1
Stable and reproducible Trans Epithelial
Electrical Resistance measurements
Evaluation of four Trans Epithelial Electrical Resistance (TEER) measuring techniques
Author: Robin Haring 29th of June 2018 Supervisors: Dr. R. Nieuwland L.G. Rikkert MSc
A bachelor thesis submitted in partial fulfilment of the requirements of a Bachelor of Science degree
2
Table of contents
Abstract ... 3
Introduction ... 4
Materials and methods ... 6
Results ... 12 Discussion ... 18 References ... 20 Appendix A ... 23 Appendix B ... 24 Acknowledgements ... 25
3
Abstract
The intestinal epithelium protects humans against the external environment. Weakening of the intestinal epithelium causes diseases like inflammatory bowel disease and necrotizing
enterocolitis . In vitro the barrier function of the intestinal epithelium is measured by trans epithelial electrical resistance (TEER), for example to study the effect of drugs. When a drug increases TEER, this drug is potentially useful to treat before mentioned diseases. To develop a stable and reproducible system, four TEER measurement techniques were evaluated, of which the EVOM2 technique proved to be most robust, with a mean coefficient of variation (CV) within a single transwell of 13%, a mean CV of 22% between transwells, and
reproducible TEER measurements between experimental runs (p=0.929; unpaired t-test). In conclusion, the EVOM2 technique is recommended to measure TEER in vitro.
4
Introduction
Body surfaces and cavities are lined with a monolayer of epithelial cells. Epithelial cells are connected by tight junctions at the lateral side of the cells (Anderson and Itallie, 2009). The paracellular space between epithelial cells is controlled by tightening and loosening of the tight junctions (Figure 1) (Shen, Weber and Turner, 2008). In this way, tight junctions control the passage of cells, ions, water and pathogens (Buzza et al., 2010). As a result, tight junctions act as a barrier which separates and protects the internal from the external environment. Weakening of this barrier function increases the epithelial permeability, thereby enabling the passage of larger particles across the paracellular space to enter the internal environment (Figure 1).
Loosening of tight junctions may contribute to inflammatory bowel disease (IBD), including Crohn’s disease and ulcerative colitis (Merga, Campbell and Rhodes, 2014; Laukoetter, Nava and Nusrat, 2008). Furthermore, necrotizing enterocolitis (NEC), the main cause of death in premature infants, is also related to increased intestinal permeability (Springer, Annibale and Aslam, 2017). An exclusive human milk (HM) diet reduces the risk of NEC (Maffei and Schanler, 2017). Whether this reduced risk is related to a change in the epithelial permeability, however, is unknown. To study the potential effects of HM and drugs on epithelial permeability, a stable and reproducible experimental set-up is essential.
Epithelial permeability can be monitored by measuring trans epithelial electrical resistance (TEER) (Chen, Einspanier and Schoen, 2015). With TEER measurements, a
Figure 1. Function of tight junctions. Tight junctions control the paracellular space between epithelial cells and in this
way control passage of cells, ions, water and pathogens. When tight junctions loosen, the epithelium becomes leaky, allowing an increased passage of cells, ions, water and pathogens.
5
monolayer of cells is cultured on a filter insert (Figure 2). One electrode is placed underneath (basolateral) and one electrode is placed above the cell monolayer (apical). An alternative current (AC) flows through the cell monolayer to close the circuit. When the AC passes the cell monolayer, the AC encounters resistance. This resistance is caused by tight junctions between the cells (Srinivasan et al., 2015). In this way, TEER allows examination of the integrity of tight junctions. When the TEER increases, the AC encounters less resistance, due to loosening of the tight junctions. Vice versa, a decrease in TEER reflects tightening of the tight junctions and thus an increase in epithelial permeability, meaning strengthening of the epithelial barrier function.
To study the change in epithelial permeability in vitro, TEER measurements have to be stable and reproducible. Therefore, four TEER measuring techniques were evaluated: (i) the
EVOM/chopstick electrode, (ii) the ECIS® 8W TransFilter, (iii) the ECIS® standard 8W1E array, and (iv) the ECIS® TEER 24. To study these TEER measuring techniques, the human epithelial colorectal adenocarcinoma Caco-2 cell line was used.
Figure 2. TEER measurement setup. For TEER meaurements, one electrode is
placed above (apical) and one electrode is placed below (basolateral) the cell layer, and an alternating current (AC) is applied between the electrodes. The AC encounters a resistance due to the tight junctions between the (Caco-2) cells. This resistance is measured in Ohms (Ω). This resistance measurement is multiplied by the surface area of the filter insert to get the TEER values (Ω*cm2)
6
Materials and methods
Cell culture
Caco-2 cells were cultured in Dulbecco’s modified Eagle medium (DMEM; Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine and 50 u/mL Penicillin Streptomycin (Thermo Fisher Scientific) at 37°C and 5% CO2. The
cells were provided by Dr. C.C. Paulusma (Tytgat Institute for Liver and Intestinal Research, Academic Medical Center, Amsterdam). The cells were cultured in T75 culture flasks
(Corning, Corning, NY) and subcultured 1:4 before TEER measurements. Caco-2 cells were subcultured in Corning® Transwell® polycarbonate membrane cell culture filter inserts with a pore size of 0.4 µm and a cell growth area of 1.12 or 0.33 cm2 (Corning). The apical side of the filter inserts were coated with fibronectin (Sigma-Aldrich, Saint Louis, MO) for 45 minutes at 37°C. After 45 minutes, the fibronectin was replaced with 200 µL culture medium and another 1 mL of culture medium was added underneath the filter insert. The transwells with culture medium above and below the filter inserts were incubated overnight at 37°C and 5% CO2. The next day, the apical 200µL culture medium was replaced with 200µL cell
suspension diluted in culture medium.
The culture medium was completely replaced at day 2 (day 0=seeding of the cells). Half of the culture medium was replaced both at the basolateral and apical side at day 4 and 6. From day 7 onwards, half of the culture medium was replaced both at the basolateral and apical side daily.
Since no filter insert was used with the ECIS 8W1E technique, the Caco-2 cells were directly subcultured in the wells (Applied Biophysics Inc, Troy, NY). Before seeding the cells, the wells were coated with 400 µL fibronectin for 45 minutes at 37°C. After 45 minutes, the fibronectin was replaced with 400 µL of culture medium and incubated overnight at 37°C and 5% CO2. The culture medium was completely replaced at day 2 (day 0=seeding of the
cells). Half of the culture medium was replaced both at the basolateral and apical side at day 4 and 6. From day 7 onwards, half of the culture medium was replaced both at the basolateral and apical side daily.
TEER measurements
All techniques generate resistance values, calculated via Ohm’s law (V=IxR). The resistance values were converted manually to TEER values by multiplying the resistance value (R[Ω]) by the area of the filter insert (A[cm2]) (TEER= R * A).
7
Three of the evaluated techniques are filter-based assays. To eliminate the contribution of the filter inserts to the resistance measurements, transwells without cultured cells were used as controls. The resistance values of the transwells without cells are subtracted from the resistance values of the transwells with cells to generate TEER values related to the tight junctions of the cell monolayer. The TEER values were compared within and between transwells, and between independent experimental runs.
After seeding the cells, a confluent cell monolayer will be formed. When a confluent cell monolayer is formed, the cells will start to differentiate. All steps cause an increase in TEER. When the cells have fully differentiated, a plateau in TEER is reached.
Cell count
Caco-2 cells were counted manually using a Bürker-Türk counting chamber (Brand, Wertheim, Germany) to calculate the appropriate cell density required for culturing. The following formula was used:
dilution factor A = (Average cells per small square ∗ dilution factor B)/volume of a square Y
Average cells per small square = sum of all cells in each square that was counted, divided by the total number of squares counted
Dilution factor B = dilution of the sample before counting. When no dilution was used, multiply by 1
Volume of one square = width * height * depth = 1*1*0.1 for the large square and 0.5*0.5*0.1 for the small square
Y =desired cell density [cells/mL]
Epithelial Volt/Ohm Meter (EVOM)/Chopstick
The EVOM2 technique (World Precision Instruments, Sarasota, FL) uses STX2 ‘chopstick’ electrodes outside the incubator to measure TEER (Figure 3). 110,000 cells were grown on 1.12 cm2 filter inserts and 46,000 cells were grown on 0.33 cm2 filter inserts. One electrode is placed above the filter insert and one is placed below the filter insert. An alternating current with a frequency of 12.5 Hz is applied between the two electrodes.
8
Since the measurements are conducted outside the incubator, TEER values can be affected by the change in environment (temperature, %CO2, humidity, etc.). To examine
whether the change in environment indeed affects the TEER values, direct TEER
measurements were compared to TEER measurements after 20 minutes outside the incubator. ECIS 8W1E
The ECIS 8W1E technique (Applied Biophysics Inc, Troy, NY) measures the resistance of a cell monolayer continuously. The technique is not filter-based, therefore both electrodes are positioned underneath the cells (Figure 4). Either 75,000 cells or 150,000 cells were grown in a single well on top of the active electrode as well as the counter electrode. Between the electrodes, an AC is applied. The resistance can be measured with 11 different AC frequencies. Which frequency is optimal to analyse tight junctions is determined by multi frequency analysis. The optimal frequency is the frequency at which the resistance value between transwells with and without cells differs most. In our experiments an AC with a frequency of 1,000 Hz was the optimal frequency (data not shown).
Since the ECIS 8W1E technique does not use filter inserts, TEER is calculated by multiplying resistance values by the area of the active electrode (4.9x10-4 cm2).
Figure 3. Schematic representation of the EVOM2/Chopstick technique setup. With the
EVOM2/Chopstick technique, cells are grown on a filter insert. The resistance is measured with STX2 ‘chopstick’ electrodes. One electrode is placed above and one electrode is placed below the cell monolayer. An alternating current (AC) with a frequency of 12.5Hz flows between the two electrodes.
9
ECIS 8W TransFilter and ECIS TEER 24
The ECIS 8W TransFilter and ECIS TEER 24 techniques (Applied Biophysics Inc, Troy, NY) have the same working principle (Figures 5 and 6). The differences between the techniques are the basolateral electrode and the number of transwells that can be taken along in one experimental run. The basolateral electrode of the ECIS 8W TransFilter partially covers the base of the transwell, while the basolateral electrode of the ECIS TEER24 technique covers the entire base of the transwell. Furthermore, the ECIS 8W TransFilter technique contains 8 transwells, whereas the ECIS TEER24 technique contains 24 transwells.
Both techniques continuously measure TEER. With the ECIS 8W TransFilter technique, 150,000 cells were grown on a 0.33 cm2 filter insert. With the ECIS TEER24 technique, 46,000 cells were grown on a 0.33 cm2 filter insert. For both techniques, culture medium is added on top of (200 µL) and below (1 mL) the filter insert. Gold electrodes are placed at fixed positions on both sides of the filter insert (Figures 5 and 6). The basolateral electrode is attached to the bottom of the transwell. The apical dipping electrode is attached to a stainless steel plate placed on top of the well array. The applied alternating current, with these techniques, can measure resistance values at 12 different frequencies. For our analysis, an alternating current with a frequency of 500 Hz was the optimal frequency (data not
Figure 4. Schematic representation of the ECIS 8W1E setup. With the ECIS 8W1E technique, cells are grown directly
on both the active and counter electrode. The cells are supplemented with culture medium on top of the cells. A continuous alternating current (AC) between the electrodes is applied, generating continuous resistance values. The obtained resistance values will be manually converted to TEER values by multiplying the generated resistance value with the area of the active electrode (4.9x10-4 cm2).
10
shown), since this was the frequency at which the resistance value between transwells with cells and without cells differed most.
Figure 5. Schematic representation of the ECIS 8W TransFilter setup. A cell monolayer is grown on a filter insert.
Above and below the cell monolayer, gold electrodes are placed. An alternating current (AC) with a frequency of 500 Hz is applied between the electrodes and encounters resistance when passing the cell monolayer. The resistance value is calculated via Ohm’s law (V=IxR). Subsequently, the resistance value is manually converted to a TEER value by multiplying the resistance value with the area of the filter insert (0.33 cm2).
Figure 6. Schematic representation of the ECIS TEER24 setup. A cell monolayer is grown on a filter insert. Above
and below the cell monolayer gold electrodes are placed. The basolateral electrode covers the entire base of the transwell. The apical electrode is attached to a stainless steel plate and therefore has a fixed position. An alternating current (AC) with a frequency of 500 Hz is applied between the two electrodes. The AC encounters resistance when passing the cell monolayer. The resistance value is calculated via Ohm’s law (V=IxR). Subsequently, the resistance value is manually converted to a TEER value by multiplying the resistance value with the area of the filter insert (0.33 cm2).
11
Statistics
GraphPad Prism 7 was used for all graphs an statistical analysis.
To determine the stability of a technique, the % coefficient of variation (%CV, SD divided by the mean * 100%) of TEER values of the last 3 days within a single transwell was calculated (Figure 7). Because the ECIS techniques generate continuous TEER values, the time scale was divided by 24 to convert the time scale from hours to days. With this time scale, the first data point of each day is appointed as ‘day x’. Subsequently, the last three days are selected and the %CV is calculated. The reproducibility of the techniques was evaluated
between transwells at one fixed time point (see Figure 7). Three fixed time points were taken
during the last three days of culture. Of those last three days, the %CV between transwells was calculated. For the ECIS techniques, the hour time scale is converted to a day time scale like discussed previously.
Also, the reproducibility of TEER values between experimental runs was determined with an unpaired t-test (Figure 7). A technique was considered not reproducible between experimental runs with a p value<0.05.
Figure 7. Determining stability and reproducibility of the technique. In both graphs, the fictive data points represent TEER values measured
during the last three days, which are used to calculate the stability of the technique by calculating the Correlation of Variation (%CV) of the TEER values within one transwell. The reproducibility of the technique is determined in two ways: the %CV between transwells at the last three days and the p-value of all transwells between experimental runs determined with an unpaired t-test.
12
Results
EVOM2/Chopstick
Figure 8 shows that the TEER values within a transwell were stable (mean %CV 1.12 cm2 filter insert=13% and mean %CV 0.33 cm2 filter insert=9%) during the last 3 days of each experimental run.
Between transwells, a considerable variation was seen (mean %CV 1.12 cm2 filter insert=22% and mean %CV 0.33 cm2 filter insert=29%) during the last 3 days in each experimental run.
Because the EVOM2/Chopstick technique was unable to reproduce TEER values
between transwells (Figure 8), it was investigated whether this can be due to the change in
environment during the TEER measurements. Therefore, direct TEER measurements and TEER measurements after 20 minutes outside the incubator were performed in the same wells. The TEER measurements after 20 minutes outside the incubator consistently showed higher TEER values than the directly measured TEER values (Figure 9). Therefore, the
Figure 8. EVOM2/Chopstick TEER measurements. (A-C) Time (days) is plotted on the X-axis against TEER (Ω*cm2) on the Y-axis. Measurements of one transwell over time are connected with a ‘guide of the eye’ line to visualize the variation within one transwell. The TEER value within one transwell over the last 3 days is stable in all graphs. The TEER value between transwells at the last three days is not
reproducible in all graphs. (A-B) A 1.12 cm2 filter inserts was used and 110,000 cells were seeded per transwell. (C) A 0.33 cm2 filter insert was used and 46,000 cells were seeded per transwell.
13
“change in environment” correlates with increased TEER values. This environmental change may explain the high mean %CV of TEER values between transwells.
Furthermore, the technique was evaluated on its ability to produce reproducible TEER values between experimental runs. The mean TEER values and SD of two experimental runs were plotted in Figure 10. Figure 10 shows the EVOM2/Chopstick technique is capable of reproducing TEER values between experimental runs (p=0.9293).
Figure 9. Effect of temperature decrease on TEER values. Time (days) is plotted on the X-axis against TEER
(Ω*cm2) on the Y-axis. Sixteen 1.12 cm2 filter inserts were used and 110,000 cells were seeded per transwell. The average TEER value with corresponding SD were plotted per day. TEER values measured after 20 minutes outside of the incubator consistently showed higher TEER values than TEER values measured directly after taking the plate outside of the incubator. Therefore, a change in the environment (inside the incubator -> outside the incubator) causes an increase in TEER.
[days]
Figure 10. Reproducibility between experimental runs. Time (days) is plotted on the X-axis against TEER
(Ω*cm2) on the Y-axis. 1.12 cm2 filter inserts were used and 110,000 cells were seeded per transwell. TEER values of run 1 (n=16) and TEER values of run 2 (n=16) overlap. Therefore, the EVOM2/Chopstick technique shows to generate reproducible TEER values between independent experimental runs. Data per experimental run (2x n=16) are shown as mean +/SD.
Run 1
Run 2
14
ECIS 8W1E
Culture medium was exchanged outside the incubator. Therefore, after reconnecting the array to the adapter and resuming the measurements, peaks due to changes in environment and medium exchange were observed (Figure 11).
Figure 11 shows that the ECIS 8W1E technique generates unstable TEER measurements within one transwell in the last 3 days (mean %CV= 22%). The TEER
measurements visualized in figure 11B, however, seems stable within one transwell in the last 3 days (%CV=4%). The expected behaviour (increasing TEER followed by a plateau in TEER) is however not observed.
TEER measurements of the transwells do not overlap in either Figure 11A or figure 11C. Also, the mean %CV of TEER values within transwells in the last three days was 31%. Therefore, TEER measurements between transwells are hard to reproduce using the ECIS 8W1E technique.
Figure 11. ECIS 8 Well 1 Gold Electrode (8W1E). (A-D) Time (hrs) is plotted on the X-axis against TEER (Ω*cm2) on the Y-axis. All experimental runs were performed using a 0.33 cm2 area filter insert. (A-B) 150.000 cells were seeded per transwell. The expected behavior, an increase in TEER followed by a plateau in TEER, was not seen. (A) The experimental run shows spiky TEER values within one transwell. Therefore, TEER measurements within one transwell were not stable. Furthermore, TEER values of the two transwells do not overlap. Therefore, TEER measurements are not reproducible between transwells. (B) Stable TEER values within one transwell was observed. (C-D) 75.000 cells were seeded per transwell. (C) The experimental run shows spiky TEER values within one transwell. Therefore, TEER measurements within one transwell was not stable. Furthermore, TEER values of the transwells do not overlap. Thus, TEER measurements between transwells are not reproducible. (D) No increase in TEER followed by a plateau in TEER was seen. Hence, the expected behavior was not observed. Furthermore, the TEER value decreases during the last three days, indicating no stable TEER measurements.
15
When comparing Figure 11A to 11B and Figure 11C to 11D, it is clear that the ECIS 8W1E technique is unable to of reproduce TEER values of two independent experimental runs (p <0.0001).
ECIS 8W TransFilter
Figure 12A misses data from time point 68.2 hrs to 90.4 hrs due to incorrect connection of the array to the adapter.
The ECIS 8W TransFilter shows to be more stable than the ECIS 8W1E technique, since there is less fluctuation between TEER measurements (Figure 12 and Figure 11). However, figure 12A and B do not show a plateau in TEER (mean %CV of TEER values within one transwell over the last 3 days = 18%) and therefore TEER measurements are unstable during the last three days.
Furthermore, the graphs in figure 12 both show that the TEER measurements between the transwells do not overlap (mean %CV = 23%). Therefore, the ECIS 8W TransFilter technique is not able to reproduce TEER measurements between transwells.
When comparing figure 12A to 12B, the ECIS 8W TransFilter is not able to reproduce TEER values between experimental runs (p<0.0001).
ECIS TEER24
Changing the culture medium outside the incubator increases TEER (Figure 13). This indicates an increase in TEER due to environmental changes. These spikes in TEER were
Figure 12. ECIS 8W TransFilter. (A-B) Time (hrs) was plotted on the X-axis against TEER (Ω*cm2) on the Y-axis. (A) The lack of data from 68.2hrs to 90.4 hrswas due to a connection error between the 8W TransFilter array and the adapter after culture medium exchange. TEER measurements within one transwell are stable. TEER values of the transwells do not overlap. Therefore, TEER measurements are not
reproducible between transwells. Also, a decrease in TEER was observed. This is not in line with the expected TEER curve. (B) Culture medium exchange outside the stove seemed to have an increasing effect on the TEER value. TEER measurements within one transwell are stable. TEER values of the transwells do not overlap. Therefore, TEER measurements are not reproducible between transwells. Also, a decrease in TEER was observed. This is not in line with the expected TEER curve.
16
ignored during analysis, since these TEER spikes were solely linked to the environmental changes
When ignoring the TEER spikes due to culture medium exchange, the stability within one transwell seems stable. This observation is in line with the mean %CV within one
transwell over the last 3 days of 17%.
Figure 13 shows a wide variation (mean %CV=22%) between TEER values in transwells in the last three days and therefore TEER measurements are not reproducible between transwells.
Due to technical issues of the ECIS TEER24 machine, only 1 experimental run was performed. Therefore, reproducibility of TEER values in two experimental runs could not be analysed.
Comparison four evaluated techniques
The EVOM2/Chopstick technique was the most stable technique when using a 0.33 cm2 filter insert (Table 1).
The only distinction based on reproducible TEER values between transwells that can be made, is that the ECIS 8W1E technique and the EVOM2/chopstick technique with a 0.33 cm2 filter insert is worst (Table 1). The other three techniques (EVOM2/chopstick technique
with a 1.12 cm2 filter insert, ECIS 8W TransFilter technique and ECIS TEER24 technique)
are equally capable of reproducing TEER values between transwells.
Figure 13. ECIS TEER24. Time (hrs) was plotted on the X-axis against TEER (Ω*cm2) on the Y-axis. After changing the culture medium, peaks in TEER were observed. Therefore, changing the culture medium outside the incubator has an increasing effect on TEER. These peaks in TEER should be neglected, since they are solely due to a decrease in temperature. When neglecting these peaks, TEER measurements within one transwell in the last three days seem stable. TEER values between transwells on the last three days however do not overlap. Therefore, TEER measurements cannot be reproduced between transwells.
17
The reproducibility of TEER values of an entire experimental run was determined for three techniques. The EVOM2/Chopstick technique was the only one able to reproduce TEER values between experimental runs.
System Mean %CV within one transwell Mean %CV between transwells p-values EVOM2/Chopstick 1.12 cm2 filter insert, 110,000 cells, n=32 13% 22% p=0.9293 (n=2) 0.33 cm2 filter insert, 46,000 cells, n=12 9% 29% ND ECIS 8W1E 4.9*10-4 cm2 electrode, 75,000 cells, n=3 32% 31% p<0.0001(n=2) 4.9*10-4 cm2 electrode, 150,000 cells, n=3 12% 26% p<0.0001 (n=2) ECIS 8W TransFilter 0.33 cm2 filter insert, 150,000 cells, n=4 18% 23% p<0.0001 (n=2)
ECIS TEER24 0.33 cm2 filter insert, 46,000 cells, n=16
17% 22% ND
Table 1. Comparison four evaluated techniques. Of all four techniques, the EVOM2/Chopstick technique with 0.33 cm2 filter inserts is most stable in measuring TEER values (mean %CV within one transwell=9%). The EVOM2/Chopstick technique with 1.12 cm2 filter inserts, the ECIS 8W TransFilter technique and the ECIS TEER 24 technique reproduce TEER values between transwells best (mean %CV between transwells=22-23%). Furthermore, only the EVOM2/chopstick technique was able to reproduce TEER values of an experimental run. ND: Not Determined.
18
Discussion
From the four evaluated techniques, the EVOM2/Chopstick technique with a 0.33 cm2 filter
insert generates the most stable TEER measurements within one transwell in the last three days (mean %CV = 9%). With a %CV of 9%, relative drug effects ≥18% can be detected with 95% confidence (Appendix A).
The EVOM2/Chopstick technique with a 1.12 cm2 filter insert, the ECIS 8W TransFilter technique and the ECIS TEER24 technique were equally good in reproducing TEER measurements between transwells (mean %CV=22-23%). With a %CV of 23%, absolute drug effects ≥46% can be detected with 95% confidence(Appendix A).
Based on the facts that (i) the EVOM2/Chopstick technique with a 0.33 cm2 filter insert produced the most stable TEER measurements within one transwell during the last three days, and (ii) because between transwells techniques were comparable, the EVOM2/chopstick technique was considered the best technique for TEER measurements.
Comparing the %CV between techniques however is infirm, because the number of repeats differs between techniques (Appendix A). Moreover, it is unclear whether the ECIS 8W1E and TransFilter techniques really measured TEER, because both techniques do not show the expected gradual increase in TEER followed by a plateau.
Nevertheless, also the EVOM2 chopstick technique is not ideal, because the TEER is measured outside the incubator. As shown in Figure 9, the environmental changes due to measuring outside the incubator clearly affects the TEER values. Because previous studies have shown that changes in the temperature affect TEER values (González-Mariscal, Ramírez and Cereijido, 1984; Srinivasan et al., 2015), we recommend to perform TEER measurements with the EVOM2/Chopstick technique on a heated plate to maintain the temperature at 37°C. Furthermore, the TEER values were also affected by the placement of the chopstick
electrodes (data not shown). To minimalize this effect, a set-up with fixed electrode position should be constructed.
TEER is also affected by the passage number of Caco-2 cells (Birske-Anderson, Finley and Newman, 1997; Senarathna and Crowe, 2015; Srinivasan et al., 2015; Lu et al., 1996). The higher the passage number, the higher the TEER values (Birske-Anderson, Finley and Newman, 1997; Senarathna and Crowe, 2015; Srinivasan et al., 2015). In our present study, the passage number of the Caco-2 cells was not taken into account.
19
Also, sub clones of Caco-2 cell lines showed to have variation in permeability
characteristics (Kissel and Walter, 1995). Therefore, the clone of Caco-2 cells will also affect TEER. This effect may explain the wide variation in Caco-2 TEER values between studies (Appendix B). Because the Caco-2 cells of one clone were used in our present study, the observed variation in TEER values cannot be explained by differences between clones.
At present, there is no standard operating procedure to culture Caco-2 cells and to perform TEER measurements. For instance, a variety in cell seeding densities is used in the field (Appendix B). Cell seeding densities influence a number of physiological and
morphological properties of differentiated Caco-2 cells (Natoli et al., 2011). Behrens and Kissel (2003) showed an average seeding density of 6*104 cells/cm2 is most favourable. Lower seeding densities led to thin monolayers with altered tight junctions and higher seeding densities to the formation of multilayer (Behrens and Kissel, 2003).
Furthermore, how long Caco-2 cells have to be cultured before fully differentiated, is obscure (Appendix B). This is a problem because full differentiation is a prerequisite when studying TEER (Hildago, Raub and Borchardt, 1989). In addition, several methods are used to monitor the extent of differentiation, and the results from these different assays differ between each other. For example, Bravo et al. (2004) showed that a maximal and homogenous
expression of the PepT1 transporter was achieved after 25 days, whereas both TEER and mannitol permeability reached a plateau after 21 days. Due to this inconsistency, performing multiple differentiation assays is recommended. Another option would be to use techniques, for instance a scanning electron microscope, to visualize the morphology of the Caco-2 cells. In this way, typical characteristics of a differentiated Caco-2 cell, like microvilli and a columnar shape, can be observed (Leitch et al., 2005).
Altogether, the EVOM2/chopstick technique performed on a heated plate, with fixed electrode positions, known passage number, 6*104 cells/cm2 and differentiated cells,
determined via multiple differentiation assay, has the greatest potential to generate stable and reproducible TEER measurements. With stable and reproducible TEER measurements, the effect of a drug on the TEER of cells can be monitored, which could speed up the discovery of a drug treating diseases related to epithelial barrier dysfunction.
20
References
Anderle, P., Rakhmanova, V., Woodford, K., Zerangue, N. and Sadée, W. (2003). Messenger RNA Expression of Transporter and Ion Channel Genes in Undifferentiated and Differentiated Caco-2 Cells Compared to Human Intestines. Pharmaceutical Research, 20(1), pp. 3-15.
Béduneau A., Tempesta C., Fimbel S., Pelleguer, Y., Janin, V., Demarne, F and Lamprecht, A. (2014). A Tunable Caco-2/HT29-MTX Co-Culture Model Mimicking Variable Permeabilities of the Human Intestine Obtained by an Original Seeding Procedure. European journal of pharmaceutics and biopharmaceutics, 87(2), pp. 290–298.
Behrens, I. and Kissel, T. (2003). Do cell culture conditions influence the carrier-mediated transport of peptides in Caco-2 cell monolayers? European journal of pharmaceutical sciences, 19(5), pp. 433-442.
Bravo, S.A., Nielsen, C.U., Amstrup, J., Frokjaer, S. and Brodin, B. (2004). In-depth evaluation of Gly-Sar transport parameters as a function of culture time in the Caco-2 cell model. European Journal
of Pharmaceutical Sciences, 21(1), pp. 77-86. doi https://doi.org/10.1016/S0928-0987(03)00205-7
Briske-Anderson, M.J., Finley, J.W. and Newman, S.M. (1997). The Influence of Culture Time and Passage Number on the Morphological and Physiological Development of Caco-2 Cells.
Proceedings of the Society for Experimental Biology and Medicine, 214(3), pp. 248-257. doi:
https://doi.org/10.3181/00379727-214-44093
Buzza, M.S., Netzel-Arnett, S., Shea-Donohue, T., Zhao, A., Lin, C., List, K., Szabo, R., Fasano, A., Bugge, T.H & Antalis, T.M. (2010). Membrane-anchored serine protease matriptase regulates epithelial barrier formation and permeability in the intestine. Proceedings of the National Academy of
Sciences, 107(9), pp. 4200-4205. doi: 10.1073/pnas.0903923107
Chen, S., Einspanier, R. & Schoen, J. (2015). Transepithelial electrical resistance (TEER): a functional parameter to monitor the quality of oviduct epithelial cells cultured on filter supports.
Histochemistry and Cell Biology, 144(5), pp. 509-515.
Deprez, S., Mila, I., Huneau, J.F., Tome, D. and Scalbert, A. (2001). Transport of
Proanthocyanidin Dimer, Trimer, and Polymer Across Monolayers of Human Intestinal Epithelial Caco-2 Cells. Antioxidants & Redox Signaling, 3(6).
Feighery L.M., Cochrane S.W. and Quinn T. (2008). Myosin Light Chain Kinase Inhibition: Correction of Increased Intestinal Epithelial Permeability in Vitro. Pharmaceutical research, 25(6), pp. 1377–1386.
Ferruzza S., Rossi C., Sambuy Y. and Scarino, M.L. (2013) Serum-Reduced and Serum-Free Media for Differentiation of Caco-2 Cells. ALTEX, 30(2), pp. 159–168.
González-Mariscal, L., Chávez de Ramírez, B. and Cereijido, M. (1984). Effect of temperature on the occluding junctions of monolayers of epithelioid cells (MDCK). The journal of Membrane
21
Hégarat, L.L., Huet, S. and Fessard, V. (2012). A co-cultured system of human intestinal Caco-2 cells and lymphoblastoid TK6 cells for investigating the genotoxicity of oral compounds.
Mutagenesis, 27(6), pp. 631-636.
Hellinger, E., Veszelka, S., Tóth, A.E., Walter, F., Kittel, Á., Bakk, M.L., Tihanyi, K., Háda, V., Nakagawa, S., Duy, T.D.H, Niwa, M, Deli, M.A. and Vastag, M. (2012). Comparison of brain capillary endothelial cell-based and epithelial (MDCK-MDR1, Caco-2, and VB-Caco-2) cell-based surrogate blood-brain barrier penetration models. European Jornal of Pharmaceutics and
Biopharmaceutics, 82(2), pp. 340-351.
Hildalgo, I.J., Raub, T.J. and Borchardt, R.T. (1989). Characterization of the Human Colon Carcinoma Cell Line (Caco-2) as a Model System for Intestinal Epithelial Permeability. Gastroenterology, 1909(96), pp. 736-749. doi: https://doi.org/10.1016/0016-5085(89)90897-4
Hilgendorf C., Spahn-Langguth H., Regårdh C.G., Lipka, E., Amidon, G.L. and Langguth, P. (2000). Caco-2 Versus Caco-2/HT29-MTX Co-Cultured Cell Lines: Permeabilities Via Diffusion, Inside- and Outside-Directed Carrier-Mediated Transport. Journal of pharmaceutical sciences, 89(1), pp. 63–75.
Kitchens, K.M., Kolhatkar, R.B., Swaan, P.W., Eddington, N.D. and Ghandehari, H. (2006). Transport of Poly(Amidoamine) Dendrimers across Caco-2 Cell Monolayers: Influence and Size, Charge and Fluorescent Labeling. Pharmaceutical Research, 23(12), pp. 2812-2826.
Kotzé, A.F., Lueßen, H.L., de Leeuw, B.J., de Boer, B.G., Verhoef, J.C. and Junginger, H.E. (1998). Comparison of the effect of different chitosan salts and N-trimethyl chitosan chloride on the permeability of intestines apithelial cells (Caco-2). American Chemical Society and American Pharmaceutical Association, 88(2), pp. 253-257.
Laukoetter, M.G., Nava, P. & Nusrat, A. (2008). Role of the intestinal barrier in inflammatory bowel disease. World Journal Gastroenterology, 14(3), pp. 401-407. doi: 10.3748/wjg.14.401
Leitch, G.J., Ward, T.L., Shaw, A.P. and Newman, G. (2005). Apical Spore Phagocytosis Is Not a Significant Route of Infection Of Differentiated Enterocytes By Encephalotizoon intestinalis. Infection
and Immunity, 73(11), pp. 7697-7704. doi: 10.1128/IAI.73.11.7697–7704.2005
Lu, S., Gough, A.W., Bobrowski, W.F. and Stewart, B.H. (1996). Transport properties are not altered across Caco-2 cells with heightened TEER despite underlying physiological and ultrastructural changes. Journal of Pharmaceutical Sciences, 85(3), pp. 270-273. doi:
https://doi.org/10.1021/js950269u
Maffei, D. & Schanler, R.J. (2017). Human milk is the feeding strategy to prevent necrotizing enterocolitis. Seminars in perinatology, 41, pp. 36-40. doi:
http://dx.doi.org./10.1053/j.semperi.2016.09.016
Merga, Y., Campbell, B.J. & Rhodes, J.M. (2014). Mucosal Barrier, Bacteria and Inflammatory Bowel Disease: Possibilities for Therapy. Digestive Diseases, 32(4), pp. 475-483. doi:
https://doi.org/10.1159/000358156
Natoli, M., Leoni, B.D., D’agnano, I., D’onofrio, M., Brandi, R., Arisi, I., Zucco, F. and Felsani, A. (2011). Cell Growing Density Affects the Structural and Functional Properties of Caco-2
22
Differentiated Monolayer. Journal of Cellular Physiology, 226(6), pp. 1531-1543. doi: 10.1002/jcp.22487
Ranaldi, G., Marigliano, I., Vespignani,I., Perozzi, G. and Sambuy, Y. (2002). The effect of chitosan and other polycations on tight junction permeability in the human intestinal Caco-2 cell line. The journal of Nutritional Biochemistry, 13(3), pp.157-167.
Sambuy, Y., Angelis, I.D., Ranaldi, G., Scarino, M.L., Stammati, A. and Zucco, F. (2005). The Caco-2 cell line as a model of the intestinal barrier: influence of cell and culture-related factors on Caco-2 cell functional characteristics. Cell biology and Toxicology, 21(1), pp. 1-26. doi:
https://doi.org/10.1007/s10565-005-0085-6
Senarathna, S.M. and Crowe, A. (2015). The influence of passage number for Caco-2 cell models when evaluating P-gp mediated drug transport. Die Pharmazie, 70(12), pp. 798-803. doi:
10.1691/ph.2015.5106
Shen, L., Weber, C.R. & Turner, J.R. (2008). The tight junction protein complex undergoes rapid and continuous molecular remodeling at steady state. Journal of Cell Biology, 181(4), pp. 683-695. doi: 10.1083/jcb.200711165
Springer, S.C., Annibale, D.J. & Aslam, M. (2017). Necrotising Enterocolotis. Medscape, retrieved from: https://emedicine.medscape.com/article/977956-overview
Srinivasan, B., Koli, A.R., Esch, M.B., Abaci, H.E., Shuler, M.L. and Hickman, J.J. (2015). TEER measurement techniques for in vitro barrier model systems. Journal of Laboratory Automation, 20(2), pp. 107-126. doi: 10.1177/2211068214561025
Thanou, M.M., Kotzé, A.F., Scharringhausen, T., Leußen, H.L., de Boer, A.G., Verhoef, J.C. and Junginger, H.E. (2000). Effect of degree of quaternization of N-trimethyl chitosan chloride for enhanced transport of hydrophilic compounds across intestinal Caco-2 cell monolayers. Journal of Controlled Release, 64(1-3), pp. 15-25.
Wahlang, B., Pawar, Y.B. and Bansal, A.K. (2011). Identification of permeability-related hurdles in oral delivery of curcumin using the Caco-2 cell model. European Journal of Pharmaceutics and Biopharmaceutics, 77(2), pp. 275-282.
Walgren, R.A., Walle, K.U. and Walle, T. (1998). Transport of Quercetin and Its Glucosides across Human Intestinal Epithelial Caco-2 Cells. Biochemical Pharmacology, 55(10), pp. 1721-1727.
Yamashita, S., Furubayashi, T., Kataoka, M., Sakane, T., Sezaki, H. and Tokuda, H. (2000). Optimized conditions for prediction of intestinal drug permeability using Caco-2 cells. European Journal of Pharmaceutical Sciences, 10(3), pp. 195-204.
Yee, S. (1997). In Vitro Permeability Across Caco-2 Cells (Colonic) Can Predict In Vivo (Small Intestinal) Absorption in Man – Fact or Myth. Pharmaceutical Research, 14(6), pp. 763-766.
23
Appendix A
%CV coupled to the needed size of a drug effect
The mean %CV can be used to estimate what size effect a drug needs to generate in order to be labelled as an actual effect with 95% confidence. For this, the following formula was used:
𝜇
𝜇0− 1.96 ∗ 𝑚𝑒𝑎𝑛 %𝐶𝑉 𝑡𝑜 𝜇
𝜇0+ 1.96 ∗ 𝑚𝑒𝑎𝑛 %𝐶𝑉 µ=mean after drug addition
µ0=mean before drug addition
μ
μ0=size of the effect after drug addition
The formula describes a 95% confidence interval. The 95% confidence interval should never include 0, since a ‘0 effect’, and therefore no effect, would then be labelled as an effect. Therefore, when the mean %CV is high, μ0μ should increase in such a way that 0 will not be included in the 95% confidence interval. This means that with a high mean %CV, the size of the effect after drug addition has to increase in order to label it as an effect with 95%
confidence. In other words, with a high mean %CV, small effects of a drug cannot be detected. With a low mean %CV, small effects of a drug can be detected. It is up to the researcher what size of effect is expected from the drug and based on this, what mean %CV is acceptable. For instance, when a researcher expects a 10% increase in TEER after drug addition, the mean %CV should not exceed 5%. However, when a 30% increase in TEER is expected, the mean %CV can be up to 15%.
When a technique measures stable TEER values within transwells, a relative drug effect can be observed. When a technique measures reproducible TEER values between transwells, an absolute drug effect can be observed. With relative drug effects, the data needs to be normalized before determining whether the drug had an effect on TEER. With absolute drug effects, the data can be directly be used to determine whether the drug had an effect on TEER. Therefore, reproducibility of TEER values between transwells would be preferable.
When the mean %CV is too high for the size of an effect the researcher wishes to analyse, the %CV can be lowered by increasing the number of repeats (n). For this, the %CV should be divided by √𝑛
24
Appendix B
Discrepancies in Caco-2 cell studies concerning cell seeding densities, days until cell monolayer differentiation, TEER values and TEER measuring techniques
Cell density [cells/cm2] Days until differentiation* TEER values [Ω/cm2] TEER measuring technique Reference 8.5*103 14 Not mentioned Millicell® ERS
voltohmmeter
(Anderle et al., 2003)
6*104 21
400-900 EVOM-2/Chopstick (Behrens and Kissel, 2003)
6*104 21 700 Millicell® ERS
voltohmmeter
(Hégarat, Huet and Fessard, 2012)
6.3*104 16 173.5 Not mentioned
(Hildalgo, Raub and Borchardt, 1989) 3.5*105 15-21 1100-1350 Millicell® ERS
voltohmmeter
(Feruzza et al., 2012) 3*105 18-21 Not mentioned Millicell® ERS
voltohmmeter
(Yamashita et al., 2000)
1.2*105 21 Not used
(transport study)
Not used (transport study)
(Yee, 1997)
8*104 21-28 Not mentioned EVOM-2/Chopstick (Kitchens et al., 2006)
8.8*104 20-30 664-1423 Millicell® ERS voltohmmeter (Lu et al., 1996) 1.1*105 21-22 300 Millicell® ERS voltohmmeter
(Wahlang, Pawar and Bansal, 2011)
1*104 23-25
Not mentioned Millicell® ERS voltohmmeter
(Thanou et al., 2000)
4*104 15
200-1,100 EVOM2/Endohm (Deprez et al., 2001) 4*105 15-17 800-1200 Millicell® ERS
voltohmmeter
(Ranaldi et al., 2002) 4.5*105 19-21 1024 EVOM-2/Chopstick (Hellinger et al., 2012)
1x105 20-35 >350 Millicell® ERS
voltohmmeter
(Walgren, Walle and Walle, 1998) 1*104 21-23 800-1000 Millicell® ERS voltohmmeter (Kotzé et al., 1998) 6.6*104 and 8.9*104
14-21 1400-2400 EVOM2/Chopstick (Hilgendorf et al., 2000) 5*105 16-21 >400 EVOM2/Endohm Feighery, Cochrane and
Quinn, 2008) 1.5*104 – 9.1*104 21-30 250 Millicell® ERS voltohmmeter (Béduneau et al., 2014)
25
