Determination of Solute Permeability of Microvascular Endothelial Cell Monolayers In Vitro

Abstract

The microvascular endothelium has a critical role in regulating the delivery of oxygen, nutrients, and water to the surrounding tissues. Under inflammatory conditions that accompany acute injury or disease, microvascular permeability becomes elevated. When microvascular hyperpermeability becomes uncontrolled or chronic, the excessive escape of plasma proteins into the surrounding tissue disrupts homeostasis and ultimately leads to organ dysfunction. Much remains to be learned about the mechanisms that control microvascular permeability. In addition to in vivo and isolated microvessel methods, the cultured endothelial cell monolayer protocol is an important tool that allows for understanding the specific, endothelial subcellular mechanisms that determine permeability of the endothelium to plasma proteins. In this chapter, two variations of the popular Transwell culture methodology to determine permeability to using fluorescently labeled tracers are presented. The strengths and weaknesses of this approach are also discussed.

1. Introduction

The exchange of solutes and fluids across the capillary and postcapillary venule endothelium is an essential function for establishing and maintaining tissue homeostasis and optimal organ function [1, 2]. During local inflammatory responses, such as to a mosquito bite on the skin, factors released at the site of injury activate nearby endothelial cells to increase microvascular permeability and facilitate transmigration of leukocytes into the local tissues, producing localized inflammation and edema that normally resolve. However, diseases such as diabetes progressively increase microvascular permeability systemically, slowly producing a pathologic inflammatory state [3, 4]. Moreover, traumatic injuries and sepsis stimulate a systemic inflammatory response (cytokine storm) that includes systemic microvascular leakage of plasma proteins into surrounding tissues. This severe level of microvascular hyperpermeability can lead to life-threatening conditions such as acute respiratory distress syndrome, abdominal compartment syndrome, and multiple organ dysfunction [5–11].

The passage of fluid and solutes across the microvascular wall is determined by (1) the net hydrostatic and osmotic pressure gradients across the semipermeable endothelial barrier and (2) the barrier properties of endothelium [1]. Under physiological conditions, the pressure gradient across the endothelium produces a small, net filtration of fluid from the blood into the surrounding tissues. The physiological, continuous formation of interstitial fluid serves to deliver nutrients to the cells of the tissues and is cleared by lymphatic drainage [12, 13]. When considering flux of this fluid across the microvascular wall, the capillary filtration coefficient introduced by Landis, a product of the hydraulic conductivity and surface area for exchange, is used to describe the rate of water passage through the endothelium [14]. For solutes, their transport across the endothelium can be described by a few different terms. First, the osmotic reflection coefficient is a term introduced by Kedem and Katchalsky into the Landis equation that indicates the fraction of proteins that are retained in plasma during transendothelial filtration of fluid, i.e., essentially the inverse of the fraction of plasma proteins that escape across the endothelial wall [15]. More commonly, terms that describe the diffusive transport of solutes across the endothelium are used. These include the solute flux ($J_{\text{s}}$), which is the amount of solute that crosses the endothelium per unit of time, and the solute permeability coefficient ($P_{\text{s}}$), a one-dimensional rate indicating how quickly solute crosses the endothelium. By convention, the units for $P_{\text{s}}$ are reported as cm/s. The relationship between $J_{\text{s}}$ and $P_{\text{s}}$ is defined in Fick’s first law of diffusion:

$$J_{\text{s}}=P_{\text{s}}S\left(C_{\text{PL}}-C_{\text{ISF}}\right)$$

This relationship indicates that the net $J_{\text{s}}$ for a particular solute across the endothelium is the product of $P_{\text{s}}$, the surface area of the endothelium available for exchange ($S$), and the concentration gradient of the solute across the endothelium, established by the difference of its plasma concentration ($C_{\text{PL}}$) minus its concentration in the interstitial fluid ($C_{\text{ISF}}$). It is important to note that Fick’s first law describes the relationship for diffusive permeability properties of the endothelium in the absence of any hydrostatic or osmotic pressure gradients between the plasma and interstitial fluid compartments. An increase in the endothelium’s $P_{\text{s}}$ to a particular solute or group of solutes will facilitate their leakage from plasma into the surrounding tissues. However, it is worth noting that in vivo, increased appearance of plasma solutes in the surrounding tissues can also be elicited by increased filtration of plasma due to increased local capillary hydrostatic pressure and capillary recruitment, both which result from increased local blood flow. The ability to assess $P_{\text{s}}$ thus represents the contribution of the endothelium’s diffusive permeability properties to microvascular leakage, independent of changes in pressure gradients across the endothelium [1, 2].

While there are several useful in vivo and isolated microvessel models that can be used to estimate or determine $P_{\text{s}}$, the establishment of culture methods for endothelial cells starting in the early 1970s [16] has facilitated development of a variety of protocols to study transport of solutes across the endothelium. One very popular method is to grow endothelial cells into a monolayer on semipermeable membranes and then place the membrane between two reservoirs, one that has tracer and the other on which accumulation of the tracer is measured (Fig. 1). The tracers may be labeled radioactively, are sometimes detected with chemical means, but most commonly are labeled with fluorophores. The data can then be presented simply as the amount of tracer that accumulates over a given time period. However, if the concentration of the tracer is measured on both sides of the endothelial monolayer, then Fick’s first law of diffusion can also be used to calculate the apparent $P_{\text{s}}$ of the tracer. While the major drawback to endothelial monolayer models is that they do not fully represent the in vivo microcirculation or intact microvessels under constant intravascular flow and perfusion pressure, they do offer the advantage of providing information about endothelial-specific mechanisms that contribute to the control of microvascular permeability. In addition, the protocol is relatively easy to perform and allows other advantages of endothelial cell culture models such as the ability to genetically modify cells or use chemical inhibitors of cell function that might not be useful in the in vivo setting.

Fig. 1.

Endothelial cell monolayer permeability cultureware and chamber systems. (a) Image of a 6-well culture plate with Costar Snapwell membranes on which the cells are grown. The Costar Transwell membranes (not shown) described in this protocol are similar, but smaller, in a 24-well plate configuration. (b) Schematic showing the configuration of the upper, “luminal” and lower, “abluminal” compartments with the endothelial cells grown on the Transwell insert in between. A tracer is added to the top compartment, and the rate of appearance of tracer in the lower compartment is measured to determine permeability of the cell monolayer. (c) Schematic of the configuration of the Ussing Snapwell chamber system, in which the “luminal” compartment is on the left and the “abluminal” compartment is on the right. Transport of the tracer from left to right across the endothelial monolayer is measured in order to determine permeability

Several considerations should be made when planning an experiment to determine the solute permeability of endothelial monolayers in vitro. While some primary endothelial cells are very easy to grow, like human umbilical vein endothelial cells (HUVEC), microvascular endothelial cells may better represent the microcirculation. In addition, endothelial cells chosen from a particular organ might be more representative of that organ’s microvascular beds and are available for a variety of organs, including the lung, heart, skin, brain, intestine, and others [5, 17–26]. The culture conditions are a second important consideration. Ideally, the semipermeable membrane upon the cells are grown should be coated with gelatin, fibronectin, or a mixture of basement membrane proteins that will promote good focal adhesion. In addition, the cells should be seeded at a density to immediately achieve confluence, and the monolayer should be allowed to grow for 5–7 days to allow junctions to mature [22, 24]. The choice of tracer is also important. Albumin that is labeled with a fluorophore is often chosen because albumin is one of the major plasma components. Choice of an albumin that is labeled with a 1:1 ratio of fluorophore is desirable to achieve a linear standard curve for fluorescence intensity and protein concentration. Some types of FITC-albumin that are commercially available have multiple FITC molecules attached, some by noncovalent binding, to make the fluorescence brighter per molecule. However, these present the problem that some FITC molecules can dissociate from the albumin, cross the endothelial barrier, and make the apparent $P_{\text{s}}$ much higher than the actual $P_{\text{s}}$. Another consideration is that certain fluorophores can alter the physiochemical characteristics of tracer molecules such as albumin [27]. FITC was shown to significantly alter the size and relative charge of albumin, and FITC-albumin flux was recorded to be significantly higher than TRITC-flux in single-perfused microvessels [27]. Some of these problems may be alleviated with the newer Alexa Fluor fluorophores, which are brighter and represent a better choice for obtaining 1:1 labeling with albumin. In addition, dextrans are labeled 1:1 with fluorophore and can be obtained in various molecular weight ranges, which allow for determining the permeability of molecules of different sizes [21, 28]. Lastly, consideration should be given to whether a protocol that provides a single $P_{\text{s}}$ per experimental group or one that compares $P_{\text{s}}$ of the same endothelial monolayers before and after an experimental intervention is needed to satisfy the study objectives. An advantage of the variation in which a single $P_{\text{s}}$ is determined is that the cultureware used requires a relatively small amount of media and tracer because the cell surface-coated membranes are small (Fig. 1b). In contrast, the variation in which multiple measurements of solute flux over time can yield values of $P_{\text{s}}$ before and after addition of test compounds utilizes a specialized chamber system that has the drawback that it requires larger amounts of media and tracer (Fig. 1c). However, this latter method has the advantage that it can provide detailed data about changes in $P_{\text{s}}$ over time. Both variations of the protocol are discussed in this chapter.

2. Materials

1. Microvascular endothelial cells and their recommended growth medium.

2. Phenol red-free endothelial basal medium (e.g., EBM-PRF from Lonza or VBM from Lifeline Cell Technology).

3. Phosphate-buffered saline (PBS).

4. 0.25% trypsin-EDTA solution.

5. Hemacytometer.

6. Gelatin solution (1.5% in 0.9% NaCl) or another matrix solution of your choice.

7. Costar Transwell Membranes (0.4 μm pores, clear polyester membrane):

   1. For single $P_{\text{s}}$ determination: Corning #3470, twelve 0.33 cm² inserts in a 24-well plate.
   2. For multiple $P_{\text{s}}$ determination: Corning #3801, six 1.13 cm² inserts in a 6-well plate.

8. For multiple $P_{\text{s}}$ determination, a special chamber system with heating manifold base assembly and a circulating water bath are required (NaviCyte Ussing vertical chamber system available from Warner Instruments; part number 66-0075 for the base assembly; part number 66-0008 for Snapwell chambers).

9. Tracer of your choice. Some examples are FITC-Albumin, Alexa Fluor-488-Albumin, FITC-Dextran-70-kDa, TRITC-Dextran-4.4-kDa, and TRITC-Dextran-150-kDa.

10. Black, solid 96-well plate, and a plate reader to measure fluorescence, such as a Molecular Devices SpectraMax L Microplate Reader.

11. Distilled or Millipore-filtered water for diluting samples in the 96-well plate.

3. Methods

3.1. Preparation of Endothelial Cell Monolayers for Study

1. Grow endothelial cells to 80–90% confluency in 60-mm or 100-mm plates in their recommended growth medium.

2. Warm growth medium, 0.25% trypsin-EDTA solution, PBS, and gelatin solution to 37 °C.

3. In a biosafety cabinet, coat the Transwell or Snapwell membrane inserts with gelatin solution. For 0.33 cm² Transwell inserts, apply 100 μL per well. For 1.13 cm² Snapwell inserts, add 500 μL per well. After applying to all of the wells, the solution may be aspirated, leaving a thin coat on the membranes.

4. Cover the plate containing the inserts, set aside, and bring the plate of cells to be trypsinized to the biosafety cabinet.

5. Aspirate the growth medium and add 2 mL of warm PBS.

6. Aspirate the PBS and add 1–2 mL of warm 0.25% Trypsin-ETDA solution.

7. Allow the cells to round up, and rap the dish on the bench if necessary to get a cell suspension.

8. Add 2 mL of warmed growth medium to the cells to inactivate the trypsin. Collect the cell suspension from the plate with a pipette and place into a 15-mL conical tube.

9. Wash the dish with 2 mL of PBS to collect any remain cells, and then pipette this suspension into the same 15-mL conical tube.

10. Take a 10 μL of the cell suspension, and place in the hemacytometer. Perform a cell count while waiting in the next step.

11. Centrifuge the cells at 300×g for 3 min.

12. Resuspend the cells in an appropriate amount of growth medium, taking into account that the seeding density for the membranes is approximately 10⁵ cells per cm² surface area (see Note 1). Seed the following:

    1. For each Transwell insert (0.33 cm²), seed 3 × 10⁴ cells in 100 μL of medium.
    2. For each Snapwell insert (1.13 cm²), seed 10⁵ cells in 500 μL of medium.
13. Add growth medium to the bottom wells supporting each insert:

    1. For the wells containing Transwell inserts, add 600 μL of medium.
    2. For the wells containing Snapwell inserts, and 2 mL of medium.

14. Place in the incubator, and allow the cells to grow for 5–7 days, performing a medium change within 24 h of seeding and then every 48 h thereafter (see Note 2).

3.2. Assay Variation 1: Transwell Membranes

1. Warm the phenol red-free basal medium to 37 °C.

2. In the biosafety cabinet, aspirate the growth medium from both the inserts (upper compartment) and the wells under them (lower compartment).

3. Replace the medium with the phenol red-free basal medium (100 μL in the upper compartment and 600 μL in the lower compartment), and place the cells in the incubator.

4. Allow the cells at least 1 h to equilibrate in the new media before starting the experiment.

5. Prepare 200 μL of 10 μg/mL FITC-albumin (or other tracers of your choice) in phenol red-free medium.

6. Depending upon the protocol, add the inhibitors or agonists that are hypothesized to increase or decrease permeability at the chosen time points before the tracer will be added (see Note 3).

7. Add 10 μL of the tracer stock solution to the each of the upper compartments. Place the plate back in the incubator and wait 30 min (see Note 4).

8. Set up the standard curve dilutions in the 96-well black plate while the cells are incubating. In column 1, row A, add 196 μL of distilled water, and then to that add 4 μL of the 10 μg/mL FITC-albumin solution. In rows B through H, add 100 μL. Then remove perform serial dilutions starting with removal of 100 μL from well A1 and transferring it to well B1, mixing, then removing 100 μL from well B1 and adding to well C1, etc. until reaching well G1. Discard the 100 μL removed from well G1. Leave well H1 as a blank. Keep the plate covered until samples are added later.

9. After 30 min, get the plate from the incubator, and move the Transwell inserts to the empty wells to stop flux of tracer from the upper compartment to the lower compartment.

10. Take 20 μL samples in duplicate from each of the lower compartments, and load them into the black, solid 96-well plate. Add 80 μL of water to each well containing a sample. From the upper compartment, take 10 μL samples in duplicate, place them in the plate, and mix with 90 μL of water.

11. Determine the concentrations of the upper compartment (luminal concentration) and the lower compartment (abluminal concentration) using the plate reader and standard curve, taking dilutions into consideration.

12. Calculate $P_{\text{s}}$ for each well using the following equation:

    $$P_{\text{S}}=\left(C_{\text{A}}/t\right)\times (\text{l}/A)\times \left(\text{V}/C_{\text{L}}\right)$$

    where:

    $C_{\text{A}}$ = albumin concentration (mg/mL)

    $t$ = time in seconds

    $A$ = area = 0.33 cm²

    $V$ = volume of the abluminal chamber (0.6 mL; see Note 5)

    $C_{\text{L}}$ = luminal concentration (mg/mL).

13. This $P_{\text{s}}$ value represents the average ${\text{P}}_{\text{s}}$ for the time period between the addition of tracer to the luminal compartment and the end of the experiment when the abluminal samples were taken.

3.3. Assay Variation 2: Snapwell Membranes

1. Warm 100 mL of phenol red-free basal medium to 37 °C.

2. Turn on the water bath that warms the Ussing chamber base assembly.

3. In the biosafety cabinet, one at a time, aspirate the growth medium from an insert, remove it from the plate, and then insert it between the Ussing chamber half-blocks, and place a retainer ring on each side (Fig. 2a). Immediately add 5 mL of warm media to each side, and place the chamber block into the base assembly. Repeat for all six inserts.

4. Attach the gas tubes that deliver 5% CO₂ to the tops of each chamber (Fig. 2b). Open the CO₂ tank, and use the adjusters on the base assembly to regulate bubbling in all of the chambers.

5. Allow the cells to equilibrate for 30 min.

6. While waiting, prepare 6 mL of a 50 mg/mL FITC-albumin in phenol red-free basal medium.

7. Set up a standard curve for FITC-albumin in solid black 96-well plate as follows: in column 1, row A, add 199.2 μL of distilled water, and then to that add 0.8 μL of the 50 μg/mL FITC-albumin solution. In rows B through H, add 100 μL. Then remove perform serial dilutions starting with removal of 100 μL from well A1 and transferring it to well B1, mixing, then removing 100 μL from well B1 and adding to well C1, etc. until reaching well G1. Discard the 100 μL removed from well G1. Leave well H1 as a blank. Keep the plate covered until more samples are added.

8. In each chamber, identify the “luminal” side on which the apical surface of the monolayers face. Remove 1 mL of medium from this side, and replace with 1 mL of the 50 mg/mL FITC-albumin solution. This will make the luminal concentration 10 mg/mL for FITC-albumin. Take two 20 μL aliquots from each luminal compartment and place in the 96-well plate.

9. To determine baseline flux, take 20-μL samples in duplicate from each abluminal compartment every 5 min (see Notes 6 and 7) for the predetermined baseline period (usually 30 min or 60 min). Keep the 96-well plate covered when not adding samples (see Note 8).

10. Add the test compound(s) to each well, and continue collecting samples for the periods predetermined by the experimental design.

11. At the end of the experiment, take two 20-μL aliquots from each luminal compartment again.

12. Bring all the volumes of the samples taken up to 100 μL by adding 80 μL water. Determine protein concentrations using a fluorescent plate reader, taking the dilutions of samples into consideration.

13. Create a spreadsheet in which the albumin protein concentrations for each chamber are plotted over time. The slope of the line between two time points multiplied by the abluminal compartment volume represents the flux of the tracer for the period between those two time points.

14. Calculate $P_{\text{s}}$ for each time point of each chamber as follows:

    $$P_{\text{s}}=\left[\left(C_{\text{A}1}-C_{\text{A}0}\right)/t\right]\times (1/A)\times \left(V/C_{\text{L}}\right)$$

    where:

    $C_{\text{A}0}$ = albumin concentration at previous time point (mg/mL)

    $C_{\text{A}1}$ = albumin concentration at current time point (mg/mL)

    $t$ = time between the two time points, in seconds,

    $A$ = area = 1.13 cm²

    $V$ = volume of the abluminal chamber (5 mL; see Note 5)

    $C_{\text{L}}$ = average luminal concentration (mg/mL) from measurements at the beginning and end of the experiment.

15. The values for $P_{\text{s}}$ calculated this way represent the average for the time period between time points. This approach allows for determining changes in $P_{\text{s}}$ over time.

Fig. 2.

Snapwell chamber assembly. (a) Attachment of the retaining ring to lock both sides of the chambers together. The yellow arrow shows the retaining ring, held by the attachment tool. A ring must be placed on both sides. Note that both compartment openings are on top. (b) Attachment of the tubes for CO₂ delivery to the chambers. The yellow arrows show attachment points, and adjustment knobs are on the other end of the tubing on the manifold