Main

Cystic fibrosis (CF) is caused by mutations in the CF transmembrane conductance regulator (CFTR) gene, which codes for the cystic fibrosis transmembrane conductance regulator (CFTR) (1–3). CFTR, a member of the adenosine triphosphate binding cassette transporter glycoprotein superfamily, functions as a chloride channel and regulates other ion channels in the apical membrane of certain epithelia. Over a thousand disease-causing mutations have been identified in the CFTR gene. These mutations have been separated into five major subtypes, which include null/decreased production (Class I), improper processing and trafficking (Class II), ion channel dysregulation (Class III), altered channel conduction (Class IV), and improper protein targeting (Class V). The most common mutation in the CFTR gene is ΔF508, a Class II mutation in which the CFTR protein is retained in the endoplasmic reticulum and targeted for degradation (4–6). ΔF508-CFTR can function as a chloride channel but is unable to reach the apical membrane (7,8).

A number of future therapeutic strategies for CF are focused on correcting or circumventing the mutant CFTR gene, with the hope that this will result in decreased pathology caused by the primary gene defect. A variety of compounds are being investigated as potential treatments in CF, for instance, 4-phenylbutyrate, trimethylamine oxide (TMAO), and curcumin, as well as flavonoids. 4-Phenylbutyrate, TMAO, and curcumin have been shown to increase ΔF508-CFTR surface expression in a number of in vitro systems (9–11), whereas once at the membrane, flavonoid compounds (genistein, apigenin) and vitamin C have been suggested to activate ΔF508-CFTR (12–14). In addition, gentamicin, an aminoglycoside antibiotic, has been shown to suppress premature termination codons and restore CFTR function in patients with the W1282X mutation (15). Other interventions, such as gene therapy using viral vectors and stem cell transplant, have also been suggested as potential interventions in CF.

Assessing the effects of the various interventions aimed at correcting the CF electrophysiologic phenotype in affected epithelia is primarily performed using in vitro electrophysiologic methods such as examining excised tissue mounted in Ussing chambers. The nasal potential difference (PD) assay is a widely accepted in vivo method, which has been used to assess functional expression of CFTR in respiratory epithelia in humans and was pioneered by Knowles et al. (16,17) in the 1980s. This method has been adapted by Grubb (18) and now allows for the investigation of many potential interventions for CF in various mouse models of CF. Similarly, a reliable in vivo assay of CFTR function in intestinal epithelia could be of great value in the investigation of pharmacologic interventions for CF mouse models. It has been suggested that the rectal potential difference (RPD) assay is a reliable in vivo method for assessment of CFTR activity (10, 11, 19). However, there are no reported studies that evaluate the reliability and reproducibility of this method.

METHODS

Mice.

Two models of gene-targeted mice for the ΔF508 mutation were used in the experiments. One of the mouse models was bred on a mixed genetic background [approximately 25% 129 and 75% Black 6 (Bl6) (20)], whereas the other ΔF508 mouse model was further backcrossed into a pure Bl6 background. We refer to these mouse models respectively as the mixed genetic background and pure Bl6 background ΔF508 mice. A transgenic CFTR−/− knockout (B6.129P2Cftrtm1Unc) was also used in the experiments as a CFTR-null model (21,22). The ΔF508-CFTR+/+ and the CFTR-null mice have severe gastrointestinal manifestations, including bowel obstruction, bowel strictures, and peritonitis. All mice were fed with Harlan Teklad 9F food (Harlan Teklad, Madison, WI) and drinking water that was supplemented with 17.5 g/250 mL of Colyte (Schwarz Pharma, Milwaukee, WI) to improve the survival of the animals (23). All three CF mouse models were bred to obtain wild-type, heterozygote, and CF-affected animals for each model system, were maintained at the Yale University Animal Facility, and were genotyped with standard protocols. All procedures were performed in compliance with relevant laws and institutional guidelines and were approved by the Yale University Institutional Animal Care and Use Committee.

In vivo potential difference measurements.

The intestinal PD was measured using a probing bridge made from PE50 tubing inserted into the rectum of the mouse. A 3 mm insertion depth for RPD and a 15 mm insertion depth for “distal colon” PD (DCPD) were used. AgCl2 voltage electrodes were connected by way of 3% agar bridges made with 3 M KCl. The insertion depths for the rectal and distal colon have been confirmed at the time of sacrifice to ensure proper placement of the tubing. The experimental electrode was connected through the probing bridge, and the reference electrode was connected through a 23-gauge butterfly needle filled with normal saline and placed subcutaneously. A microperfusion pump controlled the flow of perfusate through the probing electrode (0.5 mL/h). A separate length of PE50 tubing was inserted in the rectum and attached to a suction apparatus to minimize distension of the gut. During the procedure, the mice were maintained on a warming pad, and saline drops or artificial tears were used to keep their eyes moist. Mice were anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg) intraperitoneally before the procedure. Mice were continually observed throughout the procedure and postprocedure until they were fully recovered.

In vitro potential difference measurements.

We performed the in vitro analysis as previously described by Grubb (18). The rectum and distal colon were removed, cut longitudinally, washed of any intestinal contents, and mounted in the Ussing chamber with an aperture size of 0.04 cm2 for rectal tissue and 0.3 cm2 for distal colonic tissue (Physiologic Instruments, San Diego, CA). The PD was measured under open-circuit conditions. During all experiments, solutions were continuously gassed with 95% O2/5% CO2 and warmed to 37°C. Data were recorded continuously using the Acquire and Analyze software program (Physiologic Instruments, San Diego, CA).

Solutions.

The solutions used in this study were a Kreb's bicarbonate Ringer's solution (KBR) and a chloride-free Ringer's solution containing 5 mM barium hydroxide. Amiloride and forskolin were added to the chloride-free Ringer's solution at final concentrations of 10−4 M and 10−5 M, respectively. In a subset of experiments, a specific CFTR inhibitor, CFTRinh-172, was added to the final perfusate. All drugs and chemicals were obtained from Sigma Chemical Co. (St. Louis, MO) and J. T. Baker (Phillipsburg, NJ).

Statistics.

The t test was used to compare groups for significant differences. A p value of less than 0.05 was used to signify statistical significance. All results are reported as the mean ± SEM.

RESULTS

In vivo intestinal potential difference.

The in vivo intestinal PD was measured for 30 min. Initially, a KBR solution is perfused for a baseline measurement. It is quickly switched to the first perfusate solution that is used in the RPD assay, a chloride-free Ringer's solution containing Ba(OH)−2 and amiloride, as previously described by Fischer et al. (10,14), and the transepithelial PD is recorded for 15 min. The addition of amiloride and barium block sodium and potassium currents that could contribute to the PD. Subsequently, forskolin was added to the perfusate, and the PD was measured for an additional 15 min to activate CFTR chloride current. Data were obtained from two separate regions of the murine intestinal tract, the distal colon and the rectum (Fig. 1). The baseline PD, the mean change in the PD in response to forskolin (ΔPDfsk), and the peak forskolin PD for all CF mouse models were obtained for distal colon (Table 1) and rectum (Table 2).

Figure 1
figure 1

Representative tracings of in vivo DCPD. Intestinal PD was measured using a probing bridge made from PE50 tubing inserted 15 mm in depth, as outlined in the “Methods” section. During the initial period, the PD is measured in a chloride-free Ringer's solution containing amiloride and barium hydroxide. Subsequently, forskolin is added to the perfusing solution. (A) Representative tracing from a nonCF animal. (B) CF-affected mouse.

Full size image
Table 1A In vivo distal colon potential difference parameters from three CF mouse models
Full size table
Table 2A In vivo rectal potential difference parameters from three CF mouse models
Full size table

At baseline, in vivo intestinal PD measurements of wild-type and heterozygote mice revealed a modest hyperpolarization, which was similar in magnitude for all three models. Furthermore, in wild-type and heterozygous mice, the addition of forskolin resulted in a further hyperpolarization (a more negative PD measurement). This observation held true in all three models for both the distal colon and rectum (Tables 1 and 2). Interestingly, in both distal colon and rectum, the wild-type mice demonstrated a more negative baseline PD, a more negative peak PD in response to forskolin, and a larger ΔPDfsk than the heterozygous mice (Tables 1 and 2). However, these trends did not reach statistical significance except in the CFTR-null mouse model.

The baseline PD in CF-affected mice was significantly more positive when compared with the heterozygous and wild-type mice. This observation held true for all three CF mouse models regardless of the intestinal region studied (Tables 1 and 2). In addition, whereas all of the wild-type and heterozygous mice demonstrated a hyperpolarization of the PD in response to forskolin, the CF-affected mice demonstrated a depolarization of the PD in response to forskolin (Tables 1 and 2). These differences are similar to previously reported in vivo findings and are consistent with the CF phenotype (10).

Table 1B In vitro distal colon potential difference parameters from three CF mouse models
Full size table
Table 2B In vitro rectal potential difference parameters from three CF mouse models
Full size table

To demonstrate that the forskolin stimulated current was caused by CFTR activity, 20 μM of CFTRinh-172, a specific CFTR inhibitor, was added to the perfusate solution after forskolin stimulation in a subset of experiments (24). The addition of the blocker significantly inhibited the hyperpolarization that was observed in the wild-type and heterozygote animals in all three mouse models (n = 16). These findings suggest that the majority of the current activated by forskolin was caused by CFTR activation. No effect was observed in the CF-affected animals (n = 6) (Fig. 2).

Figure 2
figure 2

CFTRinh-172 inhibits the forskolin stimulated hyperpolarization observed in the in vivo RPD/DCPD assay in all three CF animal models. Twenty micromolar of CFTRinh-172 was added to the chloride-free Ringer's solution containing amiloride, barium hydroxide, and forskolin. Mixed background ΔF508-CFTR colony (n = 6); CFTR null colony (n = 5); pure Bl6 background ΔF508-CFTR colony (n = 5).

Full size image

In vivo ΔPDfsk is a reproducible measurement when assayed repeatedly on a single animal.

To test the reproducibility of the in vivo methods, we repeated three in vivo DCPD assays on five heterozygous and six CF-affected mice at 2 wk to monthly intervals. Although there is some variability in the absolute values, the ΔPDfsk remained similar for all of three experiments for each animal (Fig. 3) For instance, the ΔPDfsk for one of the heterozygous mice was −10.7, −11.2, and −9.6 mV, respectively, and for one of the CFTR-null mice was 3.3, 2.0, and 2.5 mV, as noted in Fig. 3.

Figure 3
figure 3

In vivo PD measurements in CF-affected and normal mice are reproducible. Repeated DCPD measurements were obtained in three separate experiments on one heterozygote (CFTR null mouse) and on one homozygote CFTR null mouse. (A) PD measurements for a heterozygous (CFTR-null) mouse are graphed over time: trial #1 (♦), trial #2 (▪), trial #3 (▴), and mean values (○) for the three experiments. Horizontal black bar represents a time length of 15 min. Black arrow denotes change to forskolin containing solution. (B) ΔPDfsk for each trial depicted in A: trial #1 (□), trial #2 (▪), trial #3 (), and the mean () ΔPDfsk for the three experiments. (C) The average ΔPDfsk for three repeated RPD assays on five heterozygote mice. (D) PD measurements for a homozygous (CFTR-null) mouse are graphed over time: trial #1 (♦), trial #2 (▪), trial #3 (▴), and the mean values (○) for the three experiments. Horizontal black bar represents a time length of 15 min. Black arrow denotes change to forskolin containing solution. (E) ΔPDfsk for each trial is depicted in D: trial #1 (□), trial #2 (▪), trial #3 (), and the mean () ΔPDfsk for the three experiments. (F) The average ΔPDfsk for three repeated RPD assays on six CF-affected mice. In all panels, error bars represent the SEM.

Full size image

In Vivo Intestinal Potential Difference versus In Vitro Intestinal Potential Difference.

Next, to test whether the in vivo PD assay results correlated with those observed in an in vitro assay, we measured the in vitro intestinal PD using the Ussing chamber. Before the in vitro assay, in vivo PD measurements were obtained on each animal. Immediately after the in vivo assay, the animals were killed, and the in vitro assay was performed. Data were again obtained from two separate regions of the murine intestinal tract, the distal colon and the rectum, after a 20-min equilibration period in KBR. The same protocol and solutions were used in the Ussing chamber as in the in vivo PD method. The mean baseline RPD and DCPD were similar in all three mouse models regardless of whether the Ussing chamber or the RPD method was used to obtain measurements (Figs. 4 and 5).

Figure 4
figure 4

Comparison of the observed change in PD in response to 10 μM forskolin (ΔPDfsk) in wild-type (□), heterozygousL (), and CF-affected mice (▪) in three CF mouse models. In vitro results are represented on the left side of each panel and in vivo results on the right side. n = 5 for all groups (wild type, heterozygote, CF affected) and both tissues (rectum and distal colon). *p < 0.05 between in vivo and in vitro results for distal colon in CF-affected groups only. (A) ΔF508-CFTR mixed background mouse model. (B) CFTR-null CF mouse model. (C) ΔF508-CFTR pure Bl6 background mouse model.

Full size image
Figure 5
figure 5

Comparison of the mean PD measured at minute intervals using the in vitro assay (closed symbols) or the in vivo (open symbols) assay. CF results are represented as circles, wild type as squares, and heterozygote as triangles. Horizontal black bar represents a time length of 15 min. Black arrow denotes change to forskolin containing solution. (A and B) Traces for the mixed background ΔF508 mice (distal colon A, rectum B). (C and D) Traces for the CFTR-null mice (distal colon C, rectum D). (E and F) Traces from the pure Bl6 background ΔF508 mice (distal colon E, rectum F).

Full size image

For all three CF mouse models, the forskolin response in both the distal colon and rectum appeared similar in magnitude when using either in vivo or in vitro methods to measure the ΔPDfsk (Figs. 4 and 5), and the inhibition of the forskolin response by CFTRinh-172 was also similar. There was only one significant difference observed between in vivo and in vitro measurements, and it was observed in the CF-affected mice. In the CF mice, there was a greater depolarization in the in vivo distal colon (ΔPDfsk) analysis compared with the in vitro results. This was true for all three CF mouse models. The distal colon ΔPDfsk measurements were 3.8 ± 0.4 mV versus. 0.9 ± 0.3 mV for mixed genetic background ΔF508 mouse model, 3.7 ± 0.3 mV versus 1.1 ± 0.2 mV for CFTR-null model, and 3.9 ± 0.6 mV versus 1.5 ± 0.3 for pure Bl6 background ΔF508 mouse model. A similar trend was also observed for the rectal PD of the CF mice; however, it was not of statistical significance. Last, a significant inhibition of the forskolin response was observed with the addition of CFTRinh-172 in both the wild-type and heterozygous tissues, similar to what observed in the in vivo assay (n = 9, data not shown). Once again, there was no effect of CFTRinh-172 on the CF-affected tissue (n = 5, data not shown).

DISCUSSION

The in vivo rectal PD assay appears to be a reliable and reproducible assay of functional CFTR expression in CF mouse models. In our study, the in vivo PD measurements correlated very closely with the PD measurements acquired using the in vitro method, suggesting that the in vivo PD assay may be a very useful alternative method for studying ion transport in CF-affected animals if serial measurements are required as part of a study design. Although the results of the two methods are nearly identical, there were interesting subtle differences when comparing techniques. Of note, in CF-affected animals, the degree of depolarization observed in response to forskolin appeared to be greater when using the in vivo assay than the in vitro assay.

In addition, in the in vivo PD assay, wild-type mice demonstrated a larger ΔPDfsk, a more negative baseline PD, and a more negative peak PD in response to forskolin compared with their heterozygote counterparts, suggesting that there is a greater degree of functional CFTR expression in wild-type than in heterozygous mice. This trend was noted for all three CF mouse models studied, in both the distal colon and rectum. Our in vivo data correlate with the observations of Gabriel et al. (25). They reported that heterozygote CFTR(−/+) mice were more resistant to secretory diarrhea after exposure to cholera toxin when compared with wild-type animals. They postulated that this heterozygote advantage was caused by less CFTR expression in the CFTR heterozygous mice (26). Interestingly, the PD measurements from the same mice using the in vitro method did not show statistically significant differences between the wild-type and heterozygous groups, which is similar to the findings of Cuthbert et al. (26) and is consistent the initial characterization of intestinal transport by in vitro methods in the ΔF508 mixed genetic background mouse model (19). We speculate that the subtle differences that are detected using the in vivo method result from the fact that the tissue is in situ and is subject to the endogenous neural-humoral supply of the gut. This is distinctly different from the removal and mounting of the same tissue in an artificial environment to maintain tissue viability.

Both methods of studying intestinal transport have their limitations. For example, serial measurements of CFTR functional expression are not possible when using the in vitro method. On the other hand, it is impossible to control the internal and external environments affecting the tissue when using the in vivo method. However, both methods used in concert can yield important details regarding the ion transport properties of the intestinal epithelia. The power of the in vivo method is its ability to measure CFTR functional expression before and after interventions aimed at correcting the CF electrophysiologic phenotype.