DISCUSSION AND CONCLUSIONS
The aim of this study was to characterize the effects of corrector RDR01752 both in cell lines that stably express F508del-CFTR or other rare CFTR mutations and in intestinal organoids that are F508del/F508del. Furthermore, we investigated the MoA of RDR01752 by evaluating its additivity with available corrector drugs, genetic revertants of F508del-CFTR, and low temperature.
Most CF drug development programs for CFTR modulators have focused on the rescue of F508del-CFTR as the most prevalent disease-causing mutant. Despite significant progress in restoring F508del-CFTR trafficking by corrector drugs, treatment of F508del-homozygous individuals with single correctors (VX-809 or VX-661) in combination with the potentiator VX-770 achieved only modest clinical improvements (Wainwright et al.,2015; Taylor-Cousar et al. 2017). Significantly greater therapeutic response was recently achieved by adding a second corrector (VX-445) to the previous combination (VX-661/VX-770) (Heijerman et al., 2019; Middleton et al., 2019), indicating that combination of correctors acting by distinct mechanisms is needed for efficient correction of F508del-CFTR and for clinical benefit of individuals with CF carrying this mutation. Although the MoA of VX-445 is still unknown, its additivity to VX-661 on F508del-CFTR rescue suggests that these compounds act by different modes.
Here, we looked into the MoA of the recently described F508del-CFTR corrector RDR01752 (Carlile et al., 2007). Our biochemical and immunofluorescence data demonstrated that RDR01752 rescues F508del-CFTR processing and PM expression in CFBE cells to levels similar to those of VX-809. However, in previous studies using baby hamster kidney (BHK) cells stably expressing F508del-CFTR, RDR01752 appeared to be less efficacious than VX-809 (Carlile et al., 2007, Sampson et al., 2011). Our functional data here, in both CFBE and FRT cell lines, also show that RDR01752 is less effective than VX-809 in restoring F508del-CFTR-mediated Cl- secretion, consistent with previous findings in human bronchial epithelial (HBE) cells (F508del/F508del) (Carlile et al., 2018). Notwithstanding, as the validation of results in patient-derived specimens is an important step to provide a better prediction of the in vivo efficacy (Amaral and De Boeck, 2019; Lopes-Pacheco, 2020), we tested here this compound in intestinal organoids (F508del/F508del). Our data revealed that RDR01752 and VX-809/VX-661 can rescue F508del-CFTR function to similar levels in this assay. The higher efficacy observed in the organoids may derive from the fact that different cell systems were used and CFTR processing and function are influenced by the cell background and polarization state (Pedemonte et al., 2010; Rowe et al.,2010; Farinha et al., 2015).
To investigate the MoA of RDR01752 we analyzed its additivity to revertants that rescue F508del-CFTR by different mechanisms. RDR01752 effects were additive to G550E and 4RK, but not to R1070W as determined by rescue of F508del-CFTR protein to the PM. G550E and R1070W were proposed to act at two distinct CFTR interdomain contact points that are disrupted by F508del (Farinha et al., 2013): while G550E likely restores the NBD1:NBD2 dimerization interface (Roxo-Rosa et al.,2006), R1070W restores the NBD1:ICL4 interaction (Serohijos et al., 2008; Thibodeau et al., 2010). Thus, RDR01752, like VX-809, might act similarly to R1070W, i.e., by restoring the anchoring of ICL4 to the NBD1 surface by filling a pocket generated by the absence of F508, as the lack of additivity of these two compounds indicates. Indeed, the replacement of an arginine with a tryptophan at position 1070 (R1070W) helps restore interactions among the aromatic residue that were impaired by the lack of F508del at the NBD1 surface (Serohijoset al., 2008; Thibodeau et al., 2010; Farinha et al., 2013). RDR01752, like VX-809/VX-661, has a fused aromatic ring in its structure that can fit into this pocket left empty by F508del (Heet al., 2013; Farinha et al., 2013). This terminal aromatic ring in RDR01752 was also found to be critical for the stabilization of isolated F508del-NBD1 with an additional stabilization effect in presence of ATP, thus suggesting that RDR01752 does not bind to the ATP binding site in NBD1 (Sampson et al., 2011). This is also in agreement with the observed additivity of RDR01752 to G550E.
On the other hand, RDR01752 was also additive to 4RK. This variant (where four arginines in the AFT are simultaneously replaced to lysines) enables some F508del-CFTR protein to traffic to the PM by escaping the ERQC (Chang et al., 1999; Roxo-Rosa et al., 2006). Notably, although RDR01752, VX-809 and VX-661 rescued F508del-4RK-CFTR PM expression, their efficacy was distinct, with the additivity of RDR01752 to 4RK being higher than that of VX-809/VX-661. In turn, the latter correctors were more effective than RDR01752 in rescuing G550E-F508del-CFTR. Altogether, these data suggest that combinations of RDR01752 with compounds that mimic the correction induced by the G550E and 4RK revertants could maximize the rescue of F508del-CFTR.
Regarding the additivity to low temperature, RDR01752, VX-809 and VX-661 were similarly additive to 27ºC incubation of F508del-expressing cells, consistent with previous reports demonstrating that VX-809 was unable to restore the thermostability of F508del-CFTR (He et al., 2013). The three correctors, which were additive to G550E and 4RK, had an even greater effect when combined to low temperature. Although none of the correctors was additive to R1070W at 37°C, additivity was observed for all three in combination to low temperature for this revertant, which was however, less pronounced for RDR01752. Altogether, these data indicate that despite the double correction effect of revertants and compounds on F508del-CFTR, there is still scope for further enhancement as indicated by low temperature data.
As with VX-809 and VX-661, RDR01752 was unable to overcome a Sec24-COPII-ER export defect of the diacidic variant DD/AA (on WT-CFTR backbone). Low temperature nevertheless enables DD/AA-CFTR to exit the ER through the conventional ER-to-Golgi pathway, since the rescued DD/AA-CFTR was found to be fully-glycosylated (Farinha et al.,2013). These data are compatible with the proposed mechanism for F508del-CFTR by low temperature (Farinha et al., 2013).
Although RDR01752 did not further enhance F508del-CFTR PM expression rescue by VX-809 or VX-661 in the immunofluorescence assay, it further increased the rescue of F508del-CFTR function when combined with VX-809versus each compound alone, which may be attributable to a weak potentiator activity of RDR01752 on F508del-CFTR channels (Sampsonet al., 2011). Notably, chronic VX-770 exposure reduced RDR01752-rescued CFTR PM expression in F508del-expressing cells, as described for VX-809 and VX-661 (Cholon et al., 2014; Veitet al., 2014). However, chronic exposure of F508del-expressing cells to a low free concentration of VX-770 prevents the negative effect on VX-809-rescued CFTR (Matthes et al., 2016). Surprisingly, rescuing of CFTR PM expression by C18, a compound with a similar chemical structure to VX-809/VX-661 (Eckford et al., 2014; Lopes-Pacheco et al., 2016), was less affected by chronic VX-770 exposure. In fact, C18 and VX-809 possibly have a different MoA since in primary HBE cells, the former failed to rescue the trafficking mutant A561E-CFTR, in contrast to VX-809 (Awatade et al., 2014).
Because there are several CFTR mutations with the same trafficking defect as F508del-CFTR, i.e., class II (De Boeck & Amaral, 2016, Lopes-Pacheco, 2020), but do not equally respond to the same CFTR corrector (Grove et al., 2011; Rapino et al., 2015, Dekkers et al., 2016; Lopes-Pacheco et al., 2017), we investigated whether RDR01752 could rescue some of these mutants. Functional assessment was initially performed by the FMP assay and then confirmed in Ussing chamber measurements of CFTR activity in polarized FRT cells stably expressing some of those CFTR mutations. As RDR01752 was demonstrated to be a weak potentiator of F508del-CFTR channels (Sampson et al., 2011), we tested its effect on R334W-CFTR, which has minimal trafficking impairment but reduced channel conductance that still allows for residual function (Sheppard et al., 1993). A slight but non-statistically significant increase in R334W-CFTR function was induced by RDR01752 whereas a greater effect was observed when combined with VX-809. We also tested RDR01752 on G85E and N1303K class II CFTR mutants which are temperature-insensitive despite having trafficking defects like F508del (Rapino et al., 2015; Lopes-Pacheco et al., 2017). N1303K was not rescued efficiently by RDR01752 or VX-809 when tested individually or in combination, consistent with previous findings in HBE cells (Awatade et al.,2014) and intestinal organoids (Dekkers et al., 2016). Regarding G85E, although it appeared to respond functionally to RDR01752 in the FMP assay, this was not confirmed in the Ussing chamber measurements. The lack of response of this mutant to several correctors has been previously reported (Grove et al., 2011; Lopes-Pacheco et al. 2017). Altogether, these data demonstrate that both G85E and N1303K trafficking defects are difficult to rescue and alternative correctors remain an unmet need for these mutants.
In conclusion, these data show that RDR01752, like VX-809/VX-661, rescues F508del-CFTR, albeit at lower efficiency, and like those two approved corrector drugs it does not rescue the G85E and N1303K traffic mutants. Our studies with revertants, aimed at understanding the MoA of this novel corrector, help explain such similarity in pharmaco-therapeutic behavior. Indeed, the data suggest RDR01752 may share a binding site on F508del-CFTR with VX-809 and VX-661, i.e., at the NBD1:ICL4 interface. The fact that RDR01752, like VX-809/VX-661, is additive to correction by the revertants G550E and 4RK and also by low temperature indicate that there is still scope for correctors to further increase the rescue of F508del-CFTR.