On the 20th of May 2022, the Lipid Energy Model (LEM) was published in an attempt to explain the drastic elevation seen in LDL-c amongst lean individuals who are on a high fat diet with low carbohydrates.
Keto Cookbook Bundle - Check it out
The Keto Box - snacks to take on the go
Sooo Ketolicious - Who said Keto wasn't delicious
The purpose of this rebuttal is to open up discussion on some aspects of the hypothesis so that understanding advances in order to validate or improve the model to reflect an accurate description of why these individuals see their LDL-c rise so drastically.
Certain effects are described which I will address with information that will highlight aspects opposing the claims made.
But first a quote from the publication summarizing the model.
- Reduction in dietary carbohydrates and depletion of hepatic glycogen stores results in a greater demand for fat as a metabolic fuel, to compensate for reduced glucose availability.
- Decreased insulin, leptin, and other changes to the hormonal milieu, result in increased hormone-sensitive lipase (HSL)-mediated lipolysis in adipocytes and greater secretion of non-esterified fatty acids (NEFAs) into the bloodstream.
- In addition to heightened use by tissues in the periphery, there is a greater rate of uptake of NEFAs by the liver. Under these conditions, there is a greater rate of synthesis of TGs from the increased fatty acid pool within hepatocytes.
- Increased rates of TG synthesis in the liver leads to increased rates of hepatic assembly and secretion of TG-rich VLDL.
- The increased VLDL secretion rates, in concert with greater LPL-mediated turnover of VLDL in peripheral tissues, and greater transfer of VLDL surface components (including free cholesterol) to HDL, result in higher plasma levels of LDL-C and HDL-C.
The highlighted parts will be the focus of this rebuttal.
Increased synthesis of triglycerides resulting in higher hepatic secretion rate of VLDL
The model describes a higher release of non-esterified fatty acids (NEFA) from adiposity. This is true as research shows us a higher circulation of those fatty acids thus it is possible that the liver receives more of these NEFA.
An argument in favor of the LEM is that indeed a higher perfusion of the liver with NEFA can lead to a higher re-esterification meaning the synthesis of triglycerides based upon these NEFA (https://www.pnas.org/doi/10.1073/pnas.1423952112).
A second argument is that insulin prevents the secretion of the ApoB lipoprotein (https://www.sciencedirect.com/science/article/abs/pii/S0006291X11002117?via%3Dihub) thus under conditions of low insulin such as a high fat low carb diet, you could reason that secretion is actually increased.
However, such studies focus on the role of insulin. Glucagon, when insulin is low, rises significantly. In a paper from Volek et all. comparing ultra-endurance athletes on high carb versus a ketogenic diet (https://faseb.onlinelibrary.wiley.com/doi/abs/10.1096/fasebj.31.1_supplement.1036.3) at baseline we see a glucagon level which is >50% higher.
Glucagon is important because the liver increases beta-oxidation in response to glucagon in favor of re-esterification (https://onlinelibrary.wiley.com/doi/10.1111/jdi.13315 ; https://www.frontiersin.org/articles/10.3389/fphys.2019.00413/full). The latter study also alludes to a possible glucagon-secretion stimulation by fatty acids but results are mixed and need further study.
By inhibiting the formation of malonyl-CoA, glucagon diverts FFAs to beta-oxidation rather than re-esterification into TGs (Figure 2).
In vitro results show that glucagon diverts re-esterification of fatty acids away in favor of ketone body production (https://aasldpubs.onlinelibrary.wiley.com/doi/abs/10.1002/hep.1840130620)
The stimulation of AMPK by glucagon is also noted elsewhere (https://diabetesjournals.org/diabetes/article/69/Supplement_1/220-LB/56113/220-LB-Glucagon-Promotes-Hepatic-Autophagy-by-AMPK). In the chain of events, AMPK will inhibit HMG-CoA reductase which results in a reduction in cholesterol synthesis (https://physoc.onlinelibrary.wiley.com/doi/10.1113/jphysiol.2006.108506). More information on HMG-CoA reductase and cholesterol follows further down. AMPK, via GPAT, also reduces the re-esterification of fatty acids.
A ketogenic diet has the potential to increase hepatic AMPK (https://www.sciencedirect.com/science/article/abs/pii/S0955286321000401 ; https://journals.physiology.org/doi/full/10.1152/ajpendo.00717.2006). When AMPK is elevated, it will reduce hepatic triglyceride content and stimulate fatty acid oxidation as we’ve just seen before (https://www.mdpi.com/1422-0067/19/9/2826/htm)
2. Higher secretion rate
A simple argument against a higher secretion rate of triglyceride-rich VLDLs is that this should be reflected in the blood. Yet part of the characteristics is that these individuals have low triglycerides.
When a blood sample is taken, it reflects what is available at that point in the artery. If low levels are measured then that artery can only deliver a low quantity of triglycerides. Sampling blood from different arteries in the blood should clarify if there is perhaps an unequal distribution to the arm or perhaps a greater or faster lipoprotein lipase activity takes place. This does go against the purpose explained in the model. Energy delivery must be possible to all parts of the body thus it should be detected in higher VLDL particle availability in blood samples if this is part of the increased distribution network.
3. VLDL assembly
As stated earlier, low insulin is a condition for increased ApoB secretion. For the liver to create and secrete a VLDL particle, it is not only depending on low insulin but also triglyceride availability and other components.
Triglycerides, if not available in sufficient quantity will result in degraded ApoB (https://www.jlr.org/article/S0022-2275(20)30605-2/fulltext). On the opposite side, high continuous availability of fatty acids may also result in reduced ApoB production although it initially stimulates higher secretion. This raises the question if the condition in our subjects falls under the chronic high availability. Yet we also have glucagon interfering with triglyceride formation.
Glycerol is required to attach 3 fatty acids to form triglyceride. The following paper (https://www.sciencedirect.com/science/article/pii/S0021915013001767) hints at glycerol availability as a factor that determines VLDL secretion. We need to be careful with interpretation as this one is done under lactate infusion. It should be kept in mind that with increased glucagon, glycerol is used as a substrate for gluconeogenesis and thus may be available in limited amounts.
In 60h-fasted subjects, they concluded that most glycerol is not taken up by the liver (https://journals.physiology.org/doi/abs/10.1152/ajpendo.1996.271.6.E1110)
Thus, in 60 h-fasted individuals, most glycerol uptake does not occur in liver, and the extent of fatty acid reesterification in liver is in doubt.
Cholesterol synthesis in the liver is in part dependent on the liver X receptor (LXR) (https://www.sciencedirect.com/science/article/pii/S0021925819362167). LXR is stimulated by insulin so we can expect that in our subjects with low insulin, there will also be a low hepatic synthesis of cholesterol.
furthermore, cholesterol synthesis depends on the HMG-Coa reductase enzyme to send HMG-Coa down that pathway. Insulin increases the mRNA and protein production while glucagon and fasting lower both. (https://linkinghub.elsevier.com/retrieve/pii/S0021-9258(19)62026-0).
4. Counter observations
When looking at VLDL secretion rates via labeled omega-3 fatty acids in pill form in obese subjects, they noted a reduction (https://academic.oup.com/ajcn/article/77/2/300/4689666). According to the LEM one would expect a higher secretion given that they were having an additional dietary fatty acid supply. Of course the effect could be different as these subjects were obese and not on a high fat low carb diet which would result in different hormonal signaling. Yet it would be interesting to have an explanation for the observation as more fatty acids reached the liver.
Under low dietary carbohydrate conditions, T3 levels tend to drop with an increase in reverse T3. Even when there is no reduction in caloric value (https://pubmed.ncbi.nlm.nih.gov/6761185/).
T3 is a regulator of metabolism, it also helps in the transcription and stabilization of the HMG-Coa reductase mRNA (https://pubmed.ncbi.nlm.nih.gov/10782041/)
To further emphasize on the importance of free T3, VLDL-TG production has been investigated and found correlated independently with fasting insulin levels and resting energy expenditure (https://www.jlr.org/article/S0022-2275(20)43420-0/fulltext) and unrelated to free fatty acid availability or body fat distribution.
6. LDL clearance
As a consequence of higher VLDL secretion rates, combined with reduction in size through lipase activity, this could indeed lead to an increase in LDL particles. However, one cannot expect a higher level of secretion without a concomitant increase in clearance or LDL particle levels would continue to rise indefinitely. The LEM does not make any comments on such balancing so it is by definition incomplete.
Because the state of elevated LDL-c in our individuals is associated with insulin, we must note that the hepatic LDL receptor is reduced under insulin stimulation via stimulation of PCSK9 (https://www.atherosclerosis-journal.com/article/S0021-9150(08)00556-X/fulltext).
This is in line with evidence that point to glucagon for degrading PCSK9 which allows LDL receptor expression (https://www.ahajournals.org/doi/10.1161/CIRCRESAHA.118.313648).
However, those are not the only determinants of LDL receptor expression. Under fasting conditions, a reduction in LDL receptor is observed, at least in part due to lower free T3, which is abolished when free T3 levels are maintained at levels equal to non-fasting (https://www.sciencedirect.com/science/article/abs/pii/S1357272597001209). The authors conclude that the increase in LDL-c results from the reduction of LDL receptors.
One other modulator of LDL receptor expression is the type of fat. The more saturated the fatty acids are, the more LDL receptor is reduced and vice versa (https://www.jlr.org/article/S0022-2275(20)37254-0/pdf).
Other studies have shown that there is a shift in the low density lipoprotein particles towards larger version of it called pattern A (https://academic.oup.com/jn/article/135/6/1339/4663837). As these LDL particles grow in size, they reduce their affinity to bind to the LDL receptor with a reduction in clearance as a result (https://www.atherosclerosis-journal.com/article/S0021-9150(01)00565-2/fulltext ; https://pubs.acs.org/doi/10.1021/bi048825z).
In order for high amounts of NEFA to reach the liver, we have to look at how NEFA reach the liver. The protein albumin has a large carrying capacity for NEFA up to the point that it is responsible for carrying most of it and depending on its concentration even reduces LDL capacity for uptake (https://pdf.sciencedirectassets.com/778418/1-s2.0-S0022227520X64315/1-s2.0-S0022227520394736/main.pdf).
Albumin is the largest carrier of NEFA in the blood stream covering 99%. It needs to be reviewed if this is still valid for our subjects on a high fat low carb diet (https://www.mdpi.com/1422-0067/22/16/8411/htm).
The length and saturation level of fatty acids is important in the distribution of the carriers albumin and lipoprotein. Long chain saturated fatty acids are competed for but other fatty acids are predominantly bound to albumin (https://pdf.sciencedirectassets.com/778418/1-s2.0-S0022227520X62717/1-s2.0-S0022227520367237/main.pdf ; https://www.sciencedirect.com/science/article/abs/pii/0005276065901293)
Greater transfer of VLDL surface components (including free cholesterol) to HDL
This is important because in order for VLDL to turn into LDL, it needs to reduce its lipid content and therefor also structural components such as cholesterol to reshape the particle. This restructuring takes place in the plasma.
Such restructuring is carried out by CETP leading to the exchange where HDL donates cholesterol to ApoB containing particles (VLDL, IDL, LDL) and receives fatty acids. Under low insulin conditions, CETP activity reduces.
As alluded to earlier, on a ketogenic diet there is a shift in LDL particle size from pattern B towards pattern A.
Austin et al in a classic paper70 found that pattern B was associated with a two-fold increase in plasma triglyceride, higher plasma apoB and IDL levels and reduced HDL cholesterol and apoA-I concentration, ie, small, dense LDL did not appear in isolation from other plasma lipid abnormalities.
When CETP activity is high, it leads to smaller LDL particles, is associated with metabolic syndrome, reduces HDL-c, increases small dense LDL (https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1365-2362.1994.tb00987.x ; https://cardiab.biomedcentral.com/articles/10.1186/s12933-016-0428-z ; https://www.sciencedirect.com/science/article/pii/S0022227520328005)
We have already seen the involvement of LXR in cholesterol synthesis but it is also responsible for CETP regulation (https://www.atherosclerosis-journal.com/article/S0021-9150(10)00316-3/fulltext). AMPK directly inhibits LXR (https://www.ijbs.com/v07p0645.htm) while it requires insulin to be stimulated.
Under low CETP activity, HDL has greater cholesterol accumulation while LDL remains loaded with its triglyceride content. This is a favorable situation as this allows HDL to become large enough to incorporate ApoE. That in turn will allow the particle to increase LCAT activity which drives the export of cholesterol from cells (such as macrophages).
It is clear that a low fasting insulin and high glucagon allows for a higher VLDL secretion rate and there are studies to show this but those studies are done on subjects of which we can expect to have some hepatic fat accumulated before the trial.
NAFLD is cleared in a couple of months abstaining from carbohydrates showing a high clearance rate. Our subjects who are marked with very high levels of LDL-c however have clear livers so that the high secretion rate cannot be sustained while at the same time, an elevated fasting glucagon diverts away from re-esterification.
With the evidence presented, there are 3 main points to consider.
First, a state in which insulin is low and glucagon is high tells the body to increase its efficiency. A system in which the liver takes up fatty acids to produce ketones, use fatty acids for its own energy provision and also uses fatty acids to produce triglycerides and cholesterol and produce ApoB protein in order to build VLDL particles and distribute fatty acids for energy could benefit from having those fatty acids delivered directly from adipose to all cells in need of that energy. This allows the liver to focus on its own energy provision and production of ketones. Fatty acids can be delivered directly via albumin which has a big capacity to carry NEFA but also by keeping LDL-sized ApoB particles in circulation longer so that they can carry out the NEFA transactions.
The amino acids required for ApoB production also need to be saved in this state as the continued production and degradation requires ATP. The VLDL assembly requires energy while, if hepatic AMPK is indeed elevated, there is a need to be conservative with energy.
Low VLDL output would be a problem if cells have a requirement to receive the NEFA from adipose but we have seen that albumin is able to carry out this task.
Second, the concept of increased production does not hold unless there is also an increase in clearance. The LEM does not make note of any such increase. It does make note of an increase in lipoprotein lipase which essentially means that the lipoprotein particle is stripped of its triglyceride content but does not remove the ApoB protein from circulation. This clearance from circulation is done by the liver LDL receptor but, in contrast to requiring higher clearance, this LDL receptor is reduced under low free T3.
Third, the premise of high VLDL secretion and lipolysis in combination with remodeling assisted by HDL, to my view, needs to be corrected towards a reduced hepatic secretion of ApoB particles in a lower size range (likely LDL) with a reduction in remodeling due to low CETP activity, combined with a longer median residence of LDL particles in circulation.
Finally, I hope that with this article there is sufficient content to reconsider the statements around increased hepatic VLDL secretion. I do encourage researchers to follow the recommendation of the LEM authors to investigate on the true reason why LDLc increases so dramatically in these individuals.
As I have boldly stated based on my own summary of the main purpose of the lipid system, the potential implications can be quite high with the potential to reverse atherosclerosis should my analysis prove to be correct.
Strength and weaknesses
Some of the studies referenced provide results obtained from animals. These are good for mechanistical results although still require validation in humans. However for example enzymatic activity in the liver would require obtaining liver samples which require invasive procedures on living humans. Such studies will not be approved.
Because the increase seen in LDL-c is framed under narrow conditions, almost none of the research is done under such conditions so one has to be careful if the results do translate correctly.
Protein content in the diet can be a differentiator as it too has an effect on the main hormones glucagon and insulin which are responsible for most of the effects. In addition protein may influence the thyroid as well.
What is not addressed, and to which dietary protein partakes, is the effects across the day with several meals. In general, a meal will stimulate insulin to some degree. A high fat meal together with sufficient insulin stimulation will accumulate lipids in the postprandial phase allowing for a temporary increased production of VLDLs until insulin has subsided to its fasting levels and the liver has cleared its accumulated fatty acids.
However, sufficient resources cover the mechanisms by which we can expect an increase or decrease in VLDL secretion. We can reasonably expect that, in these individuals, there is a low fasting level of insulin and increased fasting level of glucagon in conjunction with a decrease in active T3 and low CETP activity.
Investigating these parameters in this sub-population could help identify if these aspects are truly relevant towards the noted increase in LDL-c.
– – – T H E – E N D – – –