Your daily negative XMRV study: XMRV is PCR contamination

Jemal

Senior Member
Messages
1,031
A new day, a new negative study people.

Published: 25 February 2011

XMRV is a gammaretrovirus associated in some studies with human prostate cancer and chronic fatigue syndrome. Central to the hypothesis of XMRV as a human pathogen is the description of integration sites in DNA from prostate tumour tissues. Here we demonstrate that 2 of 14 patient-derived sites are identical to sites cloned in the same laboratory from experimentally infected DU145 cells. Identical integration sites have never previously been described in any retrovirus infection. We propose that the patient-derived sites are the result of PCR contamination. This observation further undermines the notion that XMRV is a genuine human pathogen.

http://www.retrovirology.com/content/8/1/13

Sorry I have to keep posting this stuff :eek:
 

alex3619

Senior Member
Messages
13,810
Location
Logan, Queensland, Australia
Hi Jemal, contamination is always a very real problem in nested PCR. However, if sites are identical, this could also be in our favour. This is because XMRV might have a specific site preference. If this is the case, it might tell us more about possible XMRV pathogenic mechanims. Then also don't miss the observation that 12 of the fourteen patient derived sites were not identical.

My comment does not rule out the possibility that XMRV is often found due to contamination. We still need more good science. However, the probability that XMRV is not in the human population is next to zero. The questions are who has it, and what is it doing?

I'm off to read the paper if I can get hold of it.

Bye
Alex

PS This paper concludes that they suspect it is contamination. It a nudge nudge wink wink argument, but we need more science to be sure. This paper is also from:

1MRC Centre for Medical Molecular Virology, Division of Infection and Immunity, University
College London, 46 Cleveland St, London W1T 4JF, United Kingdom
2Wellcome Trust Sanger Institute, Hinxton, Cambridge, CB10 1SA, United Kingdom

Make of that what you will.
 

urbantravels

disjecta membra
Messages
1,333
Location
Los Angeles, CA
Yes, it's THAT Greg Towers.

They sure whipped that one out in a hurry, didn't they? I wonder how the Cleveland Clinic feels now that their work is being directly attacked?
 

lansbergen

Senior Member
Messages
2,512
Whilst it is conceivable that the other 12 integration sites apparently derived from prostatic tumour tissues [7, 8] are genuine patient-derived sequences, we suspect that some or all of them may also be the result of contamination with DNA from experimentally infected DU145 cells.

They suspect. I am not impressed.
 

Jemal

Senior Member
Messages
1,031
Yes, it's THAT Greg Towers.

They sure whipped that one out in a hurry, didn't they? I wonder how the Cleveland Clinic feels now that their work is being directly attacked?

Well, that's the good news I guess. The WPI is not the only institute taking punches (if this was a boxing match).
 

kurt

Senior Member
Messages
1,186
Location
USA
Alex, the article is available here: http://www.retrovirology.com/content/pdf/1742-4690-8-13.pdf

And it does not matter who published this, this is a claim that can be easily verified by any other lab willing to do the sequencing study.

That DU145 cell line is a human prostate cancer cell line, not a mouse cell line. In other words, this suggests contamination from a human cell line due to the lengthy process used to process those cells during PCR reactions. The article mentioned 80 steps in the XMRV extraction and cloning process including exposure to air at one point as evidence that working with XMRV involves high risk of contamination.

So in theory, the previous testing for mouse DNA would not rule out this type of human cell line contamination. This is a big challenge to claims of XMRV positives by PCR. Will be interesting to see if this can be confirmed. The conclusions seemed a bit strong to me as this is a very preliminary study. IMHO, this needs to be confirmed before it should be taken too seriously.
 

ukxmrv

Senior Member
Messages
4,413
Location
London
Kurt, do you think that XMRV positives by PCR is a big challenge. The WPI seem to have moved away from PCR so I wonder how serious it will be?
 

asleep

Senior Member
Messages
184
And it does not matter who published this, this is a claim that can be easily verified by any other lab willing to do the sequencing study.

Of course it matters. Retrovirology has proven to have an extremely porous review process for XMRV negative studies.

This paper is very light on science and heavy on opinion. You are right that others could verify the modest amount of science presented in this paper, but that would do very little to support the grasping conclusions of this paper.

...human cell line contamination.

Care to explain this concept? How can a virus that is actively infecting and integrating into a human cell line be considered a contaminant? Are HIV+ humans merely suffering from a "contaminant" that has run amok in their bodies and is actively "contaminating" their bodies?
 

kurt

Senior Member
Messages
1,186
Location
USA
Care to explain this concept? How can a virus that is actively infecting and integrating into a human cell line be considered a contaminant? Are HIV+ humans merely suffering from a "contaminant" that has run amok in their bodies and is actively "contaminating" their bodies?

Yes, we are all contaminated... in some way no doubt.

Seriously though, a contaminated human cell line can create false positives, that is the problem. In other words, the patient can be negative for a virus (such as XMRV in this case) but if the PCR testing process utilized a previously contaminated human cell line, the patient can be told they are positive. The mere fact that a virus is capable of infecting human cells under some conditions does not change the risk of false positives. How would you like to be tested for HIV and told you were positive, when in fact you are negative but the lab used a test contaminated by a human cell line used in one of their reagents? That is the issue at stake here.
 

asleep

Senior Member
Messages
184
Seriously though, a contaminated human cell line can create false positives, that is the problem. In other words, the patient can be negative for a virus (such as XMRV in this case) but if the PCR testing process utilized a previously contaminated human cell line, the patient can be told they are positive. The mere fact that a virus is capable of infecting human cells under some conditions does not change the risk of false positives. How would you like to be tested for HIV and told you were positive, when in fact you are negative but the lab used a test contaminated by a human cell line used in one of their reagents? That is the issue at stake here.

You are mixing issues.

If I mix HIV- blood with HIV+ blood and report the HIV- sample as positive, then of course that would be a false positive. However, this theoretical scenario does not imply that HIV itself is a "human cell line contaminant." However, that is essentially the argument that Towers and you (by your apparent enamoredness) are making with regard to XMRV.

I've bolded part of your quote. I am saying that this bolded fact very likely makes XMRV a de facto infectious human agent. Outright and arbitrarily labeling it "contamination," when it is actively replicating, infecting, and integrating within human cells, is misleading semantics.
 

Esther12

Senior Member
Messages
13,774
I'm confused bythe idea that finding XMRV from a human cell line can lead to the conclusion that XMRV is not a genuine human pathogen.

Surely it's pretty unlikely that a virus which can't really infect humans would do so, but just enough to spread contaminant about and affect all these studies?

What am I missing?
 

kurt

Senior Member
Messages
1,186
Location
USA
You are mixing issues.

If I mix HIV- blood with HIV+ blood and report the HIV- sample as positive, then of course that would be a false positive. However, this theoretical scenario does not imply that HIV itself is a "human cell line contaminant." However, that is essentially the argument that Towers and you (by your apparent enamoredness) are making with regard to XMRV.

I've bolded part of your quote. I am saying that this bolded fact very likely makes XMRV a de facto infectious human agent. Outright and arbitrarily labeling it "contamination," when it is actively replicating, infecting, and integrating within human cells, is misleading semantics.

No, in fact you have mixed the issues, you seem to be assuming that when I say 'contaminant' I am also saying 'is not infectious to humans'. Those are two different ideas. I realize some people might think they are the same, but I certain did not say XMRV is not in humans. ANYTHING can be a contaminant. An active HIV+ virus can contaminate a lab sample, in which case it is a contaminant, in that test.

Whether or not XMRV is capable of infecting humans is not the most important question here, rather, whether XMRV infection is present and important in CFS is the question. These researchers have found evidence that points to contamination of a human cell line with a known strain of XMRV isolated in prostate cancer. They obviously rushed this to print thinking it was an important find. I think we need to let other researchers weigh in before drawing too many conclusions.
 

Mark

Senior Member
Messages
5,238
Location
Sofa, UK
The conclusions seemed a bit strong to me as this is a very preliminary study.

So what's new?

Desperately rushing out arguments against, going way beyond your findings in the conclusions you draw from them, and using all the modern techniques of media manipulation to big up your own findings and play down your opponent's...in a david and goliath situation...this all just smacks of desperation and looks suspicious.

So if you admit - and it's nearly always clearly true with these negative papers - that "The conclusions seemed a bit strong to me..." then that to me is the problem in a nutshell. We can all see that, time and again, and it doesn't inspire trust.
 

asleep

Senior Member
Messages
184
No, in fact you have mixed the issues, you seem to be assuming that when I say 'contaminant' I am also saying 'is not infectious to humans'. Those are two different ideas. I realize some people might think they are the same, but I certain did not say XMRV is not in humans. ANYTHING can be a contaminant. An active HIV+ virus can contaminate a lab sample, in which case it is a contaminant, in that test.

Well, this is the heart of the issue. "Contamination" has become a very loaded term. In some contexts it it is clearly used with the broad implication of "not infectious to humans." In other contexts, especially when backpedaling from overreaching conclusions, it is used in a much narrower sense of "something externally added."

Towers and others have freely used the broader form of "contamination" to try to publicly undermine XMRV. This broad definition of "contamination" is what underlies much of the negative press about XMRV. It was the undercurrent of Retrovirology's December Surprise. It is the thrust of this very paper. See their conclusion:

In conclusion, we believe that our findings undermine a central component of the evidence for XMRV being a human pathogen.

It's quite clear that they they are wielding the term in a manner quite different than you suggest.

Whether or not XMRV is capable of infecting humans is not the most important question here, rather, whether XMRV infection is present and important in CFS is the question.

Reread the conclusion to the paper. They are in fact attacking the very idea that XMRV is an infectious pathogen, not just its association to CFS.

I think we need to let other researchers weigh in before drawing too many conclusions.

The irony of this statement burns my brain.
 

lansbergen

Senior Member
Messages
2,512
http://www.ncbi.nlm.nih.gov/pubmed/631930

Int J Cancer. 1978 Mar 15;21(3):274-81.

Isolation of a human prostate carcinoma cell line (DU 145).

Stone KR, Mickey DD, Wunderli H, Mickey GH, Paulson DF.

Abstract

A long-term tissue culture cell line has been derived from a human prostate adenocarcinoma metastatic to the brain. The cell line, DU 145, has been passaged 90 times in vitro over a period of 2 years. The cells are epithelial, grow in isolated islands on plastic Petri dishes, and form colonies in soft agar suspension culture. Karyotypic analysis demonstrates an aneuploid human karyotype with a modal chromosome number of 64. Distinctive marker chromosomes (a translocation Y chromosome, metacentric minute chromosomes and three large acrocentic chromosomes) have been identified. Electron microscopy of the original tumor tissue and of the tissue culture cell line show a remarkable similarity in cell organelle structure.

http://www.cell-lines-service.de/content/e143/e1782/e1196/index_eng.html

DU 145 was isolated by K.R. Stone et al from a lesion in the brain of a patient with metastatic carcinoma of the prostate and a 3 year history of lymphocytic leukemia. The line is not detectably hormone sensitive, is only weakly positive for acid phosphatase and isolated cells form colonies in soft agar. Ultrastructural analyses of both the cell line and original tumor revealed microvilli, tonofilaments and desmosomes. Many mitochondria, well developed Golgi and heterogenous lysosomes. The cells do not express prostate antigen.

A cell line from a metastase to the brain from a prostate cancer in a patient that also had lymphocytic leukemia.

In my opinion if it is in the cell line the change it is a human virus is far greater then the change that it is mouse contamination.
 

SOC

Senior Member
Messages
7,849
Kurt,
Based on your responses, I suspect your background is similar to mine -- the physical rather than the biomedical sciences. In a former life, I both wrote and reviewed scientific papers in my tiny little corner of the research world. In the physical sciences, scientists are indeed highly dedicated to the scientific method, and to honest, accurate science. You can rely on scientists working in good faith.

I have been told frequently in the past couple of years that this is not necessarily true in biomedical sciences. What I've seen in the past year has led me to believe this is true. I suspect it stems from an inherent lack of understanding of mathematics and statistics, which leads to a much lower standard of rigor than physical scientists expect. This allows for what I consider to be an unconscionable amount of bias and intrusion of opinion into what ought to be pure science.

The conclusions which extrapolate far beyond the data, discussions based more on opinion than factual evidence, and major conclusions drawn without considering the statistical value of the data, have led me to a deep distrust of biomedical research papers.

While I think the science sorts itself out over time, I don't think we conclude much of anything from most of these individual biomedical research papers. The exception, in my opinion, being those that have been heavily vetted by top-ranked journals. Biomedical research seems to work more by everybody pitching their snippets into a big pile for a decade and then shaking it to see what finally falls out. It probably works.... in the long run.

I no longer go out of my way to scientifically support or condemn the little papers. For us, it's a waste of scientific effort.

The politics of these papers is an entirely different matter. Politics do seem to play a large part in biomedical research, and most researchers in that field know how to interpret it. Clinical doctors and the media, on the other hand, do not. We physical scientists probably don't either, sad to say. I leave the politics to those academically and emotionally suited to political wrangling... and it needs to be done. But it's not science and it doesn't work like science.
 

alex3619

Senior Member
Messages
13,810
Location
Logan, Queensland, Australia
Hi everyone, you should really read the following article:

http://www.biotechniques.com/news/W...n=4bff12fd25-Weekly_02242011&utm_medium=email

It is also incredibly improbable that XMRV is not a real human virus. The debate is on actual prevalence and pathogenicity. Human cell lines can indeed carry it. So can humans. One in ten lab workers may be infected. It all depends on what assumptions you make. One in ten in the population may also carry it, but I suspect in the work capable population you might find it in only about five percent. That is enough to carry it around everywhere, and contaminate every lab. For example, has the WPI tested every single one of their workers? How about the lab cleaners?

However, from that argument I go to this: how come most to all of the controls are negative in the zero zero studies? There should be a background rate. These tests seem to be failing.

The only probable issue under my hypothesis might be that the real prevalence is more like the Japanese (1.7% ?) rate, and that lab workers might have a higher rate like 4% or something, so the source of contamination could easily be the lab or other staff, and it could be raising the reported control rates to nearly 8%, and due to heavy handling the patient rate could be much higher. This only happens in labs where there are contaminated workers. In labs without contaminated workers, the prevalence in controls could also be very very low.

I think this scenario is unlikely, but I wanted to put it out there.

Ultimately it comes down to the same point I have made repeatedly - we need more science until we have a definitive answer. We are not there yet.

Bye
Alex
 
Back