Development of a 3D finite element model of lens microcirculation
Authors
Abstract
Background
It has been proposed that in the absence of a blood supply, the ocular lens operates an internal microcirculation system. This system delivers nutrients, removes waste products and maintains ionic homeostasis in the lens. The microcirculation is generated by spatial differences in membrane transport properties; and previously has been modelled by an equivalent electrical circuit and solved analytically. While effective, this approach did not fully account for all the anatomical and functional complexities of the lens. To encapsulate these complexities we have created a 3D finite element computer model of the lens.
Methods
Initially, we created an anatomically-correct representative mesh of the lens. We then implemented the Stokes and advective Nernst-Plank equations, in order to model the water and ion fluxes respectively. Next we complemented the model with experimentally-measured surface ionic concentrations as boundary conditions and solved it.
Results
Our model calculated the standing ionic concentrations and electrical potential gradients in the lens. Furthermore, it generated vector maps of intra- and extracellular space ion and water fluxes that are proposed to circulate throughout the lens. These fields have only been measured on the surface of the lens and our calculations are the first 3D representation of their direction and magnitude in the lens.
Conclusion
Values for steady state standing fields for concentration and electrical potential plus ionic and fluid fluxes calculated by our model exhibited broad agreement with observed experimental values. Our model of lens function represents a platform to integrate new experimental data as they emerge and assist us to understand how the integrated structure and function of the lens contributes to the maintenance of its transparency.
Keywords
Computational modelling Ocular lens Microcirculation Finite element Physiological opticsBackground
The microcirculation model was initially based on a combination of electrical impedance measurements and theoretical modelling [8–11]. The observation that the measured ionic currents were directed inward at the poles and outward around the equator [Figure 1A, led to the suggestion that these currents represent the external portion of a circulating ionic current that drives the internal microcirculatory system within the lens. Briefly, the working model is that the current, carried primarily by Na^{+}, enters at all locations around the lens via the extracellular space between fibre cells. Na^{+} eventually crosses the fibre cell membranes, and then flows from cell-to-cell towards the surface via an intracellular pathway mediated by gap junction channels [Figure 1B. The gap junction coupling conductance in the outer shell of differentiating fibres is concentrated at the equator [12, 13]. Hence, the intracellular current is directed to the equatorial epithelial cells. Here, the highest densities of Na^{+}/K^{+} pumps are located to actively transport Na^{+} out of the lens [14]. Thus, the intracellular current effluxes are highly concentrated at the equator causing the net current to be outward. At the poles, there is very little intracellular current so the net current is predominantly inward, along the extracellular spaces [Figure 1B. The driving force for these fluxes is hypothesized to be the difference in the electromotive potential of surface cells that contain Na^{+}/K^{+} pumps and K^{+}-channels, and inner fiber cells that lack functional Na^{+}/K^{+} pumps and K^{+}-channels and whose permeability is dominated by non-selective cation and Cl^{-} conductances [15]. This electrical connection together with the different membrane properties of the surface and inner cells causes the standing current to flow. In this model, the circulating current creates a net flux of solute that in turn generates fluid flow. The extracellular flow of water convects nutrients towards the deeper lying fiber cells, while the intracellular flow removes wastes and creates a well-stirred intracellular compartment.
Thus in this model, it is the circulating Na^{+} current that generates a circulation of fluid inside the lens. It is important to note that while the existence of the circulating ionic currents are firmly based on existing experimental data, circulating fluid flows in the lens have proven more difficult to measure. At present, these fluxes are only predicted to occur from indirect measurements and models of the measured electrical properties of the lens. The current model of the lens fluid dynamics is based on an equivalent circuit analysis of the microcirculation system [3, 16]. This analytical model inherently relied on approximate solutions and simplification of the underlying physics. To improve our understanding of the circulation system, here we have adopted a finite element modelling (FEM) approach. Using FEM, we have developed a 3D computer model of the microcirculation system that encapsulates the complex interplay of its important features and can also be solved numerically. This computer-based modelling approach allowed structural features such as fiber cell orientation, extracellular space dimensions and gap junction distribution to be included in the model. It also included functional information on the spatial differences in membrane permeability between surface and inner cells, thought to drive the circulating currents.
In this paper we describe our expansion of the equations that govern ion and fluid dynamics in the lens [2, 3, 10, 15–18] to 3D, and their subsequent implementation into a new FEM that encapsulates known structural and functional parameters of the mouse lens. This model is based on our continuous imaging and modelling iterative investigation into the fluid dynamics of the lens [19–25]. To test the ability of this computer model to reproduce the functional properties of the lens, we have used a series of experimentally derived boundary conditions [26–28] to allow it to be solved. We show that our model is capable of predicting the experimentally measured steady state lens properties and to generate circulating ion and water fluxes. Hence we believe that our computer model is a useful tool to study how lens structure and function influence its optical properties.
Methods
In this section, we first present the derivation of the fundamental mathematical equations, originally formulated by Mathias et al. to describe the microcirculation system [2, 3, 10, 15–18]. We then develop a computer mesh to represent the structure of the mouse lens to enable these equations to be implemented in 3D. Next we implement the model using the C++ programming language and solve the model using experimentally derived boundary conditions [26–28]. The resultant 3D model calculates ion concentration gradients, membrane potential gradients, and intra- and extracellular ion and water fluxes in different regions of the lens.
Derivations of general equations
Glossary of symbols used in this manuscript
Symbol |
Description |
Units |
---|---|---|
C |
concentration |
mol/ m^{3} |
D |
diffusion coefficient |
m^{2}/s |
R |
gas constant |
J/ (mol.K) |
e |
electron charge |
C |
E |
Nernst potential |
V |
F |
Faraday constant |
C/mol |
h |
extracellular cleft width |
m |
I_{max} |
maximum Na/K ATPase pump current density |
A/cm^{2} |
I_{p} |
Na/K ATPase pump current density |
A/cm^{2} |
j_{m} |
transmembrane flux |
mol/ (m^{2}s) |
g |
conductivity per membrane area |
S/m^{2} |
k_{B} |
Boltzmann constant |
J/K |
K_{0.5} |
half-maximal concentrations |
mol/ m^{3} |
L_{s} |
surface hydraulic permeability |
m^{3}/(N.s) |
L_{p} |
intercellular hydraulic permeability |
m^{3}/(N.s) |
Os |
osmolarity |
Osm/L |
p |
pressure |
Pas |
P_{m} |
membrane solute permeability |
m/s |
r |
equatorial radius of the model |
m |
a |
posterior radius of the model |
m |
b |
anterior radius of the model |
m |
s_{α} |
solute source |
mol/(m^{3}s) |
s |
fluid mass source |
1/m |
T |
temperature |
K |
V_{m} |
transmembrane potential |
V |
f |
body force |
N/Kg |
z |
valence |
- |
α |
solute species |
- |
ε_{0} |
permittivity of vacuum |
F/m |
ε_{r} |
dielectric constant |
S/m |
ζ |
zeta potential (cell membrane potential) |
V |
λ_{D} |
Debye length |
m |
Λ |
volume fraction |
- |
μ |
dynamic viscosity |
N.s/m^{2} |
ρ |
mass density |
Kg/m^{3} |
ρ_{m} |
membrane density |
1/m |
σ |
membrane reflectance |
- |
ϕ |
potential |
V |
u |
velocity |
m/s |
Fluid fluxes
Fluid flow parameters
Symbol |
Description |
Value |
Units |
---|---|---|---|
R |
Gas-constant |
8.314 × 10^{3} |
pJ/(nmol.K) |
ε_{0} |
Permittivity-of-vacuum^{5} |
8.854 |
pF/m |
ε_{r} |
Dielectric-constant-(water)^{5} |
80.4 | |
$\tilde{h}$ |
Extracellular-cleft-width^{2} |
40 |
nm |
μ |
Fluid-viscosity |
700 |
mPa.ms |
ζ |
Zeta-potential^{1} |
−15 |
mV |
λ_{D} |
Debye-length^{2} |
1 |
nm |
L_{p} |
Intercellular-hydraulic-permeability^{3} |
4.0 × 10^{−8} |
m/(Pa.s) |
L_{f} |
Fibre-cell-membrane-hydraulic-permeability^{3} |
4.0 × 10^{−8} |
m/(Pa.s) |
L_{s} |
Surface-hydraulic-permeability^{3} |
2.7 × 10^{−7} |
m/(Pa.s) |
σ |
Intercellular-membrane-reflectance^{1} |
1 | |
$\tilde{K}$ _{e} |
Extracellular-hydraulic-conductivity |
3.05 × 10^{−14} |
m^{2}/(Pa.s) |
$\tilde{k}$ _{e} |
Extracellular-electro-osmotic-coefficient |
1.45 × 10^{−8} |
m^{2}/(Vs) |
Ion fluxes
Solute transport properties
Symbol |
Description |
Value |
Units |
---|---|---|---|
kB |
Boltzmann constant |
1.380 × 10^{−11} |
pJ/K |
e |
Electron charge |
1.6 × 10^{−10} |
nC |
F |
Faraday constant |
9.648 × 10^{4} |
nC/nmol |
D_{Na} |
Free solution/cytoplasm Na^{+} diffusion^{1} |
1.39 × 10^{−6} |
mm^{2}/s |
D_{K} |
Free solution/cytoplasm K^{+} diffusion^{1} |
2.04 × 10^{−6} |
mm^{2}/s |
D_{Cl} |
Free solution/cytoplasm Cl^{-} diffusion^{1} |
2.12 × 10^{−6} |
mm^{2}/s |
D_{Nai,c} |
Intracellular Na^{+} diffusion |
1.39 × 10^{−8} |
mm^{2}/s |
D_{Ki,c} |
Intracellular K^{+} diffusion |
2.04 × 10^{−8} |
mm^{2}/s |
D_{Cli,c} |
Intracellular Cl^{-} diffusion |
2.12 × 10^{−8} |
mm^{2}/s |
D_{Nae,c} |
Extracellular Na^{+} diffusion |
1.39 × 10^{−6} |
mm^{2}/s |
D_{Ke,c} |
Extracellular K^{+} diffusion |
2.04 × 10^{−6} |
mm^{2}/s |
D_{Cle,c} |
Extracellular Cl^{-} diffusion |
2.12 × 10^{−6} |
mm^{2}/s |
g_{Na} |
Na^{+} fibre cell membrane conductivity^{2} |
2.2 |
mS/m^{2} |
g_{Cl} |
Cl^{-} fibre cell membrane conductivity^{2} |
2.2 |
mS/m^{2} |
g_{K} |
K^{+} surface membrane conductivity^{2} |
2.1 |
S/m^{2} |
Imax |
Na^{+}/K^{+} pump max. pump rate |
0.478 |
A/m^{2} |
K_{1/2Na} |
Na^{+}/K^{+} pump 1/2 max Na^{+} concentration |
9 |
mM |
K_{1/2K} |
Na^{+}/K^{+} pump 1/2 max K^{+} concentration |
3.9 |
mM |
We solved these equations [Eq. 1 to Eq. 3] for the movement of ions and water in the extracellular and intracellular spaces of the lens. These two domains were linked by cell membranes which could be crossed by ions and water from one space to another [Figure 1B]. Since the water and ions enter the lens from the extracellular space, we started with these equations.
Extracellular flux equations
The fluid flow in the extracellular clefts can be partially described by the Stokes flow equations [Eq. 1 & Eq. 2. Another part of the extracellular fluid fluxes has been shown to be due to electro-osmosis [39]. Electro-osmosis is due to the osmotic gradient created by uneven charge distribution in a fluid affected by an electric field. It has been shown that this osmosis is essential to the modelling of extracellular fluxes in the ocular lens [39].
The parameters and their units are listed in [Table 1. We didn’t consider electro-osmosis in the intracellular space portion of our model [18]. This was since the electric field gradient across cell cytoplasm was considered negligible. We then coupled the extracellular water fluxes with the ionic fluxes in this domain.
Extracellular ion fluxes
We computationally solved the advective Nernst-Plank equation [Eq. 3] for the lenticular extracellular ion fluxes. We modelled the coupled water and ion extracellular fluxes throughout the 3D model. These fluxes could then become trans-membranous flows at any point of the model, crossing the modelled cell membranes.
Trans-membrane flux equations
The parameters and their units are listed in [Table 1]. In the above equations, E is the Nernst potential. At the Nernst potential, there is no net flow of ions across the cell membrane through the channels. Hence, in our model we linked the trans-membrane water and ion fluxes by ionic concentrations (i.e. osmosis) and membrane potential (i.e. Nernst potential). After crossing the membrane, we treated the water and ion fluxes as intracellular fluxes.
Intracellular flux equations
We modelled the fluxes in the intracellular space to pass through a mixture of cells cytoplasm and membranes [44]. This was due to the large size of each element in our model compared to the lens cell volumes. It was impractical to discretely model the intracellular flow through each cell cytoplasm and membrane in a given element. Instead, we homogenized the flow equations to obtain one formula, describing the net cytoplasmic and membranous fluxes through every element.
Here ϕ is the angular degree from the equator (i.e. ϕ = 0° at equator, ϕ = 90° at anterior pole and ϕ = −90° at posterior pole). It has been determined that G _{ max } = 0.6mS/cm in the radial orientation, and in the other directions G = 0.23mS/cm [7]. We modelled the intracellular ionic fluxes to be directed by the above equations and the regional distribution of gap junctions [Eq. 15 - Eq. 17. According to the microcirculation theory, these ionic fluxes then moved towards the periphery of the lens accompanied by water flows until they reach the boundary of the model.
Surface flux equations
The parameters and their units are listed in [Table 1] and the subscript “o” points to the outside (i.e. boundary) conditions. We considered the surface fluxes as the output of the system, while the extracellular fluxes were its input. For an incompressible stable system, like the current lens model, the input and output levels should equate at all times and this conservation of mass was controlled for as part of the Stokes equations [Eq. 1].
We modelled the total surface ionic fluxes by combining the pump-generated and channel-based flows [Eq. 8, Eq. 23 & Eq. 24].
Electro-neutrality equation
We assumed the initial condition of the system to be electrically neutral. Hence, the equation above ensures the electro-neutrality of the model at any point of time. We applied this equation over all the ion species modelled here. This was based on the assumption that the ion species not modelled (e.g. calcium) had no significant influence on electro-neutrality of the model.
Solving the model
Finite Element Mesh Creation
Finite element mesh specifications for the current model
Modelling property |
Property value |
---|---|
Coordinate system |
Cylindrical Polar |
Basis function |
Cubic Hermit |
Number of nodes |
5281 |
Number of elements |
2130 |
Boundary conditions
Initial conditions at outer lens boundary for the present model
Species |
Description |
Quantity |
---|---|---|
Na_{e} (mM) |
Extracellular Na^{+} concentration |
110 |
K_{e} (mM) |
Extracellular K^{+} concentration |
8 |
Cl_{e} (mM) |
Extracellular Cl^{-} concentration |
115 |
Na_{i} (mM) |
Intracellular Na^{+} concentration |
7 |
K_{i} (mM) |
Intracellular K^{+} concentration |
100 |
Cl_{i} (mM) |
Intracellular Cl^{-} concentration |
10 |
T (K) |
Temperature |
310 |
Convergence criterion
We defined the maximum change of concentrations between two consecutive iterations (C _{ n+1 } - C _{ n }) in all the elements as the convergence parameter. We observed that using the above boundary and initial conditions, this error was decreased with each iteration cycle. We set the model’s convergence criterion to be less than 5 mM. We believed that level of iteration error was adequate, since most of the solution fluctuations were caused by [K^{ + }]_{ i } and [Na^{ + }]_{ e } fields and the enforced iteration error threshold was less than 5% of the initial field value.
We solved the 3D model on our high performance computer (HPC) at the Auckland Bioengineering Institute (ABI). Our mainframe was comprised of an IBM server with 64 processors of 1.9 GHz calculation speed, peak performance of 306.21Gfps and 256 GB of physical memory.
Displaying the data
Text formatted file of the model’s computed fields were linked via JAVA programming language format to CMGUI ( http://www.cmgui.org) [48, 49], an advanced 3D visualisation software package with modelling capabilities. CMGUI was used for model visualisation and manipulation and allowed for automated the scaling and false colouring of the data. The following presentations were all created using this method.
Results
Using the boundary conditions listed in [Table 5], we solved the model and generated 3D maps of standing fields of intracellular and extracellular ion concentrations, electrical potentials and pressure, plus circulating ionic and water fluxes. In this section we first use quarter section views of the 3D model [Figure 3C] to represent regional differences in standing electrochemical and pressure fields, allowing these predicted properties to be compared to experimentally derived values. Then we use the full model view to generate 3D vector maps that visualize the predicted ion and water fluxes in the lens for the first time.
Standing electrochemical and pressure fields
Comparison of modelled standing electrochemical gradients outcomes and the existing empirical data
Standing fields |
3D model radial range |
Experimental radial range |
3D model averaged |
Experimental averaged^{*} |
---|---|---|---|---|
[Na^{+}]_{i} (mMol) |
6.9 to 16 |
5 to 16 ^{1} |
12.5 |
13.8 ^{3}, 21.5 ^{4} |
[K^{+}]_{i} (mMol) |
91 to 100 |
Not Measured |
95 |
91.2 ^{5} |
[Cl^{-}]_{i} (mMol) |
10 to 12 |
Not Measured |
11.35 |
16.8 ^{5}, 12.3 ^{5} |
ϕ_{i} (mV) |
−57 to −64 |
−50 to −70 ^{2} |
−61.5 |
−70 ^{6} |
ϕ_{e} (mV) |
0.0 to −27 |
0 to −36 ^{2} |
−13.5 |
−30 ^{6} |
P_{i} (kPa) |
0.0 to 10 |
0.0 to 48 ^{7} |
4.8 |
Not Measured |
Circulating fluxes
Comparison of modelled circulating currents and the existing experimental mouse data
Parameter |
Model prediction |
Experimental data |
---|---|---|
Net current density (μA/cm^{2}) | ||
J_{equator} |
17, outward |
22 ^{1}, 20 ± 2.6 ^{2}, outward |
J_{anterior pole} |
12, inward |
26 ^{1}, 32 ^{2}, inward |
J_{posterior pole} |
10, inward |
7 ^{1}, 42 ^{2}, inward |
Fluid velocity (nm/s) | ||
V_{equator – intracellular} |
2.1, outward |
Not Measured |
V_{anterior pole – extracellular} |
190, inward |
Not Measured |
V_{posterior pole – extracellular} |
250, inward |
Not Measured |
Discussion
It has been proposed that the circulating currents observed experimentally in lenses from a variety of species [5, 14, 41, 52] drive an internal microcirculation system. In the absence of a blood supply the microcirculation delivers nutrients and removes metabolic waste products from inner fibre cells, while maintaining ionic homeostasis [7]. The circulating fluxes are thought to be the net result of spatial differences in the location of ion channels and transporters that determine local membrane permeability, and the density of gap junction channels that direct intercellular fluxes within the lens [3]. We first created a 3D mesh of the lens structure. We then implemented a series of equations that describe the local transport properties of cells in different regions in the lens. Finally we solved these equations using a FEM approach at each location of the model. In summary, we have produced a computer model of the lens that not only accurately predicts standing ionic concentrations [Figure 4 and electrical [Figure 5 gradients and the existence of a pressure gradient [Figure 6, but also produces the first 3D vector maps of predicted current [Figure 8 and fluid [Figure 10 fluxes inside the lens. The observed agreement between experimentally measured values and those calculated by our model [Table 6 & Table 7, suggest that our computer model is mimicking lens physiology and generates a microcirculation. However, the underestimation of the magnitude of the intracellular pressure gradient highlights the fact that the model is only as robust as its underlying assumptions and will require additional refinement and revision as new experimental data on lens structure and function becomes available. In the following sections we discuss the assumptions and limitations of our current model with the view to highlight areas where further refinements of the model are required.
The solute flux in a fluid is governed by diffusion, electro-diffusion (if the solute is charged) and advection. Diffusion is the random walk of particles due to Brownian motion. Electro-diffusion is the flux of a charged particle due to the force applied by an electric field. Advection is the transport of a solute by a fluid that is moving. The physical origins of these transport processes and the derivations of the equations have been discussed previously in the literature [31, 32]. The implementation of these equations in a computational platform is very well explained elsewhere. In this study, we implemented the driven equations and solved for a set of converged-upon 3D fields. This method produced the calculated standing ionic concentration gradients in the lens predicted by our model [Figure 4.These gradients in turn gave rise to a trans-membrane electric-potential gradient field [Figure 5, based on the Hodgkin–Huxley model. Consider a cell membrane that is not equally permeable to all the present ionic species on either side of it. Here, the permeable ions will tend to move down their concentration gradient taking their electrical charge with them as they go. Therefore, an electrical potential will be generated, which will drive them in the reverse direction. An electrochemical equilibrium will be reached when the diffusive force equals the electromotive force.
Our model has calculated a hydrostatic pressure gradient [Figure 6, the existence of which has recently been confirmed experimentally [29]. While the orientations of the calculated and measured pressure gradients are similar, they differ in magnitude (48kPa compared to 10kPa). It has been proposed that this pressure gradient is generated by the restricted flow of water from the centre to the periphery of the lens by a pathway mediated by gap junction channels; since genetic manipulations to increase or decrease gap junction numbers produces inverse changes in the pressure gradient. This illustrates that structural components of the lens can influence the magnitude of the pressure gradient. This suggests that the difference between calculated and measured pressure fields may reflect the absence of a structural feature not currently captured in our model. In this regard we have recently identified a zone in the inner cortex of the lens that exhibits a reduction in the penetration of solutes and water [19, 22] that could influence the magnitude of the calculated pressure gradient. Future updates of the model will include such newly discovered structural elements, allowing their effect on the calculated pressure fields to be assessed.
The electrochemical fields, combined with the hydrostatic pressure gradients [Figure 6 in our model generate the circulating ionic currents throughout the lens [Figure 8 & Figure 9. The ionic fluxes are accompanied by water flows in the microcirculation system. The water fluxes are generally described by the Navier–Stokes equations, which are derived from the conservation of mass, momentum, and energy principals. Also, the fluid flow in the lens can be described as slow or low-Reynolds number flow. It is also reasonable to assume that the fluid flow in a normal lens is near or at steady-state at all times [18]. These simplifications reduce the Navier–Stokes equations to the Stokes equations. The derivation and implementation of these equations into a computational framework is discussed elsewhere [16, 18]. In our model, solving this set of equations results in the calculation of the water flow velocity fields [Figure 10. Although the current model appears to accurately mimic the physiological homeostasis of the lens [Table 6 & Table 7, there are still some aspects that need future improvement.
Our model presently has been implemented with a line suture structure which runs from the anterior to the posterior of the lens, through its core [Figure 3A&B]. However the mouse has a Y-shaped suture that rotates 180 °C from the anterior to the posterior pole [42]. Hence, we are planning to improve the current model with an anatomically accurate asymmetrical 3D structure of the sutures. In our current model, the 3D extracellular solute diffusion coefficients are constant throughout. However, we [21, 53] and others [54–56] have recently identified a barrier to the movement of solutes in the lens, using variety of techniques. We strongly believe that the existence of this barrier is very important in shaping the fluid dynamics of the lens and in general its microcirculation. Hence, we will implement the imaged barrier as a part of continuous improvement of our computational model.
Another important structural feature of the ocular lens is the presence of a Gradient of Refractive Index (GRIN), which acts to correct for inherent spherical aberration to improve the optical properties of the lens [57]. This GRIN profile of the lens is directly dependant on the local water/protein concentration makeup of the lens [58, 59]. Recently, we have experimentally showed that the water/protein gradient of the lens is actively upheld by the microcirculation system [19]. Hence, another step in the improvement of the current model is to add equations to estimate the concentration gradient of water in the lens. Using that calculated field, we will be able to estimate the 3D GRIN maps of the lens. We would then use these 3D GRIN maps and optical ray-tracing software to produce a model that links lens physiology to the optical properties of the lens.
Our first generation 3D finite element model of lens structure and function describes ion and fluid dynamics in the mouse lens. We chose to model the mouse lens as ion and fluid dynamics have been extensively studied in this species [3, 4, 15, 16]. We also believe the model is an essential first step towards creating a comprehensive model of the human lens. Any model of the human lens model would need to include its more complex structural features and would need to be created so that its dimensions could be altered to study the effects of lens growth and ageing on the circulation system [60]. This future model would enable us to study the changes in lens physiology thought to underlie the initiation of age related nuclear cataract.
Conclusion
During this project, a 3D model of the flux movements inside the mouse ocular lens was designed and executed using our high performance computer (HPC). Reviewing the results of the current model, it appears that solute fluxes, accompanied by water, enter the lens via the extracellular space all around it but with larger magnitudes around the polar regions. Among these inwardly extracellular solute fluxes, the Na^{+} fluxes were seemed to be dominant followed closely by Cl^{-} fluxes. Conversely, the solute effluxes appear to be via the intracellular space and seemed to be more pronounced around the equatorial region of the lens. The K^{+} fluxes were found to be the primary intracellular fluxes, caused mainly by exterior Na^{+}/K^{+} ATPase pumps. The net effect of these influx and effluxes were thought to be best explained by the calculated net current densities. The pattern of these net current densities at the surface of the lens was similar to previous experimental findings [5, 6, 52]. Same fields modelled inside the lens found to be in agreement with the microcirculation theory of the lens [3, 9, 16, 61].
This study brings together all the available experimental and theoretical data on the fluid dynamics of the ocular lens in order to create a comprehensive 3D model of this tissue. Previous studies have investigated the links between the steady state fluxes in the lens and its physiological homeostasis [62–65]. Using our computational model, we would be able to study the connections between the biodynamic natures of these perturbations and their functional consequences.
Acknowledgments
We wish to thank Prof Richard T Mathias for his insightful inputs in the development of the current model.
Supplementary material
Copyright information
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.