#402 ‒ NMR blood analysis: how heart disease risk, insulin resistance, inflammation, and mortality risk can be assessed from a single blood sample | Jim Otvos, Ph.D.

How investigating a flawed 1986 cancer test led to the development of NMR (nuclear magnetic resonance) lipoprotein testing [A: 3:30, V: 0:11]

• Peter has followed Jim’s work for roughly 15 years
  • Tom Dayspring introduced him to it in May 2011, when Peter was immersing himself in lipidology
• Most listeners have had an LDL-P or HDL-P test done without realizing the test traces back to Jim, who created it

Tell us your story 

• Jim’s PhD is in biochemistry
• He spent 20 years in academia using NMR spectroscopy as a structural tool
• NMR machines are found in every chemistry department
• He held chemistry appointments at the University of Wisconsin, Milwaukee, then moved to North Carolina State University in 1990
• He was using NMR for the usual purpose—studying biomolecules
  • (He will explain how NMR works later)
  • NMR is very useful for a particular purpose: helping organic chemists determine the structure of molecules that they synthesize 
  • Jim was using it to understanding what was going on at the active site of zinc metalloenzymes (more challenging than understanding small molecules)

In 1986 a paper in the New England Journal of Medicine claimed a simple NMR test could tell whether someone had cancer, irrespective of the cancer type 

• The claim drew attention in the NMR field, though Jim did not normally read the NEJM
• The paper laid no mechanistic foundation; it simply measured how wide a couple of prominent signals in the NMR spectrum of blood plasma were, halfway up the signal
  • A narrow signal meant cancer
  • A signal that was not narrow meant no cancer
• Because it was in the NEJM, everyone with an NMR machine wanted to replicate it
• Jim’s chemistry department was not associated with the medical school, so he had no way to get plasma
  • He crossed the street to a hospital and talked the lab into giving him 6 leftover plasma samples from healthy people
• Half the signals were narrow and half were broad—but none of these people had cancer
  • 3 were women who had just given birth
  • Pregnancy was one false positive already noted in the NEJM paper

Jim shares, “If it wasn’t for that sort of linkage to something that seemed consistent with what was published, I probably never would’ve taken another spectrum of plasma.”

• Out of scientific curiosity he kept measuring plasma, and noticed the supposed cancer signal was NOT a clean symmetrical NMR signal—it had lumps, bumps, and shoulders
• Asking where the signal showed up and which molecules gave rise to it, it became clear the signals came from lipids and lipoprotein particles

Serendipitously Siemens Medical Systems gave him about $100,000 to investigate lipoproteins in people with and without cancer

• This was after a one-hour presentation on what he might learn with the money—a rare opportunity for a professor who had to go through a lot more hoops to get funding
• The funding let him get samples from people with and without cancer and separate the major lipoproteins—VLDL, LDL, and HDL 
• The VLDL signals always sat to the left of the LDL signals
• The LDL signals always sat to the left of the HDL signals
• It was the superposition of these signals and their relative concentrations differing that gave rise to the different shapes of this composite mixture signal that you would see in a plasma sample
  • It was obvious that this signal was coming from lipoprotein

The narrow “cancer” signal turned out to reflect higher triglycerides and lower HDL cholesterol (which together make the signal narrower)—nothing to do with cancer itself 

⇒ People with cancer have, on average, higher triglycerides and lower HDL cholesterol, a fact published 20 years earlier

• In 1991 Jim published a paper in Clinical Chemistry showing the NMR signals from isolated VLDL and HDL did NOT differ at all between people with and without cancer
• What was highly reproducible was that the VLDL, LDL, and HDL signals showed up in slightly different places 

Turning the signal into a lipoprotein measurement, and the decision to commercialize 

• Jim’s idea was to use this putative cancer signal as a source of information about lipoprotein concentrations
  • A simple low-tech NMR spectrum—obtainable on any machine—could generate a signal whose shape and amplitude let you deduce the concentrations of VLDL, LDL, and HDL
• On Siemens’ advice he filed a patent, despite knowing little about patenting or caring about commercialization
  • This NMR method seemed to have some advantage over the usual way of measuring triglycerides and LDL and HDL cholesterol via normal chemical methods
  • The patent was issued
• There were no NMR machines in clinical laboratories—and there still are none to this day—so clinical translation would require a commercial vehicle
• The idea evolved across the early 1990s to about 1995–96, supported by a couple of NIH grants for analytic development

Unexpectedly, NMR could differentiate not just VLDL, LDL, and HDL but their size subspecies—smaller, medium, and larger particles 

• Ron Krauss at UC Berkeley’s Donner Laboratory had shown (via laborious gradient gel electrophoresis) that LDL could be differentiated by size

⇒ And he found that people with a prevalence of small, dense LDL had greater cardiovascular risk at a given LDL cholesterol level than people with large LDL

• But the electrophoresis took a couple of days start to finish, so it was not clinically translatable or efficient
• Krauss sent Jim about 45 samples along with the electrophoresis tracings
• When Jim applied his analysis for decomposing the composite NMR signal into its parts, he could clearly tell pattern A (large LDL) from pattern B (small LDL) 
• After a couple of years, he refined the method to quantify small and large LDL and HDL
• The clinical promise of measuring LDL size drove Jim toward commercializing NMR testing—a step he felt very unqualified to take

“What would drive the utility of NMR testing was if it could measure something better and different… if we could measure small dense LDL pattern A and B.”‒ Jim Otvos

• It really did seem that you could not only generate the same information as a lipid panel by NMR, but you could also measure something better and different
• If we could measure LDL pattern A and B [pattern B are small dense], that would be a very useful thing clinically

⇒ LDL pattern B has a 3-fold greater risk associated with it at a given level of LDL cholesterol 

How standard lipid panels measure cholesterol and triglycerides, and why LDL cholesterol is estimated rather than directly measured [A: 15:15, V: 12:50]

A patient gets their blood drawn—what has to happen chemically to produce a basic lipid panel? 

• The basic lipid panel: total cholesterol, LDL-C, HDL-C, triglycerides all in mg/dL
• A conventional lipid panel uses standard chemistry-based assays: a reagent is added that reacts with what you are trying to measure
• The triglyceride assay is instructive because triglycerides are fatty acids esterified to glycerol
  • What actually happens in that assay is the blood is exposed to a lipase that hydrolyzes, that separates the fatty acid from the glycerol—it leaves the glycerol
  • And then something else is added to to make a color change in proportion to the amount of glycerol 
  • So the assay is actually counting the glycerol and imputing the triglyceride from the roughly 3:1 ratio

{end of show notes preview}