Journal of Theoretical Biology 332 (2013) 89–95

Contents lists available at SciVerse ScienceDirect

Journal of Theoretical Biology journal homepage: www.elsevier.com/locate/yjtbi

Modelling adipocytes size distribution H.A. Soula a,c,n, H. Julienne b, C.O. Soulage a, A. Géloën a a b c

Université de Lyon, CARMEN INSERM U1060, INSA-Lyon, F-69621 Villeurbanne, France Université de Lyon, Joliot-Curie E.N.S Lyon, F69007 Lyon, France EPI BEAGLE INRIA, F-69621 Villeurbanne, France

H I G H L I G H T S

    

Adipocytes size distribution is bimodal. Adipocytes inflate and deflate according to lipid fluxes. We model the lipid fluxes in adipocytes. We show a saddle-node bifurcation with two stable modes. We use individual-based system to model cells and obtain bimodal distribution.

art ic l e i nf o

a b s t r a c t

Article history: Received 2 July 2012 Received in revised form 18 March 2013 Accepted 22 April 2013 Available online 30 April 2013

Adipocytes are cells whose task is to store excess energy as lipid droplets in their cytoplasm. Adipocytes can adapt their size according to the lipid amount to be stored. Adipocyte size variation can reach one order of magnitude inside the same organism which is unique among cells. A striking feature in adipocytes size distribution is the lack of characteristic size since typical size distributions are bimodal. Since energy can be stored and retrieved and adipocytes are responsible for these lipid fluxes, we propose a simple model of size-dependent lipid fluxes that is able to predict typical adipocytes size distribution. & 2013 Elsevier Ltd. All rights reserved.

Keywords: Adipocytes Size distribution Lipid fluxes White adipose tissue Cellularity

1. Introduction Obesity, energy consumption and food intake have attracted a lot of attention from the modelling community (Chow and Hall, 2008; de Graaf et al., 2009). Obesity can be simply stated as an excessive white adipose tissue accumulation due to an energy imbalance (Flier, 2004; Kahn et al., 2006; Abrams and Katz, 2011). Most models have focused on the chronic energy imbalance and only few models have studied the cell responsible for the energy storage namely the adipocyte. At first glance, adipocytes are simply fat depots. Triglycerides (TG) are intracellularly stored in as a fat droplet (via a mechanism called lipogenesis) and are hydrolyzed (via a mechanism called lipolysis) whenever needed and excreted into the extracellular milieu in form of glycerol and non-esterified fatty acids (NEFAs). The overall lipid storage represents the excess energy consumption n Corresponding author at: Université de Lyon, excess of energy, CARMEN INSERM U1060, BAT IMBL, INSA-Lyon, F-69621 Villeurbanne, France. Tel.: +33 4 72 43 81 96. E-mail addresses: [email protected], [email protected] (H.A. Soula).

0022-5193/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jtbi.2013.04.025

relative to energy expenditure: an unbalanced energy intake has a direct consequence on the adipocyte population. Adipocyte size and number for instance reflect directly the volume of the lipid content. Obese subjects tend to have an increased mass of adipose tissue mainly resulting from more and bigger adipocytes (Drolet et al., 2008; MacKellar et al., 2010; Arner and Spalding, 2010). The size difference between adipocytes in the same population can be of an order of magnitude (three order of magnitude for the volume) such a variation is unique in the same organism. Contrary to most cell types, adipocytes have no characteristic size e.g. an unimodal size density function. The size distribution of adipocytes has two modes— one peak in the small size around 15 μm and one peak for bigger adipocytes generally more than 60 μm (Mclaughlin et al., 2007). Size distribution differences were linked with diabetic phenotype (Mclaughlin et al., 2007), hepatic steatosis (Kursawe et al., 2010) and inflammation (Mclaughlin et al., 2010; Liu et al., 2010). Several other studies have shown that individual adipocytes size may have an impact on its behavior like in adipokine secretion (Skurk et al., 2007), glucose metabolism (Lay et al., 2001), lipolysis (Zinder and Shapiro, 1971; Smith, 1971), lipogenesis activity (Gliemann and Vinten, 1974) and adrenoceptor expression (Lafontan and Berlan, 1995). Recent

90

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95

works have dealt with the issue of size dynamics: How adipocyte population size shifts under various dietary conditions and among different rat strains (Jo et al., 2009, 2010; MacKellar et al., 2009, 2010). All these studies focused on describing a peculiar function of adipocytes: their size. To store excess energy, organism can either create more storage – new cells – or increasing storage capacity. We propose in this paper a simple model that take into account lipid fluxes to explain size variation and ultimately size distribution. This is done by assuming that adipocytes inflate (via lipogenesis) and deflate (via lipolysis) according to their lipid content and the rate of these fluxes are also size dependent. We are easily able to obtain typical adipocyte distribution with their two main features: size width and bimodality.

100 µm

2. Background In this section, we first take a look at the data as it can be obtained in our laboratory. We obtained a couple of distribution to show the degree of precision as well as general shape of distribution. Note that this data is provided for illustration purpose. Adipocytes size can be obtained through fixation essentially as described in Etherton et al. (1977) and the use of Coulter–Counter. Briefly, 40–50 mg of retroperitoneal WAT were washed twice in NaCl 0.9% (w/v), twice in 50 mM 2,4,6-trimethylpyridine/ NaCl 0.9% and fixed in 0.12 M osmium tetroxide/50 mM 2,4,6-trimethylpyridine/ NaCl 0.9% for 96 h at room temperature. Samples were washed with NaCl 0.9% and then left in NaCl 0.9% for 24 h at room temperature They were subsequently washed with 8 M urea /NaCl 0.9% and then left in this solution at room temperature for 72 h. Samples were then rinsed through 250 μm nylon mesh with NaCl 0.9% containing TRITON X100 0.01% (v/v). Prior to analysis, supernatant was discarded and cells were resuspended in glycerol. Cells were then diluted into beakers containing Isoton II solution (Beckman coulter). Adipose cell-size distribution was assessed using a Beckman Coulter Multisizer IV (Beckman Coulter, Villepinte, France). Beckman Coulter Multisizer IV was used with a 400 μm aperture. The range of cell-sizes that can effectively be measured using this aperture is between 20 and 240 μm. Beckman Coulter Multisizer IV was precalibrated with 43 μm diameter latex beads (Beckman coulter). The instrument was set to count 1000 particles per run, and the fixed-cell suspension was diluted so that coincident counting was less than 10%. After collection of pulse sizes, the data were expressed as particle diameters and displayed as histograms of counts against diameter using linear bins and a linear scale for the cell diameter. Cell-size distribution runs were performed to obtain a minimum of 12,000 cells per tissue. Fig. 1 displays the typical result of the fixation process of 12,000 cells as indicated. Three distributions are displayed in Fig. 2 for two rat strains and a human. These distributions show the typical features of adipocytes size distributions as described in Mclaughlin et al. (2007) which are 1. distribution does not display one characteristic size but two, 2. the first mode is around (or below) the detection range of the apparatus, 3. small sizes correspond to the classical eukaryotic cells size in the 15–20 μm range, 4. a second mode of big adipocytes with cell of unusual sizes (bigger than 50 μm with extreme as high as 200 μm), 5. both groups are neatly separated as shown by a size – called the nadir – with very low density. We present a simple model of the size and fat content evolution of an adipocyte that is able to reproduce the bimodality of the distribution.

Fig. 1. White adipose cells in phase contrast microscope image after fixation process. Adipose tissue was fixed with osmium tetroxide and adipocytes were released by incubation in urea (8 M). The scale is indicated as 100 μm bar in the bottom left.

0.02 0.018 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0

0

50

100

150

200

250

Fig. 2. Several examples of adipocyte size distribution. These distributions were obtained in the laboratory for several sample tissues. WISTAR rat (green plain curve), Obesity-Resistant LOU/C rat (blue dot-dashed line), Healthy Human (red dashed curve). These data come from distribution obtained from ongoing studies and are provided for illustration purpose only. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

3. Model To model adipocyte size distribution we will make the following assumptions: The only mechanism for size variation will be through fatty acids fluxes—adipocyte will inflate or deflate according to its lipid droplet volume. There will be no mechanical constraints on the variation of size such as extra-cellular matrix (ECM) pressure, vascularization, etc. We will place ourselves at the limit of fast diffusion of fatty acids in the ECM. Extracellular fatty acids will be equally available among cells instantly at a constant concentration. These assumptions suppose that the fatty acids undergo a free diffusion in the extracellular milieu. Finally adipocytes fluxes are membrane-based phenomenons. As such fluxes will be surface limited as it was suggested by several publications (Jo et al., 2010; Gliemann and Vinten, 1974; Smith, 1971). An adipocytes of size r (radius in μm) will store an amount l of triglycerides (in nmol). The variation of the content of triglycerides

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95

4. Results

can be described by the following equation: dl L κ nr l 2 ¼ ar 2 −ðB þ br Þ ¼ T ðr; l; LÞ dt L þ κ l κ nr þ r n kt þ l |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl} lipogenesis

ð1Þ

lipolysis

The lipogenesis part is the product of three factors: one factor describing the surface limited diffusion rate (ar2), Michaelis– Menten like term for the available lipid content L around the cell (which will be the same for all cells) and a Hill-like term to describe the resistance toward indefinite intake of triglycerides ðn 4 0Þ. This Hill function uses a cutoff radius κr (in μm) above which lipogenesis rate decreases dramatically. The lipolysis part is the product of a classical dose-dependent rate B+br2 – with B being the basal lipolysis and b being the catecholamines dependent term – and of another term compensating when the lipid content is low inside the cell. This also prevents any lipid release greater than what remains inside the cell. Values for B will be taken around 125 nmol h−1 and b around −1 0:27 nmol h μm−2 . The important parameter will then be a or more crucially a/b which will be the lipid flux ratio. Note that these are hourly rates. These rates were estimated from maximal lipolysis rates in conjunction with size distributions performed previously in our laboratory and published elsewhere (Soulage et al., 2008). This first equation refers to the rapid evolution of fatty acid content. The slower variation concerns the volume V for the cell itself to adapt to cope with its content size τ

dV ¼ V l l þ V 0 −V dt

ð2Þ

with V0 the minimal volume (in μm3 ) which includes the nucleus, the cytoplasm and the cell machinery, Vl the of 1 unit of l (in nmol) and V the cell volume. This is a simple spring model where membrane production or degradation is proportional to the difference pressure between the container and the content. V0 will typically be 1:4104 μm3 which corresponds to a cell of minimal radius r 0 ¼ 15 μm. No estimation is known for the rate τ and since we are only interested in equilibrium it will be taken long −1 enough ∼10 h−τ ¼ 0:1 h . The volume taken by 1 nmol of triglycerides was taken as ¼ 1:091  106 μm3 . This value was computed from the molar mass of a typical triglyceride (884 g/mol) with density 0.81. Transforming the equation for the radius we obtain τ

dr V l þ V0 r ¼ l 2 − ¼ τRðr; lÞ dt 3 4πr

ð3Þ

For a population of cells, Eq. (1) describes the lipid fluxes in and out of cell. The variable L refers to the bathing lipid content (in nmol) that should also be modified by the fluxes. To describe this properly, let u(t,r,l) be the density function of the cells with radius between r and r+dr, triglycerides content between l and l+dl at time t. Note that the volume of the surrounding medium will be unaffected by the adipocyte size variation and will remain constant. This is the reason why L is not expressed as a concentration. We will assume that the number of cells is constant obtaining the following transport equation: ∂t uðt; r; lÞ þ ∂r ½Rðr; lÞu þ ∂l ½T ðr; l; LÞu ¼ 0 with dL d ¼− dt dt

Z Z luðt; r; lÞ dr dl ¼ −

dUðtÞ dt

91

ð4Þ

ð5Þ

U(t) is the total amount of lipid stored in the cells. Setting initial conditions: u(0,r,l)¼u0(r,l) and L(t)¼ L0+U(0)−U(t) since the overall lipid content is conserved in this case.

We expect to find using realistic parameters a simple condition to obtain size distributions such as the ones in Fig. 2. We will only look for equilibrium solution for Eq. (4). 4.1. One cell model We begin by detailing the analysis of Eqs. (1) and (3) for one cell with radius r and lipid content U(t)¼l(t) in that case. Then L(t)¼L0+l(0)−l(t). We will first assume that L is constant. We can rearrange and normalize by setting γ ¼ L=ðL þ κ l Þ. Note that 0 ≤ γ ≤ 1. A detailed study of the null-clines is available in the appendix. The main message is that since both fluxes are surface limited, all radii should be an equilibrium which amounts to look for boundary conditions. For small radius, the lipolysis rate vanished whereas lipogenesis still operates. This equilibrium depends on how fast lipolysis becomes a really surface based mechanism i.e. when l=ðκ t þ lÞ∼1. To obtain equilibrium with small radius, this condition turns into κt {

B4π V l aγ

On the side of high values for the radius, the lipogenesis is dampened as soon as n 4 0 and κ r c1 whereas lipolysis is very strong. The equilibrium on this side yields a high value of r as soon as aγ 4b stressing the obvious need for enough available lipids to fill up big adipocytes. Finally note that for high values of the lipogenesis surface rate a the condition on small sizes may not hold anymore returning back to one size equilibrium but with high values for the radius. This situation is depicted in Fig. 3 where the fixed points are displayed for increasing values of the ratio aγ=b (Fig. 4). Two equilibria situation is a generic situation which could lead to a bimodal distribution. The case where L is not a constant behaves similarly with bifurcation parameter being the initial amount of total lipid, values for both inside and outside the cell, L0+U(0). For a high enough ratio a/b and L0+U(0), two stable equilibria exist. In Fig. 5, various bifurcations are displayed for a/b and L0+U(0). One needs higher lipogenic values with less and less lipid content. 4.2. Population distribution evolution Since we want to recover population size distribution we will need to derive the evolution of a given number of adipocytes. These are all in interactions via the L parameter since any cell can take or give in a size-dependent manner lipids. In this section we will solve numerically a slightly different version of Eq. (4). Indeed, in its current form, the population behavior will not differ much from the one cell case: a population with a distribution of starting conditions like r(0) will converge to the fixed equilibrium radius depending on the position relative to the unstable fixed point. The resulting equilibrium distribution will be a linear combination two dirac function whose weights will depend on the initial distribution. To obtain realistic distribution, we must assume that some parameters of the adipocytes evolution are variable among the adipocytes. In order to limit the dependence on the parameters we will assume variability on only two parameters, namely r0 and a, the minimal size and lipogenesis maximal flux.

92

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95

Both of these parameters are potential ‘target’ for genetic and/ or epigenetic variations. This choice of parameters is arbitrary and several other parameters – such as the lipolysis rates – are also possible. We will now assume that an adipocyte labelled i has a constant minimal radius r i0 and ai lipogenesis maximal rate both values will be drawn using a normal distribution. Mean values will be equal to the typical parameters 〈r 0 〉 ¼ 15 μm for the minimal radius and 〈a〉 for the lipogenic rate. The latter will vary as above between −1 0.2 and 0:5 nmol h μm−2 . The standard deviation will be r0/3 for the minimal radius—yield a coefficient of variation (CV) of 30%. The standard deviation for a will be also a multiple of the mean with a CV ranging from 10% to 40%. We solved the Partial Differential Equation described in Eq. (4) using a Monte Carlo algorithm: a given number N of adipocytes will have the following parameters r i0 , ai, time dependent radius ri(t) and lipid content li(t) starting with a normal distributed ri(0) and an equivalent li(0) such as dr i =dtð0Þ ¼ 0. The external lipid content is set to L(0)¼L0. Using an explicit Euler method we compute each N trajectories in random order during a given dt ¼10−4 h. The external lipid content L(t) was updated after every

update of each adipocyte. We verify that the total overall lipid content LðtÞ þ ∑N i ¼ 1 li ðtÞ does not vary during the simulation. To avoid size effect, the available lipid is scaled by N so we replace κl by Nκ l in Eq. (1). By setting the same parameters described above and using N ¼20,000 cells, we tested several values of the CV for a from 10% to 40%. As shown in Fig. 6 the equilibrium distribution is remarkably similar to the ones observed in Fig. 2. Our model simply yields realistic adipocytes size distributions without the need for high CV values. All the classical features are obtained with CV of around 20% (green curve). These CV's are in the bounds of biological variation. Indeed, a closer inspection of the data obtained in Soulage et al. (2008) shows that for WISTAR rats, maximal lipolysis rate was 2:457 7 0:716 μmol=h=106 c (CV ¼29%), basal rate 0:262 7 0:144 μmol=h=106 (CV ¼ 55%), lipogenesis rate 3 0:74 7 0:26 μmole H=g=d (CV ¼35.1%). The individual cell lipogenesis seems to show similar variations in Gliemann and Vinten (1974). A more thorough study of the impact of all the parameters is needed. In particular, are we able to reproduce a given distribution —such as the ones in Fig. 2. This has been so far a real issue that there is still some work needed to be able to answer this question. In the remaining of this paper we will focus on the storing

160 140

110

120

100 90

100

80

80

70 60

60

40

50

20

40

0

30 1

1.05

1.1

1.15

1.2

1.25

1.3

1.35

1.4

1.45

1.5

20 10

Fig. 3. Radius equilibria for the one cell model defined by Eqs. (3) and (1) for increasing values of aγ=b for κt ¼ 0:05 (dashed grey—no bifurcation), κ t ¼ 0:025 nmol (black), κ t ¼ 0:01 nmol (red) and κ t ¼ 0:005 nmol (blue). Other −1 −1 parameters are: γ ¼ 1, b ¼ 0:27 nmol μm−2 h , B ¼ 125 nmol h , r 0 ¼ 15 μm, κ r ¼ 200 μm and n ¼3. ⋆ indicates solutions whose nullclines are provided in Fig. 4. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

6

5

5

4

4

3

3

2

2

1

1

20

40

60

80

100

120

140

160

180

200

1.1

1.2

1.3

1.4

1.5

1.6

1.7

Fig. 5. Equilibrium radius versus a/b with non-constant L. Two trajectories are computed using L0+U(0) as the values indicated. Branches are obtained by either setting U(0) ¼0 (low branches) or L0 ¼ 0 (high branch). Other parameters are: −1 −1 κ r ¼ 200 μm, b ¼ 0:27 nmol μm−2 h , B ¼ 125 nmol h , κ t ¼ 0:01, κ l ¼ 0:001.

6

0

1

0

20

40

60

80

100

120

140

160

180

200

Fig. 4. Nullclines of Eqs. ((3) black) and (1) – κ t ¼ 0:05 nmol (dashed grey) and κ t ¼ 0:01 nmol (red) – with corresponding close-up for radius between 15 μm and 25 μm in insert, for two values of aγ=b 1.13 (left) and 1.21 (right). This corresponds to the ⋆ in the bifurcation diagram in Fig. 3. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95 0.12

93

30 3.8 3.6

0.1

25

3.4 3.2

0.08

20

3 2.8

0.06

0

10

20

30

40

50

15

0.04

10

0.02

0

5

0

20

40

60

80

100

120

140

160

180

200

Fig. 6. Adipocyte size distribution as a function of the coefficient of variation of the random variable a—10% (blue), 20% (green), 30% (red) and 40% (cyan). In inset, the total volume occupied by the distribution. L0 ¼ 3 nm/cell. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.) 0.25 4

2 1

0.15

0

0

1

2

3

140

160

4

0.1

0.05

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Fig. 8. Maximal volume content as a function of the fluxes ratio a/b and the cutoff radius κ r . The volume is computed as nmol/cell. ⋆ indicates the corresponding equilibrium distribution displayed in inset.

leading to a bigger storage such as the proliferation of new adipose cells. Another way is possible by increasing lipogenic or inhibiting lipolytic activities. We can study the maximal storage capacity of the adipocytes by looking at the equilibrium volume with extremely high L0. This is done in Fig. 8 as a function of both the flux ratio a/b and the parameter κ r —the maximal radius cutoff. To illustrate size distribution differences, distributions with similar volume—but with different radius cutoff and flux ratio are displayed in inset. Their position in the main graph is indicated by a ⋆. For artificially high flux ratios and small radius cutoff, the adipocyte distribution is more and more unimodal—blue distribution.

x 106

3

0.2

0

5. Discussion 0

20

40

60

80

100

120

180

200

Fig. 7. Adipocyte size distribution as a function of L0. In inset, the total volume occupied by the distribution. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

capabilities of the adipocytes population and the external lipid concentration as a possible adipogenic signal.

4.3. Lipid storing capacity In Fig. 7, we represented the distribution obtained for 5 increasing external lipid contents (L0). The volume occupied increases accordingly and proportionally to the lipid volume available until a maximum volume is reached. Any further increase in the external lipid content yields no changes in the distribution. The purple and light green distribution (for L0 ¼3 and L0 ¼4 respectively) are the same but only the light green is on the foreground. Note that the growth strategy is to increase the volume via increasing both small and big adipocyte sizes. When L0 reaches a critical value, the adipocyte population reaches a maximum volume and all remaining lipid remains in the extra-cellular medium. This critical value is 3 nmol/cell in the simulation displayed in Fig. 7. Since the storing capacity is reached, the increase of the lipid concentration could be the triggering signal to several adaptation

Adipocytes are cells whose size can vary greatly throughout out their lifetime. Those variations are originated from lipid fluxes. The influx is used to store excess calories whereas the efflux is for energy retrieval purpose. Several works have suggested that these fluxes are surface-limited inducing a feedback loop in the size evolution equation. We showed here that these feedback loops can be simply modeled to reproduce size variations of adipocytes. Our simple model displays a saddle-node bifurcation with two stable fixed points. These two stable points provide an elegant explanation for the lack of characteristic size among real adipocytes and their bimodal size distribution. This bimodality is also in our model a generic trait depending on obvious physiological characteristics such as the ratio of maximal influx by maximal efflux and the available lipids. Realistic size distributions were obtained by introducing variability in the maximal rate of lipid influx as well as the minimal radius of cells. The required variability is around 20% which is well inside admitted biological bounds making our model quite realistic (Soulage et al., 2008). Our model does not take into account adipogenesis but provides a possible mechanism for the triggering of creation of new cells. However since these are likely to be small, a constant influx of newly created cells could constitute the small peak of the distribution. This could constitute an alternate explanation for the bimodality.

94

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95

In addition, our model assumes several features that need to be addressed in the future. First, no extra-cellular medium (ECM) lipid transporter (such as lipoprotein in the blood) has been reported so far. This weakens the assumption of lipid diffusion and availability in the model. But apart from direct transmission from endothelial cells, no real alternatives exist to explain how lipid moves in and out of the adipose tissue. This direct transmission would imply that virtually each adipocyte is linked to an endothelial cell. Note that in this case, the model remains the same if we replace blood by ECM. Secondly, no other hormones implicated in the fat tissue regulation as been modelled here. Molecules such as insulin will direct the flow toward more lipid influx this it blocks lipolysis and enhances lipogenesis. In the other hand, stress-related hormones such as catecholamines are known to trigger fat release. We provided here a model for equilibrium distribution assuming implicitly that hormones concentrations are averaged into the parameters. Apart from these issues, the present model is a very simple mechanistic and physical explanation to the adipocytes'size features.

solution is small. Since r∼r 0 , we can write r ¼ r 0 þ ϵ and then l¼

4πr 20 ϵ Vl 2

Additionally, we can assume κ nr þ r n ∼r n and B þ br ∼B yielding l¼

ar 2 γκ t r n0 B

leaving a value for ϵ ϵ¼

V l aγκt B4π

with the condition ϵ{1. This leads to a condition on the only remaining free parameter κl {

B4π V l aγ

Whenever both condition are met, aγ 4 b and κt {B4π=V l aγ two fixed points exist.

References Appendix A. Existence of two stable sizes We will assume in this section that L is constant using γ ¼ L=ðL þ κ l Þ. The null-clines will be l¼

4πr 3 −V 0 3V l

0 ¼ ar 2 γ

κ nr

κnr l 2 −ðB þ br Þ κt þ l þ rn

We separate now the analysis on the two branches. First case: rc1 In that case, the adipocyte will be full lcκt and lipolysis will be at its maximum rate l=ðl þ κ t Þ∼1. Then l¼

4πr 3 −V 0 3V l

0 ¼ ar 2 γ

κnr 2 −ðB þ br Þ κ nr þ r n

High solutions for r can be extracted by neglecting the basal rate and simplify further by rescaling r by r ¼ r=κ r yielding aγ ¼ 1 þ rn b Technically there is a solution as soon as aγ 4 b but for r to be high requires r not too close to zero. Assuming aγ 4 b, we need to find another solution on the lower branch for r close to r0. Second case l{κt In that case, previous equations reduce to l¼ l¼

4πr 3 −V 0 3V l ar 2 γκt κnr 2

ðκ nr þ r n ÞðB þ br Þ

The equilibrium lipid and radius is the intersection of a cubic that vanishes for l ¼r0 and a positive rational polynomial which is positive, vanishes at zero and tends to zero when r tends to ∞ since n≥1. There is always an intersection but we need to ensure that the

Abrams, P., Katz, L.E.L., 2011. Metabolic effects of obesity causing disease in childhood. Curr. Opin. Endocrinol. Diabetes Obes. 18 (1), 23–27, http://dx.doi. org/10.1097/MED.0b013e3283424b37. Arner, P., Spalding, K.L., 2010. Fat cell turnover in humans. Biochem. Biophys. Res. Commun. 396 (1), 101–104, http://dx.doi.org/10.1016/j.bbrc.2010.02.165. Chow, C.C., Hall, K.D., 2008. The dynamics of human body weight change. PLoS Comput. Biol. 4 (3), e1000045, http://dx.doi.org/10.1371/journal. pcbi.1000045. de Graaf, A.A., Freidig, A.P., Roos, B.D., Jamshidi, N., Heinemann, M., Rullmann, J.A.C., Hall, K.D., Adiels, M., van Ommen, B., 2009. Nutritional systems biology modeling: from molecular mechanisms to physiology. PLoS Comput. Biol. 5 (11), e1000554, http://dx.doi.org/10.1371/journal.pcbi.1000554. Drolet, R., Richard, C., Sniderman, A.D., Mailloux, J., Fortier, M., Huot, C., Rhéaume, C., Tchernof, A., 2008. Hypertrophy and hyperplasia of abdominal adipose tissues in women. Int. J. Obes. (London) 32 (2), 283–291, http://dx.doi.org/10.1038/sj. ijo.0803708. Etherton, T.D., Thompson, E.H., Allen, C.E., 1977. Improved techniques for studies of adipocyte cellularity and metabolism. J. Lipid Res. 18 (4), 552–557. Flier, J.S., 2004. Obesity wars: molecular progress confronts an expanding epidemic. Cell 116 (2), 337–350. Gliemann, J., Vinten, J., 1974. Lipogenesis and insulin sensitivity of single fat cells. J. Physiol. (London) 236 (3), 499–516. Jo, J., Gavrilova, O., Pack, S., Jou, W., Mullen, S., Sumner, A.E., Cushman, S.W., Periwal, V., 2009. Hypertrophy and/or hyperplasia: dynamics of adipose tissue growth. PLoS Comput. Biol. 5 (3), e1000324, http://dx.doi.org/10.1371/journal. pcbi.1000324. Jo, J., Guo, J., Liu, T., Mullen, S., Hall, K.D., Cushman, S.W., Periwal, V., 2010. Hypertrophy-driven adipocyte death overwhelms recruitment under prolonged weight gain. Biophys. J. 99 (11), 3535–3544, http://dx.doi.org/10.1016/ j.bpj.2010.10.009. Kahn, S.E., Hull, R.L., Utzschneider, K.M., 2006. Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature 444 (7121), 840–846, http://dx. doi.org/10.1038/nature05482. Kursawe, R., Eszlinger, M., Narayan, D., Liu, T., Bazuine, M., Cali, A.M.G., D'Adamo, E., Shaw, M., Pierpont, B., Shulman, G.I., Cushman, S.W., Sherman, A., Caprio, S., 2010. Cellularity and adipogenic profile of the abdominal subcutaneous adipose tissue from obese adolescents: association with insulin resistance and hepatic steatosis. Diabetes 59 (9), 2288–2296, http://dx.doi.org/10.2337/ db10-0113. Lafontan, M., Berlan, M., 1995. Fat cell alpha 2-adrenoceptors: the regulation of fat cell function and lipolysis. Endocrinol. Rev. 16 (6), 716–738. Lay, S.L., Krief, S., Farnier, C., Lefrère, I., Liepvre, X.L., Bazin, R., Ferré, P., Dugail, I., 2001. Cholesterol, a cell size-dependent signal that regulates glucose metabolism and gene expression in adipocytes. J. Biol. Chem. 276 (20), 16904–16910, http://dx.doi.org/10.1074/jbc.M010955200. Liu, A., Sonmez, A., Yee, G., Bazuine, M., Arroyo, M., Sherman, A., McLaughlin, T., Reaven, G., Cushman, S., Tsao, P., 2010. Differential adipogenic and inflammatory properties of small adipocytes in Zucker obese and lean rats. Diabetes Vasc. Dis. Res. 7 (4), 311–318, http://dx.doi.org/10.1177/1479164110386126. MacKellar, J., Cushman, S.W., Periwal, V., 2009. Differential effects of thiazolidinediones on adipocyte growth and recruitment in Zucker fatty rats. PLoS ONE 4 (12), e8196, http://dx.doi.org/10.1371/journal.pone.0008196. MacKellar, J., Cushman, S.W., Periwal, V., 2010. Waves of adipose tissue growth in the genetically obese Zucker fatty rat. PLoS ONE 5 (1), e8197, http://dx.doi.org/ 10.1371/journal.pone.0008197.

H.A. Soula et al. / Journal of Theoretical Biology 332 (2013) 89–95

Mclaughlin, T., Sherman, A., Tsao, P., Gonzalez, O., Yee, G., Lamendola, C., Reaven, G.M., Cushman, S.W., 2007. Enhanced proportion of small adipose cells in insulin-resistant vs insulin-sensitive obese individuals implicates impaired adipogenesis. Diabetologia 50 (8), 1707–1715, http://dx.doi.org/10.1007/s00125007-0708-y. Mclaughlin, T., Deng, A., Yee, G., Lamendola, C., Reaven, G., Tsao, P., Cushman, S., Sherman, A., 2010. Inflammation in subcutaneous adipose tissue: relationship to adipose cell size. Diabetologia 53 (2), 369–377. Skurk, T., Alberti-Huber, C., Herder, C., Hauner, H., 2007. Relationship between adipocyte size and adipokine expression and secretion. J. Clin. Endocrinol. Metab. 92 (3), 1023–1033, http://dx.doi.org/10.1210/jc.2006-1055.

95

Smith, U., 1971. Effect of cell size on lipid synthesis by human adipose tissue in vitro. J. Lipid Res. 12 (1), 65–70. Soulage, C.O., Zarrouki, B., Soares, A.F., Lagarde, M., Géloën, A., 2008. Lou/c obesityresistant rat exhibits hyperactivity, hypermetabolism, alterations in white adipose tissue cellularity, and lipid tissue profiles. Endocrinology 149 (2), 615–625, http://dx.doi.org/10.1210/en.2007-0317. Zinder, O., Shapiro, B., 1971. Effect of cell size on epinephrine- and acth-induced fatty acid release from isolated fat cells. J. Lipid Res. 12 (1), 91–95.

Modelling adipocytes size distribution

tissue mainly resulting from more and bigger adipocytes (Drolet et al., 2008 ... Note that this data is provided for illustration purpose. Adipocytes size can be ...

658KB Sizes 1 Downloads 74 Views

Recommend Documents

Texture and Bubble Size Measurements for Modelling Concentrate ...
of reducing the high dimensional bubble size distribution data associated with them ... the froth in a froth flotation process,"SmartFroth 5", Adams & Adams Patent Attorneys ...... pendency typically being shown on a grade-recovery curve. ... to the

Module Size Distribution and Defect Density
observation, supported by several data sets examined, is that the module sizes often follow exponential distribution. We show how the two models used together.

1 Determination of Actual Object Size Distribution from ...
as focused beam reflectance measurement (FBRM), optical fiber probe, ...... Liu, W.; Clark, N. N.; Karamavruc, A. I. General method for the transformation of.

1 Determination of Actual Object Size Distribution from Direct Imaging ...
device. As illustrated in Figure 1, the larger the distance between the object and the imaging device, the smaller the size of the object image and hence, the ... However, a reduction in DOF may not always be possible and the sampling time will be ..

1 Determination of Actual Object Size Distribution ... - Semantic Scholar
-1. ) Size (mm). Simulated (Low mag. change). Reconstructed (Low mag. change). Actual p(R) ps(y) (Low deg. of mag. change) pr(R) (Low deg. of mag. change).

Effective use of bubble size distribution measurements
1.1 Machine vision systems for froth flotation. Flotation ... to monitor the state of the froth is that such a system is ... between an apple (condition one) or an orange.

The Size Distribution of Exporting and Non-exporting ...
distribution of exporting firms among Chinese provinces, in varying degrees, are less than 1 ... in 2007. This relationship between the power law exponents of exporting and non-exporting firms is in .... details are in an online appendix Table 5.

1 Determination of Actual Object Size Distribution from Direct Imaging ...
However, a reduction in DOF may not always be possible and the sampling time will be substantially increased with a ..... For normal application of the proposed technique where DOF is an optical property, the DOF of the imaging device will ..... Mart