Chemical Potential, Diffusion, Osmosis PDF

Diffusion and osmosis The text under the slides was written by György Vámosi 2019 This instructional material was prepared for the biophysics lectures held by the Department of Biophysics and Cell Biology...

Diffusion and osmosis The text under the slides was written by György Vámosi 2019 This instructional material was prepared for the biophysics lectures held by the Department of Biophysics and Cell Biology Faculty of Medicine University of Debrecen Hungary https://biophys.med.unideb.hu 2 3 4 According to the equipartition theorem (=equal sharing), in thermal equilibrium the internal energy of a system is shared equally among its energy storing degrees of freedom on time average. These energy storing degrees of freedom are e.g. the translational motion of a molecule along the axes of the coordinate system, or the rotation around different axes. Each degree of freedom stores ½ kT energy on time average (k=1.38x10-23 J/K, Boltzmann’s constant; T (Kelvin): absolute temperature). Thus, the absolute temperature is a measure of average energy per degree of freedom. The translational motion of a molecule along the three axes stores on average 3/2 kT energy. The kinetic theory of gases predicts that the velocity of a noble gas follows Maxwell-Boltzmann distribution, and the so-called root mean square velocity (square root of the average of the squared velocities of the molecules) is 𝑣̅ = √3𝑘 𝑇⁄𝑚. This can be easily seen, given that Ekin=1/2 mv2 =3/2 kT (according to the equipartition theorem). With increasing T the whole distribution and its maximum shifts toward higher velocities. Note that the peak velocity is similar to the speed of sound in air, since sound propagates via collisions between the gas atoms. The kinetic behavior of molecules helps us understand Brownian motion, which is propelled by the thermal energy during which particles collide and exchange energy and momentum. 5 Adolf Fick (1829-1901), a German physicist and physiologist made observations about the spontaneous dispersion of particles in fluids. In his experiment he layered a dye solution below pure water, without mixing. First, the sharp boundary disappeared, and after a few weeks the dye was distributed evenly in the whole volume. Here, the solute moved from a region of high concentration to a region of low concentration spontaneously via a process called diffusion. Diffusion is the net transport of particles due to the inhomogeneity of concentration or other factors (temperature, solubility), propelled by Brownian motion (thermal motion). 6 Fick’s 1st law describes the diffusive flux under the assumption of steady state, i.e., assuming that the concentration does not change with time. It postulates that the flux goes from regions of high concentration to regions of low concentration, with a magnitude that is proportional to the concentration gradient (spatial derivative of the concentration). The negative sign expresses the direction of net diffusion of a solute (from high to low concentration). Such a situation can occur e.g. when the extracellular and intracellular space are separated by the membrane, and in the cell the transported material is consumed (or stored), while in the ec. space blood circulation replaces the material that diffused into the cell. Thus, the ic. and ec. concentrations can remain constant. 7 The question arises whether the direction of net diffusion (which way do more molecules move) is governed by some external force, somehow originating from the different concentrations, or it is determined by the random motion of particles. It turns out that we can explain the observation that net diffusion goes from regions of high concentration to regions of low concentration by simple probabilistic considerations. If the concentration in compartment 1 (on the left) is larger than in compartment 2 (right), c1>c2, and we assume that every particle can move in any of the six principal directions (±x, ±y, ±z) with equal probability, then more particles will move from left to right (from compartment 1 toward compartment 2) than vice versa, simply because there are more particles in compartment 1 than in 2 (N1>N2) originally. Thus, we do not need to assume that there is a force directing the motion of the particles, we can explain the direction of diffusion by simply assuming random motion of particles. This is the molecular (or we could say probabilistic) interpretation of diffusion. It is important to note that the motion of every particle is random; however, because of the above considerations, the direction of net flow is determined by the difference in concentrations! 8 We can also explain diffusion in terms of classical thermodynamics. According to the 2nd law of thermodynamics, the total entropy of a system plus its environment increases until reaching an equilibrium. This increase of total entropy is equivalent to the decrease of the Gibbs free energy of the system alone, if the system is at constant temperature and pressure (which holds approximately true for a living organism). Thus, the chemical potential, which is the molar Gibbs free energy, also tends to a minimum. The formula of the chemical potential is µ=µ0(T,p)+RT×ln(c), where c is the molar concentration of the solute. It can be seen from the equation that the total Gibbs free energy of the system decreases if particles move from a region of high concentration (or temperature) to a region of low concentration (or temperature or pressure). The chemical potential is a kind of potential energy, which tends toward a minimum. For neutral particles the difference in chemical potentials determines the direction of net diffusion. If particles also have an electric charge and they are in an electric field, the chemical potential is supplemented with an electric potential energy term: QΨ (product of the electric charge and the electric potential). For a mol of particles the charge is Q=zF, and the electric potential energy is zFΨ where z is the 9 charge number (e.g. +1 for Na+, +2 for Ca2+), F is Faraday’s constant (and Ψ is the electric potential (in the ic. or ec. space). The term µ=µ0(T,p)+RT ln(c)+zFΨ is called the electrochemical potential. Its gradient will determine the direction of the ion flux across ion channels in the plasma membrane. The net current of a given ion will flow from a region of high electrochemical potential to a region of lower electrochemical potential. At given values of the ec. and ic. ion concentrations the change of membrane potential (voltage across the membrane, i.e., the difference Ψic-Ψec may revert the current. The electrochemical potential is the basis of understanding ion equilibria and currents between the two sides of the membrane! 10 The diffusion coefficient characterizes the mobility of microscopic particles. It depends on the size and shape of the particle, the viscosity of the medium and the temperature. The Stokes-Einstein equation says that for a spherical particle of radius r, in a solution of viscosity η (pronounce: eta), at absolute temperature T, the diffusion coefficient is D=kT/(6πηr). More generally, for any shape of molecule, D=kT/f, where f is the form or friction factor. The name expresses that the diffusional motion is slowed by friction between the particle and the solvent. The temperature dependence is dual: it is direct through the kT term, and indirect through the viscosity, which decreases with increasing temperature (think of heated honey). The molecular mass M is proportional to the volume or r3, and D is inversely proportional to r, 3 thus, D is inversely proportional to √𝑀. This means that an 8-fold increase of the mass causes only a 2-fold decrease of D. The shape also affects D; if two molecules have the same volume, then the one with the larger surface area will have a larger friction with the solvent molecules, therefore a lower mobility and lower D. This can be seen through the example of spherical serum albumin and elongated, rod- like tropomyosin. 11 Fick’s 2nd law describes diffusion in general, when the concentration changes not only as a function of the coordinates but also in time (non-stationary diffusion). A simple example is when a droplet of dye diffuses in a bowl. The equation contains the first derivative of c with respect to time, and the second derivative of c with respect to the spatial coordinates (in the 1D case with respect to x). The solution of this so-called partial differential equation is the function c(x,y,z,t) or in 1D, c(x,t) describing the distribution of concentration in space and time. The solution depends on the initial conditions (the distribution c(x) at t=0) and the boundary conditions (defined by the shape of the reservoir in which the particles are confined, or the presence of other internal boundaries). 12 In the most simple case, when particles start diffusing from the origin and there are no boundaries present (a droplet of dye diffuses in an infinite container), the c(x,t) solution of Fick’s 2 nd law follows a Gaussian distribution. At t=0 all particles are in a narrow neighborhood of width “l” of the origin. At later time points they follow a Gaussian distribution with an increasing width (increasing standard deviation σ). The area below the Gaussian function is constant corresponding to the constant amount of diffusing particles. As can be seen from the formula of c(x,t), the denominator of the exponent, 4Dt increases with time. We know from biostatistics that for a normal distribution this term is equal to 2σ2, i.e., twice the variance of the distribution. The variance in this case is the average of the squared displacement of particles from the origin; therefore, we can write the Einstein-Smoluchowski equation for 1D: ̅̅̅ 𝑥 2 = 2𝐷𝑡, the average squared displacement is proportional to the diffusion coefficient and the time elapsed from the beginning of the diffusion process. 13 We can extend the E-S equation to describe the total displacement ̅̅̅̅̅ 𝛥𝑟 2 in 2D (membrane) and 3D (e.g. cytoplasm, nucleus). The time dependence of diffusive motion is quite different from that of directed motion with constant velocity. We can express the time needed to cover a certain distance by diffusion as: 𝑡 = ̅̅̅̅̅ 𝛥𝑥 2 ⁄2𝐷. We can see that t is proportional to the square of the distance, (not the distance itself like in case of motion at constant velocity). The difference in behavior arises from the nature of diffusive motion: it is random motion resulting from continuous collisions between molecules leading to a zig-zag path and an ever changing velocity, in stark contrast to straight line motion at constant velocity. Keep in mind that diffusion is a stochastic (random) process, we cannot predict the path of individual molecules, we can only predict the average behavior of a large ensemble of molecules. Because of the squared dependence of the average diffusion time on the distance to be traveled, diffusion is sufficiently fast for short distances, but becomes exceedingly slow for longer distances. This can be exemplified for a molecule the size of glucose in a log-log plot of average displacement vs. time: Covering a distance of 1 µm takes a few milliseconds, to diffuse 1 cm it takes about a day, to diffuse 1 m takes about a year. Therefore, diffusion is sufficiently fast for short range transport, e.g. inside the cell or from the capillaries to surrounding tissues over a few tens of µm- s, but not fast enough for long range transport. For long range transport we need the circulatory system. Even in a large cell such as a neuron, ATP-driven directed motion is needed. Axonal transport uses motor proteins (kinesins and dyneins) carrying their cargo along microtubules. 14 A typical example for the occurrence of diffusion in biological transport is the exchange of O2 and CO2 between red blood cells and alveoli in the lung. These gas molecules can easily penetrate membranes because of their small size, and cover the ~1 µm distance across the alveolo-capillary membrane. Using the Einstein-Smoluchowski equation for 1D (the path to be overcome is perpendicular to the membrane), and knowing the diffusion coefficients of the gas molecules, we can calculate the average diffusion time, which is on the order of tens or hundreds of milliseconds. The time spent by a RBC in the capillary of the alveolus is ~0.3 s, which is several orders of magnitude longer than the time needed for the diffusion process to take place. 15 Diffusion is the main mechanism of short range transport for many processes in living organisms. After circulation carries molecules close to their destination through the blood vessels, the last tens of micrometers between the capillaries and the surrounding tissues are covered by diffusion during the transport of nutrients, metabolites, gases, hormones, drugs, ligands, etc. The constituents of membranes: lipids and proteins diffuse laterally and also perform rotational diffusion in the plane of the membrane. These processes are important in the assembly of multi-subunit proteins (such as many membrane receptors or ion channels), and in transmembrane signaling processes. Diffusion inside cells is required for molecular recognition processes and biochemical reactions. Reactions are brought about by diffusion-propelled collisions between reaction partners such as an enzyme and its substrate. Transmembrane transport taking place between the extra- and intracellular space has 3 main mechanisms. Simple diffusion does not require external energy, it goes “downhill” the concentration gradient. It is the mechanism of transport for gases and uncharged hydrophobic molecules such as steroid hormones. Facilitated diffusion takes place with the assistance of transporter molecules. It is also a passive mechanism not requiring external energy, and it also goes downhill the concentration (in the presence of electric field, downhill the electrochemical potential) gradient. Molecules or ions that cannot penetrate the membrane by simple diffusion due to their charge or hydrophilic character 16 (e.g. glucose, amino acids, different cations, anions, etc.) are transported by this mechanism through the hydrophilic pore of transport proteins. Active transport also uses transport proteins; however, molecules or ions are transported “uphill”, against their electrochemical potential gradient. It requires external energy, and has two main forms. In primary active transport energy is produced directly by ATP hydrolysis by the transporter, which has ATP hydrolyzing activity. In secondary active transport the energy needed for the uphill transport of the substrate is provided by the co-transport of another substance going down its own electrochemical potential gradient. The total ΔG of this cotransport is negative, i.e. energetically favorable. In this case the energy stored in the gradient of the second substance is produced by primary active transport processes such as the activity of the Na+/K+ ATPase; thus, ATP is also needed for the long term maintenance of secondary active transport processes. 17 Membranes are made up of phospholipids forming a bilayer, proteins immersed in the lipid bilayer (or attached to it), and carbohydrates decorating these lipids and proteins (→ glycolipids and glycoproteins). Further constituents are sphingolipids and cholesterol. The membrane can be described by the fluid mosaic model put forth by Singer and Nicolson. According to this model, the constituents of the membrane are constantly performing lateral and rotational diffusion in the plane of the membrane like in a 2D fluid. Because of the larger viscosity of the membrane relative to water, the diffusion coefficient of membrane molecules is much smaller than what a molecule of the same size and shape would have in an aqueous solution. The membrane is not a homogeneous mixture because lipids segregate into domains based on their physico-chemical properties (such as degree of saturation and chain length). Phospholipid domains containing unsaturated lipids have a fluid-like consistency, whereas domains enriched in saturated sphingolipids and cholesterol are more gel-like having a higher viscosity. Membrane proteins may also be enriched in certain types of lipid domains they prefer, which may restrict their free diffusion. Other factors limiting the free diffusion of proteins are attachment to or interaction with the cytoskeleton (mainly actin) or the extracellular matrix, and aggregation of proteins. Due to these reasons the diffusion coefficient of membrane proteins varies in a wide range. 18 Lateral diffusion of membrane constituents can be measured by several fluorescence microscopy techniques. Fluorescence recovery after photobleaching is a classical method. First, molecules of interest (a membrane protein or lipid) are tagged with a fluorescent marker (antibody, organic dye, GFP, etc.). In the fluorescence microscope a laser beam illuminates a small area of the membrane and excites dye molecules; the fluoresence intensity is measured by a CCD camera/PMT. A short pulse of 1000-fold higher laser intensity partially bleaches the dyes in the illuminated spot. Fluorescence intensity drops, then recovers due to diffusion of molecules with intact dye labels into the observed area. From the recovery curve of fluorescence intensity the half-recovery time is determined, which is inversely proportional to the diffusion coefficient, T1/2  1/D. The R recovery fraction (mobile molecules) and the 1-R immobile fraction (immobile molecules) can be determined from the fluorescence recovery curve. The method characterizes the average diffusion properties of a large ensemble of molecules: their diffusion constant(s), the mobile and immobile fractions (attached e.g. to the cytoskeleton or to large protein aggregates). 19 An alternative method to investigate the diffusion properties of membrane components is single particle tracking, a single molecule detection technique. The protein or lipid of interest is tagged with a fluorescent marker (antibody, fluorescent bead, GFP) or a colloidal particle (10-30 nm gold particle attached to an antibody) that scatters/absorbs light. A fast camera attached to the microscope takes fluorescence or transmitted light image of the cell every 2-10 ms, and from this image series the trajectory of the particle is reconstructed using software that identifies the labeled particles. The squared displacement of the studied particle over different time intervals is calculated from the image series, and the average squared displacements are plotted vs. the elapsed time. From the time dependence of the squared displacement (𝛥𝑟 ̅̅̅̅̅2 𝑣𝑠. 𝑡) the mechanism of motion: free diffusion, directed motion, hindered diffusion can be revealed. I. In case of free diffusion (random motion in the membrane without obstacles) ̅̅̅̅̅ 𝛥𝑟 2 follows the 2D version of the Einstein-Smoluchowski equation: ̅̅̅̅̅̅ 𝛥𝑟 2 = 4𝐷𝑡, thus, the ̅̅̅̅̅ 𝛥𝑟 2 𝑣𝑠. 𝑡 graph is linear. II. If the motion of the particle is directed and it moves with a constant velocity (e.g. along a microtubule), the displacement itself (not its square) is proportional with time, thus, ̅̅̅̅̅ 𝛥𝑟 2 = 𝑣 2 𝑡 2 , and we see a positive deviation in the ̅̅̅̅̅ 𝛥𝑟 2 𝑣𝑠. 𝑡 graph (upward curvature from the linear graph characterizing free diffusion). 20 III. If the motion of the particle is hindered e.g. by the boundary of a membrane domain or by the underlying cytoskeleton, then within these boundaries the behavior is similar to free diffusion, but over longer distances (and times) as the particle hits the boundary, the squared displacement will have a negative deviation relative to free diffusion. From such experiments we can learn about the interactions of a membrane component, the domain structure of the membrane and interactions with the underlying cytoskeleton. 21 The movie shows the diffusion of a membrane protein, the transferrin receptor in the plasma membrane of a rat kidney epithelial cell. The attached colloidal particle is visualized by light microcopy. The motion of the particle can be divided into intervals during which it moves within “transient confinement zones”. These zones are probably defined by the underlying actin cytoskeleton (“membrane skeleton”), which interferes with the motion of the intracellular tail of the receptor. The experiment was performed in the laboratory of Prof. Akihiro Kusumi (Kyoto University). He provided an explanation for reduced diffusion speed of lipid and protein molecules in the cell membrane, based on a model of hop diffusion in which lipids/proteins are confined to a reduced space, created by membrane-skeleton-induced compartments. These compartments allow the lipids or proteins to diffuse freely in a limited region but limit the diffusion to other parts of the membrane. Getting to a new compartment is called “hop diffusion” while diffusion in the compartment is allowed by Brownian movement. The compartments of the cell are responsible for the reduced long-range diffusion speed of the lipids or proteins when compared to artificial membrane vesicles. High-speed single-molecule tracking of membrane molecules thus brought a paradigm shift of the plasma membrane concept from the two-dimensional continuum fluid to the partitioned fluid. 22 Osmosis is the spontaneous net movement of solvent particles from an area of lower solute concentration to an area of higher solute concentration through a semipermeable membrane, which is permeable to the solvent but impermeable to the solute particles. Osmosis tends to equalize the solute concentrations on the two sides. Solute particles cannot cross the semipermeable membrane; therefore, to understand osmosis, we need to deal with the chemical potential of the solvent only. In the simplest case there is pure solvent, say water, in compartment 1 (left side), and a solution containing solute in compartment 2 (right side). The net flow of water will proceed toward the side where its chemical potential (i.e., its partial molar Gibbs free energy) is lower. The chemical potential can be expressed as μw,I = μ0(p0,T) + RT lnνw,i + Vw (pi-p0) where i=1,2, μ0(p0,T) is the chemical potential at temperature T and pressure p0, (p0 is the ambient atmospheric pressure), pi and νw,i are the hydrostatic pressure and the mole fraction of water in compartment “i”, Vw is the molar volume of water and R is the gas constant. Initally, p1= p2, and 1 = νw,1 > νw,2, thus, 0 = RT lnνw,1 > RT lnνw,2; and μw,1 > μw,2. I.e., the presence of solute decreases the mole fraction and thereby the chemical potential of water in the solution, causing net diffusion of water to proceed from the pure water into the solution. Then, if the volume of compartment 2 cannot increase, the incoming water will raise the 23 hydrostatic pressure p2, and the term Vw (p2-p0) in the formula of μw,2 will also increase. When p2 is so high that μw,2 reaches the same value as μw,1, a dynamic equilibrium of water diffusion is established, which is called osmotic equilibrium. According to van’t Hoff’s law, the extra hydrostatic pressure needed to stop net osmosis is 𝑝𝑜𝑠𝑚𝑜𝑠𝑖𝑠 = 𝑝2 − 𝑝1 = 𝑖𝑐𝑅𝑇. This pressure is called the osmotic pressure of the solution. Here, c (mol/L) is the solute concentration, “i” is the number of particles the solute dissociates into (e.g. for glucose, i=1, for NaCl, i=2). Osmotic pressure is a colligative property meaning that it does not depend on the type of solute, only on the concentration of solute particles. The van’t Hoff law needs some correction for non-ideal or more concentrated solutions. This is taken into account by the osmotic coefficient, 𝜙. For dilute solutions 𝜙 converges to 1. For a solution containing different solutes the osmotic pressure can be calculated as a sum: 𝑝𝑜𝑠𝑚𝑜𝑠𝑖𝑠 = 𝑅𝑇 ∑𝑘 𝑖𝑘 𝑐𝑘 𝜙𝑘. The osmolarity or osmotic concentration of a solution is ∑𝑘 𝑖𝑘 𝑐𝑘 𝜙𝑘. 24 Since the cell membrane is permeable to water, mammalian cells are sensitive to the osmotic concentration of their environment. The osmolarity of blood is 275-300 mOsm depending on the hydration state of the body. The membrane of red blood cells is particularly permeable to water, so it is a good indicator of the osmolarity of the ec. environment. In a hypotonic or hypoosmolar environment (lower osmolarity than inside the cell) the cell swells and may burst (hemolysis). In a hypertonic or hyperosmolar (higher osmolarity) environment it shrinks, and in an isotonic or isoosmolar (equal osmolarity) environment the cell is in equilibrium with the medium. Solutions isotonic with blood: 5.5% (0.3 M) glucose, 0.9% (9 g/L, 154 mM) NaCl also called saline. Saline is used in intravenous drips (infusions), eye drops, for flushing wounds, intramuscular injections. Applying water instead of an isotonic solution may cause hemolysis (rupture) of blood cells and lysis (disintegration) of tissue cells. 25 Biological examples for osmosis: Plant roots contain hyperosmotic fluid relative to the water in the soil; thus, they can take up water. Uptake of water results in an increased hydrostatic pressure, called turgor, in the plant tissue. When the plant loses water, it becomes flaccid. The osmotic pressure in the tree is responsible for the plant being able to suck the water up the trunk against gravity (although capillarity also plays a role here). Edema is excess water in the insterstitium (leaking out from blood vessels due to several factors such as increased permeability of vessel walls, increased hydrostatic pressure of blood or decreased osmotic concentration of plasma proteins). A wet pack containing bitter salt can remove edema from swollen tissues by osmosis. Purgative salts (MgSO4, Mg-citrate) are poorly absorbed from the small intestines, they take up water from the epithelia of the large intestine by osmosis diluting the content. Cerebral edema can be treated by hyperosmotic infusion (hyperosmotic saline or mannitol) to reduce the increased pressure of the cerebral fluid by osmotic uptake of water into cerebral blood vessels. Hyperglycemia, the elevated blood glucose level in diabetic patients increases the osmolarity of blood and takes up water from tissues causing dehydration. 26 Dialysis is a purification procedure in biochemistry using a semipermeable membrane to remove dissolved ions or small molecules from a solution of macromolecules. Cutting membranes transmit particles smaller than their well-defined pore size. In medicine, (hemo)dialysis is the process of removing excess water, solutes and toxins from the blood in people whose kidneys can no longer perform these functions properly. This is called renal replacement therapy. Dialysis works on the principles of the diffusion of solutes and ultrafiltration of fluid across a semi-permeable membrane. Hemodialysis removes waste products and excess water by circulating blood outside the body through an external filter, called a dialyzer, that contains a semipermeable membrane. The concentrations of solutes normally found in the urine (for example phosphorus and urea) are undesirably high in the blood, but low or absent in the dialysis solution. The dialysis solution has levels of minerals like potassium and calcium that are similar to their natural concentration in healthy blood. 27 If two compartments containing salty sea water and fresh water are separated by a semipermeable membrane, net osmosis of water occurs from the fresh water to sea water through the membrane because the mole fraction of water and its chemical potential are lower in sea water. However, if excess hydrostatic pressure, greater than the osmotic pressure (of sea water), is applied to sea water, the chemical potential of water in the sea water will be greater than in the fresh water compartment. (Note that the chemical potential of water contains the term Vw (p-p0)). If this is the case, the direction of net diffusion reverses, it goes from the side of the sea water to the side of fresh water (toward the lower chemical potential). Reverse osmosis can remove many types of dissolved and suspended chemical species as well as biological ones (e.g. bacteria) from water, and is used in both industrial processes and the production of potable water. Reverse osmosis is most commonly used in desalting sea water to produce drinking water. 28 The capillary wall is slightly permeable to water. The blood pressure (hydrostatic pressure due to the work of the heart), tends to press out water through capillary walls. Blood plasma proteins (albumins, globulines, fibrinogenes) increase the osmotic pressure of blood, helping to retain water in the capillaries. These two pressure components and the water permeability of capillaries and small veins affect mainly the water homeostasis of the tissues. If the blood pressure is increased in the capillaries, or the osmotic pressure of blood is decreased, water leaks out from the capillary to the tissues causing edema. Increased blood vessel wall permeability (e.g. in inflammation) and obstruction of fluid clearance in the lymphatic system also contribute to edema. The different pressure components at the arterial and venous sides of the capillary are shown in the figure. At the arterial side, the blood pressure is 35 Hgmm, the hydrostatic pressure of the tissue is 2, and the osmotic pressure due to plasma proteins is 25 Hgmm. The net pressure points outward (8 Hgmm), so water leaks from the capillary into the tissue. At the venous side, the blood pressure drops to 15 Hgmm, the osmotic pressure of blood is still 25 Hgmm, the hydrostatic pressure in the tissue is 1 Hgmm, and the osmotic pressure in the tissue (due to plasma proteins) is 3 Hgmm, resulting in a net 8 Hgmm pressure pointing inward, retaining water from the tissue to the capillary. If this equilibrium is disturbed, edema can ensue. Causes include the decrease of blood plasma proteins (due to liver disease and consequent underproduction of albumin, kidney disease leading to loss of proteins via urine, low protein intake due to starvation), or increased blood pressure. 29 30