Authors Accepted Manuscript

On the thermodynamics of smooth muscle

contraction
Jonas Stlhand, Robert M. McMeeking, Gerhard A.
Holzapfel
www.elsevier.com/locate/jmps

PII:
DOI:
Reference:

S0022-5096(15)30225-8
http://dx.doi.org/10.1016/j.jmps.2016.05.018
MPS2908

To appear in: Journal of the Mechanics and Physics of Solids

Received date: 27 October 2015
Revised date: 7 May 2016
Accepted date: 15 May 2016
Cite this article as: Jonas Stlhand, Robert M. McMeeking and Gerhard A.
Holzapfel, On the thermodynamics of smooth muscle contraction, Journal of the
Mechanics and Physics of Solids, http://dx.doi.org/10.1016/j.jmps.2016.05.018
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.

On the Thermodynamics of Smooth Muscle Contraction

Jonas Stalhand1 , Robert M. McMeeking2,3 , Gerhard A. Holzapfel4,5
1

Solid Mechanics, Department of Management and Engineering

Linkoping University, 58183, Linkoping, Sweden

Department of Mechanical Engineering & Materials Department

University of California Santa Barbara, CA 93106, USA

School of Engineering, University of Aberdeen, Kings College, Aberdeen, AB24 3UE, UK

4

Institute of Biomechanics
Graz University of Technology, Stremayrgasse 16-II, 8010 Graz, Austria, and
5

Norwegian University of Science and Technology (NTNU)

Faculty of Engineering Science and Technology, 7491 Trondheim, Norway
To appear in Journal of the Mechanics and Physics of Solids
May 15, 2016
Abstract
Cell function is based on many dynamically complex networks of interacting biochemical
reactions. Enzymes may increase the rate of only those reactions that are thermodynamically consistent. In this paper we specifically treat the contraction of smooth muscle cells
from the continuum thermodynamics point of view by considering them as an open system
where matter passes through the cell membrane. We systematically set up a well-known
four-state kinetic model for the cross-bridge interaction of actin and myosin in smooth
muscle, where the transition between each state is driven by forward and reverse reactions.
Chemical, mechanical and energy balance laws are provided in local forms, while energy
balance is also formulated in the more convenient temperature form. We derive the local
(non-negative) production of entropy from which we deduce the reduced entropy inequality and the constitutive equations for the first Piola-Kirchhoff stress tensor, the heat flux,
the ion and molecular flux and the entropy. One example for smooth muscle contraction is
analyzed in more detail in order to provide orientation within the established general thermodynamic framework. In particular the stress evolution, heat generation, muscle shorting
rate and a condition for muscle cooling is derived.

Keywords.

Continuum thermodynamics; smooth muscle cell; contraction; kinetic model;

constitutive equation; entropy inequality; Gibbs free energy

1 Introduction
Smooth muscle cells (SMC) are important physiological constituents in, for example, the arterial wall which make functioning of the cardiovascular system possible. They are responsible
for controlling long-term changes in extracellular matrix turnover and short-term changes in the
lumen diameter due to mechanical, chemical and neural signals [15]. However, in most arterial
wall models smooth muscle contraction is not considered, hence the constitutive response is
only valid for the passive state of smooth muscles. Detailed quantification of the influence of
smooth muscle contraction on the mechanical behavior of arteries is an open and promising
research area. In particular the coupling of myosin kinetics with both a mechanical model for
organ deformation and a membrane model incorporating action potentials is of pressing need.
Such developments should be coordinated with the analysis of a coupled constitutive framework
for the various relevant phenomena such as the generation of mechanical stress, the diffusion
of biochemicals and ions, heat flux, and the presence of entropy in the system. In addition,
the formulation should be suitable for incorporation within numerical methods such as a finite
element analysis to simulate realistic boundary value problems (BVPs) whose solutions are of
pressing need.
SMCs are contained in organs such as blood vessels, airways, intestines, the urinary bladder and the uterus. They are spindle-shaped and are typically 100 m long with a diameter of
about 5 m except near the nucleus where they are slightly thicker. Activation of muscle contraction can be initiated through different stimuli (mechanochemical or electrical). The primary
determinant of muscle contractility is intracellular free calcium Ca 2+ . Increases in the intracellular calcium concentration, partly regulated by ion channels which can open and close, leads
to contraction (Ca2+ concentrations of 600-800 nM in fully contracted smooth muscle) while
decreases lead to relaxation (100 nM in resting smooth muscle), see [29]. Smooth muscle shortening is then accomplished by contractile apparatus consisting of actin and myosin filaments,
and the relative sliding between the filaments caused by cross-bridging is responsible for contraction. For the structure and spatial organization of actin and myosin molecules in a SMC see,

for example, [5, 4, 11, 19] and the textbook on molecular biology of the cell by Alberts et al.
[1]; the precise relative arrangement of actin and myosin within the smooth muscle still awaits
discovery. It should be noted that smooth muscle contraction occurs much more slowly than in
a skeletal muscle, starting at 5-100 ms after the initial stimulus and taking up to 10 s, [28], or
even minutes [2, 24]. Furthermore, smooth muscle can shorten more than striated muscle.
Several models have been developed to capture particular aspects of the electrophysiologic,
mechanical and mechanochemical response of smooth muscle cells. In particular, models that
are concerned with the electrophysiology of the jejunum and the uterus are documented by,
for example, Poh et al. [21] and Tong et al. [32], respectively. Mechanical models of smooth
muscle contraction are documented in, for example, Murphy [17] and Gestrelius and Borgstrom
[7]. Recently, integrated continuum models for smooth muscle have been proposed that consider chemical, electrophysiological and mechanical phenomena. For a brief review of such
mechanochemical models describing the basic features of active vascular smooth muscle cells
using a continuum mechanical framework see Murtada et al. [18]; a more recent mechanochemical model not summarized in [18] is the one by Murtada et al. [20]. In these models the active
stress is governed by phosphorylated myosin as determined by the four-state kinetic model of
Hai and Murphy [10]. A pioneering model that incorporates mechanical, chemical and electrical components of smooth muscle cell function during the generation of active stress was proposed by Yang et al. [35], with inclusion of membrane characteristics. Recently, Sharifimajd
and Stalhand [26] established a continuum model for smooth muscle contraction that includes
cell membrane excitability and the kinetics of myosin phosphorylation. On the basis of [26]
Sharifimajd et al. [27] developed a novel electro-chemo-mechanical model to replicate uterine
smooth muscle cell excitation-contraction. On the basis of nonlinear continuum mechanics, a
few additional studies have documented phenomenological models that capture smooth muscle contraction [22, 36, 25]. Such models fail to relate the sequence of coupled physiological
processes to the underlying biophysical (microscopic) structure of the cell.
In the present paper we aim to analyze and discuss the coupled thermodynamical processes
associated with smooth muscle contraction which, in the present form, have not been docu-

mented before. We construct an analytic example designed to give better understanding of heat
generation and muscle shortening. In particular, we aim to formulate models for smooth muscle
contraction within the framework of continuum thermodynamics. Thereby, the smooth muscle
cell is considered as an open system where matter passes through the cell membrane. We start
by deriving an extended version of the well-known four-state kinetic model of Hai and Murphy [10]. Therein the transition from one state to another (and its reverse) is described by two
reactions, forward and reverse. Next we provide the chemical and mechanical balance laws
in local forms. Energy balance consists of the rates of change of internal energy and internal
mechanical work, plus a thermal contribution from external sources and, finally, chemical energy emanating from external surface and volumetric sources. From an explicit expression for
the entropy change we infer the local (non-negative) production of entropy from which in turn
we deduce a reduced entropy inequality and constitutive equations for the first Piola-Kirchhoff
stress tensor, the heat flux, the ion and molecular flux and the entropy. For practical reasons the
energy equation is recast into a temperature form, and with the resulting equation cases such
as isometric tension and isotonic contraction are studied. Finally, a representative example for
smooth muscle contraction is analyzed in detail. In particular, an explicit solution for the fourstate kinematic model is derived, the stress evolution, heat generation and muscle shorting rate
is deduced from the general thermodynamic framework and a condition for muscle cooling is
provided. In an appendix is studied the relationship between the rate of change of Gibbs free energy and that of entropy occurring during the energy transformations associated with the kinetic
model. This result depends on the chemical potentials and sources of ions and molecules.

2 Constitutive Smooth Muscle Model

In this section, within nonlinear continuum thermodynamics, we derive the four-state model
for myosin kinetics, then we provide the local chemical, mechanical and energy balance laws
(given in both entropic and temperature forms) and we deduce the related constitutive equations
and the entropy inequality.

2.1 Four-state model for myosin kinetics

Smooth muscle contraction is generated in a cyclic interaction between actin and myosin filaments. The interaction starts by myosin forming strong bonds to actin, so called cross-bridges,
which activate a force generating power stroke that translates the filaments relative to each
other. After the power stroke, myosin detaches from actin and returns to its initial state where
the cycle can start over again. Myosin must be phosphorylated to attach to actin and to perform the power stroke. The phosphorylation is intimately related to the intracellular calcium
ion concentration where an increase in the concentration promotes the forward reaction making
myosin phosphorylated while a reduction has the opposite effect. Hai and Murphy [10] proposed that myosin transforms among four discrete states: detached, dephosphorylated myosin
M; detached, phosphorylated myosin M p ; attached, phosphorylated myosin AM p ; attached, dephosphorylated myosin AM commonly referred to as the latch state. The transition among the
myosin states is given in Fig. 1. The figure is based on the four-state kinetic model in Walker
et al. [33] but is extended to account also for the reverse reaction in each transformation, see
also Lazalde and Barr [14]. These extra reactions are introduced since the principle of detailed
balance requires each process to be matched by its reverse at equilibrium.
For the model in Fig. 1, let kq+ and kq , q = 1, . . . , 8, denote the forward and reverse rate
constants, respectively. In general, we take a superscribed + to denote a forward reaction while
a superscribed denotes the reverse reaction. The top transition in Fig. 1, M to M p (and its
reverse), consists of two reactions: one catalyzed by myosin phosphatase
k+

1

M + Pi


Mp ,

(1)

k1

and one catalyzed by myosin light-chain kinase

k2+

M + ATP 

Mp + ADP.

(2)

k2

Chemicals such as Ca2+ , myosin light-chain kinase and calcium-calmodulin are excluded in
these reactions since they are considered to be catalysts and not species being converted in the

Figure 1: The kinetic model for myosin transformation including the four states: detached,
dephosphorylated myosin M; detached, phosphorylated myosin M p ; attached, phosphorylated
myosin AMp ; attached, dephosphorylated myosin AM. The transition among the states is controlled by the forward and reverse rate constants kq+ and kq , q = 1, . . . , 8, respectively, while
ADP, ATP and Pi denote adenosine diphosphate, adenosine triphosphate and inorganic phosphate, respectively.
reaction. The forward and reverse reaction rates for reaction (1) are given by
r1+ = k1+ [M][Pi ],

r1 = k1 [Mp ],

(3)

respectively, and for reaction (2) by

r2+ = k2+ [M][ATP],

r2 = k2 [Mp ][ADP],

(4)

where the brackets indicate molar concentrations. Let the total number of moles of myosin be
N and take the molar fraction of the four myosin states to be given by n M , nMp , nAMp , and nAM .
The concentration of detached, dephosphorylated myosin and detached, phosphorylated myosin
are then given by [M] = NM /V = nM N/V and [Mp ] = NMp /V = nMp N/V , respectively,
where NM and NMp are the number of moles of myosin in the M and M p states, respectively,
and V is the referential volume. Further, assume that the ATP and P i are held at a fixed

concentration by their synthesis and delivery to the smooth muscle cell. Similarly, assume that
ADP is also held at a fixed concentration by the mitochondria which recycles it by gathering it
up and energetically converting it back to ATP. From (3) and (4) the forward reaction increases
the concentration of Mp according to
n Mp N
= (r1+ r1 ) + (r2+ r2 )
V

 

= k1+ [M][Pi ] k1 [Mp ] + k2+ [M][ATP] k2 [Mp ][ADP]

 nM N 
 nMp N
= k1+ [Pi ] + k2+ [ATP]
k1 + k2 [ADP]
.
V
V

(5)

Dividing both sides of (5) by N/V and introducing the rate constants k 1 = k1+ [Pi ] + k2+ [ATP]
and k2 = k1 + k2 [ADP], we arrive at
n Mp = k1 nM k2 nMp .

(6)

The forward reaction depletes detached, dephosphorylated myosin and, as a consequence,

n M = k1 nM + k2 nMp .

(7)

The other transitions according to Fig. 1 are: for Mp to AMp (or its reverse),
k+


Mp 

AMp ,

(8)

k3

and
k+


Mp + ADP + Pi 

AMp + ATP,

(9)

k4

for AMp to AM (or its reverse)

k+


AMp 

AM + Pi ,

(10)

k5

and
k6+

AMp + ADP 

AM + ATP,

(11)

k6

and for AM to M (or its reverse)

k7+

AM 

M,

(12)

k7

and
k8+

AM + ATP


M + ADP + Pi .

(13)

k8

Repetition of the steps between eqs. (3) and (7) for all transformations gives

k2
0
k7
nM
n M (k1 + k8 )

nM
n M
k
(k
+
k
)
k
0
1
2
3
4
p
p

=
,

n
0
k3
(k4 + k5 )
k6
nAMp
AMp

n AM
k8
0
k5
(k6 + k7 )
nAM

(14)

where k3 = k3+ + k4+ [ADP][Pi ], k4 = k3 + k4 [ATP], k5 = k5+ + k6+ [ATP], k6 = k5 [Pi ] +

k6 [ATP], k7 = k7+ + k8+ [ATP], and k8 = k7 + k8 [ADP][Pi ].
The system in (14) is exactly the four-state kinetic model proposed by Hai and Murphy [10],
apart from k8 which is set to zero therein; see also [30] for the use witin a mechanochemical
context. Because nj (j = M, Mp , AMp , AM) is the fraction of myosin in state j, the sum of n j
over all state must equal unity, i.e.

nj = 1.

(15)

It is only the attached states, i.e. AMp and AM, which can sustain an external load. The
difference between these states is that the AMp state generates an active force by the power
stroke while the AM state contributes to load carrying only through a passive extension of the
cross-bridges.

2.2 Local balance laws

From a thermodynamical point of view the smooth muscle cell is an open system where matter
such as ions, nutrients, and metabolic waste products from its environment pass through the
cell membrane. There are several ions and molecules associated with the regulation of smooth
muscle contraction, for example, calcium and ATP. The rate of change of the concentration of
an ion or molecule Ni inside the cell is given by the local balance law
N i = Riext + Rich DivJi ,

i = 1, . . . n,

(16)

where Riext and Rich are external and chemical reaction sources, respectively. This equation,
therefore, identifies an explicit distinction between sources R iext , that bring ions into the representative volume element from outside and sources R ich , that represent chemical reactions
converting some species into others in the representative volume element. The ion and molecular flux per unit reference area is characterized by Ji and Div() is the material divergence
of (). Note that quantities such as material density and ion concentration may vary with both
position and time even though they are defined on the reference configuration. The reference
configuration is only an arbitrarily chosen domain whose topology remains invariant. Restrictions such as mass conservation or incompressibility are introduced on physical or experimental
grounds.
The equation of motion and the symmetry condition for the stress tensor are obtained from
the balance of linear and angular momentum, respectively, see Holzapfel [13], and the results
read with the assumption that the acceleration and body force are negligible
DivP = O,

PFT = FPT ,

(17)

where P is the first Piola-Kirchhoff tensor (engineering stress) and F denotes the deformation
gradient. The first relation in (17) is referred to as the equation of equilibrium in elastostatics.
Having stated the chemical and mechanical balance laws, we turn our attention to the energy
balance. Energy in the cell is stored and transported in the form of high-energy bonds in ATP.
When these bonds are hydrolyzed, an exogenic reaction splits ATP into adenosine diphosphate
(ADP), inorganic phosphate (Pi ), and release about 50 kJ/mole (see Alberts et al. [1]). ATP
is consumed, i.e. hydrolyzed at various places in the cell in order to drive processes in an
energetically unfavorable direction, for example, by myosin to break cross-bridges after the
power stroke or by calcium pumps to remove calcium ions from the cell.
In the present study we are not concerned with growing structures so that the localized form
of the energy balance can be stated as
E = P : F DivJQ + RQ

Div(Hi Ji ) +

i=1

i=1

Hi Riext ,

(18)

where E is the internal energy per unit reference volume and P : F is the rate of internal
mechanical work per reference volume. The sum DivJQ + RQ equals the classical thermal
contribution from external sources where JQ is the heat flux per unit reference area and RQ is
the heat source per unit reference volume. The final two terms in eq. (20) represent the chemical
energy from external surface and volumetric sources, respectively, where Hi is the partial molar
enthalpy. We use the last term in the manner indicated to avoid explicitly including chemical
reactions such as those that generate ATP and dispose of it. Note that eq. (18) only includes
the chemical energy from external sources since internal chemical reactions merely convert one
form of the internal energy to another. For later use, we prefer to rewrite the internal mechanical
work in terms of the symmetric second Piola-Kirchhoff stress tensor S by using the connection
P = FS, [13], so that
P : F = FS : F = S : FT F = S :

C
,
2

(19)

where S = ST has been used together with the right Cauchy-Green tensor C = F TF, while C
denotes the material time derivative of C. Hence, by using (19), eq. (18) can be rewritten as
n
n

E = S : DivJQ + RQ
Div(Hi Ji ) +
Hi Riext .
2
i=1
i=1

(20)

Next, we introduce the Helmholtz free energy per unit reference volume with the Legendre
transformation
= E ,

(21)

where is the entropy (amount of disorder) per unit reference volume and is the temperature.
We specify the Helmholtz free energy such that
= (C, Ni , , i ),

(22)

where Ni is the set of concentrations of ions such as calcium and molecules such as ATP or
ADP, while i are internal variables. Note that i is not restricted to a scalar quantity; it can
also be a vector or a tensor quantity, depending on the situation. Proteins such as actin are

10

not included in this set of concentrations because we assume the absence of remodeling and,
therefore, such concentrations remain constant.
Substitution of eq. (21) and (22) into (20) gives, after a rearrangement of terms, we get



C

: +

=
2

n
n
n
m

ext
DivJQ + RQ
Div(Hi Ji ) +
Hi Ri
 , i ,
Ni
Ni
i
i=1
i=1
i=1
i=1


S2
C

(23)

where ,  indicates the scalar product. This notation is chosen to accommodate internal
variables of various kinds (scalar, vector, tensor) in the mathematical exposition.
Using (16), we deduce from (23)





n 

: +
DivJQ + RQ
Div(Hi Ji )
DivJi

= S2
C
2

Ni
i=1

n 
n
m

ch

ext
Hi
Ri
R
 , i .
(24)
+
Ni
Ni i
i
i=1
i=1
i=1
Next, introducing the chemical potential i of of the ith component, i.e.
i =

,
Ni

(25)

and applying the divergence identity

Div(i Ji ) = i DivJi + Ji Gradi ,

(26)

together with the relationship H i i = i where i is the partial entropy [3], we find that
eq. (24) becomes






n

= S2
: +
Div JQ +
i Ji
C
2

i=1
n
n
n
m

ext
ch
Ji Gradi +
i Ri
i Ri
 , i .
+RQ
i
i=1
i=1
i=1
i=1

(27)

Finally, divide both sides of eq. (27) by and apply the divergence identity (26) with i and Ji

11

replacing and i Ji , respectively. After a rearrangement of terms, the result reads


n
n
JQ

RQ

+
+
= Div
i Ji +
i Riext

i=1
i=1


n

C
1
1

1
S2
: 2 JQ Grad
+
Ji (Gradi + i Grad)

C
2

i=1


n
m

1

1
1

ch
+

i Ri
 , i .

i=1
i=1 i

For an interpretation of the term ni=1 i Rich see the Appendix.

(28)

2.3 Constitutive equations, entropy inequality

In this section we use eq. (28) to obtain restrictions on the constitutive equations for muscle
contraction by expressing the entropy rate in terms of its external supply and internal production.
The terms in the first row of (28) are the entropy supplied from external sources while the terms
in rows two and three are the internal entropy production. The second law of thermodynamics
implies that the local production of entropy D int must be non-negative, i.e.
Dint 0,

(29)

for all admissible processes, see, for example, Holzapfel [13] and Gurtin et al. [8]. Hence, the
following inequality must hold for smooth muscle contraction


n

C
1
1
1

S2
: 2 JQ Grad
Ji (Gradi + i Grad)
Dint =

C
2

i=1


n
m

1

1
1

ch
+

i Ri
 , i  0,

i=1
i=1 i

(30)

heat fluxes JQ , ion and molecule fluxes Ji , and chemical reaction

for all admissible rates C,
rates Rich .
Assume that smooth muscle contraction is a decoupled process where each thermodynamic
force in eq. (30) is independent of the fluxes in other terms. The first term in eq. (30) is associated with the work done by the applied loads and we assume no viscous stress so that
S=2

.
C

(31)

12

For the second and third terms, we assume

JQ = k Grad,

(32)

Ji = ki (Gradi + i Grad) ,

(33)

and

where k and ki are positive parameters. Equation (32) is the classical Fourier law which forbids
heat from traveling up a temperature gradient. Similarly, eq. (33) forbids material traveling up
a chemical potential gradient, qualified by the gradient of temperature. Finally, for the entropy
term, we take
=

(34)

and back-substitute eqs. (31) to (34) into (30) to obtain the reduced entropy inequality

i Rich

i=1

 , i  0,
i
i=1

(35)

which must be satisfied in a smooth muscle contraction process.

2.4 The energy balance in temperature form

The energy equation can be recast into a temperature form. This is convenient for practical
reasons since thermal properties of different smooth muscles have been measured. To this end,
start by substituting the constitutive equations (25), (31) and (34) into (24). The result after a
rearrangement of terms reads


n
n
n

ext
= Div JQ +
Hi Ji + RQ +
(Hi i ) Ri
i Rich
+

i=1

i=1

i DivJi
 , i .
i
i=1

i=1

i=1

(36)

Next, using eq. (34) and the Helmholtz free energy (22), it is straightforward to compute
n
m

2
2

2

:C
 , i 

=
Ni
C
Ni
2
i
i=1
i=1
n
m

2
1

=
: C +
i N i + cv
 , i ,

i
i=1
i=1

13

(37)

Figure 2: Illustration of the kinematics associated with smooth muscle contraction. The total deformation gradient F between the reference and the deformed states is multiplicatively
decomposed into an active part Fa , linked to filament sliding, and an elastic deformation F e ,
related to the deformation of cross-bridges.
where the partial molar entropy i = /Ni and the specific heat cv = 2 /2 > 0 (see
Holzapfel [13]) have been introduced. Back-substitution of eq. (37) 2 into (36) together with
(16) gives

2
: C
C
i=1
i=1
n
n
m
m

ch
, i 
(i + i ) Ri +
(i + i ) DivJi +

 , i . (38)

i
i
i=1
i=1
i=1
i=1

= Div JQ +
cv

Hi Ji

+ RQ +

(Hi i i ) Riext +

Using the relationship H i = i + i and eq. (26), where i is replaced by Hi , we obtain

n
n

2
ch

:C
cv = DivJQ + RQ +
Hi Ri
Ji GradHi
C
i=1
i=1
m
m

, i 
+

 , i .
i
i
i=1
i=1

(39)

2.5 Isometric and isotonic contraction

Inspired by the kinematics used in nonlinear plasticity, we assume that the total deformation
gradient F can be multiplicatively decomposed into an elastic part F e and an active part Fa . We
postulate that [31]
F = F e Fa ,

(40)

14

see Fig. 2. The elastic deformation Fe is related with cross-bridge stretching while the active
deformation Fa is linked to filament sliding. Smooth muscle contraction is essentially a onedimensional process along the muscle fiber. Let the fiber direction at a point in the reference
configuration be given by the unit vector N and assume that the active deformation maps this
vector according to
Fa N = a N,

(41)

where N is a unit vector in an intermediate configuration and a > 0 describes the (pre)stretch
caused by muscle contraction. A simple choice for the active deformation which satisfies
eq. (41) is
Fa = a N N.

(42)

We model smooth muscle as having one internal variable namely its active deformation
(1 =)Ca = FT
a Fa . Its conjugate variable is the stress S a /2 = /C a and the work rate
per unit volume done by the smooth muscle becomes S a : C a /2.
By rewriting eq. (39) we obtain
n
n

S C
:
Hi Rich
Ji GradHi
2
i=1
i=1
a
Sa C a
C
:
+ Sa :
,

2
2

= DivJQ + RQ +
cv

(43)

where eq. (31) has been used. We note that Sa = S and therefore eq. (43) becomes

C
a)
= DivJQ + RQ + S : 1 (C
Hi Rich
cv
2
i=1
n

Ca
.

Ji GradHi + S :
2
i=1
n

(44)

Let us now study the consequence of (44) for smooth muscle activity. First we assume that
the smooth muscle is internally homogeneous such that GradH i = 0 and that there are no
external heat sources so that RQ = 0. We also assume that the behavior is adiabatic which gives
= 0. When the smooth muscle
DivJQ = 0. Now consider an isometric tension. For this case C

15

has reached steady state C a = 0 and (44) specializes to

=
cv

Hi Rich .

(45)

i=1

Equation (45) shows that the temperature change in an isometric contraction is due to the energy
transformation associated with the chemical reactions taking place. The net enthalpy change in
an isometrically contracting smooth muscle is negative (Lonnebro and Hellstrand [16]) and the
term on the right-hand-side of eq. (45), including the preceding negative sign, is positive and
the smooth muscle temperature will increase.
Next, consider a contraction or extension against a constant load, i.e. an isotonic behavior.
We assume that isotonic conditions imply a constant elastic cross-bridge deformation with C
a = 0, [9]. Hence, (44) becomes
C
n

Hi Rich .
cv = S :
2
i=1

(46)

In eq. (46), the situation is somewhat different. The first term on the right hand side is the
work rate of the applied load. Consider the case that the chemical reaction is not occurring.
In that situation only stretching of the smooth muscle under tension is possible. Therefore,

the term S : C/2

represents the work done by the applied load while stretching the smooth
muscle against the resistance of the myosin motors. As a consequence this work heats up the
smooth muscle and the temperature increases as given by the left hand side of (46). Now let
the chemical reaction take place such that the smooth muscle contracts. The second term of
the right hand side of (46) is then positive while the first term of the right hand side is negative
because the smooth muscle is under tension and S is performing work on (lifting) the load.
The reduced entropy inequality (35) shows that in this circumstance the right hand side of
(46) is greater than or equal to zero. The smooth muscle therefore heats up, i.e. its temperature
increases according to (46), but by an amount less than the heat released by the chemical reaction. The balance of the heat released by the chemical reaction is absorbed by the mechanical
work done by the smooth muscle against its isotonic load.

16

3 Representative Example
We now give a representative example for the contraction of the smooth muscle cells encompassing the remodeling of their myosin proteins, their generation of contractile stress and the
heat that is given off in the processes. To this end, we consider a uniaxial and shear free contraction, where the muscle fibers are aligned with the axial direction. Hence, the rate of mechanical
where P is the first Piola-Kirchhoff (engineering)

work is given by S : C/2

= P : F = P ,
stress and is the stretch.

3.1 Solution of the four-state kinetic model

We first give a solution to eq. (14) commencing from a down regulated state of the smooth
muscle cells in which all myosins are in the detached, dephosphorylated state, i.e. n M = 1.
The smooth muscle is then up-regulated at time t = 0, and phosphorylation of myosin and its
cross-bridging commences. We express nM from eq. (15) and insert it into eq. (14), then omit
the top row of the equation to obtain

(k1 + k2 + k3 ) k4 k1
k
n
n
k1
Mp 1
Mp

nAMp + 0 . (47)
n AMp =
k3
(k4 + k5 )
k6

k8
n AM
k8
k5 k8 (k6 + k7 + k8 )
nAM
For guinea pig taenia coli, we choose the relationships k 1 = k6 = 2k4 , k2 = k3 = k5 = 3k4 and
k7 = k8 = 0, [19]. These ratios are not identical to those used by Murtada et al. [19], but are
somewhat similar and have the advantage of enabling an exact solution to eq. (47). The resulting
behavior of this exact solution is qualitatively similar to the numerical solution obtained in [19]
and thus serves our purpose for the representative example. With initial conditions n Mp =
nAMp = nAM = 0 and thus nM = 1. The closed form solution reads

nMp

0.1

0.214

0.1

0.214

nAMp = 0.3 + 0.214 expk4 t + 0.3 exp5k4 t + 0.214 exp8k4 t . (48)

0.643
0.45
0.3
0.107
nAM

17

3.2 Stress evolution, heat generation and muscle shortening rate

We propose that the isometric first Piola-Kirchhoff stress, say P i , for these smooth muscles is
given by
Pi = (nAMp + nAM )Po ,

(49)

where Po is the maximum possible isometric stress. The current isometric stress is thus proportional to the sum of the fraction of the myosins that are actively cross-bridging or in the latch
state. We note that from (48) we get
nAMp + nAM = 0.75(1 expk4 t ) 0.107(expk4 t exp8k4 t ).

(50)

We represent this as
nAMp + nAM = 0.75(1 expk4 t )

(51)

since we expect the second term on the right hand side of eq. (50) to be unimportant as t grows,
and with eq. (51) we will obtain the relevant insights into the system behavior. As a consequence
of using eq. (51), the isometric stress P i becomes
Pi = 0.75Po(1 expk4 t ).

(52)

Walker et al. [33] note that smooth muscle cell behavior is adequately described by the Hill
relationship between load and rate of shortening, as proposed in [12], and which we provide for
our purposes in the form
(P + )(Pi + o ) = o (Pi + ),

(53)

where is the coefficient of shortening heat and o is the free contraction rate of smooth muscle
fibers, in our case a negative quantity since we define shortening to be a negative strain rate.
Given (52), we can rewrite (53) as
o (1 expk4 t ) + o ] = o [0.75Po(1 expk4 t ) + ].
(P + )[0.75P

18

(54)

The elastic response originates within the cross-bridges, and the elasticity of the cross-bridges
in smooth muscle is linear elastic (see, e.g., [6], [34]). This has also been noted in [23] that the
serial elasticity in muscle cells acts more like a spring rather than a rubber band. We therefore
assume a linear elastic behavior for the cross-bridges. Since the elastic cross-bridges deformation takes place with respect to the intermediate configuration, the stress tensor will be defined
therein. By using a pull-back operation, the cross-bridge stress can be transformed to the reference configuration, and reads [9]
1
T
1
T
S = F1
= C1
a C : Ee Ca ,
a C : (Fa Ee Fa )Fa

(55)

where C is the linear fourth-order stiffness tensor, and Ee is an Euler-Lagrange-like elastic strain
defined as
1
(C Ca ) .
2

Ee =

(56)

Equations (55) and (56) are specialized to uniaxial contraction by using their projections along
the fiber vector N. If the Poisson effect is negligible, the elastic strain and stress read, respectively,
Ee =


1 2
2a
2

(57)

and
S=

Y Ee
,
4a

(58)

where Y is an elastic modulus. For later use, by utilizing eq. (57), we also compute the strainrate of Ee , i.e.
E e = a a .

(59)

According to Hill [12], eq. (53) is accompanied by the heat generation Q per unit reference
volume given by
Q = Qi

a a
,

(60)

19

where Qi is the rate of heat generation in the isometric state and a a / is the heat due to
shortening, see Appendix B. The term a a / is thus the heat generation during shortening
in excess of that associated with isometric smooth muscle response. The term is a positive
constant and the negative sign in (60) is required because we define shortening to give rise to
negative values of a a /, in contrast to Hills usage that defines shortening to give rise to a
positive strain rate [12].
We assume now that the isometric heat generation Q i is associated only with cross-bridging
myosins that are actively cycling, so we introduce
Qi = Qo nAMp = Qo [0.3(1 exp5k4 t ) + 0.214(exp8k4 t expk4 t )],

(61)

where (48) has been used. The positive constant Qo gives the isometric heat generation per unit
reference volume when all available myosins are actively cross-bridging and cycling.
From eq. (60) and the mechanical work, the total rate of energy output of the smooth
muscles per unit volume is given with (58), (59) and (61) 2 as


a
a

a a
a
a
P
= Qi (P + )
S E e
= Q P = Qi (P + )

a a
Y d(Ee )2
= Qo [0.3(1 exp5k4 t ) + 0.214(exp8k4 t expk4 t )](P + )
4
,

2a dt

(62)

where the connection S = P/ has been used, [13].

We first consider this in the final steady state condition after the transients have died away,
with temperature uniform and fixed. The smooth muscles will be in an isometric state at constant stress and thus
= 0.3Qo .

(63)

Since the smooth muscle state is unchanging and the temperature is constant, this heat must
be conducted away from the smooth muscle, and in steady state must also equal the energy
supplied to the smooth muscle in the form of splittable ATP to fuel myosin cross-bridge cycling.
To provide closure to the equations, we assume that this rate of energy supply is constant and
determines the rate at which the smooth muscles can contract. As before, we assume constant

20

temperature and thus eqs. (63) and (62)4 together give

(P + )

a a
= Qi 0.3Qo = Qo [0.214(exp8k4 t expk4 t ) 0.3 exp5k4 t ],

(64)

where we have assumed that the rate of change of elastically stored energy is negligible due to
the stiffness of the smooth muscle fibers, i.e. (Y /2 4a)(d(Ee )2 /dt) 0. Now we combine the
Hill relationship (53) with (64) 1 to deduce the results


0.3Qo Qi
Pi ,
P = 1+
o

a a =

Qi 0.3Qo
o . (65)
(Pi + ) o (Qi 0.3Qo )Pi

Given the explicit expressions (52) and (61) 2 , these become




Po (1 expk4 t ),
P = 0.75 1

(66)

and
a a =

[0.75Po (1

expk4 t )

o ,
+ ] o 0.75Po (1 expk4 t )

(67)

where the abbreviation

= Qo [0.214(exp8k4 t expk4 t ) 0.3 exp5k4 t ]

(68)

has been introduced. In the case of the stress in eq. (66), it starts at zero and rises in a complicated manner to the final isometric value of 0.75Po , while the shortening rate in eq. (67)
commences with the value 0.3Qo / and rises in a complicated manner to the final isometric
state of zero. Note that the shrinkage rate 0.3Qo / is the value required so that the heat output
exactly equals the energy supplied at the beginning of the process, as can be confirmed from
eq. (51), since at the very beginning of up-regulation Q i = 0.

3.3 Requirement for muscle cooling

We now consider the energy balance in temperature form as in (39). Since the temperature is
held fixed, the entire right hand side sums to zero. We assume that the smooth muscles are
cooled fast enough so that the temperature remains uniform as well. We take 2 /C and

21

 2 /i to be zero for the want of any information on these derivatives. The result from
eq. (39) is then that the cooling must be given by
RQ =

Hi Rich P a .

(69)

i=1

The summation term accounts for all chemical reactions, including those remodeling the myosins
and controlled by eq. (48) plus the chemical reactions involved in cyclic cross-bridging of the
myosin motors. If all reactions are exothermic in the direction in which they proceed, the summation term will always be negative, indicating that cooling is required in the isometric state.
The second term on the right hand side is always positive, showing that shortening relieves
some of the need for system cooling, because the relevant energy is being removed from the
smooth muscles by their mechanical work.
If we consider that there are j 1 reactions, numbered from 1 to j 1 that play a role in the
operation of cross-bridge cycling, their contribution to the sum on the right hand side of eq. (69)
is the constant energy supply rate of 0.3Qo . Let the reactions numbered j through n be those
associated with myosin remodeling as controlled by eq. (48). Equation (69) can therefore be
rewritten as
RQ =

Hi Rich 0.3Qo P a.

(70)

i=j

If we assume that the energetic demand for phosphorylation per myosin is equal to the energetic
demand of cross-bridge cycling per myosin, then we deduce that
n

Hi Rich = Qo

i=1

n Mp + n AMp
,

(71)

where is the rate in cycles per second at which myosin cross-bridge cycling takes place in the
isometric state. Since eq. (48) gives us
n Mp + n AMp = 2k4 exp5k4 t ,

(72)

we conclude from (70) that the requirement for cooling is

RQ =

2Qo k4
exp5k4 t 0.3Qo P a ,

22

(73)

where the product P a is given by multiplying the right hand side of (66) with the right hand
side of (67). The product is zero at the beginning of up-regulation and in the steady isometric
state. Therefore, the demand for cooling is at first RQ = (0.3 + 2k4 /)Qo and falls in
magnitude by the time of isometric steady state to a value of 0.3Q o . However, the transient
may be non-monotonic because of the positive contribution of the term P a . Due to the
complexity of P a , and due to the many assumptions we have made in getting to eq. (73), we
have not investigated this detail of the cooling process.

References
[1] B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, and P. Walter. Molecular Biology
of the Cell. Garland Science, New York, 5th edition, 2008.
[2] A. Arner. Mechanical characteristics of chemically skinned guinea-pig taenia coli. Eur. J.
Physiol., 395:277284, 1982.
[3] H. B. Callen. Thermodynamics and an Introduction to Thermostatistics. John Wiley &
Sons, New York, 2nd edition, 1985.
[4] P. H. Cooke, F. S. Fay, and R. Craig. Myosin filaments isolated from skinned amphibian
smooth muscle cells are side-polar. J. Muscle Res. Cell Motil., 10:206220, 1989.
[5] R. Craig and J. Megerman. Assembly of smooth muscle myosin into side-polar filaments.
J. Cell Biochem., 75:990996, 1977.
[6] L. E. Ford, A. F. Huxley, and R. M. Simmons. Tension responses to sudden length change
in stimulated frog muscle fibres near slack length. J. Physiol., 269:441515, 1977.
[7] S. Gestrelius and P. Borgstrom. A dynamic model of smooth muscle contraction. Biophys.
J., 50:157169, 1986.
[8] M. E. Gurtin, E. Fried, and L. Anand. The Mechanics and Thermodynamics of Continua.
Cambridge University Press, New York, 2010.

23

[9] P. Haupt. Continuum Mechanics and Theory of Materials. Springer-Verlag, Berlin Heidelberg, 2002.
[10] C.-M. Hai and R. A. Murphy. Cross-bridge phosphorylation and regulation of latch state
in smooth muscle. J. Appl. Physiol., 254:C99106, 1988.
[11] J. Herreros. Coronary surgery. developments in the last decade. indications and results.
Rev. Esp. Cardiol., 58:11071116, 2005.
[12] A. V. Hill. The heat of shortening and the dynamic constants of muscle. Proc. R. Soc.
Lond. B, 126:136195, 1938.
[13] G. A. Holzapfel. Nonlinear Solid Mechanics. A Continuum Approach for Engineering.
John Wiley & Sons, Chichester, 2000.
[14] C.A. Lazalde and L. Barr. Four-state models and regulation of contraction of smooth
muscle. I. Physical considerations, stability, and solutions. Math. Biosci., 112:130, 1992.
[15] Q. Li, Y. Muragaki, I. Hatamura, H. Ueno, and A. Ooshima. Stretch-induced collagen synthesis in cultured smooth muscle cells from rabbit aortic media and a possible involvement
of angiotensin II and transforming growth factor-. J. Vasc. Res., 35:93103, 1998.
[16] P. Lonnbro and P. Hellstrand. Heat production in chemically skinned smooth muscle of
guinea-pig taenia coli. J. Physiol., 440:385402, 1991.
[17] R. Murphy. Mechanics of vascular smooth muscle. In Handbook of physiology, Vol II.
Waverley Press, Baltimore, 1980.
[18] S. Murtada and G. A. Holzapfel. Investigating the role of smooth muscle cells in large
elastic arteries: a finite element analysis. J. Theor. Biol., 358:110, 2014.
[19] S. Murtada, M. Kroon, and G. A. Holzapfel. A calciumdriven mechanochemical model
for prediction of force generation in smooth muscle. Biomech. Model. Mechanobiol.,
9:749762, 2010.

24

[20] S. Murtada, S. Lewin, A. Arner, and J. D. Humphrey. Adaptation of active tone in

the mouse descending thoracic aorta under acute changes in loading. Biomech. Model.
Mechanobiol., 2016. in press.
[21] Y. C. Poh, A. Corrias, N. Cheng, and M. L. Buist. A quantitative model of human jejunal
smooth muscle cell electrophysiology. PLoS One, 7:e42385, 2012.
[22] A. Rachev and K. Hayashi. Theoretical study of the effects of vascular smooth muscle
contraction on strain and stress distributions in arteries. Ann. Biomed. Eng., 27:459468,
1999.
[23] J. A. Rall. Mechanism of Muscular Contraction. Springer-Verlag, New York, 2014.
[24] C. M. Rembold and R. A. Murphy. Latch-bridge model in smooth muscle: [Ca 2+ ]i can
quantitatively predict stress. Am. J. Physiol., 259:C251C257, 1990.
[25] A. Schmitz and M. Bol. On a phenomenological model for active smooth muscle contraction. J. Biomech., 44:20902095, 2011.
[26] B. Sharifimajd and J. Stalhand. A continuum model for excitation-contraction of smooth
muscle under finite deformations. J. Theor. Biol., 355:19, 2014.
[27] B. Sharifimajd, C. J. Thore, and J. Stalhand. Simulating uterine contraction by using an
electro-chemo-mechanical model. Biomech. Model. Mechanobiol. 2016. in press.
[28] H. A. Singer and R. A. Murphy. Maximal rates of activation in electrically stimulated
swine carotid media. Circ. Res., 60:438445, 1987.
[29] A. P. Somlyo and A. V. Somlyo. Smooth muscle structure and function. In H. A. Fozzard,
E. Haber, R. B. Jennings, A. M. Katz, and H. E. Morgan, editors, The Heart and Cardiovascular System. Scientific Foundations, pages 12951324, New York, 1992. Raven
Press.

25

[30] J. Stalhand, A. Klarbring, and G. A. Holzapfel.

Smooth muscle contraction:

mechanochemical formulation for homogeneous finite strains. Prog. Biophys. Mol. Biol.,
96:465481, 2008.
[31] J. Stalhand, A. Klarbring, and G. A. Holzapfel. A mechanochemical 3D continuum model
for smooth muscle contraction under finite strains. J. Theor. Biol., 268:120130, 2011.
[32] W. C. Tong, C. Y. Choi, S. Kharche, A. V. Holden, H. Zhang, and M. J. Taggart. A
computational model of the ionic currents, ca 2+ dynamics and action potentials underlying
contraction of isolated uterine smooth muscle. PLoS One, 6:e18685, 2011.
[33] J. S. Walker, C. J. Wingard, and R. D. Murphy. Energetics of crossbridge phosphorylation
and contraction in vascular smooth muscle. Hypertension, 23:11061112, 1994.
[34] D. M. Warshaw and F. S. Fay. Tension transients in single isolated smooth muscle cells.
Science, 219:14381441, 1983.
[35] J. Yang, J. W. Clark Jr., R. M. Bryan, and C. Robertson. The myogenic response in isolated
rat cerebrovascular arteries: smooth muscle cell model. Med. Eng. Phys., 25:691709,
2003.
[36] M. A. Zulliger, A. Rachev, and N. Stergiopulos. A constitutive formulation of arterial
mechanics including vascular smooth muscle tone. Am. J. Physiol. Heart Circ. Physiol.,
287:H13351343, 2004.

Appendix A: Energy transformation within the kinetic model

Let us study the energy transformations associated with the four-state kinetic model and derive
the related rate of change in Gibbs free energy G. During a set of reactions it is the total change
in Gibbs free energy which determines whether or not the entire reaction sequence can occur.
To that end, we start with the first transformation from M to M p as given by the reactions (1)

26

and (2). The rate of change in the Gibbs free energy in this transformation ( G MMp ) is given
by
G MMp = M N M + Mp N Mp + ATP N ATP + ADP N ADP + Pi N Pi .

(A.1)

By using that NM = nM N together with the analogue expressions for the other species in (A.1),
we get


G MMp = N M n M + Mp n Mp ,

(A.2)

where it is used that n ATP = n ADP = n Pi = 0 since the concentrations of ATP, ADP, and Pi
are assumed constant. Substituting eqs. (6) and (7) into (A.1) and rearranging the terms gives



G MMp = N Mp M k1 nM k2 nMp .

(A.3)

Assuming ideal thermodynamics, the chemical potential for each species is given by
M = 0M + R ln([M]),

Mp = 0Mp + R ln([Mp ]),

(A.4)

where 0M and 0Mp are the standard chemical potentials for M and Mp , respectively, and R =
8.314 JK1mol1 is the gas constant. Substituting (A.4) into (A.3) gives





n
M
p
0
0
k1 nM k2 nMp .
G MMp = N Mp M + R ln
nM

(A.5)

At equilibrium G MMp = 0 and since both the reaction constants and the mole fractions in
eq. (A.5) are non-zero, equation (A.5) implies
eq 
nMp
,
0Mp 0M = R ln
neq
M

(A.6)

eq
where neq
Mp and nM are the mole fractions at equilibrium.

Further, at equilibrium n Mp = 0 and eq. (6) gives the equilibrium constant

K1eq

neq
Mp
neq
M

k1
.
k2

(A.7)

Alternatively, using eq. (7), the same result can be obtained with n M = 0 at equilibrium. Backsubstitution of eq. (A.7)2 into (A.6) and the result in (A.5) gives



k2 nMp 

k1 nM k2 nMp .
GMMp = NR ln
k1 nM

27

(A.8)

Repeating the calculation for the other three transformations, the total rate of change in the free
energy becomes



 


k4 nAMp 
k2 nMp 

k1 nM k2 nMp + ln
k3 nMp k4 nAMp
G = NR = ln
kn
k3 nMp



1 M 

k8 nM
k6 nAM 
k5 nAMp k6 nAM + ln
(k7 nAM k8 nM ) . (A.9)
+ ln
k5 nAMp
k7 nAM
The only chemical reaction taking place herein is the transformation of myosin in eq. (14). As
a consequence, we can identify
n

i Rich = G.

(A.10)

i=1

Appendix B: Rate of energy output during contraction

The heat generation Q during shortening is given by
,
Q = Qi W

(B.1)

is the extra shortening heat

where Qi is the rate of heat generation in the isometric state and W
due to contraction. In analogy with eq. (19) 3 , we define the shortening heat to be

: Ca ,
=S
W
2

(B.2)

is a material (second Piola-Kirchhoff) stress tensor which generates contraction. Using

where S
the definition Ca = FT
a Fa , where Fa = a N N, see eq. (42), the stretch rate in eq. (B.1) may
be computed as
a = 2a a N N.
C

(B.3)

Back-substitution of eq. (B.3) into (B.2) gives

=S
: a a N N = a a N SN.

(B.4)

Finally, define the contractile stress along the muscle fiber to be

= ,
N SN

(B.5)

28

and substitute (B.4) and (B.5) into (B.1), we get the requested relation (60). Note that is a
first Piola-Kirchhoff-like stress, see eq. (53), which must be divided by to obtain the material
as used in (B.4).
stress N SN,

29