Clinical proteomics: Is this the end of the beginning?

Publish a paper recently? Share it here. Or discuss other proteomics-related publications.
m/z
Angiotensin Member
Angiotensin Member
Posts: 47
Joined: Thu Oct 17, 2013 5:06 am

Clinical proteomics: Is this the end of the beginning?

Postby m/z » Fri Oct 18, 2013 10:34 am

There is a really nice paper in Science Translational Medicine this week. In brief, these guys have developed an MRM assay that meassures 13 proteins in blood(!) that potentially could be used to discrimimate between bengin and malignant lung tumors. Could this finally be the beginning of proteomics entering the clinic? I hope so.

A Blood-Based Proteomic Classifier for the Molecular Characterization of Pulmonary Nodules
Sci Transl Med 16 October 2013:
Vol. 5, Issue 207, p. 207ra142
Sci. Transl. Med. DOI: 10.1126/scitranslmed.3007013
http://stm.sciencemag.org/content/5/207/207ra142

RDUnwin
Phosphoserine Member
Phosphoserine Member
Posts: 12
Joined: Sun Jul 24, 2011 11:29 pm

Postby RDUnwin » Thu Feb 27, 2014 12:40 am

It's not only this paper - there are a few that have been published in the last year where SRM panels have been used to get some form of classification of a decent number (>100) of patient samples. Still some big hurdles before these hit clinic, though. Chief among them is the complexity of the sample prep required for proteomics/MS, which make these assays a) expensive, b) require a skilled and experienced operaor (further impact on cost), and c) yet to be determined how they peform in 1000+ samples over years of use. They also limit throughout. Crack those , and you've got an assay that can REALLY be used widespread in the clinic.

In the meantime, and to play devil's advocate, I suspect the next step for panels such as these is the development of a multiplex ELISA....

m/z
Angiotensin Member
Angiotensin Member
Posts: 47
Joined: Thu Oct 17, 2013 5:06 am

Postby m/z » Sun Mar 02, 2014 2:41 am

You are right in being the devils advocate :-) But I think this study takes their candidate markers further down the biomarker pipeline than most studies I have ever seen - and still the markers appear promising. It could be one the first examples of a "proteomics biomarker" that actually gets to demonstrate clinical impact. Even if is translated into an ELISA assay it was discovered by proteomics...

tsbatth
Ubiquitin Member
Ubiquitin Member
Posts: 58
Joined: Thu Jul 12, 2012 7:26 pm

Postby tsbatth » Tue Mar 04, 2014 5:58 am

It would seem you would have to do a cost benefit analysis depending on which condition, and how many biomarkers you are screening for. If you need monitor 10-30 biomarkers to determine a state with high confidence one could reason it's cheaper with a LC/MS based approach (factoring in time, sample prep, solvent costs) vs say 1-6 candidates which can give same answer with a traditional ELISA assay, but I'm not sure how expensive ELISA can be. I personally don't think sample prep and expertise will be a big issue since MS/GCMS/triple quads are the gold standard for many tests done routinely by the 1000's on a daily basis in labs everywhere whether it's in health, environmental etc, so the transition shouldn't be that much difficult.

RDUnwin
Phosphoserine Member
Phosphoserine Member
Posts: 12
Joined: Sun Jul 24, 2011 11:29 pm

Postby RDUnwin » Wed Mar 05, 2014 1:29 am

Of course, you would need to do a cost benefit analysis, but I still think technical challenges exist specific to plasma proteomics.

From my point of view, the major factors impeding getting a proteomics-SRM into the clinic are generally around throughput. We've got some good data on target proteins in >100 samples but that took us 2-3 weeks - fine if you are using the markers for a rare disease (<20 samples/week) and/or can wait a few weeks for the result, not so good if you are looking at common cancers, diabetes or Alzheimer's (or, as in our case, pregnancy). The assays were also done with lots of QC by an experienced Post-Doc. The MS assays currently in the clinic are small molecules one (solvent crash sample, put supernatent into a vial, inject and run a 20min gradient). Certainly for plasma, even doing a protein precipitation, overnight digest and a single one hour gradient may lack reproducibility compared to such an assay, and will certainly lack sensitivity compared to antibody-based approaches. Again, for our study, we tried such a method and we couldn't see our targets without performing abundant protein depletion, adding time, money and expertise to our sample prep. SISCAPA or similar may overcome this, but adds cost.

The cross lab SRM comparisons that have been done are fantastic and very promising, but has anybody ever done a peptide SRM-based method on >1000 plasma samples and looked at robustness over many months in many labs, yet? That's the key, I feel.

Clinical proteomics IS starting to deliver - the thread title is correct I think that we are starting to get to the end of the beginning. We can now use such methods to identify promising candidates and use them to do small-scale (n= 100-200) verifications of such. This means that we could NOW translate such assays into central labs running assays in rare conditions. In that respect we're not far off next generation sequencing-type genotyping assays, and I don't think we, as a field, shout about that enough. The technical challenge now is in the translation of such methods into 1000s or 10,000s of samples in a high throughput manner...

m/z
Angiotensin Member
Angiotensin Member
Posts: 47
Joined: Thu Oct 17, 2013 5:06 am

Postby m/z » Thu Mar 06, 2014 3:41 am

... while we are talking about clinical proteomics - we are currently setting up a a proteomics method for LCMS profiling of fat tissue from patients with amyloidosis. This method has already been implemented in several institutions, and is used in routine clinical diagnostics. But! this is just conventional label-free proteomics profiling and NOT targeted MS and there is no demand for absolute quantification (basically the aim is to provide a substitute for immune-histochemistry testing) and, also , amyloidosis is a very rare disease so throughput is not a big issue.
You can read about it here from one of the pioneering groups (not our group):

Reliable typing of systemic amyloidoses through proteomic analysis of subcutaneous adipose tissue.
Brambilla F, Lavatelli F, Di Silvestre D, Valentini V, Rossi R, Palladini G, Obici L, Verga L, Mauri P, Merlini G.
Blood. 2012 Feb 23;119(8):1844-7. doi: 10.1182/blood-2011-07-365510. Epub 2011 Sep 13.


Return to “Literature”

Who is online

Users browsing this forum: No registered users and 1 guest