Circulation August 7, 2018 Issue

Dr Carolyn
Lam:               
Welcome to Circulation on the Run, your weekly podcast summary
and backstage pass to the journal and its editors. I'm Dr Carolyn
Lam, associate editor from the National Heart Center and Duke
National University of Singapore.


                                               
Can proteomic biomarkers distinguish between subtypes of aortic
stenosis even years before surgery? Well, to find out more, stay
tuned. That's coming right up after these summaries.


                                               
The first original paper this week adds to the evidence that
smoke-free policies are associated with a lower risk of
cardiovascular disease. First and corresponding author, Dr Mayne
from Northwestern University Feinberg School of Medicine, and her
colleagues linked smoke-free policies to participants of the
Coronary Artery Risk Development in Young Adults, or CARDIA
study, which has a follow-up of up to 20 years. They found that
smoke-free policies in workplaces were associated with
significantly lower risk of incident cardiovascular disease after
controlling for a wide range of covariants. Results were weaker
for bar and restaurant bans, but in the same direction.


                                               
Preventive fractions range from an impressive 24 to 46%. Thus,
smoke-free policies may improve cardiovascular health through
reducing population exposure to tobacco smoke. However, we should
remember that much of the US population remains unprotected by
smoke-free policies. Thus, taken together with prior ecological
work, these results support the continued expansion of smoke-free
policies in indoor public places.


                                               
Most phase-3 randomized control trials feature time-to-first
event end points for their primary analysis. In chronic diseases,
however, a clinical event can occur more than once and recurrent
event methods have been proposed to more fully capture the
disease burden, as well as to improve statistical precision and
power.


                                               
However, is this really the case? This question was examined by
first author, Dr Brian Claggett, corresponding author, Dr Scott
Solomon, from Brigham Women's Hospital in Boston, Massachusetts,
and their colleagues, who sought to better characterize factors
that influence the statistical properties of recurrent events and
time-to-first event methods in the evaluation of randomized
therapy.


                                               
They performed repeated simulated trials with 1:1 randomization
of 4000 patients to active versus control therapy. Through
simulation, they varied the degree of between-patient
heterogeneity of risk as well as the extent of treatment
discontinuation. They then compared their findings with those
from the actual randomized control trials.


                                               
The authors found that the statistical power of both recurrent
events and time-t- first event methods were reduced by increasing
heterogeneity of patient risk, a parameter that's not usually
included in conventional power and sample size formulae.
Furthermore, data from real clinical trials were consistent with
the simulation studies, confirming that the greatest statistical
gains from the use of recurrent events methods occurred in the
presence of high patient heterogeneity and low rates of study
drug discontinuation.


                                               
The next paper uncovers a novel biomarker and therapeutic target
of pulmonary arterial hypertension, and that is selenoprotein P.
First author Dr Kikuchi, corresponding author, Dr Shimokawa, from
Takaoka University Graduate School of Medicine in Japan and their
colleagues performed micro-array analyses using pulmonary
arterial hypertension, pulmonary artery smooth muscle cells, and
found a 32-fold up regulation of selenoprotein P compared with
controls.


                                               
Selenoprotein P promotes cell proliferation and apoptosis through
increased oxidative stress and mitochondrial dysfunction. Using
five strains of genetically modified mice, the authors
demonstrated a pathogenic role of selenoprotein P in the
development of hypoxia-induced pulmonary hypertension.


                                               
Furthermore, sanguinarine, which is an orally active small
molecule identified by throughput screening reduced selenoprotein
P expression and pulmonary arterial smooth muscle cell
proliferation and ameliorated pulmonary hypertension.


                                               
In summary, this study shows that selenoprotein P plays a crucial
role in the pathogenesis of pulmonary arterial hypertension and
may be a useful and novel biomarker and therapeutic target in
this disorder.


                                               
Familial hypercholesterolemia is known to be associated with a
high risk of ischemic heart disease, including myocardial
infraction, but what about the risk of ischemic stroke? Well,
first author, Dr Beheshti, corresponding author, Dr Nordestgaard,
from Copenhagen University Hospital and their colleagues examined
the associations of familial hypercholesterolemia and high LDL
cholesterol with ischemic stroke in both causal, genetic, and
observational analyses using more than 106000 individuals from
the Copenhagen General Population Study, and more than 10000
individuals from the Copenhagen City Heart Study.


                                               
They used a Mendelian randomization design to test whether high
LDL per se had a causal effect on ischemic stroke risk using a
combination of the familial hypercholesterolemia causative
mutations and common genetic variants associated with high LDL.


                                               
The authors found that there was no association between familial
hypercholesterolemia mutations and ischemic stroke risk. In the
Mendelian randomization analysis, also including common genetic
variants, there was also no causal effect of high LDL on the risk
of ischemic stroke.


                                               
These findings imply that the predominant goal of targeting LDL
lowering in those with and without familiar hypercholesterolemia
is likely to reduce myocardial infractions, rather than ischemic
stroke. Well, that wraps it up for our summaries. Now for our
feature discussion.


                                               
Circulation publishes numerous papers regarding circulating
biomarkers. We talk about biomarkers in the diagnostic,
prognostic sense, but what about in a pathophysiologic sense, and
especially in a disease as important as aortic stenosis? Well,
that's what our featured paper this week is all about and I'm so
excited to have with us corresponding author, Dr Stefan
Söderberg, from Umeå University in Sweden, as well as our
associate editor, Dr Peipei Ping from UCLA. We will be discussing
the paper entitled “Proteomic Biomarkers for Incident Aortic
Stenosis Requiring Valvular Replacement.” Stefan, could you tell
us a bit about what made you look at this very interesting
question, and perhaps the unique resources you had in Sweden to
look at this?


Dr Stefan Söderberg:      I'm a
practicing cardiologist, and I have been working a lot with
aortic stenosis over the years. It's frustrating that we can't do
anything to stop the process. In many cases, the patients are old
and frail, and if you could find the means to stop the process
long before they need surgery, it will be of great benefit for
the human and for the society.


                                               
Also, knowing that the interventions on the statins, for example,
have been unsuccessful, we thought that there must be better ways
or other biomarkers. Furthermore, that many of these studies, the
phenotype of aortic stenosis has been very poorly described and
there is probably much more behind just aortic stenosis than
just, for example, calcium deposits in an X-ray, et cetera, et
cetera.


Dr Carolyn
Lam:               
You used some very unique resources in Sweden to therefore look
at the proteomic signatures of aortic stenosis. Could you
describe that and simplify perhaps the results so we can
understand it?


Dr Stefan Söderberg:      First of all,
I got this idea from other studies done up in northern Sweden. If
you have an absolutely unique setting, the combination of huge
population-based studies in 30 years back, we have a huge biobank
with examples of extraordinary good quality from each of these
participants. For example, for each participant, the blood has
been spun and put into freezer, deep freezer, within one hour for
30 years, and they are now, as I said, about 100000.


                                               
Furthermore, I'm working as a cardiologist at a university, and
up here, you do all of the aortic surgery for the whole northern
Sweden. That is, we can combine the names of the person
undergoing surgery together with these population-based surveys
and we can get details from all those who have participated in
the surveys long before they did the surgery, and they can go and
retrieve samples from cases and match controls from the freezers.
It's a unique setup. Then, when we were designing the study, we
got the chance to get these analyses done by our friends at the
university to get the proteomic analysis via a unique data
technique.


Dr Carolyn
Lam:               
Wow. Could you describe your results?


Dr Stefan Söderberg:      The results
that we found in the first set of 334 patients who underwent
surgery is 10 years after their first sampling, we found six
proteins. Then, we got the question back from Circulation to
establish a validation cohort, and we were able to do so to
include all those new cases in the last 2 years, and there we
could replicate five of these proteins.


                                               
The interesting thing that the pattern is completely different
between those having coronary artery disease from those without.
That kind of phenotyping has not been done throughout other
aortic stenosis studies. Therefore, this study is unique. We have
had two papers in the last year in the Journal of American Heart
Association from the cohort, as well, showing the thing that
happened.


                                               
For example, lipoprotein little A is only associated with future
aortic stenosis valve replacement only in those with concomitant
coronary artery disease. There are many unique things, the
prospective design, and the phenotype differentiating those with
and without coronary artery disease.


Dr Carolyn
Lam:               
Yeah, and if I may just reiterate that the population base that
you work with is just enviable and just so that the audience
realizes, these are biomarkers that were collected 11 years
before the aortic stenosis surgery, isn't it? You really had a
long follow-up.


                                               
Also, just to let everyone know, it was a proximity extension
assay that you used for the discovery, and the six proteins were
growth differentiating factor 15, or GDF15, galectin-4, von
Willebrand factor, interleukin 17 receptor A, transferrin
receptor protein 1, and PCSK9, so very interesting. Peipei, you
have a way of putting things into context so beautifully. Could
you tell us your thoughts when you saw this paper?


Dr Peipei
Ping:                  
I thought this is a very high-quality study that actually
benefited from the long-term established, well-controlled cohort
in northern Sweden, as Dr Söderberg just shared with us. On the
other end, it married a technology platform that's very
well-established and -validated, and this situation targeted
proteomics platforms using multi-proximity extension assays with
carefully controlled markers and screened 92 cardiovascular
candidate markers.


                                               
This is the kind of approach that provides semi-quantitative as
well as quantitative outputs and is capable to offer validated
screens on large population clinical subsets. A study of such
with a high value cohort combined with a validated and
well-controlled technology platform offered results that clearly
have clinical significance, as well as setting up examples for
other studies to follow. The enthusiasm from the editorial
boards, as well as the reviewers, have been substantially high
and supportive.


Dr Stefan Söderberg:      Fantastic. I'm
very glad to hear this.


Dr Carolyn
Lam:               
Stefan, you also mentioned that a very unique element was the
separation of aortic stenosis with and within coronary artery
disease, or at least established or visible coronary artery
disease. Could you explain how that provided pathophysiologic
insights?


Dr Stefan Söderberg:      First, I
should say we were very, very strict. Our routine is that
everyone was 100% undergoing, aortic valve replacement, they
undergo a coronary angiogram before. If we saw any sign of
atheromatosis, it was not enough that they had the significant
stenosis, but any signs, they were put into the group of coronary
artery disease. Those without, we couldn't see anything there.
Radiograph here reported absolutely clean coronary arteries. Of
course, we cannot exclude if there were aortic changes within
them all, of course.


                                               
We believe that this is a very important message that in order to
further study aortic stenosis, we should be very careful in
phenotyping the disease. We hope the growing cohort will be able
to do this further. For example, cuspid versus tricuspid valves,
women versus men, et cetera.


                                               
My answer in short is phenotype. Let me take one example which I
found very, very exciting. That is the finding of PCSK9, which is
closely related not only to cholesterol symptoms, but also to
lipoprotein little A emphasis. As you know, the first strong
finding in aortic stenosis was the LP little A. This is related
to that genetic finding, and that was in the huge study from
Canada. They did not have the same phenotyping, so we had
information to his important findings. That's one example.


                                               
Another example is the transferring receptor, where data has
shown that bleeding acutely in the valvular tissue causes damage,
and this relates iron metabolism to the formation of the aortic
valve. Obviously, it seems that the process in the aortic valve
is very much similar to the vessel arteriosclerosis. It seems to
be different. This is the indication that when we formulate new
studies or new drugs on aortic stenosis, we must be very careful
to use the right drug for the right type of valvular disease.


Dr Carolyn
Lam:               
Those are great points. Peipei, do you think that's the main
clinical take-home message, beyond that great comment you made
earlier that this paper's just a great example of the use of
tools, modern tools, that we have in proteomic characterization
like the proximity extension assay to provide pathophysiologic
insights when you have a really well-phenotyped cohort? What's
the critical take-home message, though? Is there one now?


Dr Peipei
Ping:                  
The take-home message is marriage of amazing high value cohort's
data sets with that of the well-controlled clinical study using
target proteomics approaches. In this particular study, one main
critical innovation is the study is capable of providing insights
regarding molecular signatures that have predicted values. As
stated in the manuscript, the circulating proteins that found
critically important, their alterations took place years before
the surgery were associated with aortic stenosis. That is of
value, clinical value, to many other clinical studies to follow.


Dr Carolyn
Lam:               
Wow. That's wonderful. Thank you so much for putting these
findings in context for us. Thank you, listeners, for joining us
today. Don't forget to tune in again next week.