Let’s define what Lp(a) is: 2 major proteins stuck together like conjoined twins, apolipoprotein(a) and apolipoprotein B-100. The ratio of each is 1:1- if you know how much apoB is on Lp(a) then you know how much apo(a) is on it
2/17 ApoB-100 is a lipoprotein, it has a specific folded structure that can allow it to be loaded with lipid molecules, which don’t dissolve in water, so that these can be sent throughout the body for various needs- precursor to hormones, adrenal metabolites, cell membranes, etc.
3/17 Think of apoB-100 as pack mules that can load and offload cargo- VLDL, LDL, IDL, small dense LDL, OxLDL and OxPL
Apo(a) is not lipoprotein, so when its on apoB-100 it sticks out in the liquid phase of blood, think of it like a pool sweeper
4/17 Now how did it happen that apo(a) gets physically stuck on apoB and nothing else? Lp(a) was discovered by Berg in 1963, but apo(a) many years later. In 1987, HUGE surprise, apo(a) was highly homologous to plasminogen (75-99%). What does this mean?. Its 75-99% identical!
5/17 Plasminogen is present in all mammals, evolved about 60 million years ago, and is a pro-enzyme that is activated for function. When a subject has a stroke, tPA is given to activate plasminogen to plasmin and this is how it works, the tPA itself has no effect on blood clots.
6/17 Plasminogen is not present on apoB-100. Plasminogen is made of 5 kringles, proteins that are stuck together at 3 sites and they look like Danish pastry. The apo(a) gene copied itself from plasminogen and is oriented in the opposite direction on chromosome 6.
7/17 Apo(a) changed during evolution, it only copied kringle 4 and 5 and lost its ability to be a protease. It also made 10 different copies of Kringle 4, with multiple identical copies of K4-2. Apo(a) acquired a free cysteine and this allows it to bind to apoB-100 to make Lp(a)
8/17 Interestingly, apo(a) looks like it evolved twice. Once in the hedgehog ~35 million years ago and then ~12 milion years ago in African apes and monkeys and humans. Hedgehogs have 37 copies of K3 only and no others. Apes and monkeys have different versions of apo(a)
9/17 The uniqueness of human apo(a) is it’s the only version that has OxPL, @OxPL_apoB. Perhaps this is a link to its link with CVD.
10/17 What does all this mean for humans:
1-There are over 40 different flavors of apo(a), and thus Lp(a), depending on how many kringles are inherited
2-Each parent contributes 1 kringle to progeny, most people have 2 different sized kringles
11/17
3-People with small number of kringles have higher number of Lp(a) particles, as the liver can make more per unit time
4-The number of apo(a) molecules is much lower than plasminogen, so its not clear if it inhibits fibrinolysis in vivo as the math does not add up
12/17
5-The kringle size difference accounts for 25-50% of Lp(a) plasma levels
6-The kringles themselves don’t seem to cause disease directly, but only by affecting plasma levels. Most of the “action” of apo(a) seems to be in K4-9, K4-10 and protease-like domain
13/17
7-To date, measuring kringles in patients has not shown definitively to add more value than knowing the plasma Lp(a) levels.
8-Early works suggests @OxPL_apoB present on KIV9, @MarlysLPA and @MBBoffa are deciphering which amino acids bind it
14/17
9-Measuring kringle number should be considered a research test and not a clinical test right now
10- Danish pastry goes well with coffee but has lots of carbs
15/17
Quiz- 5 points. Which is not correct?
1-Apo(a) is highly related to plasminogen
2-Apo(a) carries OxPL and cholesterol
3-The number of kringles on apo(a) is inversely related to the plasma Lp(a) levels
4-K4-type 2 mainly mediate plasma levels of Lp(a)
16/17
Bonus question: 1 point Which donkey represents VLDL, LDL and OxPL? Quiz next page
17/17; Choose correct answer from image in prior page. Which donkey represents VLDL, LDL and OxPL?
• • •
Missing some Tweet in this thread? You can try to
force a refresh
When one looks at your lab report, note the LDL-C says “calculated”. So, believe it or not, the most important risk factor for CVD is actually an estimated value. Hard to believe but true. 1/17
Sometimes it says direct LDL, which is an actual measurement. LDL is a complex particle containing apoB, cholesterol, cholesteryl esters, phospholipids and a bit of triglyceride. LDL-C only reflects the total cholesterol in LDL, that is why there is a “C” at end of it. 2/17
LDL particles are made of about 50% cholesterol by mass (i.e. the human body is 60% water). Lp(a) on other hand is 20-45% cholesterol. The gold standard for measuring LDL-C is ultracentrifugation. 3/17
I suspect many of my non-physician followers will be confused, so this needs a response. I think we agree that risk rises with higher Lp(a). I do agree if ones Lp(a) is 50 risk is low. I don't agree that saying a risk of >50 mg/dL is either trivial or a slogan.
I think a 20% higher risk (going from a 10% risk of dying, having a heart attack or stroke) to 8% (over 3 years only like in the trial, not a lifetime of 40-80 years!!) is significant to those whose own Lp(a) is elevated, needed one or more procedures or lost a loved one.
This 20% risk in 3 years in a trial will add up substantially as the years go on and curves will continue to separate relative to low Lp(a) patients. We have to look at lifetime risk, not what risk is in a short term trial. The 10 vs 8% could end up being 50% vs 30% in 20 years.
Its not 50, but (>)50, its like saying LDL >130 is associated with higher risk. Of course, the higher the value the higher the risk. The Novartis trial will use >70 mg/dL, which represents 15% of population at risk, ~1B. Point is no matter what metric you use, it a huge number.
Note also, the >50 mg/dL is in patients on statins, ncbi.nlm.nih.gov/pubmed/30293769. In subjects not on statins risk starts at >30 mg/dL (>75 nmol/L) Old papers show this and see most recent Madsen et al in subjects without prior event: ahajournals.org/doi/abs/10.116…
Madsen data: MACE incidence rates per 1000 person-years were 29 for individuals with Lp(a)<10 mg/dL, 35 for 10 to 49 mg/dL, 42 for 50 to 99 mg/dL, and 54 for ≥100 mg/dL. I think above 10-30 mg/dL one can consider risk fairly linear, like LDL-C.
Interesting thread of comments, I would like to make some clarifications, as I see a lot of close but inaccurate numbers being thrown out on the internet. Now that Lp(a) is getting traction clinically, there is a lot of quoting data that is not entirely correct.
1- PCSK9 inhibitors lower Lp(a) 20-30%, but in patients with high Lp(a) (>50 mg/dL), the effect is only 14%. This is the number that should be quoted in this context. See specific paper on this. ncbi.nlm.nih.gov/pubmed/30561610
2- ASOs lower Lp(a) >99% depending on the dose. In the phase 3 trial about to start the dose is 80 mg monthly and the mean Lp(a) reduction is expected to be 80% based on phase 2 data. That means some will get 70% reduction and some 90% reduction. No need to get everyone to zero
Sorry for my tardiness, but here is a brief summary of AHA Lp(a) highlights: 1- elevated Lp(a) is associated with reduced overall longevity 2- In Odyssey outcomes, elevated Lp(a) was associated with both peripheral arterial disease (PAD) and deep venous thrombosis (mainly PAD)
This effect seems to be modified by alirocumab and incidence was reduced in this group vs placebo 3- Pts with MD on statins have higher residual risk if Lp(a) is elevated- no surprise 4- ApoE2 genotype, which has weak affinity for LDL receptor is associated with lower Lp(a)
This was known ncbi.nlm.nih.gov/pubmed/28062489, but now linked to lower risk in subjects with FH
5-Elevated Lp(a) is associated with more aortic valve calcium in subjects with FH 6- OxPL, @OxPL_apoB are elevated in patients with elevated Lp(a) and there are differences in racial groups
A few more details on the FOURIER Lp(a) paper ncbi.nlm.nih.gov/pubmed/30586750:
1-33% of pts had Lp(a) >120 nmol/L (>50 mg/gL) 2- The median % Lp(a) eduction was 27%, but only 16% in 4th highest quartile, thus the higher the Lp(a) the less effective PCSK9i is in lowering it,
3-each doubling of Lp(a) was associated with 8% higher risk for CVD 4- The pts benefited irrespective of LDL-C levels. A along with our recent Lancet paper, it puts to end the 20-yr old faulty hypothesis that one does not need to worry about Lp(a) when LDL-C is controlled
5- the pts that benefited the most were those with highest Lp(a), despite getting a less effective Lp(a) lowering 6- lowering Lp(a) by 37 nmol/L (~15 mg/dL) is predicted to lead to a 20% RRR, similar to ODYSSEY ACC presentation,