Q&A with Jose Montoya of Stanford about Cytokines, T cells, Journal Reviewers and NIH
Recently, I spoke to Dr. Jose Montoya about his paper on cytokines that appeared in the journal Proceedings of the National Academy of Sciences (http://www.pnas.org/content/early/2017/07/25/1710519114.abstract) July 31. We also delved into the unpublished research discussed by his collaborator, Stanford immunologist Mark Davis, at the Open Medicine Foundation’s conference in Palo Alto on August 12.
We began our conversation with one of the most important aspects of the cytokine paper: Montoya’s decision to stratify patients by severity into three groups and compare them with matched controls rather than simply comparing patients and controls.
HJ) One of the really interesting things about this paper was the way the stratification of mild, moderate and severe patients yielded so much information about correlation of symptoms with cytokine activity. Similar stratification has not been done before. When you began this work, were you planning to do several different analyses--starting with a comparison of controls and patients--or was stratification even part of the plan initially?
JM) Our stratification and analysis by severity was a plan from the beginning. It was our priority. But first we had to compare cases and controls. We were not expecting to see this strong, staggering correlation (between cytokines and severity). It hit us in the face. I will never forget the moment when I saw that correlation—I knew what we had in our hands immediately.
And everything that we’ve looked at can be pinpointed to the fact that cytokines are driving the symptoms. Lots of patients’ complaints have to do with inflammatory symptoms. Patients tell us it’s like a flu that never goes away, they describe pain, joint paint, sore throats.
HJ) Columbia’s Mady Hornig, in a 2015 paper (Mol Psychiatry.) found differences between patients and healthy people only when she analyzed her data by duration of illness. Three years was the apparent cut off; after that, she hypothesized, an “exhaustion” of the immune system set in and cytokine production dropped off. But you did not find a connection between duration of illness and cytokine expression. Can you elaborate on that?
JM) Duration? This is very important. There could be two reasons we couldn’t replicate the Hornig paper. A big one is, yes, we had only thirty patients (out of nearly 200) who had been ill less than three years, so that is too small a sample size from which to draw any conclusions.
But the second big thing that people are overlooking is that we found a problem with our Luminex assay that has to do with non-specific binding that we were lucky to discover. It can make results look more prominent than they actually are.
We were able to correct for that. For instance, we originally found twenty-three rather than seventeen cytokines (out of 51). When we corrected for the non-specific binding, it went down to seventeen.
On the Hornig paper, we don’t know if the finding was corrected on non-specific binding, and if it were to be corrected, would most of the findings go away?
The two studies are not contradicting, but we did our analysis in every way possible. We did it by duration alone and by severity alone. And then we put it together and duration always disappeared when we corrected for non-specific binding.
HJ) In the lay press coverage, your colleague Mark Davis talked about this data being meaningful enough to result in a diagnostic test. Are you enthusiastic about that possibility?
JM) If we are talking only of the technical issues, it could be months. If we talk about the process you have to go through—through the Food and Drug Administration—to make a diagnostic test clinically available, it could be one year. But, technical issues alone? A matter of months. It’s a matter of a company seizing the opportunity. We have been approached by different people. I am trying to stay away from the commercial part. So many doctors have profited from this disease and I don’t want to be among them.
There is another path you can take where a certified lab can actually have the test, validate it, and it can be used by hospitals and doctors as long as there is only one lab that does the test.
HJ) The cytokine data in this paper is a year old, as I understand the timeline. It took a year to go from lab to publication.
JM) Part of why it took us that long to get it to the public is that we wanted to be very cautious in what we published. The one thing we are very proud of is that we spend a lot of time, hours, just thinking about what ME-CFS can be, ultimately, based on the reality of what see in the patients and what has been done. We spend hours thinking, ‘If this is what we’ve seen, then how do we solve?’
And we want to make sure that the data will speak for itself. We don’t want to have to defend it. We saw this data and were so excited, but we had to do a lot of pre-analysis. We were so fortunate to see that everything was coming together. When non-specific binding came along, we thought this was all gone. Then, when it was corrected we went from twenty-three to seventeen.
The journal received our submission on November 16, 2016, and it was accepted seven months later. So, it was seven months for them to assess the paper. That is what I expected. I do have the experience that when I submit in areas that I am considered an authority the review process goes quickly. Every paper that I submit in ME-CFS takes this long.
HJ) The paper was reviewed by Anthony Komaroff, Gordon Broderick and Ben Katz of Harvard, the University of Montreal and Northwestern University respectively. All have published before on cytokines in ME.
JM) I was very sad that through the peer review process they asked us to drop one table that I thought was very important. You have to be humble in this process! But, in that table we showed a group of symptoms that correlated with a group of specific cytokines.
For instance, people with ME have hypersensitivities to drugs, chemicals and overall have allergic symptoms. One cytokine was very suggestive—it has to do with allergic symptoms. Another thing patients complain of is nerve pain—neuropathy. We found that nerve growth factor was another cytokine that correlated with that symptom.
So we were really excited because this cytokine data is correlated not just overall, but specific cytokines are also correlated with specific symptoms. It suggests even more that these cytokines are significant drivers of the patients’ symptoms.
So, here, you start to see that our findings are really strong and if you go to a deeper layer of specificity, you can say the cytokines correlate with specificity. You can say, for instance, that these patients have symptoms that are allergic in nature, and lo and behold, we found cytokines that drive allergic reactions.
But, this is part of the natural peer review process. You cannot win everything. Please remember, we’ve seen our data for at least a year before we submitted it. We know way more from the inside what we saw over and over again for many months. People see the same things differently. And—the worst thing you can do is fight with reviewers!
HJ) It seems a shame that the table wasn’t included.
JM) Let’s see if I can find a middle ground with you regarding the PNAS reviewers. It is their prerogative—it’s their right to do their peer review. The review process makes the paper much better.
It would have been nice to have the table, but the fact that we had the paper published in such an important journal means that over time, we give lectures, we talk to the press and we bring the table with us.
HJ) Did the peer review process result in any other deletions?
JM) When we saw that transforming growth factor-beta was in patients, independent of severity, we were pretty excited about it. I sat down for weeks, literally weeks, looking at everything about the data and, my God, I saw that these are not ME studies, they are cancer studies! We were so excited that in separate work TGF beta has been proposed as such an important driver of lymphoma. We thought, ‘Could this be the missing link?’ And in the original paper we had the TGF-beta proposal with lymphoma in the abstract. The reviewer forced us to take it out of the abstract.
HJ) Given that there is such a preponderance of women with this disease your leptin findings are provocative.
JM) When you correct for everything else, leptin is higher in females than males, that is a biological fact. Jarred Younger (at the University of Alabama) has also found leptin correlated with daily fatigue.
So, leptin is higher in females and we also found it correlated with severity, so you can see a picture here. Could leptin be the reason the disease is more common than women? We are proposing that our finding related to leptin may explain why ME affects more women.
The other thing is that leptin is proinflammatory. It’s in adipose—fat tissue—and women have different distributions of fat tissue, so we also wonder if that’s part of why they have more ME.
In addition, in mice, leptin appears to disrupt the blood-brain barrier and puts neutrophils in the brains of mice. We have wondered if the cognitive abnormalities are due to leptin.
One lead we’re following is in the right arcurate fasciculus.
(Montoya published in 2014 on this finding in Radiology. His team found a distinct abnormality, a thickening of the gray matter, at two areas of the brain that are connected by the right arcurate fasciculus. The greater the abnormality, the sicker the patient, according to the study.)
Even if our work suggests this is a biomarker, we cannot get government funding. We get ridiculous critiques (from NIH grant reviewers)— ‘That has not been done before.’
But, that’s one lead we’re following and we’re reaching out to private funders.
HJ) Your colleague Mark Davis talked about problems with the NIH grant application process during his talk. He said, ‘The NIH assembles a jury that supposedly has the same knowledge as those who are applying for the grant. But imagine if Elon Musk had applied for a grant in Detroit for a Tesla ten years ago? That’s analogous to where we are.’
JM) It’s very sad, very true. ‘This has not been tried before.’ “This is novel.’
The other critique that’s very hurtful to ME being funded is that they usually recognize our names, but they ding us because we are including in our grants the names of highly recommended Stanford faculty who have not worked in this field before. But, we must attract different people! I would argue that people who have been doing the work should be abolished because they have not solved this in thirty years!
I can confirm to you without giving you names that our grant applications have been downgraded with criticisms that some of the members of my team, even though they are Stanford professors, don’t have experience in ME. It’s a great mistake by reviewers at NIH because very few scientists are interested in this field—a field with very little progress in thirty-five years!
Further, many who are involved are either dying or retiring. There should be a conscious effort, a stimulus. I understand that you must have people working with you who have some knowledge of the illness, but (NIH) has reviewers who have no idea of ME. It’s funny they critique the grant applicants when in fact the reviewers have no knowledge of the disease.
HJ) I’ve seen the list of reviewers for the center grants and it seems many of the applicants are better schooled in the disease than the NIH reviewers.
JM) There is no one in that place—my God, they have no idea what ME is. There should have been a more creative way to find reviewers. Maybe the grant review process should have been in front of them orally so they could have advanced questions to applicants.
HJ) Your collaborator Mark Davis, in his presentation at the Open Medicine Foundation, gave a talk that was purportedly going to be about autoimmunity in this disease, but he ended up saying that the evidence for autoimmunity was “low.” Quite honestly, he seemed more enthusiastic about the clonal expansion you and he are finding in a specific T cell class in ME, and the possibility that further research in this area might lead to discovery of a pathogen.
JM) We have to distinguish between the work we are doing against work others are doing that does suggest autoimmunity. So, the role of autoimmunity cannot be excluded. Mark Davis and I are not finding immunity, yet that does not mean that there is no autoimmunity there.
(The T cell work Davis described) is a separate project from the PNAS paper. We’re trying to find clever ways—if there is a pathogen—to identify it. This is an approach that involved looking at a T-cell receptor in CD8 T cells, and that T-cell receptor is specific for a pathogen. In this case, we were able to see there was a group of ME patients who clearly had a T cell receptor that was not present in controls.
We were shocked. But it was a happy surprise. We weren’t shocked because it didn’t make sense but because if it was confirmed, it could be a huge hit. A huge discovery. Here we had the possibility of one pathogen. It should have been negative—but it was not. We are excited, but similar to the approach we’ve applied to other work, we’re applying all the cautions.
This is ongoing work. There are several techniques, several technologies involved.
We strive to get closer to the truth. We are not looking for fame. We apply the scientific method. And if I understand something about the method, I go to people who are even more expert in that method. We let the data drive our work, not our preconceived notions. It’s the data standing alone, and data telling us what is the next step.