Extracellular phosphate enhances the function of F508del-CFTR rescued by CFTR correctors

Highlights • CFTR correctors rescue the plasma membrane expression of F508del-CFTR.• Extracellular phosphate enhances F508del-CFTR function rescued by CFTR correctors.• Cystic fibrosis airway epithelia express the phosphate transporter SLC34A2.• Extracellular phosphate levels might contribute to variable drug responses.


Introduction
The genetic disease cystic fibrosis (CF) is caused by mutations in the epithelial anion channel cystic fibrosis transmembrane conductance regulator (CFTR) [ 1 , 2 ]. Based on the mechanisms of CFTR dysfunction [ 1 , 2 ], CFTR modulators have been developed to rescue the plasma membrane expression, stability and function of CF therefore crucial to identify and understand the molecular players that influence the response to CFTR modulators.
Genome Wide Association Studies (GWAS) and in vitro functional investigations have identified several modifier genes that affect CF disease severity and/or the response to CFTR modulators [10][11][12][13][14] . Interestingly, three of the identified modifier genes, SLC6A14, SLC9A3 and SLC26A9, are members of the Solute Carrier (SLC) transporter superfamily ( https://www.bioparadigms.org/ ). Although understanding of why polymorphisms in SLC transporter genes have modulatory effects is incomplete, recent work has provided important insights for SLC6A14, an electrogenic, neutral and cationic amino acid transporter found at the apical membrane of airway and intestinal epithelial cells. L-Arginine transport by SLC6A14 modulated bacterial attachment to airway epithelia, CFTR activity and the response of F508del-CFTR to CFTR modulators [15][16][17][18] . These data suggest that targeting the L-arginine pathway might be beneficial to CF patients treated with CFTR modulators.
We speculated that other SLC transporters might influence CFTR activity and/or the response of CF mutants to CFTR modulators. We were intrigued to learn that mutations in the SLC34A2 gene, which encodes an apically-located Na + -dependent phosphate transporter, were linked to pulmonary alveolar microlithiasis [19] , as a consequence of SLC34A2 transporter dysfunction [ 20 , 21 ]. This work identified an important role for SLC34A2 to remove phosphate from airway surface liquid (ASL). Since cytosolic phosphate levels influence CFTR function [22] , we hypothesised that phosphate might be an unrecognised parameter governing the response to CFTR modulators. To investigate this possibility, we tested the effects of extracellular phosphate on CFTRmediated Cl − currents in Fischer rat thyroid (FRT) epithelia heterologously expressing F508del-CFTR and fully differentiated human airway epithelial cells (genotype: F508del/F508del), following rescue of F508del-CFTR with CFTR correctors (lumacaftor [ Our results demonstrated that rescued F508del-CFTR activity was markedly dependent on extracellular phosphate, but this was not the case for either low temperature-rescued F508del-CFTR or wildtype CFTR. These data demonstrate that luminal phosphate has a hitherto unknown stimulatory effect on CFTR corrector-rescued F508del-CFTR and suggest that phosphate has the potential to impact the outcome of CFTR corrector therapies.

Cells and cell culture
Primary non-CF and CF (genotype: F580del/F508del) human Airway Epithelial Cells (hAECs) and Fischer Rat Thyroid (FRT) epithelial cells heterologously expressing wild-type (WT) and F508del-CFTR were supplied, cultured and used as described in the Supplementary material .

RNA extraction, PCR and real-time quantitative PCR analysis
Total RNA extraction and real-time quantitative PCR (qPCR) were performed as described in the Supplementary material.
In other experiments, F508del-CFTR was trafficked to the plasma membrane by growing cells at 27 °C for 48 h.

Statistical analysis
Results are expressed as means ± SD of n observations with statistical analyses performed as described in the Supplementary material.

The phosphate transporter SLC34A2 is expressed in primary cultures of human airway epithelial cells
The Human Protein Atlas ( https://www.proteinatlas.org/ ENSG0 0 0 0 0157765-SLC34A2/tissue ) indicates that SLC34A2 (Na +dependent phosphate cotransporter 2B; NaPi-2B) mRNA and protein are found in the lung, including prominent expression in alveolar type II cells. To learn whether SLC34A2 is expressed in respiratory airway epithelial cells, we studied primary cultures of CF (genotype: F508del/F508del) and non-CF airway epithelial cells. Fig. 1 A demonstrates that SLC34A2 was expressed in CF and non-CF human airway epithelial cells, while Fig. 1 B reveals that the relative levels of SLC34A2 mRNA did not differ between CF and non-CF human airway epithelial cells. Fig. 1 C and D demonstrate that FRT cells express rat slc34a2 and its expression did not differ between FRT cells heterologously expressing wild-type and F508del-CFTR. These data suggest that FRT epithelia heterologously expressing human CFTR might be used to examine the effects of SLC34A2-mediated phosphate transport on CFTR function.

Phosphate enhances the function of F508del-CFTR rescued by CFTR correctors
To investigate whether extracellular phosphate modulates the activity of F508del-CFTR after its rescue by CFTR correctors, FRT epithelia heterologously expressing F508del-CFTR were treated with either lumacaftor (VX-809; 3 μM) or vehicle (DMSO 0.1% v ·v -1 ) for 48 h at 37 °C before mounting in Ussing chambers. To maximise CFTR-dependent currents, FRT epithelia were bathed in a basolateral to apical Cl − gradient ([Cl − ] basolateral , 119.8 mM; [Cl − ] apical , 6 mM) in the absence or presence of phosphate (1.24 mM K 2 HPO 4 and 2.4 mM KH 2 PO 4 ) ( Fig. 2 A and B). Under these conditions, there was no significant difference in resting shortcircuit current (I sc ) or transepithelial electrical resistance (R t ) between the four different conditions (P > 0.05) ( Fig. 2 C and D). As expected, when compared to the DMSO controls, F508del-CFTRexpressing FRT epithelia that were pre-treated with VX-809, displayed markedly larger responses to the cAMP agonist, forskolin (Fsk, 10 μM) and the CFTR potentiator P5 (5 μM), which were fully blocked by the CFTR inhibitor, CFTR inh -172 (I172, 20 μM) ( Fig. 2 A, B and E), confirming they were the result of F508del-CFTR that had trafficked to the apical membrane in response to VX-809 [9] . Strikingly, the magnitude of CFTR-mediated Cl − current generated by F508del-CFTR-expressing FRT epithelia pre-treated with VX-809 was markedly reduced when phosphate was absent from the recording solutions ( Fig. 2 A, B and E). Indeed, CFTR inh -172sensitive currents were reduced ~4-fold by the absence of phosphate.
To test whether this dependency on extracellular phosphate was also observed in epithelial cells that endogenously express F508del-CFTR, experiments were repeated using fully differentiated   normalised by subtraction of the steady-state current after amiloride addition prior to F508del-CFTR activation by forskolin. ( C -E ) Summary data show the magnitude of baseline I sc and R t before amiloride addition and the change in I sc ( I sc ) for the indicated conditions. Symbols represent individual values and lines are means ± SD (VX-809: + phosphate, n = 9; -phosphate, n = 11; DMSO: + phosphate, n = 10; -phosphate, n = 11); * , P < 0.05; * * , P < 0.01; † † † , P < 0.001 (Repeated Measure two-way ANOVA with Sidak's multiple comparison test). epithelia of CF hAECs homozygous for F508del-CFTR ( Fig. 3 ). In contrast to F508del-CFTR-expressing FRT epithelia ( Fig. 2 C), baseline I sc was higher in the presence of phosphate ( Fig. 3 C), but R t was not affected ( Fig. 3 D). Importantly, similar to F508del-CFTRexpressing FRT epithelia, extracellular phosphate markedly improved the response of VX-809-rescued F508del-CFTR to forskolin and P5. Indeed, the resulting CFTR inh -172-sensitive currents were ~2.5-fold greater than those measured in CF hAEC epithelia pretreated with VX-809 studied in the absence of phosphate ( Fig. 3 A, B and E), confirming the results obtained with F508del-CFTRexpressing FRT epithelia ( Fig. 2 ).
To learn whether the response to phosphate is specific to VX-809, experiments were repeated using both cell models with the CFTR corrector, C18 an analogue of VX-809 which produces comparable levels of F508del-CFTR correction to VX-809 [23] . Overall, we observed a similar effect of extracellular phosphate on the response to forskolin and P5 in the C18-pre-treated F508del-CFTRexpressing FRT epithelia ( Supplementary Fig. 1) and the C18-pretreated CF hAEC epithelia ( Supplementary Fig. 2), to that found with VX-809-pre-treated epithelia. Thus, extracellular phosphate enhances F508del-CFTR function rescued by the CFTR correctors VX-809 and C18.

Phosphate fails to enhance the function of F508del-CFTR after correction by low temperature incubation
We also investigated whether the response to phosphate was evident using low temperature to rescue F508del-CFTR. For these experiments, F508del-CFTR-expressing FRT epithelia were incu-bated at 27 °C for 48 h to traffic F508del-CFTR to the apical membrane [24] before epithelia were mounted in Ussing chambers to measure I sc at 37 °C. Supplementary Fig. 3A shows representative I sc traces of rescued F508del-CFTR-expressing FRT epithelia in the absence and presence of phosphate. The presence of phosphate did not affect baseline I sc nor R t (Supplementary Fig. 3B and C), but, in marked contrast to VX-809-and C18-treated F508del-CFTRexpressing FRT epithelia ( Fig. 2 and Supplementary Fig. 1), there was no effect of phosphate on the magnitude of the forskolinstimulated and P5-potentiated I sc , nor the CFTR inh -172-sensitive I sc (Supplementary Fig. 3D). As a result, when compared to VX-809-and C18-corrected F508del-CFTR, low temperature-rescued F508del-CFTR showed a much larger forskolin-stimulated and P5potentiated I sc in the absence of phosphate (~4-fold greater) ( Fig. 2 E and Supplementary Figs. 1E and 3D). This result suggests that low temperature rescue eliminated the ability of extracellular phosphate to modulate F508del-CFTR activity.
To learn whether the effect of extracellular phosphate was restricted to misfolded/mutant CFTR, we studied wild-type CFTR. Supplementary Fig. 4 demonstrates that phosphate was without effect on wild-type CFTR heterologously expressed in FRT epithelia. Extracellular phosphate had no effect on baseline I sc , R t, the forskolin-stimulated P5-potentiated I sc nor the CFTR inh -172sensitive I sc (Supplemental Fig. 4). Similarly, Supplementary Fig. 5 reveals that phosphate was without effect on native CFTR in non-CF hAEC epithelia. Taken together, these data suggest that the ability of extracellular phosphate to enhance F508del-CFTR function only occurs in cells where mutant CFTR is rescued with CFTR correctors. forskolin. ( C -E ) Summary data show the magnitude of baseline I sc and R t before amiloride addition and the change in I sc ( I sc ) for the indicated conditions. Symbols represent individual values and lines are means ± SD (elexacaftor-tezacaftor-ivacaftor (ETI): + phosphate, n = 9; -phosphate, n = 9; DMSO: + phosphate, n = 9; -phosphate, n = 9); * P < 0.05; * * * , P < 0.001; † , P < 0.05 vs. -phosphate (Two-way ANOVA with Sidak's multiple comparison test).

Phosphate enhances the function of F508del-CFTR rescued by elexacaf tor-tezacaf tor-ivacaf tor
In 2019, elexacaftor-tezacaftor-ivacaftor (Trikafta), the combination of two CFTR correctors, VX-445 (elexacaftor) and VX-661 (tezacaftor) with the CFTR potentiator VX-770 (ivacaftor), was approved for clinical use in people with CF carrying the F508del mutation on at least one allele ( https://www.fda.gov/news-events/ press-announcements/fda-approves-new-breakthrough-therapycystic-fibrosis ), following very positive clinical trial results [ 25 , 26 ]. In vitro studies demonstrate that VX-445 synergistically rescued F508del-CFTR processing when used in combination with VX-661, and that the triple combination of VX-445, VX-661 and VX-770 restored ~62% of wild-type CFTR function to nasal epithelia expressing native F508del-CFTR [27] . To learn whether the effect of extracellular phosphate was also observed when F508del-CFTR was rescued by elexacaftor-tezacaftor-ivacaftor (ETI), fully differentiated epithelia of CF hAECs homozygous for F508del-CFTR were incubated at 37 °C for 24 h with the triple drug combination [27] , before epithelia were mounted in Ussing chambers ( Fig. 4 ). Pretreatment with elexacaftor-tezacaftor-ivacaftor caused a striking increase in baseline I sc , which was not dependent on phosphate ( Fig. 4 C), but R t was unaffected ( Fig. 4 D). Importantly, the response to forskolin was noticeably improved in the presence of extracellular phosphate in elexacaftor-tezacaftor-ivacaftor rescued F508del-CFTR epithelia, which was mirrored in the magnitude of the CFTR inh -172-sensitive current ( Fig. 4 E), consistent with the results obtained with VX-809 ( Fig. 3 ). However, unlike epithelia pre-treated with VX-809, the results in Fig. 4 , demonstrate that elexacaftor-tezacaftor-ivacaftor caused a substantial increase in corrected F508del-CFTR function in the absence of forskolin stimulation, consistent with previous results [27] .

Sodium-dependence of the enhancement of elexacaf tor-tezacaf tor-ivacaf tor -rescued F508del-CFTR function by phosphate
Our results demonstrate that extracellular phosphate noticeably improved the function of corrector-rescued F508del-CFTR. To investigate whether this effect of phosphate involves a sodiumdependent phosphate transporter, we repeated the elexacaftortezacaftor-ivacaftor experiments in the absence of sodium in the solution bathing the apical membrane ( Fig. 5 ). As expected, the removal of sodium from the apical bathing solution greatly reduced baseline I sc (compare Fig. 4 C with Fig. 5 B). Importantly, in the absence of extracellular sodium, but in the presence of phosphate, forskolin no longer augmented elexacaftor-tezacaftorivacaftor-rescued F508del-CFTR function ( Fig. 5 D). This result demonstrates that the modulatory effect of phosphate is sodium dependent.

Acute addition of phosphate is sufficient to enhance the function of F508del-CFTR after correction by elexacaftor-tezacaftor-ivacaftor
For the studies described thus far, the effect of extracellular phosphate was tested by removing all phosphate from the recording solutions and comparing responses to epithelia with phosphate present throughout. To investigate the kinetics of the response to phosphate, elexacaftor-tezacaftor-ivacaftor-treated CF hAEC epithelia were mounted in Ussing chambers and bathed in phosphatefree solutions. After exposure to amiloride, phosphate was added directly to the phosphate-free solution bathing the apical membrane. The total magnitude of the acute F508del-CFTR-mediated I sc was determined (~10 min after phosphate addition) and compared to results obtained when phosphate was present for the entire experiment (~50 min total time), with each condition studied simultaneously ( Supplementary Fig. 6). The acute addition of phosphate caused a rapid increase in I sc which stabilised within 1-2 min (Supplementary Fig. 6A). Importantly, the total magnitude of F508del-CFTR-mediated I sc was identical in epithelia that were acutely exposed to phosphate, compared to those epithelia that had phosphate present throughout the experiment (Supplementary Fig. 6D). These results demonstrate that a relatively short exposure to phosphate is sufficient to improve the activity of F508del-CFTR rescued by elexacaftor-tezacaftor-ivacaftor.

Discussion
This study demonstrates that the activity of CFTR correctorrescued F508del-CFTR was dependent on extracellular phosphate. This novel effect of phosphate was observed in FRT epithelia het-erologously expressing F508del-CFTR and fully differentiated human airway epithelial cells (genotype: F508del/F508del), suggesting that it was independent of cell context. However, extracellular phosphate was without effect on low temperature-rescued F508del-CFTR and wild-type CFTR, suggesting that the response to phosphate might be specific for mutant CFTR whose trafficking (folding) defects were rescued by the CFTR correctors VX-809, C18 and elexacaftor-tezacaftor-ivacaftor [ 9 , 23 , 25 , 26 ].
Under the experimental conditions used, the effect of extracellular phosphate on CFTR corrector-rescued F508del-CFTR might result from three possible mechanisms: (i) increasing the number of channels in the apical membrane by either stimulating trafficking/insertion of F508del-CFTR into the membrane or decreasing its retrieval to enhance channel stability; (ii) acting as a potentiator, thereby increasing the open probability (P o ) of F508del-CFTR and (iii) augmenting anion flow through individual F508del-CFTR Cl − channels. However, these three mechanisms are not mutually exclusive and more than one might be involved.
For two reasons, the response to phosphate is unlikely to be explained by the recruitment of F508del-CFTR Cl − channels to the apical membrane or enhancing their plasma membrane stability. First, acute addition of phosphate to the solution bathing the apical membrane increased CFTR-mediated Cl − currents by a comparable amount as sustained exposure (Supplementary Fig. 6). Second, phosphate was present in the culture medium and was only absent from control experiments for the duration of Ussing chamber recordings. The fact that the response to acute addition of phosphate was rapid suggests an effect on channel gating and, hence P o . CFTR activity is tightly regulated to control the hydration and pH of epithelial secretions [28] . The channel is primarily activated by protein kinase A (PKA)-dependent phosphorylation of its regulatory domain (RD) [29] . Then, cycles of ATP binding and hydrolysis at two ATP-binding sites located at the interface of the nucleotidebinding domain (NBD) dimer controls channel gating and therefore, P o [29] . Among the factors that modulate channel gating are the products of ATP hydrolysis, ADP and phosphate. ADP competitively inhibits CFTR gating with its major effect at ATP-binding site 2 [ 29 , 30 ]. By contrast, in the presence of ATP, raising the cytosolic phosphate concentration increased the P o of wild-type CFTR by accelerating channel opening [22] . Phosphate did not change anion flow through individual CFTR Cl − channels and was without effect on the number of active channels [22] . Because phosphate stimulated a CFTR construct lacking most of the RD ( R-S660A-CFTR) and was without effect on unphosphorylated wild-type CFTR [22] , its enhancement of channel gating likely reflects an effect at ATP-binding site 2. Thus, extracellular phosphate likely enhances the function of CFTR corrector-rescued F508del-CFTR by modifying channel gating.
A plausible explanation for why extracellular phosphate was without effect on wild-type CFTR is the gating defect of F508del-CFTR [1] . Under the experimental conditions used, insufficient phosphate was likely transported into cells to noticeably increase the already high P o of wild-type CFTR and hence, the magnitude of CFTR-mediated Cl − current. Because the single-channel behaviour of F508del-CFTR is equivalent after rescue by either VX-809 or low temperature [31] , a similar explanation does not account for the different effects of extracellular phosphate after F508del-CFTR correction by these treatments. Instead, the lack of effect of extracellular phosphate on low temperature-rescued F508del-CFTR might be explained by the more extensive correction of the mutant protein by low temperature than individual CFTR correctors [24] , leading to greater plasma membrane expression with low temperature. Although this idea is supported by greater F508del-CFTR function in FRT epithelia heterologously expressing F508del-CFTR incubated at low temperature than rescued with either VX-809 or C18, it is not supported by the effect of phosphate on the function of native F508del-CFTR rescued by elexacaftor-tezacaftor-ivacaftor. Like the action of low temperature [24] , elexacaftor and tezacaftor robustly reverse F508del-CFTR misfolding [27] . Future studies should therefore directly compare the effect of extracellular phosphate on F508del-CFTR rescued by either low temperature or elexacaftortezacaftor-ivacaftor.
For extracellular phosphate to modulate CFTR channel gating at its NBDs, transmembrane phosphate transport is required. Previous work demonstrates that the sodium-dependent phosphate transporter SLC34A2 is expressed in the respiratory airways [19] . Building on these data, we showed that SLC34A2 expression is equivalent in CF and non-CF human airway epithelial cells, and that the modulatory effect of phosphate required extracellular sodium. However, the present results do not exclude the participation of other sodium-dependent phosphate transporters. Although expression of the closely related phosphate transporters SLC34A1 and SLC3A3 is largely restricted to the renal proximal tubule, SLC20 transporters are ubiquitously expressed at the mRNA level and SLC17A2 is found in the lung [ 32 , 33 ]. To understand better the role of SLC34A2 and other phosphate transporters in airway epithelia will require the development of specific pharmacological tools and/or the manipulation of transporter gene expression.
The requirement for extracellular phosphate, likely mediated by SLC34A2, observed in the present study is comparable to that recently described for SLC6A14 [16] . Uptake of luminal L-Arginine by SLC6A14 led to the stimulation of CFTR function, rather than alteration of CFTR expression in the plasma membrane [16] . However, and in contrast to our results, L-Arginine also stimulated 'resting' CFTR function (after correction), in the absence of PKA and CFTR potentiators [16] . Consistent with previous results [22] , in the present study extracellular phosphate was without effect on the 'resting' activity of F508del-CFTR after its rescue by CFTR correctors ( Figs. 3 C and 4 C).

Conclusion
The phosphate transporter SLC34A2 is expressed in human airway epithelial cells and its relative abundance is similar in CF and non-CF cells. Luminal phosphate stimulates the activity of F508del-CFTR rescued by the CFTR correctors VX-809, C18 or VX-445 + VX-661 with VX-770 and its action was sodium-dependent. These and other data [20][21][22] suggest that in vivo , SLC34A2 function regulates both the amount of phosphate in ASL and the local cytosolic concentration of phosphate, thereby enhancing the function of CFTR corrector-rescued F508del-CFTR by altering channel gating. Since mutations in SLC34A2 which reduce phosphate transport are linked to lung disease [19][20][21] , we speculate that any SNPs which alter SLC34A2 activity might indirectly affect F508del-CFTR function and potentially impact the outcome of CFTR corrector therapies.