Characterization of Nasal Potential Difference in cftr Knockout and F508del-CFTR Mice

Background Treatments designed to correct cystic fibrosis transmembrane conductance regulator (CFTR) defects must first be evaluated in preclinical experiments in the mouse model of cystic fibrosis (CF). Mice nasal mucosa mimics the bioelectric defect seen in humans. The use of nasal potential difference (VTE) to assess ionic transport is a powerful test evaluating the restoration of CFTR function. Nasal VTE in CF mice must be well characterized for correct interpretation. Methods We performed VTE measurements in large-scale studies of two mouse models of CF—B6;129 cftr knockout and FVB F508del-CFTR—and their respective wild-type (WT) littermates. We assessed the repeatability of the test for cftr knockout mice and defined cutoff points distinguishing between WT and F508del-CFTR mice. Results We determined the typical VTE values for CF and WT mice and demonstrated the existence of residual CFTR activity in F508del-CFTR mice. We characterized intra-animal variability in B6;129 mice and defined the cutoff points for F508del-CFTR chloride secretion rescue. Hyperpolarization of more than -2.15 mV after perfusion with a low-concentration Cl- solution was considered to indicate a normal response. Conclusions These data will make it possible to interpret changes in nasal VTE in mouse models of CF, in future preclinical studies.


Introduction
Cystic fibrosis (CF) is a lethal autosomal recessive disease that affects one in 2500 newborns in Caucasian population [1]. This disease is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene, resulting in the production of a defective CFTR protein. CFTR is the main chloride (Cl -) channel in secretory epithelia and also acts as a regulator of sodium (Na + ) transport, through inhibition of the ENaC Na + channel [2]. Mutations in the cftr gene lead to the synthesis of a non-functional CFTR, causing dehydration of the airway surface liquid, thereby impeding mucociliary clearance and creating a favorable microenvironment for bacterial infections. The most frequent mutation results in deletion of the phenylalanine residue in position 508 (F508del-CFTR). This mutation leads to the retention of the F508del-CFTR protein in the endoplasmic reticulum and impaired function of any F508del-CFTR reaching the apical membrane [3]. There is currently no curative treatment for CF. Several strategies are currently being investigated for direct correction of the mutated CFTR defects, by rescuing trafficking defects [4][5][6] or rendering the mutated CFTR functional [5,7]. However, these approaches must be tested in animal models. CF mice display nasal epithelium ionic transport abnormalities similar to those observed in humans with CF: abnormally high levels of Na + absorption and an absence of Clsecretion in response to perfusion with a low-concentration Clsolution or a solution lacking this anion [8].
Transepithelial nasal potential difference (V TE ) measurement is the most appropriate method for the in vivo exploration of ionic transport in CF [9]. This technique has been used in phase II clinical trials, as a means of assessing the restoration of CFTR function [10,11]. It may also be very useful for preclinical studies assessing the efficacy of CFTR correctors or potentiators [6,[12][13][14].
However, V TE measurement protocols differ between studies. Data have been obtained from pooled mice of different backgrounds [15,16], for small numbers of mice and not for all V TE parameters [15,17]. Only two backgrounds are well characterized [14,[18][19][20]. Moreover, few data are available concerning variability within and between animals and no threshold for a significant, drug-related change has been validated. The aim of our study was i) to establish typical V TE values, in the FVB and B6;129 backgrounds, for F508del-CFTR and cftr -/mice, respectively, ii) to determine the repeatability of V TE measure-ments, iii) to determine threshold V TE values distinguishing between the CF and WT electrophysiological responses in F508del-CFTR mice. These data should improve the use of CF mice in preclinical studies.

Mouse models
We studied male and female B6;129-CFTR tm1-Unc (cftr -/-) mice, FVB mice homozygous for the F508del-CFTR mutation (F508del-CFTR) and their respective wild-type (WT) control littermates. Mice were obtained from CDTA (Orléans, France) and housed at the SPF Animal Care Facility of Necker University. The mice were 8 to 16 weeks old and weighed 20 to 28 g. All mice were fed a fiber-free diet. Colopeg (17.14 g/l; Bayer Santé Familiale, France) was administered to CF mice to prevent intestinal obstruction. Animal protocols were approved by the local ethics committee dealing with animal welfare and conformed to European Community regulations for the use of animals in research (authorization no. P2.AE.092.09).

V TE measurements
The method for nasal potential difference measurement was adapted and miniaturized from that developed by our group for use in young children [21]. Mice were anesthetized by an intraperitoneal injection of ketamine (133 mg/kg; IMALGENE 1000, MERIAL, France) and xylazine (13.3 mg/kg; Rompun 2%, BayerPharma, France). Mice were positioned on a 45u tilt board and a paper pad was placed under the nose to prevent the mice suffocating.
Transepithelial potential was measured between an Ag/AgCl reference electrode and an Ag/AgCl exploring electrode. The two Ag/AgCl electrodes were connected to a high-impedance voltmeter (LOGAN Research Ltd, United Kingdom). The reference electrode was connected to a subcutaneous needle with an agar bridge. The exploring Ag/AgCl electrode was connected to the nasal mucosa through a double-lumen polyethylene catheter (0.5 mm in diameter) inserted into the right nostril to a depth of 4 mm. Recordings were made every second, during continuous flow, at a rate of 0.15 ml/h, of the initial Clsolution (140 mM NaCl, 6 mM KCl, 10 mM HEPES, 10 mM glucose, 1 mM MgCl 2 , 2 mM CaCl 2, pH adjusted to 7.4 with NaOH) through the lumen directly connected to the exploring electrode. The following solutions were perfused through the second lumen at a flow rate of 1.5 ml/h: (1) Clsolution for basal measurement, (2) 100 mM amiloride in Clsolution (Sigma-Aldrich, USA), to block ENaC Na + absorption (3) a low-Clsolution (140 mM sodium gluconate, 6 mM potassium gluconate, 10 mM HEPES, 10 mM glucose, 1 mM MgCl 2 , 6 mM calcium gluconate , pH adjusted to 7.4 with NaOH) containing amiloride (100 mM) to drive Clsecretion. Each solution was perfused for at least three minutes. Stability for at least one minute was required before each change in perfusion. The values analyzed were the means of the last 30 seconds.
Three parameters were investigated during transepithelial nasal potential difference (V TE ) measurements: (1) the stable maximal baseline V TE , which was obtained after the equilibration of transepithelial ion transport with Clsolution, and the successive net voltage changes between (2) baseline V TE and the Clsolution containing 100 mM amiloride (DV TE Amil ); (3) Clsolution with amiloride and low-Clsolution with amiloride (DV TE LowCl -).

Inhibitors and activators
CFTR is a cAMP-dependent channel. We therefore used forskolin (Sigma-Aldrich, USA), an adenylate cyclase activator, in B6;129 and FVB WT mice, for the specific activation of CFTR. The V TE -based outcomes were net voltage differences between 100 mM amiloride in low-Clsolution and 100 mM amiloride plus 10 mM forskolin in low-Clsolution (DV TE Forsk. ).
We used various inhibitors to identify the channels participating in the low-Clresponse in B6;129 WT mice: the CFTR-specific inhibitor thiazolidone (Inh-172) (Calbiochem, Germany), 5 mM; niflumic acid, a calcium-dependent Clchannel inhibitor (Sigma-Aldrich, USA), 100 mM and zinc chloride, a voltage-dependent Clchannel inhibitor, (Fluka, USA), 50 mM. These inhibitors were tested in low-Clsolution containing amiloride (100 mM), and the V TE -based outcomes recorded were net voltage changes after perfusion of the given inhibitor in low-Clsolution (DV TE Inh. ).
We assessed the effects of solvents (DMSO, ethanol, NaOH and KHCO 3 ) on voltage. None of the solvents altered Clsecretion at the concentration used to dissolve the various inhibitors (data not shown).

Statistical analysis
The values obtained were not normally distributed. The electrophysiological norms for DV TE parameters are therefore expressed as medians with interquartile ranges (IQR).
We compared the V TE values obtained between groups in Mann-Whitney tests.
We assessed the effects of inhibitors and activators in Wilcoxon paired signed-rank tests comparing V TE data recorded after the perfusion of low-Clsolution and V TE data recorded after the perfusion of low-Clsolution plus inhibitor/activator, except for DIDS effects for which V TEInh-172 and V TEDIDS data were compared.
The difference between two measurements for a single mice group was evaluated with the non parametric Wilcoxon paired signed-rank test. Repeatability was evaluated by the Bland-Altman method [22]. For each mouse, the difference between two measurements was calculated and plotted against the mean of the two measurements. We determined whether the differences were normally distributed in Kolmogorov-Smirnov and Shapiro-Wilk normality tests. The bias was estimated by calculating the mean of all differences between the two measurements. If this mean is not close to zero, the two assays are considered to give different results. The limits of agreement were defined as the bias 6 1.96 SD.
A case-control analysis was carried out to optimize the discrimination between control mice and F508del-CFTR mice. Cutoff points were determined from the receiver operating characteristics (ROC) curve. For each parameter, we ranked the values for WT and F508del-CFTR mice. The percentage of the WT mice effectively included in each rank indicates the sensitivity of the test; the percentage of F508del-CFTR mice effectively not included in each rank indicates the specificity of the test. Cutoff points were defined as the rank associated with the best positive likelihood ratio of sensitivity/(1-specificity), favoring specificity.
Forskolin induced no significant increase in Clsecretion in either B6;129 (n = 10) or FVB (n = 9) WT mice (Table S1). Furthermore response to forskolin perfusion did not discriminate between WT and CF mice (Table S1). We therefore decided not to test forskolin after perfusion with a low-Clsolution.

Nasal potential difference values
Typical values in B6;129 WT and CF mice. Transepithelial nasal potential difference (V TE ) measurements were performed in 50 WT and 50 cftr knockout mice (cftr -/-); representative recordings are shown in Figure 1. Sex had no effect on any of the V TE parameters in either of these groups. The cftr -/mice had higher levels of sodium transport than WT mice, as shown by the higher baseline V TE and much more pronounced response to amiloride perfusion, and an absence of chloride transport, as shown by the lack of response to perfusion with low-Clsolution (Table 1). WT mice displayed strong hyperpolarization during perfusion with the low-Clsolution (-7.8 mV (IQR = 3.8 mV)), which was inhibited by 20% (p,0.0001) with Inh-172 and an additional 26% (p,0.0001) with DIDS. The cftr -/mice did not respond to Inh-172, but displayed additional depolarization, by 3.1 mV (IQR = 3.1 mV, p,0.0001) after DIDS perfusion.
Nasal potential difference in FVB and B6;129 mice. Transepithelial nasal potential difference (V TE ) measurements were performed on 50 F508del-CFTR and 25 WT FVB mice (Table 1). No V TE difference was observed between males and females. Like cftr -/mice, F508del-CFTR mice had higher levels of sodium transport and absent or lower levels of chloride secretion than their WT littermates. WT B6;129 and FVB mice had similar levels of Na + transport, as shown by their similar baseline V TE and DV TE Amil values. By contrast, WT FVB mice had significantly lower levels of Clsecretion, as shown by their DV TE Low Cl values, which were lower than those of B6;129 mice by a factor of about 1.5 (p = 0.0025). The contribution of the CFTR was similar in mice of both backgrounds, because Inh-172 treatment resulted in significant inhibition of 1.6 mV (IQR 3.1, p = 0.0039) in FVB mice and 1.6 mV (IQR 2.3) in B6;129 mice. In WT FVB mice, DIDS treatment resulted in an additional inhibition, by 2.4 mV (IQR 2.8, p = 0.0078), corresponding to 51% inhibition of chloride secretion.
F508del-CFTR mice had lower levels of sodium transport than cftr -/mice, as shown by their significantly lower baseline V TE and DV TE Amil values (p , 0.0001 and p = 0.002). As in cftr -/mice, no chloride secretion was observed.  Intermeasurement repeatability Figure 2 shows the results of two measurements taken one to four weeks apart in 22 WT mice (A) and 21 cftr -/mice (B).
There was no significant difference between the two series of measurements for baseline V TE , DV TE Amil and DV TE Low cl -, in either WT or cftr -/mice, as assessed by Wilcoxon paired signedrank tests.
The repeatability of the test was assessed by the Bland-Altman method for baseline V TE , DV TE Amil and DV TE Low cl -, in both WT ( Figure 2C) and cftr -/mice ( Figure 2D). Differences between the two measurements were normally distributed and did not vary as a function of their arithmetic values. Intra-animal variability was defined by the limits of agreement, both in WT and cftr -/mice, and for all V TE parameters ( Table 2). Bias -i.e. the mean differences -were close to zero in both WT and cftr -/mice ( Table 2).

Discussion
In this study, we established values for nasal potential difference endpoints in mice. We studied baseline V TE , response to amiloride and to low-Clsolution in 129;B6 cftr -/and FVB F508del-CFTR mice, and their respective controls. In both CF models, ENaC activity was much higher than in the corresponding WT (Baseline V TE : , 3.5 times higher and DV TEAmil : , 5 times higher) and Clsecretion was abolished. We assessed the variability of this test on Bland-Altman plots and determined the first cutoff points for distinguishing between WT and CF mice. These cutoff points are important for the evaluation of CFTR transport restoration in preclinical studies evaluating CFTR correctors or potentiators [6,[12][13][14].

Protocol implementation
Protocols for V TE determinations in mice vary considerably, with differences in perfusion flow rate of perfusion (from 0.018 ml/h to 3 ml/h), the position of the catheter in the nostril (2 mm to 5 mm into the nostril) and the nature of the solution used (Ringer's solution, Krebs solution, other phosphate or HEPES solutions) [14][15][16][17][18][19][20]26,27]. The sequences of the solutions used also differ markedly between protocols: i) Some groups initially perfuse with the Clsolution until stabilization is achieved, before recording baseline V TE [19,20], whereas other groups do not carry out this perfusion step [17,18]. ii) The use of a CFTR activator (forskolin, isoproterenol) and CFTR-specific inhibitors during perfusion with the low-Clsolution is not systematic [14,17,18]. We defined a protocol taking into account the following points: i) WT mice displayed depolarization after initial Clsolution perfusion. We therefore decided to perfuse the epithelium with this solution until stabilization was achieved, before recording baseline V TE . ii) As no additional hyperpolarization was observed after the perfusion of forskolin solution, in either of the WT backgrounds, we decided not to use forskolin. Similar results were reported by Brady et al. for mice of the BALB/ cJ and C3H/HeJ backgrounds [15]. Moreover, the response to forskolin (which increase cAMP level), had been shown to be small in WT mice and cannot reliably used to distinguish between WT and CF genotypes because CF mice display a small response similar to that of WT mice [15,16,18,28].
We further characterized Clsecretion with various inhibitors. Inh-172, a specific inhibitor of CFTR, and DIDS, a broadspectrum inhibitor, had significant effects. Neither inhibitors of Ca 2+ -(niflumic acid) nor inhibitors of voltage-dependent Clchannels (zinc chloride) affected Clconductance. This suggests that neither Ca 2+ -dependant Clchannels (CaCC) nor voltagegated channels were active in the murine nasal epithelium in basal conditions.

VTE values in mice of the two backgrounds
The lack of well characterized backgrounds for V TE measurements led us to investigate the FVB and B6;129 backgrounds in detail. Baseline V TE was significantly higher in CF than in WT mice, consistent with the differences observed between CF patients and healthy people [21,29]. The response to amiloride perfusion was also about five times stronger in the CF models, cftr -/and F508del-CFTR mice, than in WT mice. However, this difference was smaller in F508del-CFTR mice than in cftr -/mice, suggesting that the F508del-CFTR protein may have retained some of its ENaC channel-regulating activity. The corresponding WT had similar baseline V TE and DV TEAmil values. Thus, genetic background is not responsible for effects on sodium transport. In terms of Clsecretion, mice of the two backgrounds were similarly sensitive to Inh-172 and DIDS. However, an interesting difference between these two backgrounds was that chloride secretion was almost entirely inhibited by Inh-172 plus DIDS in FVB mice, whereas secretion was inhibited by less than 50% in B6;129 mice (residual V TE = 0.7 mV vs. 4.2 mV, respectively). Thus, in addition to CFTR, there is a DIDS-insensitive Clpathway in mouse nasal epithelia, but the contribution of this pathway differs considerably between genetic backgrounds.
Hyperpolarization by at least -1.9 mV was observed in 10 % of FVB F508del-CFTR mice, but no B6;129 cftr -/mice, in response to perfusion with a low-chloride solution. This demonstrates that CFTR activity is therefore responsible for chloride secretion in wild-type B6;129 mice and also that F508del-CFTR retained a residual Cltransport activity.

Repeatability of V TE values
Repeatability between two series of measurements within the same group was good. We used Bland and Altman plots to define the limits of agreement in B6;129 WT and cftr -/mice, making it possible to distinguish between simple variability and changes due to treatment. These data are essential for interpreting treatment effects, taking intra-animal variability into account.
We did not calculate coefficients of variance (%CV) for the V TE data because the standard deviation did not vary with the mean and some means were close to zero, making interpretation unreliable [30,31]. We did not determine correlation coefficients either, because these coefficients was not appropriate to this kind of analysis [22].

Cutoff points for V TE values
We were able to determine cutoffs for each of the V TE parameters from ROC curves. We favored specificity over sensitivity. These cutoff points make it possible to classify V TE measurements as belonging to a WT or CF profile, with a high degree of discrimination. This tool is crucial for preclinical studies of new drugs for cystic fibrosis treatment, particularly given the difficulties involved in interpreting the effects of treatment due to the potential residual activity of F508del-CFTR. The low-Clcutoff, -2.15 mV, is the most relevant cutoff because it directly reflects correction of the CFTR defect. This is the first attempt, to our knowledge, to determine V TE endpoints for preclinical studies.

Applications for these cutoff points
We recently used the low-Clcutoff to demonstrate the effect of keratin-8 siRNA treatment to restore F508del-CFTR activity [6]. It was hypothesized that keratin-8 interacts with F508del-CFTR and that disruption of this interaction would restore CFTR activity. Clsecretion exceeded the -2.15 mV cutoff for 50% of the treated mice and none of the control mice, establishing proof-ofconcept for the treatment. F508del-CFTR rescue can further be demonstrated by inhibition of the response with a specific inhibitor, such as Inh-172.
In summary, we report here typical V TE values for mice of two backgrounds not previously investigated: B6;129 and FVB. We show that our protocol for V TE measurement is repeatable and we have determined V TE cutoff values for distinguishing between CF and WT responses. This study constitutes an advance in the investigation of F508del-CFTR correctors/potentiators or ENaC hyperabsorption suppressors.

Supporting Information
Table S1 Forskolin response in WT and CF mice. Values are voltage differences between 100 mM amiloride in low-Clsolution perfusion and 100 mM amiloride plus 10 mM forskolin in low-Cl-solution perfusion (DV TE Forsk ). (DOC)