Antibody finds, wipes out prostate cancer in mice(physorg.com)
physorg.com
Antibody finds, wipes out prostate cancer in mice
http://www.physorg.com/news181245375.html
8 comments
I think that the 20th and the 21st centuries will go down in history as the best time to be a mouse on this planet.
This is both funny and sadly true -- if we tolerated as much failure in people as we did in mice who knows where medicine would be today.
Any bio-nerds want to express opinions about this? Sounded super intriguing to me.
Well, first of all:
US researchers have found an antibody that hunts down prostate cancer cells in mice...
That's mice, folks. Let's review the guidelines for titling your cancer-related article: If the result is in vitro, you must use the words in vitro; if it is in mice, you must use the word mice.
(Why? Because mice are not like humans. See my earlier musings on that subject here: http://news.ycombinator.com/item?id=838160 )
Meanwhile, like all crappy popular articles, this one provides no citation. It doesn't even mention the names of any of the researchers, poor dears. But Google and I can remedy this -- this seems to be the paper:
http://www.pnas.org/content/early/2009/12/15/0911397107.abst...
Salient facts:
This is, as much as one can tell from an abstract, a fine preliminary result. But don't buy any stock yet.
Not only is this in mice, but this result is about attacking human prostate cancer that has been transplanted into mice. Now, this is a darn sight more impressive than killing cells in a dish, but it's still not the same problem as killing the cells in the human body where they originally evolved. The mouse is a nude mouse; it's immunodeficient (otherwise it would reject the human cancer). Part of the reason that the antibody may be working so well -- attacking the cancer, leaving other cells alone -- is that the cancer cells aren't just cancer, but also from a completely different part of the mammalian family tree.
The other big problem with these studies is that it's one thing to be able to label "85%" of metastatic cancer cells in an animal model, but that's not necessarily a cure. For one thing, among the remaining 15% may be the line of cells that is resistant to your antibody. Now you need another antibody!
Worse, antibodies are hard to deliver to every cancerous cell. They're not especially tiny molecules. They have a relatively easy time being delivered to normal tissues via the bloodstream, because normal tissues are very carefully engineered such that every cell is within a short distance of a capillary. But tumors, having gotten where they were by randomly flipping switches on their genomes like a toddler trying to launch the Space Shuttle, are pretty poorly organized. There are bits of the interior of a tumor that are so far from blood vessels that the cells there are asphyxiating. Unfortunately, because of this, one of the things that cancer cells tend to learn really early is how to survive in such low-oxygen conditions, and when an antibody comes along in the bloodstream and kills the easily-accessed cancer cells, it is likely to leave behind some of these orphan cells, which sit around waiting for the day that they will come across some more oxygen and start growing again. Then, a few months or years later, your cancer comes back.
(This, by the way, is also a big problem with classical chemotherapy and every other cancer treatment.)
So, this result is a relatively early report from a work in progress. Such work takes a great deal of time to come to fruition.
US researchers have found an antibody that hunts down prostate cancer cells in mice...
That's mice, folks. Let's review the guidelines for titling your cancer-related article: If the result is in vitro, you must use the words in vitro; if it is in mice, you must use the word mice.
(Why? Because mice are not like humans. See my earlier musings on that subject here: http://news.ycombinator.com/item?id=838160 )
Meanwhile, like all crappy popular articles, this one provides no citation. It doesn't even mention the names of any of the researchers, poor dears. But Google and I can remedy this -- this seems to be the paper:
http://www.pnas.org/content/early/2009/12/15/0911397107.abst...
Salient facts:
This is, as much as one can tell from an abstract, a fine preliminary result. But don't buy any stock yet.
Not only is this in mice, but this result is about attacking human prostate cancer that has been transplanted into mice. Now, this is a darn sight more impressive than killing cells in a dish, but it's still not the same problem as killing the cells in the human body where they originally evolved. The mouse is a nude mouse; it's immunodeficient (otherwise it would reject the human cancer). Part of the reason that the antibody may be working so well -- attacking the cancer, leaving other cells alone -- is that the cancer cells aren't just cancer, but also from a completely different part of the mammalian family tree.
The other big problem with these studies is that it's one thing to be able to label "85%" of metastatic cancer cells in an animal model, but that's not necessarily a cure. For one thing, among the remaining 15% may be the line of cells that is resistant to your antibody. Now you need another antibody!
Worse, antibodies are hard to deliver to every cancerous cell. They're not especially tiny molecules. They have a relatively easy time being delivered to normal tissues via the bloodstream, because normal tissues are very carefully engineered such that every cell is within a short distance of a capillary. But tumors, having gotten where they were by randomly flipping switches on their genomes like a toddler trying to launch the Space Shuttle, are pretty poorly organized. There are bits of the interior of a tumor that are so far from blood vessels that the cells there are asphyxiating. Unfortunately, because of this, one of the things that cancer cells tend to learn really early is how to survive in such low-oxygen conditions, and when an antibody comes along in the bloodstream and kills the easily-accessed cancer cells, it is likely to leave behind some of these orphan cells, which sit around waiting for the day that they will come across some more oxygen and start growing again. Then, a few months or years later, your cancer comes back.
(This, by the way, is also a big problem with classical chemotherapy and every other cancer treatment.)
So, this result is a relatively early report from a work in progress. Such work takes a great deal of time to come to fruition.
Having worked in Oncology drug development before, I don't waste my time with articles where the headline is "X kills cancer in (mice|cell lines|rats)". We've cured cancer so many times in rodents, if I were a mouse, I'd take up smoking unfiltered Marlboros while working at a nuclear waste disposal site in the Ukraine.
The primary reason there are so many false starts in cancer therapy is that there are zero reliable model organisms for doing cancer research. The closest things we have are: cancer cell lines (human cells derived from tumors and coaxed to grow forever in cell culture) and mice (often given tumors by embedding human cancer cell lines). Both of these scenarios are hugely artificial.
The cancer cell lines are so mutated and freakish that they grow outside a living organism and replicate forever in culture. So there's always the question when you have something kills them: is it because you're knocking out some basic cancer mechanism? Or instead is it that you're knocking out whatever it is that lets them live on piece of glass? Their DNA is certainly even more messed up than garden variety in vivo human cancer.
Mouse models are problematic because the mouse immune system, while similar to humans, is not the same. Plus if you're embedding human cancer cell lines into the mouse to give it cancer, what exactly are you curing?
Finally, there is the issue that we know that human cancers are highly tissue-specific. So while we talk about "cancer" as if it's one disease, it's actually a multitude of different diseases with the same rough outward phenotype. This (along with many other factors) is one of the reasons we've gotten really good at breast cancer treatment: many of the modern drugs target a particular feature of breast cancer to shut down tumor cells. But even in that case, there are other varieties of breast cancer where those drugs don't work at all.
Long story short: it all sucks a whole lot. This is why there is over a 90% (I forget the exact figure) failure rate of cancer therapies in early clinical trials. There's a huge opportunity for an enterprising biotechnologist/bioengineer to improve that number by even 10%.
The primary reason there are so many false starts in cancer therapy is that there are zero reliable model organisms for doing cancer research. The closest things we have are: cancer cell lines (human cells derived from tumors and coaxed to grow forever in cell culture) and mice (often given tumors by embedding human cancer cell lines). Both of these scenarios are hugely artificial.
The cancer cell lines are so mutated and freakish that they grow outside a living organism and replicate forever in culture. So there's always the question when you have something kills them: is it because you're knocking out some basic cancer mechanism? Or instead is it that you're knocking out whatever it is that lets them live on piece of glass? Their DNA is certainly even more messed up than garden variety in vivo human cancer.
Mouse models are problematic because the mouse immune system, while similar to humans, is not the same. Plus if you're embedding human cancer cell lines into the mouse to give it cancer, what exactly are you curing?
Finally, there is the issue that we know that human cancers are highly tissue-specific. So while we talk about "cancer" as if it's one disease, it's actually a multitude of different diseases with the same rough outward phenotype. This (along with many other factors) is one of the reasons we've gotten really good at breast cancer treatment: many of the modern drugs target a particular feature of breast cancer to shut down tumor cells. But even in that case, there are other varieties of breast cancer where those drugs don't work at all.
Long story short: it all sucks a whole lot. This is why there is over a 90% (I forget the exact figure) failure rate of cancer therapies in early clinical trials. There's a huge opportunity for an enterprising biotechnologist/bioengineer to improve that number by even 10%.
Recomment(?) from a reddit thread on the same story from someone who claims to work in cancer research:
The chances that this antibody will work in humans is EXTREMELY small.
The model that would have to be used here is injecting human prostate cancer cells into nude mice (this is a strain that is totally lacking in its own immune system so foreign cancer cells can actually survive and not be killed within minutes), allowing the tumor to grow, and then injecting the antibody or even ust the antibody fragment int the tail vein. You then palpate the tumor over the next few weeks. You'd also want to sample the tumor and maybe do some TUNEL to check for actual rates of apoptosis.
Chances are very slim that an antibody that fights this specific strain of lab grown cancer (lab grown cancers have generally been grown for so long that they are heavily mutated to say the least) will be effective in humans. Humans have many, many, many different strains of prostate cancer and unless these researchers have accidentally stumbled upon the holy grail of surface markers that somehow indicate internal genetic expression...I am skeptical.
Additionally, you've linked to an article that is so vacuous that it doesn't even feel the need to cite the actual results which it is based upon. VPNed through my department account and I can't even find the article.
http://www.reddit.com/r/science/comments/ajcaq/an_antibody_h...
The chances that this antibody will work in humans is EXTREMELY small.
The model that would have to be used here is injecting human prostate cancer cells into nude mice (this is a strain that is totally lacking in its own immune system so foreign cancer cells can actually survive and not be killed within minutes), allowing the tumor to grow, and then injecting the antibody or even ust the antibody fragment int the tail vein. You then palpate the tumor over the next few weeks. You'd also want to sample the tumor and maybe do some TUNEL to check for actual rates of apoptosis.
Chances are very slim that an antibody that fights this specific strain of lab grown cancer (lab grown cancers have generally been grown for so long that they are heavily mutated to say the least) will be effective in humans. Humans have many, many, many different strains of prostate cancer and unless these researchers have accidentally stumbled upon the holy grail of surface markers that somehow indicate internal genetic expression...I am skeptical.
Additionally, you've linked to an article that is so vacuous that it doesn't even feel the need to cite the actual results which it is based upon. VPNed through my department account and I can't even find the article.
http://www.reddit.com/r/science/comments/ajcaq/an_antibody_h...
Yeah, this is from a cancer researcher. As a former cancer researcher, I recognize the symptoms. ;)
All of this is basically true, though I hesitate to be so harsh without actually reading the article...
All of this is basically true, though I hesitate to be so harsh without actually reading the article...
First thing that comes to mind when reading titles like this. Mice is a model for human biomedical research, but far from a perfect one, especially in the cancer field. There are substantial differences that make it hard to extrapolate these results in mice to humans.
As a medical biotechnology grad student I read these claims often and it's a shame that most of the time it's impossible to find a reference to the original article. Hard to know than if it's an exaggeration generating false hope or a genuinely promising cure with consequences for treatments in humans.