wounds in that area/anywhere? Corresponding to how other treatments nerf anywhere you need rapidly dividing cells for any reason?

one possible scenario is that immune system is over-enthusiastic about sending signals that there is a wound there (instead of the cell being over-enthusiastic about receiving them)

Of course, this is all not completely confirmed.

But isn't the problem (or at least one of the problems) with the cells themselves, if, if[1] when they break off, they start growing elsewhere.

[1] I can't decided if those nested ifs are horrendous or interesting :)

having a bunch of miniature pancreases in your lungs is very bad for you, even if they are quite effective pancreases!

Oh yes, what a wonderful description. I guess its, how does a cell know which sort of cell it is? Is it based on messages from what cells it's near? Or what cells it was birthed from? Or (I assume) all sorts of different things its impossible to explain briefly.

The weirdest tumours are teratomas

Oh yes. I was wondering about that, when you were saying tumours typically didn't grow as fast as foetuses. That idea always squicks me, I didn't even quote your (great) description.
(Reply to this comment) (Up thread) (Parent) (Thread)
redbird: default
From:redbird
Date:February 1st, 2008 12:59 pm (UTC)
1 hours after journal entry, 08:59 am (redbird's time)
(Link)
Thanks.

This has me wondering whether there are any data on cancer incidence with people whose immune systems are suppressed--either through HIV or other diseases, or because of medication, including the anti-rejection drugs given after organ transplants.
(Reply to this comment) (Thread)
livredor: p53
From:livredor
Date:February 1st, 2008 01:39 pm (UTC)
2 hours after journal entry, 02:39 pm (livredor's time)
(Link)
Good question, thank you. People with AIDS get some really weird kinds of cancer that HIV negative people never get. And a lot of the mice used for experiments are immune deficient; this is partly because the immune system complicates things, but also because it is at least partially protective against some kinds of cancer. So I over-simplified when saying that the immune system can't fight cancer because cancer isn't really foreign; the immune system does have some positive effects in fact. But as with a lot of other conditions, the immune response is a double-edged sword.
(Reply to this comment) (Up thread) (Parent) (Thread)
lethargic_man: default
From:lethargic_man
Date:February 1st, 2008 01:40 pm (UTC)
2 hours after journal entry, 02:40 pm (lethargic_man's time)
(Link)
That is indeed cool, and interesting. I wonder what avenues for treatment-directed research that might open up.

(And my reaction to reading the name "Bob Weinberg" was the same as that when you first told me who your former boss was, to wit: "Wow! Er... what is it he's famous for, again?")
(Reply to this comment) (Thread)
livredor: p53
From:livredor
Date:February 1st, 2008 02:00 pm (UTC)
2 hours after journal entry, 03:00 pm (livredor's time)
(Link)
The treatment possibilities Weinberg suggested were things like: giving Herceptin immediately after a mastectomy, even though it delays wound healing, to have a better chance of preventing recurrence. Looking at ways to inhibit the foetal transcription factors that make benign tumours malignant (they come from Drosophila so they have silly names like Twist and Snail). Since adult cells don't need those regulators at all, that could be really low-toxicity, but of course it's totally science-fictional at this point. Remote possibilities include things like messing with the immune system directly, expanding some of the current work that doesn't bother to try to stop the main tumour mass from growing, but blocks it from acquiring new blood vessels so it will eventually collapse anyway.

I did explain what Weinberg was famous for, didn't I? He did the clever experiments where you put normal cells into mice and they don't do anything, but you put cells containing viral T-antigen (to block p53 and RB), and the Ras oncogene, and activated telomerase, and they grow into tumours. So he more or less defined what an oncogene and a tumour suppressor are and tied the definitions to an empirical test as opposed to the rather speculative ideas that had been around previously.
(Reply to this comment) (Up thread) (Parent) (Thread)
lethargic_man: default
From:lethargic_man
Date:February 1st, 2008 02:14 pm (UTC)
2 hours after journal entry, 03:14 pm (lethargic_man's time)
(Link)
I did explain what Weinberg was famous for, didn't I?

You did, but I was referring to my initial reaction before I'd read that far.
(Reply to this comment) (Up thread) (Parent) (Thread)
addedentry: stoplooklisten
From:addedentry
Date:February 1st, 2008 01:51 pm (UTC)
2 hours after journal entry, 02:51 pm (addedentry's time)
(Link)
Thank you for this. As it happens, Weinberg's new book, "The biology of cancer", arrived on my desk for cataloguing this morning.

If the glass is half-empty, it's depressing that humanity spent 60 years and unimaginable sums of money without making any measurable progress.

No measurable progress on reducing mortality rates, perhaps, but have the unimaginable sums of money generated 60 years' worth of progress in cytology and immunology and so on?
(Reply to this comment) (Thread)
livredor: p53
From:livredor
Date:February 1st, 2008 02:07 pm (UTC)
2 hours after journal entry, 03:07 pm (livredor's time)
(Link)
have the unimaginable sums of money generated 60 years' worth of progress in cytology and immunology and so on?
Hard to answer that. Certainly it feels like we know a whole lot more about how cells work than we did in the 30s, which is obviously how people tend to justify all that effort. But there was a massive hiatus in making the jump from knowing how stuff works, to actually developing new treatments that were as much more effective as you would expect them to be. I would also say that the really impressive drugs of the past 15 years didn't come out of nowhere; they would not have been possible without all the prior effort.

Is knowledge meaningful if you can't do anything practical with it? I would argue that it is meaningful, but it's also true that you can't be certain that your model is correct until you can actually tweak the things the model would predict you can tweak and get the results your model would predict you'd get (in this case, making people with cancer better!)
(Reply to this comment) (Up thread) (Parent) (Thread)
synergetic: default
From:synergetic
Date:February 1st, 2008 02:10 pm (UTC)
2 hours after journal entry
(Link)
That was really interesting, thank you.
(Reply to this comment) (Thread)
hairyears: default
From:hairyears
Date:February 1st, 2008 08:02 pm (UTC)
8 hours after journal entry, 09:02 pm (hairyears's time)
(Link)
That's a fascinating idea: metastases are using an embryonic mechanism for migrating through tissues - a mechanism which surely has an 'off' switch. This is a target for chemotherapy which, unlike cell division, can be fully exploited in adults.

BTW, there is a third circumstance in which cells are "supposed" to grow rapidly and relatively independently: the immune response. It's not the astonishing proliferation that interests me: it's the fact that leukaemia is so rare. The conclusion I draw from this is that there's at least one and probably several 'off' switches for immune cell proliferation.

[EDIT]
Also, the 'wound' theory recalls conversations from the 1970's; a close family friend had exploratory surgery for a lung complaint. A large and complex cancer was discovered and excised but the prognosis was poor because cancer 'goes mad' or is 'provoked' by surgery. That may sound like a childish oversimplification but, regrettably, that was the limit of our understanding in those days. He died within a matter of weeks.
[/EDIT]
(Reply to this comment) (Thread)
livredor: p53
From:livredor
Date:February 2nd, 2008 12:00 pm (UTC)
1 days after journal entry, 01:00 pm (livredor's time)
(Link)
Yes, you are quite right that an embryonic pathway makes a really attractive target. Transcription factors (proteins which directly control which genes are switched on) are regarded as hard targets pharmacologically, but hard isn't the same as impossible, and the benefit of being able to attack something which you know normal adult tissues don't need would be huge.

Immune cells certainly grow rapidly, but they do not grow independently. It's not so much that they have an off switch as a dead man's lever; the process which triggers them to grow rapidly also drives them three quarters of the way into the suicide pathway. So they need constant override signals from supporting cells and the presence of their intended target, and any interruption in that will let them complete the suicide mechanism. (Practically, they're a pig to work with because they will in fact die if you look at them funny.)

That story about your family friend is very sad as these stories always are. (With lung cancer it's tempting to wonder if it isn't merciful for it to happen quickly.) But it's also an illustration of why surgeons need to talk to molecular biologists more.
(Reply to this comment) (Up thread) (Parent) (Thread)
wesleysgirl: default
From:wesleysgirl
Date:February 21st, 2008 12:27 am (UTC)
19 days after journal entry, February 20th, 2008 08:27 pm (wesleysgirl's time)
(Link)
Hi!

I was LJ-hopping this evening and ended up here, and I wanted to say thank you for sharing this really interesting information even if some of it was way over my head.
(Reply to this comment) (Thread)



Contribute something
View all comments chronologically