Should I be surprised this article from November 2015 has received no attention (citations) yet? Well, no, I am not. The article discusses about fasting – water fasting, to be more precise – and this seems to be non-taboo.
It was published in the Journal of Clinical Investigation (a reputable publication, imho) and it followed the physiologic adaptations of human subjects in a 10 day fasting experiment. 
The article did, however, receive recognition by David Holmes in a review in Nature in January 2016 . And Nature is the #1 scientific publication as of March 2016, according to the h5 index (a measure of the impact and the productivity of a journal).
So, what was the purpose of the study?
Their primary goal was to see how the activity of FGF21 changes in prolonged fasting.
But, why is this important?
According to the NIH (2016), the protein encoded by FGF21 is part of the FGF (fibroblast growth factor) family. This family of proteins deals with tissue repair, cell growth, tumor growth, morphogenesis, and other processes. 
Data from mouse models (FGF21 transgenic mice) attribute various functions to FGF21:
– induction of ketogenesis and gluconeogenesis    
– growth hormone resistance (preventing energy expenditure on growth) 
– disruption of the HPO axis (minimizing energy expenditure on reproduction) 
Researchers explain :
FGF21 transgenic mice are also resistant to a high-fat/high-carbohydrate diet; despite eating more food than do their WT littermates, they gain less weight, an effect that appears to be at least partially mediated by induction of thermogenesis. Why a hormone that is upregulated during starvation would enhance weight loss and induce thermogenesis is unknown.
They think that it is important to study the metabolic effects of FGF21 to determine how the mechanisms observed in mice apply to humans. If these mechanisms remain consistent, they would provide rationale for further developing FGF21 mimetics for treating diabetes and obesity.
Previous human studies, according to the researchers , have not shown consistency in these mechanisms:
One small study that measured serum FGF21 before and after a 7-day fast in patients with rheumatoid arthritis demonstrated a modest 74% increase in FGF21 levels. Subsequently, however, numerous studies have demonstrated no change in FGF21 levels with fasting for up to 72 hours. Moreover, women with anorexia nervosa, a state of chronic nutritional deprivation, have reduced or similar levels of FGF21 as compared with levels in normal-weight controls.
Thus, they wanted to test the hypothesis that fasting increases circulating levels of FGF21 and see if the mechanisms of actions of FGF21 observed in mice also apply to humans (sorry for the repeat).
FGF21 – And Ten Days of Water Fasting
They recruited 8 women and 3 men (n=11, small sample – this may be a drawback) with a mean age of 31.5 years, normal and slightly overweight (100% to 130% of ideal bodyweight). These were healthy subjects with no history of eating disorders or chronic illness.
They had an initial health assessment 10 days before the fast (blood, weight, height, etc).
10 days after this assessment, the subjects (having fasted for at least 8 hours – overnight fast, I suspect) were admitted to the Center for Clinical Investigation at Brigham and Women’s Hospital for a 10-day fast. 
While fasting, subjects only consumed:
– water (ad libitum)
– daily multivitamin
– daily 20 mEq potassium chloride (to prevent hypokalemia)
– daily 200mg allopurinol (a synthetic drug which inhibits uric acid formation, used to treat gout and related conditions – this is interesting) 
9 out of the 11 subjects completed the experiment.
Researchers used DXA to study their body composition. They also used PET and MRI technology to study their brown adipose tissue. In terms of blood markers, they measured: FGF21, ketones, adiponectin, T3, BCAAs, glucagon, insulin, HOMA-IR, ALT, AST, NEFA (nonesterified fatty acids), and several others.
Measurements were taken: pre-visit (10 days before fasting), at baseline (day 0), and days 1,3,5,7,9 and 10 (final day) of the fast. 
Results and Interpretation
Probably the most important finding is in the abstract :
Unlike mice, which show an increase in circulating FGF21 after only 6 hours, human subjects did not have a notable surge in FGF21 until 7 to 10 days of fasting.
As seen in the figure above, FGF21 levels decreased during the initial phase of the fast, and it markedly increased in the late phase, days 7-10.
What points me to say that this is a well conducted experiment is the fact that researchers considered potential errors in their measurements when they observed the surged increased FGF21 levels :
We also considered the possibility that we were measuring the accumulation of inactive FGF21. Therefore, we re-measured our samples using an assay specific for the intact protein. These measurements were consistent with those of the initial finding, therefore providing independent confirmation for the late, fasting-mediated surge in functional FGF21.
Another finding of the study involves the correlation between ketogenesis and FGF21 levels. They found no correlation, which is somewhat different from what was previously seen in mouse studies :
FGF21 is not a causal mediator of the ketogenic response in humans. Importantly, there was no significant difference between FGF21 levels at baseline compared with FGF21 levels at the time point at which ketone levels increased by at least 10-fold (P= 0.43).
Many folks looking into and trying to interpret research studies often blame mice experiments as being inaccurate and not relevant to human studies. The same folks appeal to other mice studies when trying to prove their beliefs. The irony…
Dr. Collin Champ (and I, of course) thinks that mice studies are appropriate for testing hypothesis in eukaryotic models. If such tests lead to important findings, the experiments may be further conducted in humans. When doing human trials, diligent researchers will account for the metabolic, physiologic, circadian and other differences between mice and humans (controlling for variables). And these researchers seem to have been doing just that :
Because mice are known to rapidly lose weight during fasting, this raises the question of whether interspecies differences in weight loss rates could explain the difference in FGF21 dynamics. Indeed, the C57BL/6 mice used in this study lost an average of 17.4% ±1.6% (SD) of their BW over a 24-hour fasting period, whereas the human subjects lost 9.2% ±0.93% (SD) of their BW over the 10-day fast.
They also studied brown adipose tissue (BAT) and the possible correlation between FGF21 and thermogenesis :
Fasting-induced FGF21 does not lead to induction of thermogenic fat in humans. Although seemingly paradoxical, FGF21 has consistently been implicated in the induction of adaptive thermogenesis in WAT; therefore, we considered the possibility that fasting-induced circulating FGF21 would drive thermogenesis.
They also used qPCR (quantitative polymerase chain reaction) on fat biopsy samples (periumbilical) to determine the expression of genes that are usually involved in thermogenesis 
We did not reliably detect uncoupling protein 1 (UCP1) at any of the time points, which may be due to the fact that thermogenic activity is not typically seen with FDG imaging at the sampling site, including in the present study. We did, however, observe downregulation of other key transcriptional regulators of the thermogenic program.
They advise for caution when interpreting data, given the site from which the sample was taken (abdomen). Hence, I suspect they may have seen different results if fat was sampled from other (more BAT rich) areas of the body .
These data must be interpreted with caution, given the sampling site; however, the effect of fasting on key transcriptional regulators of the thermogenic program is consistent with the PET imaging data in demonstrating a fasting-mediated suppression of thermogenesis, while coinciding with the reduced energy expenditure observed during starvation (Figure 2C)
The substantial amount of raw data that came out of this study allows for studying other markers and how their secretion and their dynamics change during prolonged fasting.
For example, they measured how FGF21 correlates with T3 during prolonged fasting :
Given that the fasting-mediated reduction in energy expenditure has been linked to decreased thyroid hormone levels, we also examined the thyroid axis by measuring triiodothyronine (T3) levels and the transcription of thyroid hormone effector genes to determine whether these were associated with changes in FGF21.
They found that, similarly to other studies :
– T3 levels decrease with fasting, dropping by day 3 of the fast and remaining at that level throughout the fast :
All subjects had final-fast-day T3 levels that were lower than their baseline levels.
From what I see, they wanted not only to measure how FGF21 changes during fasting but also how it correlates with other biomarkers.
Similarly to ketones, markers like BCAAs and glutamine did not correlate with FGF21 levels. However, :
In examining serially measured variables with biologically plausible relationships with the metabolic adaptation to starvation, we observed that only insulin, HOMA-IR, nonesterified fatty acids (NEFA), and serum transaminases displayed late changes grossly resembling the dynamics of FGF21 (Figure 4). – see the graphic above
Even though they did not find association between total weight loss and FGF21, they observed significant positive association between absolute weight loss and FGF21 levels :
Unexpectedly, subjects with the most weight loss had the most delayed upregulation of their FGF21 levels.
From my point of view, this is a well conducted study which, so far, received little to no attention. First of its kind (to my knowledge) it observed how FGF21 levels change in prolonged fasting (they like to call it starvation) :
Not only did it take the full 10 days to demonstrate a statistically meaningful induction of FGF21, but we also observed an initial decline in FGF21 levels in the majority of subjects.
While other positive adaptations such as autophagy can take place earlier with fasting, this study goes to show that markers such as FGF21 take longer until entering the scene. Similar studies are required to see if these findings are replicated.
We need more studies also because of the potential applications of FGF21 mimetics (compounds to increase the activity of FGF21) in pathologic states such as diabetes and obesity.
These mimetics should/may not require the need of using prolonged fasting for the activity of FGF21 to be increased; but this begs the question whether or not increased FGF21 activity in the non-fasted state is similar to that in prolonged fasting. A similar question applies to using exogenous ketones to enter ketosis when consuming a carbohydrate rich diet (is there a competition between fuels? what other metabolic adaptations occur).
For the question at hand (regarding FGF21), researchers speculate :
For example, it is possible that the administration of FGF21 to humans in the fed state may promote increased energy expenditure and uncoupled thermogenesis. If FGF21 or related mimetics continue to advance through the drug development pipeline, it may be that their efficacy and any potential off-target effects will be highly dependent on coexistent nutritional status.
They probably talk about this study from 2013 in which researchers used FGF21 in the form of a compound called LY2405319 in a human double-blind proof-of-concept trial. They found :
- Obese patients with type 2 diabetes were treated for 1 month with an FGF21 analog
- Improvements in LDL-C, HDL-C, triglycerides, and body weight were demonstrated
- FGF21 treatment produced increases in adiponectin and β-hydroxybutyrate
- This study reports FGF21 pharmacodynamic effects in humans
Other recent studies worth mentioning (that you may read):
Pharmacokinetics and pharmacodynamics of PF-05231023, a novel long-acting FGF21 mimetic, in a first-in-human study.
Fibroblast growth factor 21 regulates energy metabolism by activating the AMPK-SIRT1-PGC-1alpha pathway
Understanding the Physiology of FGF21
In light of these finding, I reiterate on the importance of studying the mechanisms of FGF21, since it may have potential health benefits. Developing compounds to promote its activity may be critical as well, especially if further research shows that FGF21 is up-regulated only with prolonged water fasting and under no other circumstances. And prolonged water fasting, imho, is an unapproachable strategy for most folks even though, ironically, most of us are geared (equipped by evolution) to go without eating for long periods of time.
Please see ‘further reading’ if you’re interested about this.
- Fazeli, P. K., Lun, M., Kim, S. M., Bredella, M. A., Wright, S., Zhang, Y., … & Steinhauser, M. L. (2015). FGF21 and the late adaptive response to starvation in humans. The Journal of clinical investigation, 125(12), 0-0.
- Holmes, D. (2016). Metabolism: Fasting induces FGF21 in humans. Nature Reviews Endocrinology, 12(1), 3-3.
- FGF21 – Fibroblast Growth Factor 21 – Homo Sapiens
- Badman, M. K., Pissios, P., Kennedy, A. R., Koukos, G., Flier, J. S., & Maratos-Flier, E. (2007). Hepatic fibroblast growth factor 21 is regulated by PPARα and is a key mediator of hepatic lipid metabolism in ketotic states. Cell metabolism, 5(6), 426-437.
- Potthoff, M. J., Inagaki, T., Satapati, S., Ding, X., He, T., Goetz, R., … & Burgess, S. C. (2009). FGF21 induces PGC-1α and regulates carbohydrate and fatty acid metabolism during the adaptive starvation response. Proceedings of the National Academy of Sciences, 106(26), 10853-10858.
- Fisher, F. M., Estall, J. L., Adams, A. C., Antonellis, P. J., Bina, H. A., Flier, J. S., … & Maratos-Flier, E. (2011). Integrated regulation of hepatic metabolism by fibroblast growth factor 21 (FGF21) in vivo. Endocrinology, 152(8), 2996-3004.
- Liang, Q., Zhong, L., Zhang, J., Wang, Y., Bornstein, S. R., Triggle, C. R., … & Xu, A. (2014). FGF21 maintains glucose homeostasis by mediating the cross talk between liver and brain during prolonged fasting. Diabetes, 63(12), 4064-4075.
- Inagaki, T., Lin, V. Y., Goetz, R., Mohammadi, M., Mangelsdorf, D. J., & Kliewer, S. A. (2008). Inhibition of growth hormone signaling by the fasting-induced hormone FGF21. Cell metabolism, 8(1), 77-83.
- Owen, B. M., Bookout, A. L., Ding, X., Lin, V. Y., Atkin, S. D., Gautron, L., … & Mangelsdorf, D. J. (2013). FGF21 contributes to neuroendocrine control of female reproduction. Nature medicine, 19(9), 1153-1156.
- Gaich, G., Chien, J. Y., Fu, H., Glass, L. C., Deeg, M. A., Holland, W. L., … & Moller, D. E. (2013). The effects of LY2405319, an FGF21 analog, in obese human subjects with type 2 diabetes. Cell metabolism, 18(3), 333-340.
Further (recommended) readings:
Kharitonenkov, A., Shiyanova, T. L., Koester, A., Ford, A. M., Micanovic, R., Galbreath, E. J., … & Gromada, J. (2005). FGF-21 as a novel metabolic regulator. The Journal of clinical investigation, 115(6), 1627-1635.
Lee, P., Linderman, J. D., Smith, S., Brychta, R. J., Wang, J., Idelson, C., … & Kebebew, E. (2014). Irisin and FGF21 are cold-induced endocrine activators of brown fat function in humans. Cell metabolism, 19(2), 302-309.
Foltz, I. N., Hu, S., King, C., Wu, X., Yang, C., Wang, W., … & Gupte, J. (2012). Treating diabetes and obesity with an FGF21-mimetic antibody activating the βKlotho/FGFR1c receptor complex. Science translational medicine, 4(162), 162ra153-162ra153.
Gälman, C., Lundåsen, T., Kharitonenkov, A., Bina, H. A., Eriksson, M., Hafström, I., … & Rudling, M. (2008). The circulating metabolic regulator FGF21 is induced by prolonged fasting and PPARα activation in man. Cell metabolism, 8(2), 169-174.
Hotta, Y., Nakamura, H., Konishi, M., Murata, Y., Takagi, H., Matsumura, S., … & Itoh, N. (2009). Fibroblast growth factor 21 regulates lipolysis in white adipose tissue but is not required for ketogenesis and triglyceride clearance in liver. Endocrinology, 150(10), 4625-4633.
Laeger, T., Henagan, T. M., Albarado, D. C., Redman, L. M., Bray, G. A., Noland, R. C., … & Morrison, C. D. (2014). FGF21 is an endocrine signal of protein restriction. The Journal of clinical investigation, 124(9), 3913-3922.
Images: adapted from here