TWiV 113: Alan Rein on XMRV

When Alan says he’s surprised that it only takes a small amount to contaminate – well yes, that is what you call contamination.

Knowing the CFS community well, I think something needs to be cleared up, even though it has been repeatedly stated by many.

CFS patients, expert doctors, and careers, have not jumped on the XMRV research believing it is the cause of this disease. The support for research is about getting the answer right whether it is the cause or not. CFS research discoveries rarely gets followed up on, funding is small, and not much biomedical research gets funded. Jonathan Stoye has said twice that ‘it is in no one’s interest to not find this virus’. It would be lovely to think so, but that is not the reality that this research field has experienced. Therefore the scientific community has lost the trust of those patients, expert doctors and careers. The press release from the Wellcome Trust this last week is a good example of the reality of this field. There are too many powerful people who will take these political steps in order to have things their way. This must stop. It is unscientific, and helps no one.

Thanks guys for a great TWiV that advanced my understanding of the detection of XMRV. I’ve always shaken my head over that mouse mitochondrial DNA assay, thinking it’s just not convincing enough, but I like the sounds of this retrotransposon assay much better. Thanks also for Alan Rein’s thoughts on the sequences found in the Lo/Alter paper. I also noticed the disparity between the comments in the Hue paper (pear reviewed) and the media.

I have wondered for some time if sequencing all the virus or viral sequences found in XMRV positive patients was the way out, and although that question wasn’t asked, from the comments about sequencing, it sounds as if it’s not the way out of the current difficulties.

I would like to hear some discussion of WPI’s culture assay in the future. I know that this assay will not be practical where high throughput is needed, but if it’s as good as WPI says it is, could it be a way to get consistent results among different labs and at least find out if the virus is really there?

I’d like to see Ila Singh given a chance to comment on the criticisms of her work expressed in this episode.

This does not answer the question if all samples were handled in the same manner. Specifically, it does not address the collection of the blood samples. In the Science study, the samples from CFS patients were not freshly collected for the study but were already stored.

The the unpublished UK study, the samples might have been freshly collected, but were they also collected at the same time, by the same phlebotomist and were they handled in the exact same manner?

People tend to criticize the negative studies, but there is one thing I don’t understand: if I were Judy Mikovits and I were sure to differentiate between CFS patients and controls convincingly, I’d call Myra McClure or John Coffin or Frank van Kuppenveld to collect and blind the samples for/to me.

There is no way they could remain sceptical of my results if I were able to pick the sick needles from the blinded hay.

Some reports indicate that samples from controls weren’t handled in exactly the same manner as the samples from patients. And if blinding is indeed just blinding, that is one more great argument to ask a competing lab to do it. If not for a methodological reason, do it to convince the community and therefore to help the patients around the world as fast as possible.

So far, in a small test with blinded samples that were collected after the Science paper and were sent around to several labs, Mikovits and Ruscetti haven’t been able to differentiate between pedigreed positive samples and pedigreed negative samples in a convincing manner. And until they can, self reported results will not be very persuasive in solving the matter.

Can contamination be cultured? Will viral fragments multiply?

The claims you are making are outrageous, and illogical considering that labs who are able to find the retrovirus are involved in the BWG and the Lipkin study. What possible reason would any of them have to get this wrong? How would that save the life of anyone’s own daughter? The WPI have from the beginning offered to work and swap samples with other labs, such as van Kuppleveld and McClure (those we know of) The responsibility lies elsewhere.

Phase I conclusions:
1) “Five assays (CDC, FDA-Lo, NCI, WPI, Gen-Probe) for whole blood and all six assays for plasma were found to have no substantial differences in terms of sensitivity”
2) “The overall similarity of results suggests sensitivity of assays cannot explain differences in XMRV detection in clinical samples reported by the participating laboratories”
http://www.cfids.org/webinar/slides-121710.pdf

The supposedly well pedigreed negative sample cannot with certainty be known to be negative. There is no official test yet. If the samples had been contaminated in Phase IIa by the WPI collecting them, then the NCI would have found those samples positive as well. This therefore does not support the contamination THEORY. Furthermore the CDC methodology was different here than it was in Switzer et al.

Phase IIb:
Mikovits confirmed at the Blood Products Advisory Committee Meeting, that there is complete concordance between Ruscetti’s (NCI) and WPIs results by serology. The WPI serology results were not published. You are making accusations based on your own personal prejudices. Other member of the BWG were in the room, they would most certainly have spoken up if she was in any way wrong. Again, you cannot state with certainty that a sample is negative, even if it was pedigreed. False negative are to be very much expected at this time.

- Swapping samples is totally different from letting a competing lab collect and blind samples for you. The WPI was offered to take part in the latter by Dr. Enlander but were reportedly not interested. I would sure be interested if I was confident enough of my results. The only time WPI have been sent independently collected samples (outside of the BWG) they performed pretty poor: 2 out of 7 CFS patients and 1 out of 3 healthy people tested positive with samples supplied by Van Kuppenveld.

- Other people in the BWG did not speak up because Mikovits was not spreaking about results from the BWG. If those results were truly part of the BWG, they would have been presented at the meeting. WPI did not even complete all of the PCR results in time for the meeting and did not offer to do serology for phase IIb. Only CDC and NCI (Ruscetti) checked for antibodies. Mikovits was speaking of checking each other’s samples outside of the (blinded) setting of the BWG. I certainly look forward to WPI putting their serology assay to the (blinded) test of the BWG’s (Phase III) and Lipkin’s study.

- I (or anyone else) cannot state with certainty that the well pedigreed negative sample was negative. But because it was a healthy donor that was repeatably tested for PCR, culture and serology and because the donor was bled only once, the chances of this donor being actually positive for XMRV was well below 1%. You are saying that there is “only” a 99+% chance that there is something fundamentally wrong with Ruscetti’s positive serology result of the healthy and pedigreed negative control?

- If contamination occured in Phase IIa, you would not expect all samples but just a fraction of samples to be contaminated. Furthermore, some (CDC’s and WPI’s) XMRV PCR tests may be more sensitive than others (Coffin’s NCI lab) in picking up these trace amounts of contaminants. The combined results of Phase IIa and IIb are thus in line with some kind of contamination, while they certainly are not in line with the original Science findings. However, I will admit that the sample size is far too small to draw definitive conclusions based on these results alone.

- Your quote does not support your previous statement. From the transcript of the Phase I meeting at the FDA web page: “it looks like the CDC lab performed particularly well, the other three [including WPI] slightly less well, but still being able to detect to the end of the panel. The FDA Lo lab had one false positive, which turned out to be upon subsequent sequencing a nonspecific band of human genomic collagen. The WPI lab also had one out of six [pedigreed negatives] positive.” You can also see the chart which showed that the CDC performed better than WPI in both sensitivity and specificity on the FDA (but also on the CFDIS) web page.

Calling contamination a theory is too much credit btw, even in my view. In science, a theory is much stronger than a single hypothesis. See: http://wilstar.com/theories.htm. But thanks…

It is too funny to imagine my posts on this blog could influence the matter in the slightest.

How can the Van Kuppenveld results be just ‘about right’ when the incidence of XMRV in healthy controls was higher than in CFS patients (33% vs 29%)? That just doesn’t make sense, no matter were those healthy controls came from. And in Phase IIb of the BWG, Mikovits and Ruscetti collectively detected precisely as much XMRV or XMRV antibodies in pedigreed positive patients (8 out of 24) as in pedigreed negative samples (2 out of 6). This last task should be even easier, I might add. What are the odds of getting it accidentally wrong twice when you are actually right?

Also, from the slides that you have provided, it can be seen that the four positive samples were well pedigreed as well. While the general hypothesis that detection through PCR is troublesome can be accepted, these four subjects were tested multiple times by PCR beforehand, 17 times in total. On 16 of those 17 occasions the PCR(-only) result was positive for XMRV. It then does not make much sense that these very same patients gave such problems in detecting XMRV using PCR in the setting of the BWG. One patient even tested PCR positive 5 out of 5 times beforehand, but WPI could still detect XMRV only 1 out of 4 times in blinded fashion, which is the same result as they had with the pedigreed negative. Even in the unlikely event the pedigreed negative was a positive, you would still expect WPI to differentiate between this then weak positive sample and a sample that was tested as PCR postive 5 out of 5 times.

However, despite all the above, I am all for further investigation in the matter through the Lipkin study and Phase III and IV of the BWG. Given all the facts I just do not have high hopes anymore for the people that wish the existence of an XMRV->CFS link. The present evidence does not support such a link IMO, but of course that can always change through new data. There is still enough reason for further investigation into the matter, but Mikovits and Ruscetti should come with convincing results, which shouldn’t be too hard if the original report is indeed true, instead of falling back to their original results whenever they cannot reproduce these results themselves in blinded fashion. Again, it cannot be hard, when the Science and also the PNAS paper are indeed correct.

Finally, I quoted from the transcript of the Phase I results, which can be found here:
http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/BloodVaccinesandOtherBiologics/BloodProductsAdvisoryCommittee/UCM225388.pdf
It’s on pages 105/106. Also, on the slides you yourself have presented, you can clearly see that CDC performed better than WPI and NCI on whole blood sensitivity.

Here was you implying these comment had some influence:

“I think you want to jump the gun on this and put an end to the research.”

Oh well. The other loose ends:

- I already quoted “looks like”, so please do not act like I have quote mined. More importantly, the extra text you quoted is about the relation to the actual dilution, not in relation to other labs. Furthermore, I pointed you to the slide that showed the actual whole blood results of Phase I, which showed that the CDC performed better than WPI when confronted with identical samples. I’ll conclude this topic with a quote from the CFDIS web page: “CDC’s whole blood assay was the most sensitive under these conditions, while WPI was the only lab reporting an unexplained false positive result on a negative sample.”

- Van Kuppenveld did use other criteria, but your comment equally applies to your earlier argument of where the controls came from. You just wouldn’t expect the same results from controls and patients.

- They are “continuing their work”, which is exactly why that it is so strange that they haven’t cleared this mystery yet. If the Science results were as reproducable as the authors assert they are, this controversy would be long over. If I tomorrow report that apples fall twice as hard from my roof as pears do, nobody could ever definitely prove in the future that my findings of 12/29/2010 were wrong. It is not only up to me to report some miraculous findings, but it also my duty to supply methodology that makes my observations/results reproducable for others.

- I cannot be certain that a well pedigreed negative is positive, just as you cannot be certain that there is no invisible elephant on Judy Mikovits’ shoulder. Both are just very unlikely. The problem is that you reason backwards: because a “good” lab found a positive in a pedigreed negative, shows that the pedigreed negative was maybe not actually negative. Perhaps you can enlighten me as to why Mikovits and Ruscetti haven’t even tried to play this card as explanation for their false positive results – only patients on forums have?

Coffins IAP assay will also amplify human IAPs in people who have a retroviral infection and or autoimmune symptoms, wont they?

Alan Rein Summary:

1) Alan says that Lo/Alter did not find exogenous MLV’s but found mouse contamination due to using faulty(contaminated) assays. Recommends Harvey to go back and use the IAP probe. He says the MLV sequences alter found is exactly the same as the sequences in mouse(contamination).

2) Alan did not find XMRV in his own prostate tumor study.

3) When Ila Singh gave some of her XMRV+ samples to Alan, he found those same samples to be negative for XMRV when tested with his antiserum.

4) Alan says mouse contamination is so abundant that another researcher found XMRV mouse contamination in different cereal boxes.

5) Alan says it is not likely that geographical differences account for producing both positive and negative XMRV prostate studies(because he believes there is no XMRV since he did not find any in his prostate cancer study).

6) Alan says there is mouse contamination(nucleic acid) in water that even a .2micron reverse osmosis filter cannot filter out mouse nucleic acid (contamination) from water. Thus, all the water used in the labs have mouse contamination.

7) When asked about the antibodies to XMRV found by WPI proving infection, Alan didn’t have any comment but that it has to be reproduced by other labs.

8) Alan says he does not know nor can he explain why patient samples have more contamination then control samples(when BLINDED).

9) Alan strongly says that it is a big possibility that XMRV is just a contaminant and that it doesn’t exist.

10) Alan says * if * there are viruses(XMRV) in prostate tumor tissues, it is in so low a number that it doesn’t have anything to do with disease and it is just a passenger.

Alien Rein is either 100% right or 100% wrong.