Adipose Lipolysis and Free-Fatty-Acid Feedback

From Insulin to Circulating Fatty-Acid Supply

Adipose tissue is a regulated source of circulating non-esterified free fatty acids (FFAs). Lipolysis hydrolyzes stored triglyceride, releasing fatty acids and glycerol. Insulin acutely suppresses this process; when insulin falls, adipose FFA release can rise. The resulting change in circulating FFA exposure can influence glucose metabolism: fatty-acid oxidation can support hepatic glucose production, while sustained or elevated FFA exposure can reduce insulin-stimulated glucose uptake in muscle. These are linked processes, but a compact model should not imply that one circulating concentration reveals every underlying tissue mechanism.

The model below isolates this pathway while treating other regulators, such as catecholamines, substrate availability, and tissue blood flow, as fixed or summarized by parameters. It describes circulating FFA supply and its effects on glucose fluxes; it does not model lipoprotein-particle production or clearance.

Definition: Let $I$ be plasma insulin concentration, $L$ adipose lipolytic FFA release rate, $F$ plasma FFA concentration, $R_F$ the rate at which adipose-derived FFA enters the modeled circulating pool, $V_F$ its apparent distribution volume, and $CL_F$ its effective clearance. A minimal insulin-suppressed lipolysis model is $$ L(I)=L_{\min}+\frac{L_{\max}-L_{\min}}{1+(I/IC_{50,L})^{h_L}}. $$ Here $L_{\min}$ and $L_{\max}$ are the lower and upper rates under the modeled conditions, $IC_{50,L}$ is the insulin concentration halfway between those rates, and $h_L>0$ controls curve steepness.

The decreasing function of insulin represents suppression: as $I$ increases, the denominator increases and $L(I)$ moves toward $L_{\min}$. The parameters are conditional on the physiological context. For example, a change in adrenergic drive could alter the apparent lipolysis rates even if insulin is unchanged. The equation is therefore not a universal biochemical law, but a useful input-output representation when those other influences are held fixed or are not resolved.

Worked Example: Insulin Suppression of Lipolysis

Suppose, hypothetically, $L_{\min}=0.40$ mmol/min, $L_{\max}=4.00$ mmol/min, $IC_{50,L}=50$ pmol/L, and $h_L=2$. Calculate the release rate at $I=100$ pmol/L and $I=25$ pmol/L.

At $100$ pmol/L, $$ L(100)=0.40+\frac{4.00-0.40}{1+(100/50)^2} =0.40+\frac{3.60}{5} =1.12\ \text{mmol/min}. $$

At $25$ pmol/L, $$ L(25)=0.40+\frac{3.60}{1+(25/50)^2} =0.40+\frac{3.60}{1.25} =3.28\ \text{mmol/min}. $$

Under this model, the lower insulin concentration is associated with a higher adipose FFA release rate. These values are hypothetical model outputs, not clinical measurements.

Circulating FFA as a Dynamic State

Adipose release does not instantaneously determine plasma FFA concentration. FFA also leaves the circulating pool through tissue uptake and metabolism. A one-pool mass balance makes that distinction explicit. Let $f_{\mathrm{sys}}$ be the fraction of adipose lipolytic release that enters this modeled pool. With $L$ in mmol/min, $V_F$ in L, and $CL_F$ in L/min, the balance is:

\frac{dF}{dt}=\frac{f_{\mathrm{sys}}L(I)}{V_F}-\frac{CL_F}{V_F}F.

This representation assumes a well-mixed pool, constant volume, and first-order effective removal over the range considered. The parameter $CL_F$ combines removal processes rather than assigning them to particular tissues. At a steady state, input equals removal. Thus, if $I$ and the other inputs remain constant,

F_{\mathrm{ss}}=\frac{f_{\mathrm{sys}}L(I)}{CL_F}, \qquad \tau_F=\frac{V_F}{CL_F}.

The steady-state concentration depends on the ratio of systemic FFA entry to clearance. The time constant $\tau_F$ describes the approach to steady state in this linear one-pool model: after an input step, the difference between $F(t)$ and its new steady state declines exponentially with time constant $\tau_F$.

Worked Example: FFA Concentration and Response Time

Continue with the hypothetical release rate $L=1.12$ mmol/min. Suppose $f_{\mathrm{sys}}=0.75$, $V_F=40$ L, and $CL_F=8.0$ L/min. Calculate the steady-state concentration and time constant. If the initial concentration is $0.060$ mmol/L, also calculate its initial rate of change.

Systemic entry is $f_{\mathrm{sys}}L=0.75(1.12)=0.840$ mmol/min. Therefore, $$ F_{\mathrm{ss}}=\frac{0.840\ \text{mmol/min}}{8.0\ \text{L/min}} =0.105\ \text{mmol/L}, \qquad \tau_F=\frac{40\ \text{L}}{8.0\ \text{L/min}}=5.00\ \text{min}. $$

At the initial concentration, input per unit volume is $0.840/40=0.0210$ mmol/L/min, while removal per unit volume is $(8.0/40)(0.060)=0.0120$ mmol/L/min. Hence $$ \left.\frac{dF}{dt}\right|_{F=0.060} =0.0210-0.0120 =0.00900\ \text{mmol/L/min}. $$

The positive rate means the concentration is initially rising toward $0.105$ mmol/L. A higher release rate does not by itself specify the concentration unless clearance and the pool assumptions are also considered.

Connecting FFA Exposure to Glucose Fluxes

FFA can affect hepatic glucose production through more than one route. Fatty-acid oxidation supplies energy and supports gluconeogenesis; glycerol released during lipolysis is a gluconeogenic substrate. FFA exposure can also be associated with reduced hepatic responsiveness to insulin. These pathways differ in mechanism and timescale. A compact model can summarize their net effect, but it cannot distinguish them unless additional states or measurements are available.

One phenomenological representation adds a saturable FFA-dependent term to an insulin-suppressed hepatic glucose production model:

H(I,F)=H_0\left(1-E_{\max,H}\frac{I}{K_{I,H}+I}\right) +k_H\frac{F}{K_{F,H}+F}.

Here $H$ is hepatic glucose production, $H_0$ is its reference rate, $E_{\max,H}$ is the maximum fractional insulin suppression in the absence of the added FFA term, and $K_{I,H}$ is the insulin concentration for half-maximal suppression. The positive coefficient $k_H$ has units of glucose production rate; $K_{F,H}$ has concentration units. The FFA term summarizes a net increase in modeled production over the range studied. It should not be interpreted as proof that circulating FFA acts through one specific hepatic mechanism.

For muscle, a useful compact extension is to let FFA exposure attenuate the insulin-stimulated component of glucose uptake:

U_m(I,F)=U_{m,0} +\frac{V_{m,I}\,I/(K_{I,m}+I)} {1+\gamma_m F/(K_{F,m}+F)}.

Here $U_m$ is muscle glucose uptake, $U_{m,0}$ is the modeled insulin-independent component, $V_{m,I}$ is the maximum insulin-stimulated increment, and $K_{I,m}$ is its half-maximal insulin concentration. $K_{F,m}$ is the FFA concentration for half-maximal saturation of the modifier, and $\gamma_m\geq0$ sets its strength. This form reduces the insulin-stimulated increment as FFA rises while leaving the baseline term unchanged. It is a modeling choice, not a claim that FFA can never affect basal uptake.

Worked Example: FFA-Associated Changes in Glucose Fluxes

Consider hypothetical rates at fixed insulin $I=100$ pmol/L. Let $H_0=10.0$ mg/kg/min, $E_{\max,H}=0.70$, $K_{I,H}=50$ pmol/L, $k_H=4.00$ mg/kg/min, and $K_{F,H}=0.100$ mmol/L. For muscle, let $U_{m,0}=1.00$ mg/kg/min, $V_{m,I}=5.00$ mg/kg/min, $K_{I,m}=50$ pmol/L, $K_{F,m}=0.100$ mmol/L, and $\gamma_m=1.50$. Compare $F=0.050$ and $F=0.150$ mmol/L.

The insulin fraction is $100/(50+100)=0.667$, so the insulin-suppressed hepatic component is $$ 10.0\left[1-0.70(0.667)\right]=5.33\ \text{mg/kg/min}. $$ At $F=0.050$ mmol/L, $F/(K_{F,H}+F)=0.050/0.150=0.333$; thus $H=5.33+4.00(0.333)=6.67$ mg/kg/min. At $F=0.150$ mmol/L, the fraction is $0.150/0.250=0.600$; thus $H=5.33+4.00(0.600)=7.73$ mg/kg/min.

For muscle, the insulin-stimulated increment before FFA modification is $5.00(0.667)=3.33$ mg/kg/min. At $F=0.050$, the modifier denominator is $1+1.50(0.333)=1.50$, giving $U_m=1.00+3.33/1.50=3.22$ mg/kg/min. At $F=0.150$, the denominator is $1+1.50(0.600)=1.90$, giving $U_m=1.00+3.33/1.90=2.75$ mg/kg/min. In this specified model, higher FFA is associated with higher hepatic production and lower muscle uptake at the same insulin concentration.

Justification, Assumptions, and Model Checks

The steady-state result follows directly from the mass balance. At steady state, $dF/dt=0$, so input and removal are equal:

0=\frac{f_{\mathrm{sys}}L(I)}{V_F}-\frac{CL_F}{V_F}F_{\mathrm{ss}} \quad\Longrightarrow\quad f_{\mathrm{sys}}L(I)=CL_FF_{\mathrm{ss}} \quad\Longrightarrow\quad F_{\mathrm{ss}}=\frac{f_{\mathrm{sys}}L(I)}{CL_F}.

If $f_{\mathrm{sys}}$, $CL_F$, and the lipolysis parameters are positive and fixed, then increasing insulin decreases $L(I)$, and therefore decreases $F_{\mathrm{ss}}$. This follows because

\frac{dL}{dI} =-\frac{(L_{\max}-L_{\min})h_L(I/IC_{50,L})^{h_L}} {I\left[1+(I/IC_{50,L})^{h_L}\right]^2}<0 \quad\text{for } I>0,\ L_{\max}>L_{\min}.

The direction of the modeled downstream effects is also explicit: with $k_H>0$, the added hepatic term rises with $F$; with $\gamma_m>0$, the muscle insulin-stimulated increment falls with $F$. These are assumptions encoded by the chosen functions. Data may support, weaken, or fail to distinguish them.

Useful checks include confirming that simulated concentrations and fluxes remain nonnegative, that $L_{\min}\leq L(I)\leq L_{\max}$, and that the predicted time course is compatible with observed timing. If measured FFA changes lag far behind or precede the one-pool prediction, an instantaneous insulin-to-lipolysis relation or a single well-mixed pool may be inadequate. Residual patterns that vary systematically with insulin, time, or feeding state can indicate omitted regulation or a missing delay. Parameter estimates can also be weakly identifiable: similar FFA profiles may result from different combinations of $f_{\mathrm{sys}}$, lipolysis rates, distribution volume, and clearance. Concentration data alone do not generally identify every flux separately.

A more mechanistic alternative would represent adipose insulin signaling, triglyceride hydrolysis, fatty-acid re-esterification, and tissue-specific uptake as separate processes. That may be warranted when those mechanisms are measured or are the scientific question, but it requires more parameters and data. The compact functions above are easier to estimate and use for system-level hypotheses, at the cost of mechanistic resolution. In particular, the model’s circulating FFA pool is not a model of VLDL triglyceride production or clearance.

Common Pitfalls and Limitations

Equating lipolysis with plasma FFA concentration. Release is a flux; concentration also depends on distribution and clearance. Use a mass balance when the time course matters.
Treating the FFA term as a proven direct cause. A fitted association can summarize substrate supply, oxidation, insulin resistance, or omitted covariates. The equation alone does not separate those mechanisms.
Mixing units. $L$ and $R_F$ are rates, $F$ is concentration, $CL_F$ is volume per time, and $V_F$ is volume. The steady-state ratio $R_F/CL_F$ must have concentration units.
Assuming one response timescale. The circulating concentration may respond quickly while downstream metabolic adaptations occur over a different timescale. Add delays or states only when the data and question support them.
Extrapolating beyond the modeled range. Saturating functions impose limiting behavior. Their predictions at extreme insulin or FFA concentrations may not represent physiology outside the calibration range.

Key takeaway: Model adipose lipolysis as an insulin-regulated FFA input, calculate circulating FFA by balancing input and removal, and represent downstream glucose-flux effects as explicit, testable functions. Keep the distinction between a useful phenomenological link and a uniquely established mechanism.

Check Your Understanding

Use the equations and assumptions in this tutorial. The answers follow each question.

If $I=IC_{50,L}$, what is $L(I)$?
Solution: Since $(I/IC_{50,L})^{h_L}=1$, $L=L_{\min}+(L_{\max}-L_{\min})/2$, the midpoint of the two limiting rates.
If systemic FFA entry is $0.60$ mmol/min and clearance is $6.0$ L/min, what is the steady-state FFA concentration?
Solution: $F_{\mathrm{ss}}=0.60/6.0=0.100$ mmol/L.
In the one-pool model, what happens to the steady-state concentration if clearance doubles while entry stays fixed?
Solution: It halves, because $F_{\mathrm{ss}}=R_F/CL_F$.
Does a higher $F$ necessarily prove that adipose lipolysis increased?
Solution: No. Higher FFA can reflect increased entry, reduced clearance, or both; the concentration balance does not identify the source of change by itself.
For the muscle model, what happens to the insulin-stimulated increment as $F$ increases when $\gamma_m>0$?
Solution: The denominator $1+\gamma_mF/(K_{F,m}+F)$ increases, so the increment decreases, while $U_{m,0}$ remains unchanged in this chosen formulation.