A randomized 3-way crossover study indicates that high-protein feeding induces de novo lipogenesis in healthy humans (Charidemou et al., 2019)
First it was protein will kill your kidneys and bone, then it kicked you out of ketosis or was converted straight to glucose. Now, not only does it stimulate de novo lipogenesis (DNL), but it is the amino acids themselves that are literally converted to fat.
Do the biochemical pathways exist? Sure. More so on paper because I doubt very much of the fat stored in your body literally came from protein. I can see if you increased protein enough to drive a big energy surplus, then some more dietary fat would get stored as body fat and you might see increased DNL, but even some older protein overfeeding studies show this doesn't really happen that much*
*it's tricky to define the magnitude of a surplus if expenditure keeps going up; with protein, the surplus on paper is usually overestimated unless you're constantly measuring expenditure and adjusting intake
Which brings us to today's study. Great study design (3-way crossover), single meal prepared & provided to the participants. Small sample size. Weird nutritional controls, although I suspect this decision was more practical than scientific. The high fat diet was very high fat and the high protein diet was... higher than control, but if I were designing this study trying to show 'protein -> DNL -> fat,' I'd've gone with more than 40g.
Indeed, this study was very nuanced...
no conflicts of interest or nefarious funding source.
There's a lot of good (nerdy) info about NAFLD & DNL in the intro. But for context, when they're trying to determine how much dietary carbs are used in DNL, they're giving 175 grams (Timlin and Parks, 2005), 385 grams (Sanders et al., 2018), etc., not 40 grams of protein :/
They fed the test meals and took blood samples for five hours. Most of the analysis was on the later time points because the early time points would reflect fatty acids from the diet in chylomicrons, which come from the intestine and we're interested in hepatic DNL here. Also, even if they isolated liver-derived VLDL/LDL (which they did), DNL still takes a while and we wouldn't expect to see evidence of it instantly.
They weren't assessing DNL with tracers. They were using surrogates palmitate (because it is a common product of DNL) and linoleate (because it isn't)... using a methodology that was validated by feeding almost 400 grams carbs in their study of 40 grams protein. Not the most sensitive of techniques.
Aaaaand here's the money shot:
No difference in plasma fatty acids, and more palmitate in LDL/VLDLs. And if you divide the palmitate with a slight decrease in linoleate, you get the figure on the right.
For what it's worth, there was less linoleate in the high protein diet so we'd expect linoleate to be lower in this group. Full stop.
It's starting to get very murky.
A technique that was validated in a very different context; and it's difficult to conceive perfect dietary controls anyway -- in this study, the control group was 15% protein and 45% carb while the HP group was 32% protein and 35% carb and we get this:
No difference in insulin, so that couldn't explain it. And fewer carbs in the HP group, so that doesn't explain it, but double the protein so it was the glutamate! The protein-DNL increased the palmitate/linoleate ratio, oh no wait, it increased the palmitate in LDL/VLDL triacylglycerols but linoleate was lower because there was less in the meal, but we still get that hot 16:0/18:2 graph.
To their credit, the authors did follow-up with some cell culture studies showing hepatocytes are capable of converting glutamate into palmitate, butttttttttttttttt
That's all for now!
For personalized health consulting services: [email protected].
If you want the benefits of ‘shrooms but don’t like eating them, Real Mushrooms makes great extracts. 10% off with coupon code LAGAKOS. I recommend Lion’s Mane for the brain and Reishi for everything else.