GOOD AFTERNOON EVERYONE.
WELCOME TO THE WEDNESDAY
AFTERNOON LECTURE.
A VERY INTERESTING APPLICATION
OF TECHNOLOGIES AND PROTEOMICS
AND OTHER PARTS OF ANALYTICAL
CHEMISTRY TO FASCINATING ISSUES
THAT RELATE TO IDENTIFICATION OF
BIOMARKERS THAT MAY BE
PREDICTIVE OF TRANSLATIONAL
SUCCESS.
SOMETHING THAT WE ARE ALL
INTERESTED IN AND CERTAINLY HERE
AT NIH WITH THE INCREASED
CONVERSATION WE ARE HAVING ABOUT
HOW TO PLAY AS EFFECTIVE A ROLE
AS POSSIBLE IN HELPING PROJECTS
GET ACROSS THE VALLEY OF DEATH.
FROM AN ORIGINAL GOOD IDEA TO
SOMETHING YOU CAN PUT INTO A
HUMAN PATIENT WITH SOME
CONFIDENCE IT MIGHT WORK, THE
TECHNOLOGY WILL YOU HEAR ABOUT
FROM STEVE CAR IS HIGHLY
RELEVANT.
I'M NOT SURE WHETHER WHEN STEVE
GOT HIS Ph.D. IN ANALYTICAL
CHEMISTRY AT MIT ABOUT 30 YEARS
AGO HE WOULD HAVE IMAGINED
COMING TO THE NIH TO GIVE A
SPECIAL LECTURE OF THIS SORT.
I WOULD BET THAT WASN'T QUITE
PART OF THE LIFE PLAN.
THIS IS A -- IT IS A GREAT
EXAMPLE OF WHICH IN THE PAST IT
DIDN'T INTERSECT ALL THAT MUCH
ARE NOW INTERSECTING A LOT IN
THE SERVICE OF MEDICAL RESEARCH
AND THAT'S GREAT TO SEE THOSE
BOUNDARIES BREAKING DOWN AND
OFTENTIMES SOME OF THE MOST
INTERESTING SCIENCE SEEMS TO
HAPPEN AT THOSE INTERFACES.
STEVE HAS BEEN WORKING AT THOSE
INTERFACES THROUGHOUT A VERY
INTERESTING CAREER AFTER THAT
Ph.D. FROM MIT AND A POSTDOC
AT MIT, HE WENT ON TO THREE
YEARS AT HARVARD MEDICAL SCHOOL
AND MOVED TO THE PRIVATE SECTOR
AND SPENT MANY YEARS AT
SMITHKLINE WHICH THEN BECAME
SMITHKLINE BEECHAM, WHICH THEN
BECAME GLAXOSMITHKLINE.
THESE THINGS TEND TO HAVE THEIR
EVOLUTIONARY PROCESS.
AND THEN HE ULTIMATELY MADE A
LEAP 10 YEARS AGO INTO A NEW
COMPANY CALLED MILLENNIUM,
FAMILIAR TO MANY OF YOU, WHICH
NO DOUBT WAS A WILD RIDE IN THE
EARLY DAYS OF THAT ENTERPRISE.
THEN AFTER A YEAR AT
NORTHEASTERN AT THE BARNETT
INSTITUTE, HE JOINED THE BROAD
INSTITUTE, JUST GETTING UNDERWAY
AT THAT POINT, AN INSTITUTE THAT
MANY ARE VERY FAMILIAR WITH,
IT'S A REMARKABLE ENGINE OF
DISCOVERY IN BOSTON THAT BRINGS
TOGETHER LOTS OF EXCITING IDEAS
AND INVESTIGATORS AND TECHNOLOGY
AND THERE STEVE SERVES AS THE
PLATFORM DIRECTOR FOR PROTEOMICS
AND BIOMARKER DISCOVERY,
OVERSEEING A TEAM OF 20
INDIVIDUALS WHO ARE INVOLVED IN
A WIDE RANGE OF INTERESTING
PROJECTS THAT CROSS MANY
DIFFERENT DISEASES AND IF YOU
WANTED TO TALK ABOUT ALL OF
THOSE, WE WOULD BE HERE FOR MANY
HOURS.
HE IS GOING TO TALK ABOUT
QUANTITATIVE BIOLOGY AND
BIOMARKER DISCOVERY WITHOUT
IMMUNOASSAYS MAKING IT CLEAR
THAT WHEN YOU DON'T HAVE TO HAVE
AN ANTIBODY TO BE ABLE TO
MEASURE A PROTEIN AND THAT'S A
GOOD THING BECAUSE WE KNOW WE
DON'T HAVE ENOUGH ANTIBODIES OR
NOT ENOUGH THAT ARE ALL THAT
EFFECTIVE IN THEIR SPECIFICITY
AND SENSITIVITY AND MAS
SPECTROMETER COMES TO THE RESCUE
AND STEVE HAS BEEN A LEADER IN
FIGURING OUT WAYS HOW TO DO THAT
IN A HIGHLY MULTIPLEX SENSITIVE
FASHION IN A FASHION THAT ALSO
ALLOWS YOU TO ASSESS THE
QUANTITATIVE LEVELS OF LOTS OF
PROTEINS IN BIOLOGICAL FLUIDS
DERIVED FROM REAL HUMAN BEINGS.
SO WE ARE DELIGHTED TO HAVE
STEVE HERE WITH US THIS
AFTERNOON.
PLEASE JOIN ME IN WELCOMING
DR. STEVE CAR.
[APPLAUSE]
>> THANK YOU VERY MUCH FOR THAT
TERRIFIC INTRODUCTION AND THANK
YOU EVERYONE WHO IS HERE TO HEAR
MY LECTURE THIS AFTERNOON.
I'M JUMP RYAN IN.
I THINK FRANCIS HIT IT ON THE
HEAD.
WE HAVE LOTS OF THINGS WE WOULD
LIKE TO MEASURE, SO, GIVEN THE
FOCUS OF MY INSTITUTE IN PLANE
PLACES ON DOING GENOMIC-BASED
DISCOVERIES, I JUST WANT TO
REMIND EVERYBODY THE ROAD, THE
PATH, THE PHENOTYPE, GOES
THROUGH PROTEINS IN A LARGE
NUMBER, IF NOT IN ALL CASES.
AND SO WE HAVE LARGE NUMBERS OF
PROTEINS WE WOULD LIKE TO BE
ABLE TO MEASURE, JUST THE
UNANNOTATED IS ON THE ORDER OF
20,000 PROTEINS PREDICTED AND
THE MODIFICATION STATUS FOR
THESE IS ENORMOUS.
100,000 IS THE TOTAL GUESS.
BUT IT'S A BIG NUMBER.
AND WE WOULD LIKE TO BE ABLE TO
MAKE THESE MEASUREMENTS OF
PROTEINS AND THEIR MODIFICATION
STATES IN A VARIETY OF DISEASES,
STATES OF DEVELOPMENT, TREATMENT
WITH DRUG AND GENOMIC
PERTURBATION, SUCH AS A KNOCKOUT
OR KNOCKIN.
WHERE WE ARE AT ARE EXPERIMENTS
THAT ARE YIELDING LARGE NUMBERS
OF HYPOTHESIS OF UNCLEAR
RELEVANCE AND I'LL COME BACK TO
THAT POINT IN A MOMENT.
SO WE HAVE THIS GAP AND THIS
TERMINOLOGY VALLEY OF DEATH CAN
BE APPLIED IN A NUMBER OF
CONTEXT AND I WOULD CONJECTURE
OR VENTURE TO SAY THAT IT
REPRESENTS THE VALLEY OF EFFORT
FOR THESE MEASUREMENTS.
WE HAVE AN UNMET NEED IN ORDER
TO ACHIEVE MEASUREMENT OF ALL
THESE RELEVANT CHANGES USING
ASSAYS THAT ARE HIGHLY SPECIFIC,
SENSITIVE, PRECISE AND CAN BE
MULTIPLEXED IN THE BIOMARKER
AREA.
IT IS, IF YOU'RE IN THE
DISCOVERY MODE, IT ISN'T
PRACTICAL TO BE TAKING 100
MICROLITERS AND MEASURING ONE
ANO LIGHT WHEN YOU WANT TO
MEASURE ACROSS DIFFERENT
PROTEINS.
IT'S NOT ETHICAL, IT'S NOT
PRACTICAL.
BUT OUR GOAL HERE IS TO GET THE
QUANTITATIVE BIOLOGICAL
KNOWLEDGE.
SO, THERE IS A GAP AND THAT GAP
HAS LANGLEY BEEN FILLED NOW FOR
DECADES WITH IMMUNOASSAYS AND
THESE ARE CRITICALLY IMPORTANT
REAGENTS.
THEY WILL REMAIN CRITICALLY
IMPORTANT REAGENTS.
HOWEVER, THE SITUATION THAT WE
FIND OURSELVES IN NOW IN TERMS
OF NEW PROTEINS BEING IDENTIFIED
IN OUR EXPERIMENTS WHETHER
COMING OUT OF GENOMIC BASED
IDENTIFICATION OR FROM OTHER
OHMICS DISCOVERY METHODS IS THAT
THE NUMBER OF USEFUL ANTIBODIES
WE HAVE TO MEASURE THESE
MOLECULES IS RELATIVELY SMALL.
AND MAKING NEW IMMUNOASSAYS
ANTIBODIES THAT ARE REALLY
SPECIFIC AND SENSITIVE.
IT'S A FAIRLY TIME CONSUMING
TASK.
IT WOULD BE NICE IF WE HAD NEW
APPROACHES TO THIS PROBLEM AND
THAT'S WHERE I'LL FOCUS A LOT OF
MY TALK TODAY.
THERE HAVE BEEN ADVANCES MADE IN
THE APPLICATION OF EXISTING
TECHNOLOGIES TO INCREASE THE
MULTIPLEX BUILTY AND THROUGH PUT
OF THESE APPROACHES, SUCH AS THE
USE OF ANTIBODY ARRAYS.
THERE IS USES PARTICULARLY
THROUGH THE COMPANIES AND LARRY
GOLD'S EFFORTS AND WORKING OR
DISPLAYING PEPTIDES, PROTEINS
AND PEPATOIDS, SYNTHETIC AMYLOID
OF PEPTIDES, TO LOOK AT THOSE
THAT ARE BOUND.
THESE METHODOLOGIES ARE LIMITED
BY CONTENT.
THE AVAILABILITY OF HIGH QUALITY
REAGENTS TO PUT DOWN ON THESE
WERE APT MERES AND THE LACK OF
DEMONSTRATED SPECIFICITY PLAGUED
THESE APPROACHES.
SO I'M NOT GOING TO TALK TODAY
MUCH AT ALL ABOUT THE DISCOVERY
ARM OF THE PROTEOMICS PIPELINE
BUT I WANT TO EMPHASIZE A PLACE
WHERE THIS LACK OF REAGENTS IS
MOST ACUTE IS IN BIOMARKER
DISCOVERY.
AND A FUNCTIONING PIPELINE FOR
DOING THIS WORK REQUIRES YOU
HAVE TO START WITH SOME KIND OF
DISCOVERY METHOD, THIS COULD BE
PROTEOMICS AND IT CAN BE OTHER
SOURCES OF INFORMATION BE IT THE
LITERATURE OR GENOMIC BASED
METHODS.
BUT YOU END UP FROM YOUR
DISCOVERY EFFORTS NOT WITH
BIOMARKERS BUT WITH CANDIDATES
NEED TO BE EVALUATED.
HERE IN LIES THE KEY NEED.
TECHNOLOGIES TO BRIDGE THIS GAP
FROM THE DISCOVERY TO
PRECLINICAL VALIDATION.
AND WE'LL COME BACK ON
TO -- WE'LL TALK ABOUT PRIMARILY
THE CENTRAL ASPECT AND BRIEFLY
TOUCH ON THE FORMER.
SO DISCOVERY.
DISCOVERY HAS GOTTEN A REALLY
BAD NAME IN PROTEOMICS AND IN
PART BECAUSE OF WHERE A LOT OF
THIS WORK WAS STARTED.
AND TRYING TO DO DISCOVERY WORK
USING PLASMA OR SERUM, IF IT CAN
BE AVOIDED, IS A TERRIBLY
DIFFICULT PLACE TO START.
AND THAT IS -- JUST LOOK AT THE
RESULTS EVEN TODAY, THE BEST
EFFORTS TO DATE HAVE YIELDED
FEWER THAN 2000 PROTEINS
IDENTIFIED OUT OF PLASMA OR
SERUM.
THIS IS THE PROBLEM HERE REALLY
IS THE ENORMOUS DYNAMIC RANGE OF
THE PROTEINS THAT ARE PRESENT IN
BLOOD, ESTIMATED TO BE ON THE
ORDER OF 10 TO THE 11th.
MANY PROTEINS OR A SMALL NUMBER
EXTREMELY HIGH ABUNDANCE.
AND THE DYNAMIC RANGE OF THE MAS
SPECTROMETER IS LIMITED RELATIVE
TO THIS RANGE.
SO IT'S A PROBLEM HERE IN TRYING
TO ANALYZE BLOOD.
BETTER STARTING PLACE ARE
TISSUES OR PROXIMAL FLUIDS,
PROXIMAL FLUID IS A SURROGATE OF
THE TISSUE.
PROXIMAL FLUIDS SUCH AS SYNOVIAL
FLUID OR EVEN CSF OR
BRAIN-DERIVED PROTEINS, IN THIS
CASE THIS PROXIMAL FLUID IS
TAKING A TUMOR TISSUE, DICING IT
UP AND PUTTING IT INTO SAIL
ININE AND THEN AN INCUBATOR TO
BE ABLE TO CREATE A
PSEUDOPROXIMAL FLUID OF THE
ACTIVELY SHEDDING SECRETED
PROTEINS AND WE ARE USING THIS
PROCESS IN OUR BIOMARKER
DISCOVERIES IN A VARIETY OF
SOLID TUMORS.
WHEN YOU DO THIS, IT'S QUITE
STRIKING THE CONTRAST IN TERMS
OF THE DEPTH OF COVERAGE ONE
GETS SO RATHER THAN 2000
PROTEINS, YOU TYPICALLY GET
6-8,000 AND YOU NEED PROTEINS
GROUPS CONFIDENTLY IDENTIFIED.
SO YOU'RE GETTING INTO POSSIBLY
ASPfç MUCH AS 50% OF THE ACTIVELY
EXPRESSED PROTEINS ARE
DETECTABLE IN THE SAMPLES.
AND IMPORTANTLY, IN THE CASE OF
OUR VARIOUS CANCER STUDIES, WHEN
WE LOOK AT THE LIST, THEY ARE
HIGHLY ENRICHED PROTEINS WHICH
HAVE BEEN PREVIOUSLY IDENTIFIED
AS ASSOCIATED WITHl[gb CANCER.
IS THAT PROOF THAT THIS IS WORK
SOMETHING NO.
THAT IF THAT WEREN'T THE CASE, I
WOULD SAY FULL STOP.
NOW, ONE OF THE PROBLEMS WITH
ANY REAPPROACH, WE'LL COME BACK
ON TO THIS, IN ORDER TO GET THIS
DEPTH, YOU HAVE TO DO A LOT OF
SAMPLE PROCESSING.
WE TYPICALLY WILL DEPLETE
ABUNDANT PROTEINS FRACTIONATE AT
THE PEPTIDE AND POSSIBLY THE
PROTEIN LEVEL AND THEN LENGTHY
UNINSTRUMENT ANALYSIS.
A TYPICAL ANALYSIS IN OUR LAB
TAKES ESSENTIALLY TWO WEEKS PER
SAMPLE RUNNING 24-7 ON THE
MACHINE.
WHEN YOU DO THAT, YOU GOAT THESE
SORTS OF NUMBERS AND TYPICALLY
WHAT YOU'LL OBSERVE IS THAT
5-15% OF THE PROTEINS ARE 5-FOLD
OR HIGHER OVER EXPRESSED IN THE
TUMOR VERSUS THE CONTROL.
SO YOU'RE TALKING ABOUT HUNDREDS
TO UPWARDS OF 1,000 PROTEINS
THAT ARE CANDIDATES COMING OUT
OF THE EXPERIMENTS.
THAT'S THE PROBLEM.
THERE AREN'T 1,000 ANTIBODIES
THAT YOU CAN GO TO THE CATALOG
AND USE TO ANALYZE WHETHER OR
NOT THESE PROTEINS HOLD UP IN
LARGER NUMBERS OF PATIENT
SAMPLES.
SO JUST TO REFLECT ON THIS FOR A
MOMENT.
OUR DISCOVERY OHMICS MINIMALLY
CREDENTIALED MARKERS FOR
TESTABLE HYPOTHESIS NOT
BIOMARKERS.
WE GET HUNDREDS OUT OF THESE
EXPRESSION ANALYSIS EXPERIMENTS.
DEEP INTERROGATION AS I SAID,
INCREASES ANALYSIS TIME WHICH
LIMITS THE PRACTICAL NUMBER OF
SAMPLES YOU CAN ANALYZE IN THESE
DISCOVERY EXPERIMENTS.
AND AS WE ALL KNOW, USE OF FEW
SAMPLES IN A DISCOVERY MODE IS
AGAINST VERY HIGH DATA
DIMENSIONALITY IS THE RECIPE FOR
HIGH FALSE DISCOVERY RATE.
THAT DOESN'T MEAN -- IT DOESN'T
MEAN THAT THE UNDERLYING
TECHNICAL REPRODUCIBILITY OF THE
MASS SPECTROMETERS FAULTY.
IN FACT, I DON'T HAVE TIME TO GO
INTO THIS BUT RICHARD SMITH'S
LAB AND MY LAB HAVE COLLABORATED
TO COMPARE WITH THE TECHNICAL
REPRODUCIBILITY IN EXACTLY THIS
HIGHLY COMPLEX FRACTIONATION AND
ANALYSIS PARADIGM.
WE FIND 80% OF THE PROTEINS WE
IDENTIFIED ARE COMMON BETWEEN
THE LABORATORIES ACROSS MULTIPLE
PATIENT SAMPLES AND MORE
IMPORTANTLY, 75% OF THE PROTEINS
THAT ARE IDENTIFIED AS
DIFFERENTIALLY ABUNDANT, 5 FOLD
OR HIGHER ARE ALSO IN COMMON.
SO THESE INSTRUMENTS ARE
PRODUCING SIMILAR RESULTS.
THE PROBLEM IS THAT THERE IS
GREAT INTERINDIVIDUAL
VARIABILITY FROM PATIENT TO
PATIENT AND A HIGH PROTEIN LEVEL
IN ONE PATIENT IS NOT -- IT'S A
BIOMARKER.
IT COULD BE NORMAL VARIATION IN
THAT PROTEIN ABUNDANCE.
AND FINALLY, WE TYPICALLY WILL
WORK -- WE ARE NOT SO PROVINCIAL
WE ONLY LOOK AT OUR PROTEOMICS
DATA.
WE TYPICALLY WILL MINE
MICROARRAY EXPERIMENTS OF HIGH
QUALITY.
OF COURSE WE GO TO THE
LITERATURE AND SO REALLY TO
REACH OUR CANDIDATE WITH THE
RESULTS OF THESE EXPERIMENTS FOR
WHICH THERE IS NO INFORMATION
ABOUT THE PRESENCE OR
DETECTABILITY OF THOSE
CANDIDATES AND APRIL BLOOD WHICH
IS ULTIMATELY -- PERIPHERAL
BLOOD WHERE OUR BIOMARKER TEST
HAS TO BE DEPLOYED.
WE NEED ROBUST QUANTITATIVE
METHODS TO HELP US PRIORITIZE
THESE LENGTHY LIST OF
CANDIDATES.
THEY ARE COMING OUT OF OUR
DISCOVERY EXPERIMENTS TO WHITTLE
DOWN TO A LIST THAT IS IS WORTH
EVALUATING IN LARGE NUMBERS OF
PATIENT SAMPLES.
AND I'M GOING TO DESCRIBE TO YOU
NOW OUR TWO APPROACH THAT IS WE
HAVE DEVELOPED THAT ARE A REALLY
KIND OF INTER1THAT SEQUENTIALLY
STEP YOU THROUGH PRIORITIZATION
AND THEN QUANTITATIVE ASSAYS FOR
THESE PROTEINS.
THAT'S THIS VERIFICATION ARM AND
WE ARE GOING TO HAVE A ACRONYM
SOUP HERE WHICH I'M NOT GOING TO
DWELL ON BUT WALK YOU THROUGH
EACH TECHNOLOGY IN A MOMENT.
IMMUNOAS AS REMAIN ON THIS LIST
BECAUSE WE WILL BE OPPORTUNISTIC
IF GOOD QUALITY REAGENTS ARE
AVAILABLE.
SO STEP ONE OF THIS IS THE USE
OF A SEMIQUANTITATIVE LABEL-FREE
TARGETED APPROACH TO DETECT
PROTEIN CANDIDATES IN PROTEIN
PLASMA AND WE CALL THIS -- WHAT
WE ARE DOING A VERY SMALL SCALE
REDISCOVERY EXPERIMENT, THE
SAMPLE PROCESSING NOW WE MOVE
FROM OUR WHATEVER WE DISCOVERED
IT, WHETHER IT BE TISSUE OR
MOCKS MALL FLUID, WE WILL LOOK
IN BLOOD AND ASK AND ANSWER THE
QUESTION WHEREVER WE DISCOVER
THIS PROTEIN, IS IT DEDUCTIBLE
IN PERIPHERAL BLOOD?
IN CASES AND IS IT
DIFFERENTIALLY ABUNDANT IN CASES
VERSUS CONTROLS?
SO WE DEPLETED PROTEINS, CUT THE
PROTEINS UP TO PEPTIDES AND HAVE
A SMALL NUMBER OF FRACTIONS THAT
WE INTERROGATE USING THE SAME
INSTRUMENT WE USED FOR DISCOVERY
BUT NOW USING IT IN A TARGETED
FASHION.
SO, THESE INSTRUMENTS ARE PRETTY
MARVELOUS THESE DAYS.
TELL THEM YOU CAN TRADE LIST OF
PEPTIDES TO GO AFTER THAT AS
LONG AS 2000 PEPTIDES IN LENGTH.
SO WE USE OUR DISCOVERY
EXPERIMENTS KNOWING WHAT
PEPTIDES WE OBSERVED FOR OUR
DIFFERENTIALLY ABUNDANT PROTEINS
AND POPULATE OUR INCLUSION LIST
WITH THE MASSES OF THESE
PEPTIDES.
AND THEN, THE INSTRUMENT LOOKS
CONSTANTLY EVERY FEW
MICROSECONDS AT THAT LIST AND
ASKS, AM I DETECTING ANYTHING
WITHIN A FEW PART PER MILLION
THAT IS -- AND THIS IS FOR
PEPTIDE AMASS 1,000, THAT'S IN
THE FOURTH DECIMAL PLACE IS
WHERE WE ARE LOOKING.
SO A PEPTIDE THAT HAS A MASSIVE
IN THE TOLERANCE IN THE LIST, IF
I SEE IT, THEN AND ONLY THEN DO
I TRIGGER COLLECTION OF THE
SEQUENCING EXPERIMENT.
THAT'S THE EXPERIMENT.
SO IN THIS WAY, WE USE THE CYCLE
TIME IN THE MACHINE ONLY TOW
FOCUS IN ON THINGS THAT ARE ON
THAT LIST AND NOT LOOKING AT
ANYTHING ELSE.
THIS IS -- WE GET BETTER
SENSITIVITY AND THIS IS HIGHLY
SPECIFIC BECAUSE WE ARE USING
HIGHER RESOLUTION CORRECTION OF
THE PRECURSORS.
WE CAN DO UP TO 2000 PEPTIDES
MONITORING IN A RUN.
SO A WAY TO THINK ABOUT THIS IS
A HIGHLY MULTIPLEX MASS
SPECTROMETRY WESTERN.
IT'S NOT QUANTITATIVE.
IT'S SEMIQUANTITATIVE.
IT'S USEFUL FOR ASKING AND
ANSWERING THIS QUESTION.
THIS IS TO GIVE YOU A RESULT
WITHOUT WALKING YOU THROUGH THE
DATA IN THE CASE OF BREAST
CANCER BIOMARKER PROJECT WHERE
WE HAD MORE THAN 6,000 UNIQUE
PROTEINS IDENTIFIED IN FLUIDS.
WE HAD 1200 OUT OF THE 6,000
PROTEINS UP REGULATED MORE THAN
5 TIMES IN THE DISCOVERY DATA.
WE WERE ABLE TO OBSERVE 250 OR
20% OF THE 1200 IN OUR
PERIPHERAL PLASMA OF PATIENT
CASES.
SO THIS REDUCES OUR RISK QUITE
SIGNIFICANTLY AND IT'S ABOUT 260
WE WOULD START THIS ADDITIONAL
WORK ON.
IN ANOTHER STUDY THAT I'LL WALK
THROUGH IN MORE DETAIL, THIS IS
A MODEL OF MYOCARDIAL
INFARCTION, PMI.
IN HUMANS WE IDENTIFIED 1200
PROTEINS IN OUR DISCOVERY
EXPERIMENT.
THIS WAS DONE IN BLOOD.
WE HAD TO START THERE.
OF THOSE 1200, ROUGHLY 10%, 122,
WERE UP REGULATED 5 FOLD AND 52
OF THOSE QUALIFIED IN PERIPHERAL
BLOOD FROM A DIFFERENT SET OF
PATIENTS UNDERGOING THIS
PROCEDURE.
SO AGAIN, THESE NOW BECOME IN
BLUE, THE TARGETS FOR SUBSEQUENT
WORK.
THE NEXT STEP IN THIS PROCESS IS
QUANTITATIVE DEVELOPMENT OF
TARGETED MS-BASED ASSAYS TO
MEASURE THESE THINGS IN LARGE
NUMBERS OF PATIENT SAMPLES.
AND HERE WE ARE GOING TO USE
SOMETHING CALLED STABLE ISOTOPE
DILUTION, MULTIPLE REACTION
MONITORING OR SELECTION REACTION
MONITORING.
LC-MST IS TECHNOLOGY THAT WE
DIDN'T INVENT BUT WE COOPTED
FROM THE CLINICAL CHEMISTRY
COMMUNITY THAT HAVE BEEN USING
THIS PROBABLY FOR 30 YEARS OR
LONGER, FOR THE ANALYSIS OF
SMALL MOLECULES.
WE ARE JUST USING THAT SHAME
GENERAL APPROACH NOW TO MEASURE
QUANTITATIVELY -- THAT SAME
GENERAL APPROACH -- BASEDDA
PEPTIDES DERIVED FROM THOSE
PROTEINS IN COMPLEX SAMPLES.
THE WAY WE DO THIS IS, HERE ARE
OUR CANDIDATES, THESE ARE THE
PROTEINS THAT WERE ABUNDANT
DERIVED FROM THE LITERATURE.
WE DEFINED SIGNATURE PEPTIDES.
THESE ARE PEPTIDES THAT WE
EXPECT WE EITHER OBSERVED FROM
THESE PROTEINS OR EXPECT TO
OBSERVE USING COMPUTATIONAL
APPROACHES.
WE CAN TALK ABOUT IN THE Q&A,
THERE ARE WAYS OF DOING THIS.
WE SYNTHESIZE INTERNAL STANDARDS
WHICH ARE HEAVY STABLE
ISOTOPICALLY VERSIONS OF THESE
PEPTIDES AND SPIKE THEM IN WITH
OUR SAMPLE AND ANALYZE THEM
TOGETHER WITH THE ENDOGENOUS OR
LIGHT VERSION OF THAT PEPTIDE.
THESE STABLE PEPTIDES THAT WE
ADD ARE FIZZIO CHEMICALLY
IDENTICAL TO THE NATURALLY
DERIVED PEPTIDE.
SO THE ONLY DIFFERENCE IS THAT
THEY WEIGH MORE FROM THE
PEPTIDE.
THAT IS NATURALLY OCCURRING IN
THE SAMPLE.
AND TYPICALLY, WE DO THIS BY
LABELING THE C TERMINAL LYSINE
OR ARGININE WITH CARBON 13 AND
15 ISOTOPES.
BUT EVERY OTHER PROPERTY IT'S
RETENTION TIME.
IT'S IDENTICAL TO THE LIGHT
PEPTIDE.
AND THEN AGAIN THIS IS THE MRM
TECHNOLOGY.
DON'T WORRY ABOUT THIS.
I DON'T WANT YOU TO UNDERSTAND
THE FOLLOWING.
SAMPLES BEING SPRAYED OUT OF THE
LIQUID GRAPH, SELECT SPECIFIC
SPECIES IN THE SAMPLE, THESE ARE
THE PEPTIDES, HEAVY AND THE
LIGHT VERSION.
WE BREAK THEM APART AND THEN WE
ANALYZE ONLY A SMALL SET OF THE
FRAGMENTS, NOT ALL FROM THESE
PEPTIDES AND ALL OF THIS IS DONE
TO INCREASE OUR CYCLE TIME, OUR
THROUGH PUT.
AND THEN WE RATIO IN ORDER TO
GET THE RATIO OF THE ANNA LIGHT
PEPTIDE TO THE SPIKE PEPTIDE.
PRETTY SIMPLE.
WHAT'S IMPORTANT HERE IS THAT
THESE RATIOS GIVE US PRECISE
RELATIVE QUANTIFICATION.
WE'LL COME BACK TO THIS.
THIS IS NOT ACCURATE
QUANTIFICATION.
THERE ARE WAYS OF GETTING TO
THAT WHEN YOU NEED TO.
THIS IS JUST HIGHLY PRECISE
RELATIVE QUANTIFICATION WHICH IS
WHAT WE NEED FOR MEASURING
ACROSS PATIENT SAMPLES AND
LOOKING AT CHANGES.
IMPORTANTLY, WE CAN DO THIS IN A
HIGHLY MULTIPLEX FASHION WHERE
WE CAN DO IN A MOMENT 100 OR
MORE PEPTIDES SIMULTANEOUSLY CAN
BE QUANTIFIED.
THEY GIVE US 100 TIMES MORE
SENSITIVITY THAT CONVENTIONAL
APPROACHES.
THIS WORK IS NOW PUBLISHED IN
NATURE BY A TECHNOLOGY IN 2009
BUT REPRESENTED A LARGE SCALE
REALLY THE FIRST LARGE SCALE
INTERLABORATORY EVALUATION OF
THIS TECHNOLOGY.
IT WAS DRIVEN THROUGH THE
NCICPTAC PROGRAM AND
DEMONSTRATED FOR THE FIRST TIME
THESE METHODOLOGIES COULD BE
MULTIPLEXED AND ASSAYS COULD BE
CONFIGURED AND DEPLOYED ACROSS
NON-EXPERT LABS WITH HIGH
REPRODUCIBILITY AND BOTH
INTRAAND INTER-LAP CVs OF 20%O
LORER.
THAT MIGHT STRIKE"cuG
Wñ YOU BUT
REALIZE THESE ARE NOT CLINICAL
ASSAYS OF THE THEY DON'T NEED TO
BE CLINICAL ASSAYS TO SATISFY
THE NEED FOR THIS VERIFICATION
STEP.
AND IF WE CAN TALK ABOUT WAYS TO
GET THESE DOWN RELEVANT TO THE
CLINICAL ASSAY SUB-10% RANGE
WHEN WE NEED TO.
IMPORTANTLY, WE CREATED REAGENTS
AND METHODS AND MULTI-LABORATORY
DATA SETS AVAILABLE ONLINE IN
PUBLIC REPOSITORS AND THIS IS
FACILITATED OTHER LABORATORIES
GETTING INTO THIS BUSINESS.
I SHOULD POINT OUT THIS IS THE
GOAL THE STUDY WAS NOT TO
ACHIEVE HIGH SENSITIVITY.
GETTING 1-3 MICROGRAMS AT THE
PROTEIN LEVEL PER MILL IS NOT
PARTICULARLY SENSITIVE.
THAT'S THE POINT OF SUBSEQUENT
STUDIES.
I SHOULD SAY THIS IS A
SUBSEQUENT STUDY UNDERWAY.
IT'S LOOKING AT 125 PLEX MRM
ASSAY NOW DONE IN DEPLETED
PLASMA.
THE REASON FOR USING DEPLETED
PLASMA IS TO GET THE LOQs DOWN
INTO THE NANOGRAM PER MILL
RANGE.
A MORE REASONABLE RANGE FOR A
LOT OF POTENTIAL BIOMARKERS OF
INTEREST.
THIS IS A LOT OF WORDS.
BUT WE ARE DOING IS THIS FOR 34
DIFFERENT PROTEINS AND WE ARE
DOING BLINDED VERIFICATION STUDY
WHICH WILL ASSESS THE
REPRODUCIBILITY FOR MEASUREMENTS
OF UNKNOWNS ACROSS MULTIPLE
LABORATORIES AND MULTIPLE
INSTRUMENT PLATFORMS.
THIS IS JUST AN EXAMPLE OF WHAT
THIS DATA LOOKS LIKE.
SO THIS IS 334 PEPTIDES, 1,000
TRANSITIONS BEINGS MONITORED IN
A SINGLE RUN.
IT'S REALLY QUITE FEASIBLE TO DO
THIS NOW ON MODERN MASS SPECK
INSTRUMENTATION.
SO BACK TO THIS QUESTION OF
SENSITIVITY.
AND THIS IS PLASMA.
THIS IS THE FAMOUS SLIDE SHOWING
THE DYNAMIC RANGE OF PROTEINS
EVERY ONE OF THESE MARKS IS A
CLINICALLY MEASURED PROTEIN.
ALL THE WAY DOWN TO THE BOTTOM
OF THE RANGE WHICH IS THE
INTERLUKEINS, THIS IS THE 10 TO
THE 11th ORDER OF MAGNITUDE
RANGE WHERE A MICROGRAM PER MILL
WAS HERE AND THINGS LIKE CRP
WORK AND THEN HERE IS THE
NANOGRAM PER MILL WHERE YOU HAVE
IN THE COURSE OR WHEN YOU HAVE
DISEASE, PSA AND CEA AND THE
ELEVATIONS INTO THIS RANGE.
SO THERE ARE ACTUALLY A LOT OF
PROTEINS THAT ARE WELL WITHIN
THIS ONE MICROGRAM PER MILL
RANGE WHERE MRM TECHNOLOGY IN
CURRENT FORM COULD BE DEPLOYED
IF YOU CARED TO DO IT.
AND ACHIEVE THESE HIGHLY
MULTIPLEXED AMERICAMENTS.
BUT THE REALITY IS THAT
MOST -- MEASUREMENTS -- MOST
CREDENTIALED BIOMARKERS EXIST IN
THIS NANOGRAM PER MILL RANGE AND
LOWER.
WE NEED TO FIGURE OUT WAYS OF
ACHIEVING OR GETTING US INTO
THIS RANGE.
AS I SAID ON THE SLIDE TITLE, WE
VIEWED THIS AS THE PLASMA IN THE
NEW YORK CITY OF MATRIX.
IF YOU CAN MEASURE IT THERE, YOU
CAN MEASURE IT ANY WHERE.
THE WAY WE DO THIS IS BY AGAIN A
SIMPLIFIED PROCESSING APPROACH
THAT INITIALLY DEPLETES ABUNDANT
PROTEINS AND THEN ENRICHES FOR
PEPTIDES WITH A FRACTIONATION
THAT PRODUCES 6-8 FRACTIONS AND
THE WAY YOU DO THIS DOESN'T
MATTER.
IT COULD BE EXCHANGE, IT COULD
BE REVERSE PHASE CHROMATOGRAPHY.
THAT DOESN'T MATTER.
YOU THEN INTERROGATE EACH ONE OF
THESE FRACTIONS FOR YOUR N
PEPTIDES, COULD BE UP TO 100 PER
FRACTION.
WE DEMONSTRATED THAT USING THIS
SIMPLIFIED PROCESSING WE COULD
PUSH THE SENSITIVITY FROM A
MICROGRAM PER MILL DOWN TO THE
BOTTOM OF THE NANOGRAM PER MILL
RANGE, EFFECTIVELY FOR ANY
PROTEIN OF INTEREST.
SO THIS IS A GENERAL APPROACH
FOR MEASURING PROTEINS IN THE
CONTEXT OF PLASMA OR ANY OTHER
BIOLOGICAL MATRIX YOU LIKE TO
LOOK IN.
I'LL JUST ILLUSTRATE THIS VERY
QUICKLY IN THE CONTEXT OF THIS
MODEL THAT I MENTIONED WHICH IS
THE PLANNED MYOCARDIAL
INFARCTION MODEL.
IT IS A THERAPEUTIC PROCEDURE IN
WHICH A CATHETER IS INSERTED
INTO A PATIENT WHO HAS
CONGESTIVE OR CONJUNCTION OF THE
VENTRICALS, BLOOD FLOW IS
RESTRICTED THROUGH THE HEART.
WE INJECT, THE CLINICIANS FROM
MASS GENERAL HOSPITAL, INSERT
THE CATHETER AND INJECT ALCOHOL
INTO THE ENLARGED REGION OF THE
HEART.
THIS KILLS THE TISSUE IN THAT
REGION, RELAXING IT AND ALLOWING
BLOOD FLOW TO BE RESTORED.
WHILE THIS PROCESS IS GOING ON,
WE CAN SAMPLE USING A SECOND
CATHETER IN THE CORONARY
SCIENCE, THE DRAINAGE CANAL FOR
THE MYOCARDIUM.
WE CAN SAMPLE THE BLOOD IN THAT
CORONARY SINUS PRE-OBLATION,
PREINJECTION OF THE ALCOHOL AND
10 MINUTES AND THEN 60 MINUTES
POST-INJECTION OF THE ALCOHOL.
AND THEN WE CONTINUE TO SAMPLE
IN THE PERIPHERY OF THE PATIENT.
BUT THE IRB REQUIRES US TO
REMOVE THE CATHETER, SAMPLING
CATHETERS AT ONE HOUR TIME
POINT.
WE DID OUR DISCOVER NETHIS AND I
MENTIONED THAT'S WHERE WE GOT
THE 1200 PROTEIN LIST FROM THE
120 DIFFERENTIALLY ABUNDANT
PROTEINS.
AND THEN WE ARE GOING TO MOVE TO
THE SEMIQUANTITATIVE
VERIFICATION STAGE, WHICH IS THE
ACCURATE INCLUSION MASS
SCREENING THROUGH THE SAMPLE NOW
IS NOT THE CORE NARROW SINUS
BLOOD BUT THE PLASMA OF OTHER
CASES.
AND THEN WE ARE GOING TO MOVE
INTO A QUANTITATIVE VERIFICATION
STAGE I WILL ILLUSTRATE USING
THOSE ASSAYS TO MONITOR
PERIPHERAL PLASMA OF BOTH
PLANNED MYOCARDIAL INFARCTION AS
WELL AS SPONTANEOUS PATIENTS AND
CONTROLS WHO ARE UNDERGOING
RETEEN CATHETERIZATION.
SO HERE IS THE TOP END OF THE
LIST OF 55 QUALIFIED PEPTIDES BY
NAMES.
THERE WERE ANTIBODIES AVAILABLE
FOR THE LIST OF 55 ONLY
FOR -- THIS SHOULD SAY 6-55.
WE CHOSE 4 OF THESE CANDIDATES
THAT HAD INTERESTING TEMPORAL
PROFILES IN TERMS OF CHANGE AND
ABUNDANCE.
THEY WENT UP EARLY AND TENDED TO
STAY UP, THE THINGS WE LIKE TO
MONITOR FROM A BIOMARKER
STANDPOINT AND WE DECIDED TO
MAKE ASSAYS FOR THOSE.
WE ALSO WERE ASSAYING FOR OTHER
PROTEINS IN OUR DISCOVERY LIST
THAT WERE WELL-KNOWN, WELL
ESTABLISHED MARKERS OF
CARDIOVASCULAR DISEASE.
I SHOULD SAY THIS IS ONE OF THE
GREAT THINGS OF TRYING TO USE
THIS.
THERE ARE SOME THINGS YOU BETTER
FIND IF YOU'RE WORKING ON
MYOCARDIAL INFARCTION.
SO NONSPECIFIC MARKERS LIKE CRP,
WHICH GO UP IN VIRTUALLY AGAIN
DISEASE.
THE MORE SPECIFIC THINGS, NEW
MARKERS AND THESE NOVEL
CANDIDATES THAT CAME OUT OF OUR
DISCOVERY EFFORT.
SO WE USED MULTIPLE PEPTIDES FOR
PROTEIN AS WE AGAIN THIS IS A
CONFIDENCE-BUILDING.
NOT JUST A SINGLE PEPTIDE BUT
MULTIPLE ONES AND CONSTRUCTED
THIS 42-PLEX ASSAY.
THIS IS DATA FOR -- PUBLISHED
ALREADY FOR ABUNDANT PROTEIN
MARKNERS THESE PATIENTS.
THIS IS A SUBSET NOW OF THE 10
OR MORE PATIENTS THAT WE HAVE
DONE THIS IN.
SO THIS IS CRP AND NPO SHOWING
THE RED AND GREEN -- SORRY.
THE PINK AND THE GREEN REPRESENT
BIOLOGICAL REPLICATE ANALYSIS OF
COMPLETE PROCESS REPLICATES
GOING BACK TO THE PATIENT SAMPLE
SHOWING REPRODUCIBILITY OF THE
MRM AND THE GREEN IS THE -- WE
HAPPEN TO HAVE ELIASA AVAILABLE
FOR ALL 4 PROTEINS AND I'M
SHOWING YOUcq)á?eç MRM DATA TO ELIASA.
THERE IS NOT 100%
CORRESPONDENTS.
NOR WE EXPECT THERE TO BE.
WHAT IS IMPORTANT IS THE
TEMPLATE TRENDS ARE CONSISTENT
WITH OUR INTERASSAY CDs ARE
BELOW 25%.
THESE ARE THE -- AND I DON'T
MEAN FOR YOU TO BE ABLE TO SEE
THIS VERY WELL.
BUT YOU CAN SEE THAT THESE ARE
BASICALLY SIX DIFFERENT PATIENTS
OUT OF THE 10 WE HAVE DONE SO
FAR.
TRACKING THESE FOUR NOVEL
CANDIDATES AND THE SHOWING A
TEMPORAL TRANSFER AT FOUR
DIFFERENT TIME POINTS FROM
PREINJECTION, 10 MINUTES, 60
MINUTES AND 24 HOURS ALL
MEASURED IN THE PERIPHERAL BLOOD
OF THESE PMI PATIENT SAMPLES.
SO YOU CAN SEE THAT THERE IS
VERY GOOD TRENDS.
OBVIOUSLY SOME PATIENTS RESPOND
DIFFERENTLY THAN OTHERS.
THERE ARE SOME CONSIDERABLE
DIFFERENCES, FOR EXAMPLE, THIS
PATIENT VERSUS THIS PATIENT.
THAT'S NOT TO BE UNEXPECTED.
SO JUST INDIVIDUAL VARIABILITY.
ALL OF THESE PROTEINS WERE DOWN
AT THE BOTTOM OF THE NANOGRAM
PER MILL RANGE AND THE CV FOR
THESE MEASUREMENTS WERE
REASONABLY TIGHT, UNDER 20%.
NOW IN THE LAST COUPLE OF
MINUTES, I'M GOING TO MOVE TO
A -- REALLY THE NEXT STAGE OF
THIS TECHNOLOGY.
SO I HAVE FOCUSED ON THE FACT
THAT WE ARE NOT USING PROTEINS
DIRECTED ANTIBODIES TO DO AN
INITIAL ENRICHMENT STEP.
INSTEAD WE ARE USING SEPARATION
METHODS TO GET US TO AN INITIAL
DECOMPLEX SAMPLE WE CAN MAKE OUR
MEASUREMENTS IN.
NOW I'M GOING TO TELL YOU ABOUT
THE USE OF ANTIBODIES DIRECTED
AGAINST THE PEPTIDE AND THAT'S
IMMUNOPRECIPITATE THE PEPTIDE.
NOT MAKING ANTIBODIES AGAINST
THE PEPTIDE AND THEN TRYING TO
PULL OUT THE PROTEIN, WHICH IS
THE MORE COMMON THING THAT IS
DONE.
LET ME TELL YOU WHO THIS WORKS
AND I'LL TELL YOU WHY WE WANT TO
GO THIS WAY.
SO HERE IS OUR PEPTIDE IN A
DIGESTED PLASMA SAMPLE FOR
EXAMPLE, AS WE HAVE DONE BEFORE,
WE ADD OUR HEAVY LABELED
STANDARD SIGNATURE PEPTIDE.
NOW INSTEAD OF BEING A
FRACTIONATION, WE DIRECTLY
CAPTURE OUT OF THIS DIGEST, WITH
ANTIPEPTIDE ANTIBODIES ON
MAGNETIC BEADS.
AND SO WE ARE SIMULTANEOUSLY
CAPTURING THE CARB BON 12
VERSION OF THE PEPTIDES FROM THE
ANNA LIGHT OF INTEREST AS WELL
AS THE SPIKED IN HEAVY INTERNAL
STANDARD FOR THAT PEPTIDE AND WE
GO THROUGH THE SAME EXPERIMENT.
WHERE NOW WE WASH TO GET RID OF
NONFIZZ LOGICALLY BOUND PEPTIDES
AND THEN GO DIRECTLY INTO THE
MAS SPECTROMETER.
YOU CAN SEE THIS ELIMINATES THE
DEPLETION WHICH IS WHICH IS
LABOR AND TIME SAVINGS.
AND WE END UP WITH EFFECTIVELY
THE SAME RESULTS.
SO THE ADVANTAGES HERE, THIS IS
SIMPLER HANDLING THAN WHAT I
JUST SHOWED YOU AND USING THIS
APPROACH WE DIRECTLY REACH THE
SAME SENSITIVITY AROUND A
NANOGRAM PER MILL STARTING WITH
ROUGHLY 10-30 MICROLITERS OF
PLASMA.
IMPORTANTLY THIS REQUIRES A
SINGLE ANTIBODY.
THE MASS SPECTROMETER
EFFECTIVELY IS THE SECONDARY
ANTIBODY HERE.
IT IS GIVING US THE SPECIFICITY
WE WOULD USE A SECONDARY
ANTIBODY FOR AS WELL AS THE
MEASUREMENT.
SO, WHAT ARE THE THINGS THAT
GIVE US THAT SPECIFICITY?
IT'S THE MASSIVE PEPTIDE
MEASURED BY THE INSTRUMENT AND
THE FRAGMENTATION OF THAT
PEPTIDE AND THE RELATIVE RATIO
TO THE FRAGMENTS OF ONE ANOTHER
AND THE RETENTION TIME IN THE
SYSTEM.
SO THERE IS MULTIPLE BITS OF
EVIDENCE THAT SAY THIS IS VERY
HIGH SPECIFICITY YOU WHAT THINK
YOU'RE MEASURING IS WHAT YOU'RE
MEASURING.
AND THEN WE CAN DO THE RATIO.
AND IMPORTANTLY, CREATING AN
ANTIPEPTIDE ANTIBODY THAT CAN
PULL OUT THE PEPTIDE OF INTEREST
TURNS OUT TO BE A RELATIVELY
STRAIGHTFORWARD PROCESS.
AND I'LL COME BACK TO THIS IN A
FEW SLIDES.
THIS IS MUCH MORE AMENABLE TO
AUTOMATION AND PROCEDURE I
SHOWED YOU ON THE PREVIOUS
SLIDE.
SEMIAUTOMATEED THIS USING BEAD
HANDLING ROBOT.
MOVING TO LIQUID HANDLING
SYSTEMS TO MAKE THIS AS ROBUST
AND AUTOMATED AS POSSIBLE.
AND THIS WILL ALSO INCREASE OUR
MEASUREMENT CV.
GOING BACK TO THIS SYSTEM, OUR
PLANNED MYOCARDIAL INFARCTION
MODEL, THIS IS CAPPA ASSAYS
DIRECTED AGAINST PEPTIDES AND
WELL-KNOWN MARKER OF CARDIAC
INFARCTION AND CARDIOVASCULAR
DISEASE AND A NOVEL MARKER STILL
BEING CRERBLED INTERLEUKIN 33.
IN BOTH CASES, THIS IS IN PLASMA
ABLE TO HIT SUBNANOGRAM PER MILL
LIMITS OF DETECTION AND
QUANTIFICATION.
THIS IS THE USE OF THE CARDIAC
COMPONENT MEASUREMENTS IN
PATIENT SAMPLES FOR FIVE
DIFFERENT PATIENTS SHOWING THAT
REMOVED AT LEAST THIS INTO
ACTUAL ASSAY USE.
THE IMPORTANT POINT IS THIS IS
CAPPA ASSAY RELATIVE TO
CONVENTIONAL SANDWICH AMINO
ASSAYS IS THEY ARE NOT SUBJECT
TO THE SAME INTERFERENCES, OF
WHICH THERE ARE MANY, THAT PLAGE
CONVENTIONAL IMLONG CALL ASSAYS.
THESE ANTIVIDEO TYPIC RESPONSES.
IN LARGE PART, THIS IS BECAUSE
THESE ANTIBODIES ARE DIGESTED
WHEN WE DIGEST THE PLASMA.
SO THEY ARE NOT PRESS WENT WE
ARE MAKING THE MEASUREMENT.
AND I SHOULD POINT OUT THAT THIS
IS -- IF YOU'RE INTERESTED IN
THIS FURTHER, I DIRECT TO YOU
THIS ARTICLE IN THE JOURNAL OF
THE IMMUNOLOGICAL METHOD THAT IS
DISCUSSES THIS IN MORE DETAIL.
NOW I TALKED ABOUT HITTING
NANOGRAM PER MILL LEVELS USING
10-30 MICROLITERS OF PLASMA.
MORE RECENTLY WE PUSHED THIS TO
SAY, WHAT HAPPENS IF WE -- HERE
IS AN EXAMPLE AGAINST 9 CANCER
RELEVANT TARGETS.
SORRY YOU CAN'T READ THAT.
I CAN'T EITHER.
IT'S NOT THAT IMPORTANT.
THE POINT IS, THAT IN THE
CONTEXT OF THE PLASMA, YOU CAN
HIT A NANOGRAM PER MILL FOR
EVERY ONE OF THESE TARGETS.
THIS IS A 9-PLEX CAPTURE AND
WITH GOOD CVs.
IF YOU NOW AS PROOF OF
PRINCIPAL, DIGEST A MILL OF
PATIENT PLASMA AND USE UP THE
AMOUNT OF ANTIBODIES FOR
CAPTURE, YOU CAN HIT 10 PICO
GRAM PER MILL DETECTION LIMITS
FOR THESE SAME 9 PROTEINS.
I WOULD SHOW YOU THERE IS REALLY
RELATIVELY FEW QUALIFIED
IMMUNOASSAYS AVAILABLE THAT KIND
OF SENSITIVITY GENERALLY MORE IN
THE 58-100 PICO GRAM PER MILL
RANGE.
SO THIS IS A DRAFTING TOOL THAT
CAN THINK ABOUT THIS AS A
DRAFTING TOOL FOR ASSAYS THAT
CAN GIVE US THIS SENSITIVITY IF
WE NEED IT.
IT'S NOT ALL THAT
STRAIGHTFORWARD AND IT'S A
COMPLICATION FOR THIS.
BUT EVEN WITH 100 MICROLITERS OF
PLASMA THAT PUSHES US INTO THE
PICO GRAM RANGE.
WE HAVE BEGUN TO ASSESS THE
INTERLABORATORY REPRODUCIBILITY
OF THESE ASSAYS AND IT'S QUITE
GOOD.
IT'S KIND OF IN THE RANGE OF
7-15%.
THERE IS A PAPER THAT IS WORKING
ITS WAY THROUGH THE REVIEW
PROCESS FROM MY LAB AND MANDY
POWELL'S LAB, WHO IS MY PARTNER
IN ALL OF THESE ACTIVITIES.
IT'S REALLY LOOKING AT A GROUP
AT THE UNIVERSITY OF VICTORIA
LOOKING AT THE REPRODUCIBILITY
OF THESE ASSAYS.
WE ARE WELL INTO THE PROCESS OF
CREATING AND QUANTIFYING NOW
THESE NUMBERS THAT ARE NO LONGER
CORRECTIVE.
250 CAPPA ASSAYS.
THIS IS AGAINST ROUGHLY ABOUT
110 DIFFERENT PROTEINS SO
MULTIPLE CAPPA ASSAYS PER
PROTEIN TARGETS GET.
THIS SHOWS YOU THAT WE EVALUATED
THESE 216 RISK ASSAYS IN A VERY
ROUGH 4 POINT CURVE TO GET THE
CAPTURE EFFICIENCY.
THIS IS HOW WE EVALUATE EACH OF
THESE ANTIBODIES.
AND SORT OF GRADE THEM.
AND THE IMPORTANT POINT IS THAT
95% OF THE PROTEINS HAVE AT
LEAST ONE WORKING ASSAY OUT OF
THESE 1PLUS PROTEINS.
AND 45% OF THE PEPTIDES PERFORM
THIS CAPTURE.
AND FOR ROUGHLY THEY NUMBER WE
HAVE TWO OR MORE ASSAYS PER
PROTEIN.
THESE ASSAYS, I HAVE SHOWN YOU 9
PLEX.
MORE RECENTLY, WE PUSHED THE
PLEX LEVEL UP TO 50 PLEX.
SO THIS IS SHOWING YOU PANELS OF
THE 216 ASSAYS CONFIGURED AS
GROUPS OF 10, 20, 30, UP TO 50.
AND THIS IS FOR ONE OF THE
PEPTIDES IN EACH OF THOSE
CONTEXT YOU WILL SEE THE
PERFORMANCE REALLYSf
R
CHANGE VERY MUCH.
IT'S NEARLY INDEPENDENT OF THE
PLEX THAT IT HAPPENS TO BE IN.
THERE ARE SOME BAD ACTORS.
SO THERE ARE SOME SOURCES OF
INTERFERENCE.
THIS REPRESENTS FEWER THAN 15%
OF THE ASSAYS SHOW THIS KIND OF
BEHAVIOR WHERE WHEN WE START TO
LOOK AT HIGH PLEX LEVELS, WE
BEGIN TO SEE INTERFERENCE
CREEPING IN.
SO, I WANT TO POINT OUT THAT WE
HAVE APPROXIMATELY 230 OF THESE
ANTIPEPTIDE ANTIBODIES MADE.
ANOTHER 150 WILL BE EVALUATED
BEFORE THE END OF THIS YEAR.
WE HAVE A LARGE NUMBER OF
FRACTION MRM ASSAYS AND
PROGRESS.
SO THERE IS A LOT OF WORK IN
THIS AREA AND WE ARE MOVING
RAPIDLY INTO PATIENT SAMPLE
ANALYSIS.
I MENTIONED BRIEFLY THAT THE
POTENTIAL FOR CLINICAL
UTILIZATION OF THIS, I JUST WANT
TO POINT OUT THE FIRST PROTOTYPE
CLINICAL CAPPA ASSAY HAS ALREADY
BEEN GENERATED BY OUR COLLEAGUE
AT THE UNIVERSITY OF WASHINGTON.
THIS IS THE REFERENCE AND IF YOU
CAN'T WRITE IT DOWN OR IF YOU'RE
INTERESTED, E-MAIL ME.
BUT THE POINT IS, THIS IS A
CLINICALLY MEASURED
CANCER-RELATED PROTEIN FOR
THYROID CANS THEY'RE HAS LOTS OF
INTERFERENCES.
THIS IS CAPPA ASSAY LARGELY GETS
AROUND WITH INTERFERENCES AND
GIVES PERFORMANCE THAT IS NEARLY
AS GOOD AS THE CURRENT CLINICAL
ASSAY.
NOW, I TALK ABOUT ALL THE
POSITIVE FEATURES OF THIS.
I HAVEN'T SHOWN YOU A SINGLE
MASS SPECTRUM THIS ENTIRE TIME
BUT I COULDN'T RESIST SHOWING
YOU SOME DATA.
AND THIS IS TO MAKE A POINT.
THAT IS THAT YOU CAN'T JUST JUMP
TO THIS TECHNOLOGY AND ASSUME
EVERYTHING IS GOING TO WORK.
PART OF IT IS THAT THE SOFTWARE
ISN'T THERE YET TO HELP THE
ANALYST FIGURE OUT WHETHER THEY
HAVE AN ASSAY THAT IS OR IS NOT
DETECTING THE RIGHT THING.
THIS HAS TO BE WITH THE FACT
THAT THERE ARE INTERFERENCES
THAT WILL SCREW UP.
THE ACCURACY OF YOUR
QUANTIFICATION IF YOU DON'T
RECOGNIZE THEM.
BECAUSE THESE ARE WELL BEHAVING
PEPTIDES.
PAY ATTENTION TO THE
CORRESPONDENCE OF THE RED AND
BLUE LINES HERE IN TERMS OF
RELATIVE RATIOS FOR VARIOUS
SPIKE LEVELS.
YOU'LL NOTICE THESE ARE ALL
PRETTY GOOD.
THAT'S THE RATIO OF THE HEAVY TO
THE LIGHT AN LIGHTósxFcWl SO xñWE WANT
 TO
SEE THAT.
WE WANT TO SEE THE FRAGMENT
LINES FROM THE PEPTIDE ADDING
THE SAME RATIOS AS WE ADD
INTERNALLY.
HERE IS THE CASE WHERE THE
PEPTIDE AND THE HEAVY ISOTOPE,
LOOK AT THE RATIO.
SO HERE IS THE HEAVY ISOTOPE
THAT SHOULD, THE BLUE SHOULD
HAVE THE SAME RATIOS RELATIVE
RATIOS, AS THE RED.
AND IT DOES NOT.
SO, IF YOU ARE PAYING ATTENTION,
YOU WOULD SAY THIS ISN'T THE
PEPTIDE OR IT HAS A TREMENDOUS
INTERFERENCE PRESENT.
AND I NEED TO DO OTHER THINGS.
SO WE HAVE WITNESSED A PROGRAM
CALLED, AUDIT.
IT ALLOWS US TO GO FROM HAVING
TO ANALYZE ALL OF THE DATA BY
EXPERTS TO A AUTOMATED SYSTEM
WHICH TELLS US WHEN THERE IS AN
INTERFERENCE PROBLEM.
IT'S AN OPEN ACCESS SOFTWARE
WHICH LOOKS AT THE RELATIVE WAY
OF THESE PEAKS AND ALSO LOOKS AT
THIS CV OF THE ANNA TREATS GIVE
THE ANALYST INSIGHT INTO WHICH
PEPTIDES HAVE BAD PERFORMANCE
AND WHICH ONES ARE OKAY.
YOU DON'T HAVE TO LOOK AT THE
DATA.
I'M ABOUT TO WRAP UP IN TWO
MINUTES.
I WANT TO TELL YOU SOME THINGS
THAT ARE ON THE NEAR HORIZ KNOW.
SO ON THE NEAR HORIZON, ROUTINE
100 PLEX ASSAYS USING THE
FRACTION MRM-BASED APPROACH FOR
ANY MATRIX OF ISSUE.
WE ARE GOING TO GET TO IMPROVED
SENSITIVITY THROUGH AUTOMATION
OF THIS PEPTIDEQ+[m3 ANTIBODY
APPROACH.
ROUTINE 30 PLEX IS CAPPA.
IF WE WANT TO MOVE THIS INTO
CLINIC, CAN WE SPEED UP THE
INJECT CYCLE?
THE ANALYSIS TIME.
THE ANSWER TO THAT IN
PRELIMINARY DATA IS YES.
WE CAN WE THINK GET THE ANALYSIS
TIME IN 50 PLEX.
WHICH WOULD MEAN 1,000 PROTEINS
PER DAY.
BEGINS TO LOOK LIKE THE CLINICAL
STANDPOINT.
AND EQUALLY IMPORTANTLY THE
SENSITIVITY OF MASS SPECTROMETER
INSTRUMENTS AND THE SPECIFICITY
OF THESE MACHINES IS INCREASING
ALL THE TIME AND IS GOING TO
TRANSLATE DIRECTLY IN OUR LIMITS
AND DETECTION OF QUANTIFICATION.
AND WE HAVE BEGUN TO EXPLORE
WITH THE FDA THROUGH A MOCK 510K
PROCESS, WHAT THE ISSUES ARE
THAT THE FDA WILL BE LOOKING AT
IF AND WHEN THE ASSAYS MOVE
FORWARD.
I SHOULD SAY THIS IS NOT THE
PRIMARY FOCUS OF MY LABORATORY
RIGHT NOW.
WE ARE REALLY FOCUSED ON THIS
VERIFICATION PIECE BUT NO REASON
FOR US TO CONSIDER THE BRIGHT
FUTURE OF WHAT ABOUT USING
CLINICAL SETTINGS?
TODAY THE ANSWER IS USE
VERIFICATIONS DOWN TO A SMALL
NUMBER OF HIGHLY CREDENTIALED
CANDIDATES AND MOVE THEM FORWARD
ON TO EXISTING CLINICALLY
DEPLOYED ANALYZER USING
MONOCLONAL ANTIBODIES.
IT CAN BE QUITE A CHANGE.
BUT I THINK THIS IS SOMETHING
WHICH WE SHOULD PAY ATTENTION TO
FOR THE FUTURE.
I FOCUSED ALMOST ENTIRELY ON
UNMODIFIED VERSIONS OF THEVGóçç
PROTEIN.
IT SHOULD GO WITHOUT SAYING ONE
CAN USE THIS TO MONITOR FOR ANY
VERSION OF THE PROTEIN YOU MIGHT
WANT TO GO AFTER.
SO THERE IS A COUPLE OF THINGS
WHERE WE ARE ONLY BEGINNING TO
LOOK AT THIS FROM A ANALYSIS
POST TRANSLATION MODIFICATION
AND A MUTATION STANDPOINT, SOME
WORK FROM JOHN, WHO IS DOWN IN
FLORIDA AND ALISA FROM A LAB
WHICH IS USED TECHNOLOGY TOW
LOOK AT EITHER PTMs WITH
PHOSPHOPEPTIDES SPECIFICALLY OR
UNIQUE VARIANTS OF PROTEINS IN
THE CONTEXT OF CANCER.
AND OF COURSE FOR IT TO WORK,
YOU WOULD HAVE A PATHOGENOMIC
PROTEIN, SOMETHING THAT
DETECTION ALONE, BECAUSE IT'S
UNIQUE, COULD BE DIAGNOSTIC AND
YOU COULD USE IT FOR ANY PTM OF
INTEREST.
SO, I THINK THIS TECHNOLOGY IS
REALLY DISCOVERY-BASED
LARGE-SCALE EXPERIMENTS AREN'T
GOING TO GO AWAY.
BUT I THINK THAT THESE
METHODOLOGIES I FOCUSED ON WILL
HAVE A TRANSFORMATIVE EFFECT ON
PROTEINS AND MOVING IT INTO A
MORE PRECISE, QUANTITATIVE AND
COMPREHENSIVE SCIENCE.
THIS REQUIRES US TO HAVE
BASICALLY DATABASES OF PROTEINS
AND PEPTIDES FROM THEM AND
EXPERIMENTALLY DERIVED
PROPERTIES.
THIS IS HAPPENING THROUGH
EFFORTS AT IFB, AT THE BROAD
INSTITUTE AND OTHER LOCATIONS.
REPOSITORIES OF REAGENTS IS ALSO
HAPPENING SLOWLY THROUGH FUNDED
EFFORTS TO NCI AND NHLBI AND
VARIOUS FOUNDATIONS AND ASSAY
DEVELOPMENT AND APPLICATION
LAPTORIES.
AND I WOULD SAY THAT THIS REALLY
BEGINS TO LOOK LIKE THE
INGREDIENTS OF A HUMAN PROTEIN
DETECTION AND QUANTIFICATION
PROJECT.
I'LL FINISH WITH THIS SLIDE
BECAUSE MUCH MORE OF A PLEA THAN
ANYTHING ELSE.
ALL OF THE WORK WE ARE DOING
WILL MEAN NOTHING IF WE DON'T
START WITH GOOD BIOSPECIMENS.
AND THERE IS REALLY AN URGENT
NEED OF COLLECTIONS OF
HIGH-QUALITY CASES AND
CONTROLLED SAMPLES.
THEY ARE ACCESSIBLE FOR THESE
DISCOVERY EXPERIMENTS.
THESE COLLECTIONS DO EXIST.
BUT THEY ARE VERY RESTRICTIVE IN
TERMS OF ACCESS TO THEM.
AND FOR DISCOVERY PURPOSES, WE
NEED GENERALLY MORE MATERIAL AND
THEY ARE WILLING TO GIVE.
I THINK THE TIME IS NOW TO
CONSIDER HOW WE BUILD THIS UP.
VERIFICATION PROGRAMS HAVE BEEN
HAMPERED BY THIS.
SOME OF THEM ARE REALLY QUITE
BAD.
THEY HAD BEEN HANDLED
DIFFERENTLY AND INTRODUCE BIAS.
THIS IS A BIG PLAGE OF THE
BIOMARKER PROTEOMICS AREA WHICH
WE KNOW HOW TO DEAL WITH BUT
REQUIRES US TO ADDRESS THIS
PROBLEM WITH THE LACK OF
BIOSPECIMEN.
AND STUDY BY STUDY, COLLECTION,
THE FUNDAMENTALLY WHAT WE ARE
FORCED TO DO RIGHT NOW BUT IT'S
VERY INEFFICIENT.
AND IT CAUSES US TO REBUILD
THESE CLINICAL COLLECTIONS WHICH
I THINK IS REALLY UNNECESSARY.
SO WE NEED TO HAVE SOME
ATTENTION PAID TO THEM.
I HOPE I HAVE SHOWN YOU THE
QUALIFICATION METHODS AND THEY
ARE STILL OF TREMENDOUS VALUE.
BUT THEY HAVE LIMITATIONS.
NOT ONLY IN NUMBERS BUT IN TERMS
AND IN CONTENT AND TERMS OF
VARIABILITY TO BE MULTIPLEXED
THAT THESE NEW APPROACHES DO
HAVE SUFFICIENT SPECIFICITY,
SENSITIVITY AND ARE SUFFICIENTLY
QUANTITATIVE TO MAKE
MEASUREMENTS IN ANY LYING CALL
CONTEXT.
AND THAT IF WE -- BIOLOGICAL
CONTEXT.
IF WE INTEGRATE, YOU CAN MOVE
FROM DISCOVERY PARADIGMS,
SYSTEMATICALLY TO REL
CREDENTIALED MARKERS THAT ARE
WORTH STUDYING IN LARGER
NUMBERS, PATIENT SAMPLES,
CLINICAL VALIDATION.
AND THIS IS A LARGE GROUP OF
PEOPLE THAT MAKE THINGS WORK IN
MY LAB AND I'D LIKE TO POINT OUT
MIKE HAS BEEN INSTRUMENTAL IN
THE MD-Ph.D. IN THE LAB AND
INSTRUMENTAL IN A VARIETY OF
EXPERIMENTS AS HAS MGH AND
CENTRAL TO ALL THE QUALIFICATION
WORK, AND A VARIETY OF OTHER
FOLKS.
MY COLLABORATORS IN THE PMI
EXPERIMENTS I TALKED ABOUT ARE
INCLUDING OF ROBERT AND THE MARK
AND THE CANCER RELATED WORK IN
ALL OF THIS, EFFORTS ARE DONE
COLLABORATIVELY WITH AMANDA AT
THE FRED HUTCHINSON CANCER
RESEARCH CENTER AND HER GROUP.
LEE ANDERSON HAS BEEN PART OF
OUR TEAM HELPING TO GIVE US
ADVICE AND GUIDANCE AND A GROUP
AT UNIVERSITY OF VICTORIA THAT
HAS BEEN HELPING US IN THAT
EFFORT AND NONE OF THIS WORK
WOULD BE POSSIBLE WITHOUT THE
GENEROUS FUNDING AND SUPPORT OF
THE NCI AND NHLB I AND VARIOUS
FOUNDATIONS.
THANK YOU VERY MUCH FOR YOUR
ATTENTION.
[APPLAUSE]
>> IF THERE ARE QUESTIONS, THERE
ARE PIKE MICROPHONES IN THE
AISLE.
IT'S GOOD TO GO THERE SO WE CAN
HEAR THE QUESTION.
GIVE YOUR NAME AND GO AHEAD.
>> IN TERMS OF DIAGNOSIS OF
MYOCARDIAL INFARC USING
CONVENTIONAL APPROACH, WHAT ARE
THE ADVANTAGES YOU HAVE AND DO
YOU HAVE ANOTHER PROTEIN ON
BIOMARKER YOU COULD USE EQUALLY
SENSITIVE OR BETTER?
>> SO I'M HAVING A LITTLE
TROUBLE UNDERSTANDING THE
QUESTION.
COULD YOU REPEAT JUST REPEAT IT?
>> I AM TRYING TO ASK A QUESTION
REGARDING THE SENSITIVITY OF
YOUR APPROACH USING INVASIVE
VERSUS CONVENTIONAL AND DO YOU
HAVE ANOTHER PROTEIN BIOMARKER
THAT COULD BE MORE SENSITIVE
THAN CP GAMMA B OR OTHERS AS YOU
MENTION ED?
>> SO, AGAIN THIS IS -- WE ARE
NOT FAR ENOUGH LONG TO SAY WE
HAVE A BIOMARKER THAT IS MORE
SENSITIVE AND SPECIFIC THAN
EITHER OF THOSE OTHER MARKERS.
NOT YET.
THIS IS A BIOMARKER PATHWAY
BEGINS WITH DEVELOPING THAT
HIGHLY CREDENTIALED SET OF
POTENTIAL OF CANDIDATE MARKERS
THAN PUTTING IT TOGETHER IN
PANELS AND MEASURING THEM IN
PATIENT SAMPLES AND LOOKING AT
THE RECEIVER OPERATOR CURVE
PERFORMANCE IN TERMS OF
SENSITIVITY AND SPECIFICITY
THE PANEL, FOR THE INDIVIDUAL
MARKERS AND THE PANEL.
WE ARE MOVING IN THAT DIRECTION.
THAT IS A MAJOR FOCUS OF OUR
CARDIOVASCULAR PROTEOMICS CENTER
WHICH WAS JUST RECENTLY FUNDED.
AND WE ARE JUST IN THE FIRST FEW
MONTHS OF THAT PROGRAM.
SO, TODAY NO WE DON'T.
DO WE HAVE HOPE WE WILL GET
THERE?
YES.
BUT I THINK YOU HAVE TO VIEW
BIOMARKER DISCOVERY IN THE SAME
LIGHT AS ONE VIEWS DRUG
DISCOVERY.
YOU START OFF WITH REASONABLE
HYPOTHESIS AND GOOD TARGETS.
YOU MARCH DOWN THE PATH OF
BUILDING ADDITIONAL INFORMATION
AND SPECIFICITY AND BIOLOGICAL
IMPACT OF THAT PARTICULAR DRUG
AND THEN YOU GO INTO HUMAN
PATIENTS AND LARGE NUMBERS.
AND THAT'S WHERE THINGS CAN FALL
APART.
AND THE TIMELINES FOR THIS CAN
BE QUITE LENGTHY AND I THINK
THAT WE SHOULD PUT BIOMARKERS
DISCOVERY IN THE SAME CONTEXT.
IT WILL TAKE THE SAME SORT OF
TIME LINES AND THE FAILURE RATE
OR THE ATTRITION RATE OF
CANDIDATES IN THIS PROCESS IS
PROBABLY GOING TO BE VERY
SIMILAR.
>> THANK YOU.
>> NICE INTERESTING, PROVOCATIVE
TALK, STEVE.
YOU TALKED ABOUT A LITTLE BIT
ABOUT THE IMPLEMENTATION AND
SUGGESTED THAT IN SOME CASE IT
IS MAY BE GOOD TO GO TO PURELY
ANTIBODY-BASED ASSAYS AT THE
ENDPOINT.
I'M WONDERING ABOUT THE SSKAPA
ANTIBODIES AS YOU DEVELOP
SPECIFICALLY FOR THE MASS
SPECK-BASED ASSAYS.
DO THESE PEPTIDE-DIRECTED
ANTIBODIES WORK FOR WHOLE
PROTEINS AND CAN THEY IN GENERAL
BE USED TO DEVELOP NONMASS
SPECK-BASED ASSAYS?
>> THIS IS A TERRIFIC QUESTION,
MARK.
SO, I WILL POINT OUT, I SHOULD
START OUT BY POINTING OUT THAT
THE CRITERIA FOR PEPTIDE
SELECTION HAS NOTHING TO DO WITH
CALCULATED IMMUNOGENIC
PROPERTIES OF THE PEPTIDE.
SO THAT IS THE STARTING POINT.
WE SORT OF BIASED AGAINST THAT
POSSIBILITY POTENTIALLY.
OUR CRITERIA FOR SELECTION HAVE
EVERYTHING TO DO WITH THE
RESPONSE OF THAT PEPTIDE AND ITS
FRAGMENTATION BEHAVIOR.
WE WANT PEPTIDES THAT GIVE THE
STRONGEST ELECTROSPRAY AS
POSSIBLE.
SO THAT'S THE STARTING CRITERIA
FOR THIS.
IT IS QUITE CONCEIVABLE THAT
SOME NUMBER OF THESE PEPTIDES
WILL BE SURFACE EXPOSED, HAVE
THE RIGHT PROPERTIES AND BE
SUITABLE FOR ANTIBODIES TO THEM
BE SUITABLE FOR CAPTURE OF
EITHER THE NATIVE OR THE
UNFOLDED PROTEIN.
WE DON'T -- THE SMALL NUMBER OF
CASE WHE
IT'S NOT LOOKING VERY GOOD.
BUT MANDY IS IN THE PROCESS OF
DOING THIS IN A LARGER NUMBER
TAKING THOSE ANTIBODIES AND
PROBING A MUCH LARGER NUMBER OF
IN TACT PROTEINS TO SEE WHETHER
WE CAN IP.
BUT THE EARLY DATA SUGGESTS IT
WON'T BE A GENERAL ROUTE.
I SHOULD SAY ALL OF THESE
ANTIBODIES ARE POLYCLONALS AND
FOR REAL ULTIMATE UTILITY, WE'LL
HAVE TO MOVE THESE INTO MONOS
AND THE WAY WE DO THIS IS BY
SAVING THE SPLEEN OCYTES FROM
THEITS AND WE HAVEYQ%Kí@?ç MADE
MONOCLONALS OUT OF A SUBSET IN
THE TARGETS IN THE PROGRAM, A
PANEL OF ABOUT 12 OF THESE, THAT
WILL BE IN LARGE ENOUGH PER
PETTUAL QUANTITES FOR PEOPLE TO
BE ABLE TO USE.
ONE CAN ALWAYS MOVE ALL OF THOSE
INTO MONOCLONALS IF YOU WANTED
TO.
>> STEVE, TO THE PREVIOUS
QUESTION, IN MY LAB, WE DID MANY
YEARS AGO WORK ANALOGOUS TO
THIS.
IT WAS ABOUT THE SAME TIME,
2005.
WE DEVELOPED POLYCLONAL
ANTIBODIES TO TREAT HUMAN --
[INDISCERNIBLE]
AND OUT OF THOSE THREE, TWO
WORKED FOR PROTEINS.
WE USED THEM WHEN WE COMPARED ON
VESSELS.
NOT ONLY WORKED FOR THE SAME
CAPTURING PEPTIDES.
THEY WORKED FOR THE WHOLE
PROTEINS TOO.
BUT NOT ALL OF THEM.
OKAY?
BUT SOME.
AND LATER, WHEN WE WORKED WITH
HUMAN SAMPLES JUST THOSE
PEPTIDES WERE JUST -- HAPPILY
FOUND.
THEY CORRESPONDED TO THE ONES
THAT WERE ABLE TO IDENTIFY ON
WHOLE PROTEINS.
>> THAT IS VERY ENCOURAGING.
I THINK IT IS WORTH THE TIME TO
TEST, WHICH ONES WORK AND WHICH
ONES DON'T.
BECAUSE THE BIFUNG AT REAGENTS
WOULD BE A BIG WIN.
-- BIFUNCTIONAL --
>> YOU COMMENTED ON THE FACT
THAT IT'S DIFFICULT TO USE PLA
MA OR BLOOD FOR DISCOVERY.
COMMENT ON URINE AS A PROXIMAL
FLUID AND PROBLEMS IN USING
URINE IN DISCOVERY AS WELL?
I'M A NEVERROLOGIST SO IT'S A
FLUID DEAR TO MY HEART.
>> I THINK --
>> SECOND FAVORITE FLUID.
>> YOUR CLEG JUST ASK THE
QUESTION, IT'S PROBABLY BETTER
EXPERT ON URINARY PROTEOMICS
THAN I AM.
WE ARE NOT DOING URINARY
PROTEOMICS AS A MOLECULAR
EPIDEMIOLOGY INFECTIOUS DISEASE
PROTEOMICS PROJECT.
URINE IS A GOOD SOURCE FOR
MARKERS AND HAS A LOT OF
COMPLICATIONS RELATIVE TO PLASMA
OR SERUM.
AND ONE OF THOSE IS THE VERY
HIGH DEGREE OF VARIABILITY IN
THE ABSOLUTE ABUNDANCE OF
PROTEINS IN THE FLUID AND IN THE
RELATIVELY OVERALL CONTENT OF
PROTEINS.
THERE IS JUST TECHNICAL ISSUES
FROM A PROCESSINGGT
BUT IT'S A PERFECTLY, IF YOU'RE
LOOKING AT DISEASES OF THE
KIDNEY, OKAY, AND I WOULD THINK
URINE WOULD BE A TERRIFIC
STARTING POINT.
NOW MARK AND I HAD A
CONVERSATION ABOUT EXOSOMES THIS
MORNING, AND I THINK THERE IS A
LOT -- THIS IS PART OF THE
SPECTRUM OF METHODOLOGIES TO
ENRICH FOR PROTEINS OF INTEREST.
SO ANYTHING YOU CAN DO THAT
WOULD COMPARTMENT ALLIES
POTENTIAL MARKERS, WILL ENRICH
FOR SUBSETS OF PROTEINS YOU'RE
INTERESTED IN, BE IT EXOME,
LOOKING AT PROXIMAL FLUIDS,
THOSE ARE THE STRATEGIES I WOULD
DEPLOY.
I THINK URINARY PROTEOME FROM
THE ANALYSIS I HAVE SEEN IS
ESSENTIALLY AS COMPLICATED AS
THE PLASMA PROTEOME JUST LESS
CONCENTRATED.
>> THANK YOU.
>> I HAVE TWO TECHNICAL
QUESTIONS ABOUT MRM ASSAYS WITH
INTERNATIONAL STANDARDS.
SO HOW DO YOU DETERMINE THE
CONCENTRATION OF YOUR INTERNAL
STANDARDS GIVEN HIGH DYNAMIC
RANGE.
CRPMPR WAS LIKE TWO ORDERS
HIGHER IN JAPANESE AND OF COURSE
IT IF YOU SPIKE TOO MUCH IT WILL
BE NOT GOOD -- AND THE SECOND
QUESTION IS, IF YOU VERIFY THE
COMPLETENESS OF YOUR DIGEST
BECAUSE THIS ALSO WILL EFFECT
YOUR --
>> ALL OF THESE ARE VERY GOOD
QUESTIONS.pBxé
SO, WE BEGIN BY DOING KIND OF AN
INITIAL SURVEY OF WHAT
THE -- GET SOME NOTION OF WHAT
THE RANGE IS THAT WE WANT TO BE
MAKING THE MEASUREMENT IN.
IS IT HUNDREDS OF NANOGRAMS PER
MILL OR THE BOTTOM OF THE
NANOGRAM PER MILL RANGE?
WE WILL ADJUST THE SPIKE LEVEL
TO BE WITHIN A FACTOR OF 20-50
OF THE CONCENTRATION WE EXPECT
TO BE ABLE TO MEASURE.
WE DON'T HAVE TO BE ANY CLOSER
THAN THAT.
BUT IT'S NOT GOOD TO BE AT
CONCENTRATIONS, DIFFERENTIALS
LARGER THAN THAT.
IT STARTS NEAR THE PHYSICS OF
THE INSTRUMENTATION START TO
PREVENT YOU FROM MAKING ACCURATE
MEASUREMENT.
SO THAT'S ONE OF THE QUESTIONS.
THE OTHER WAS THE COMPLETENESS
OF THE DIGESTION.
>> BECAUSE THE SPIKED PEPTIDES,
THEY ARE -- THE SPIKED PEPTIDES,
THE INTERNAL STANDARDS, THEY ARE
LIKE COMPLETE DIGESTED PEPTIDES,
RIGHT?
>> YES.
SO I DIDN'T AGAIN HAVE TIME TO
TALK ABOUT THIS.
THERE ARE A NUMBER OF WAYS OF
SKINNING THIS PARTICULAR CAT.
THE PROBLEM, AS YOU'RE ELUDING
TO, IS THE DIGESTION EFFICIENCY,
THE RELEASE EFFICIENCY OF ANY
GIVEN TRYPTIC PEPTIDE FROM A
PROTEIN IS GOING TO VARY WIDELY.
AND IT DOES.
EVEN UNDER OPTIMAL DIGESTION
CONDITIONS, WE SEE FACTOR OF 10
DIFFERENCE BETWEEN OPTIMALLY
RELEASED PEPTIDE AND ONE THAT IS
POORLY CLEAVED.
THAT COULD BE THE RESULT OF
MULTIPLE CLIP SITES BECAUSE
THERE ARE RAGGED OR MULTIPLE
BASIC SITES WHERE IT'S CLEAVING
AT BOTH.
OR IT COULD BE A PROTECTED
REGION OF THE PROTEIN AND IT'S
VERY HARD FOR THE TRIPS TOINE
GET AT IT.
SO THIS IS EVEN UNDER
CIRCUMSTANCES WHERE WE ARE
HIGHLY DENATURING WITH UREA OR
WHATEVER YOU HAPPEN TO BE USING.
SO I KEPT SAYING PRECISE
RELATIVE QUANTIFICATION.
AND WHAT MATTERS HERE IS THE
REPRODUCIBILITY OF THE DIGEST.
IT DOESN'T MATTER SO MUCH
WHETHER YOU HAVE 10% RELEASE OF
A PEPTIDE OR 100% RELEASE OF THE
PEPTIDE IF THE DIGEST IS
REPRODUCIBLE.
AND IF THE CVs FOR THE
BIOLOGICAL PROCESS DIGESTION
REPLICATES SHOWS THAT THE DIGEST
ARE REASONABLY REPRODUCIBLE,
BECAUSE THE OVERALL CV FOR THE
ENTIRE PROCESS INCLUDING
DIGESTION, IS LESS THAN 25% IN
THE WORST CASE.
NOW, CAN WE MAKE THAT BETTER?
YES.
CAN WE GET TO ACCURATE
QUANTIFICATION?
WHICH I AVOIDED SAYING IN THIS
TALK FOR VERY SPECIFIC REASONS.
BECAUSE WE DON'T ACCURATELY
MEASURE THE LEVELS.
AND THE ANSWER TO THAT IS, YES,
YOU CAN.
SO ONE OF THE WAYS YOU CAN DO
THIS AND WE ARE DOING IT IN THE
CONTEXT OF THIS STUDY IS TO USE
LABELED PROTEINS AS THE STANDARD
SO THEY ARE BEING -- THIS IS
WITH OUR COLLABORATORS SPIKING
IN FULLY N15 LABELED PROTEIN.
THOSE ARE GETTING DISECTIONS
SIMULTANEOUSLY TO THE ANNA LATE
PROTEIN OF INTEREST AND
THEREFORE YOU EXPECT THE
ACCURACY TO BE GREATER AND THE
DATA SAYS THAT IS ABSOLUTELY
TRUE.
AND YOU DON'T WANT TO HAVE TO
NECESSARILY MAKE SYNTHETICAINA
LOGS OF EVERY PROTEIN OF
INTEREST -- ANALOGUES.
THERE ARE OTHER WAYS OF DOING
THIS.
SANDY MARKEY IN THE AUDIENCE IS
MAKING CAT MERES OF PEPTIDES
FROM THESE PROTEINS WITH REGIONS
BOUNDED BY THE NATURAL SEQUENCE
IN THE PROTEIN.
I THINK THIS IS A VERY GOOD
IDEA.
WE ARE MAKING PEPTIDES THAT HAVE
WINGS.
IT'S THE SAME NOTION.
SO WE EXTEND THE PEPTIDES ON
BOTH ENDS WITH ANYWHERE FROM 3-6
AMINO ACIDS WITH A NATURAL
SEQUENCE PRESENT.
SO WHEN THESE THINGS DIGEST, THE
NOTION HERE IS THAT HOW THE
PEPTIDE DIGEST WITH WINGS IS OR
HOW A PROTEIN DIGEST, ONCE IT'S
DENATURED IS DRIVEN PRIMARILY BY
THE PRIMARY SEQUENCE CONTEXT AND
BECAUSE YOU HAVE ELIMINATED
SECONDARY AND TERTIARY
STRUCTURE.
THIS IS NOT ENTIRELY TRUE BUT IT
DOES WORK VERY WELL AND IT GETS
YOU MUCH CLOSER TO AN ACCURATE
MEASUREMENT.
AND I WOULD SAY FROM A CLINICAL
STANDPOINT, IF THESE ASSAYS GET
DEPLOYED, THAT WOULD BE ONE OF
THE WAYS OF APPROACHING
DEVELOPING CALIBRATION CURVES IN
CLINICAL LABS WE WOULD BE USING
THIS INTERNAL STANDARDS.
>> THANK YOU.
>> VERY INTERESTING DISCUSSION
AND WE CAN CONTINUE THAT IN THE
LIBRARY OVER COFFEE AND COOKIES.
BUT LET'S THANK OUR SPEAKER ONE
MORE TIME.
THANK YOU, STEVE.
[APPLAUSE]