Gain- and Loss-of-Function CFTR Alleles Are Associated with COVID-19 Clinical Outcomes

Abstract

Carriers of single pathogenic variants of the CFTR (cystic fibrosis transmembrane conductance regulator) gene have a higher risk of severe COVID-19 and 14-day death. The machine learning post-Mendelian model pinpointed CFTR as a bidirectional modulator of COVID-19 outcomes. Here, we demonstrate that the rare complex allele [G576V;R668C] is associated with a milder disease via a gain-of-function mechanism. Conversely, CFTR ultra-rare alleles with reduced function are associated with disease severity either alone (dominant disorder) or with another hypomorphic allele in the second chromosome (recessive disorder) with a global residual CFTR activity between 50 to 91%. Furthermore, we characterized novel CFTR complex alleles, including [A238V;F508del], [R74W;D1270N;V201M], [I1027T;F508del], [I506V;D1168G], and simple alleles, including R347C, F1052V, Y625N, I328V, K68E, A309D, A252T, G542*, V562I, R1066H, I506V, I807M, which lead to a reduced CFTR function and thus, to more severe COVID-19. In conclusion, CFTR genetic analysis is an important tool in identifying patients at risk of severe COVID-19.

Keywords: CFTR complex alleles; COVID-19; post-Mendelian model.

Figure 1

CFTR alleles associated with mild or severe COVID-19. LASSO logistic regression analysis on the boolean representation of ultra-rare variants (the presence of at least two variants, al2), (A) and rare variants (the presence of at least two variants, al2), (B) of all autosomal genes (see Fallerini et al. 2022 for complete representations). The upward histograms (positive weights) represent the features associated with severe COVID-19, whereas the downward histograms (negative weights) represent the features associated with mild COVID-19. (A) The presence of at least two, in cis or trans, CFTR ultra-rare loss-of-function variants was picked up as one of the most important features associated with COVID-19 severity. The loss-of-function variants were defined on the basis of a residual global CFTR activity reduction between 50–79%. (B). The gain-of-function CFTR complex allele [G576V;R668C] was picked up as one of the most important features associated with COVID-19 mildness.

Figure 2

The complex allele [G576V;R668C] increases the activation and maturation of the CFTR channel. The functional and biochemical analysis of the [G576A;R668C] CFTR complex allele on heterologous expression systems. (A). Increase in CFTR activation. The bar graphs show the activity of [G576A;R668C] CFTR and, for comparison, WT-CFTR transiently expressed in CFBE41o- (upper panel) or FRT (lower panel) cells stably expressing HS-YFP. The CFTR activity was determined as a function of the YFP quenching rate following the iodide influx in cells stimulated with FSK alone (20 µM; light gray) or with FSK plus ivacaftor (1 µM; dark gray). The data are means ± SD (n = 3). The asterisks indicate statistical significance vs. the WT-CFTR protein: **, p < 0.01. (B). Increase in CFTR maturation. Representative Western blot images showing the electrophoretic mobility of [G576A;R668C] and, for comparison, wild-type, G576A, R668C, and F508del CFTR transiently expressed in CFBE41o- (upper panel) or FRT (lower panel) cells. The whole lysates derived from cells not expressing CFTR (null cells) are shown as controls for antibody specificity. In the case of F508del, the cells were treated for 24 h with DMSO alone (vehicle) or Elexa/Teza (3 µM/10 µM) to correct the mutant misfolding. The arrows indicate the complex-glycosylated (band C) and core-glycosylated (band B) forms of the CFTR protein. (C). CFTR band C densitometry of the Western blot experiments. The data are means ± SD (n = 3). The asterisks indicate statistical significance vs. WT-CFTR protein: **, p < 0.01. (D). CFTR half-life evaluation. Representative Western blot image showing the CFTR expression pattern of [G576A;R668C] CFTR and, for comparison, WT-CFTR transiently expressed in CFBE41o- cells at different time points (T = 0 and 24 h) following the CHX-induced block of protein synthesis. The whole lysates derived from cells not expressing CFTR (null cells) are shown as controls for antibody specificity. (E). The quantification of the normalized CFTR band C (left graph) following CHX treatment (T = 0 h, light gray; T = 24 h, dark gray) and band C fold-increase (right graph) over 24 h CHX treatment, obtained from experiments detailed in D, normalized with the initial value of band C. The data are means ± SD (n = 3). The asterisks indicate statistical significance vs. WT-CFTR protein: **, p < 0.01.

Figure 3

Functional analysis of CFTR ultra-rare variants on heterologous expression systems identified loss-of-function alleles with reduced activity. (A) Functional analysis of trans CFTR alleles. (B) Functional analysis of cis CFTR alleles. The bar graphs show the activity of the different CFTR variants under investigation and, for comparison, WT-CFTR transiently expressed in CFBE41o- cells stably expressing HS-YFP. CFTR activity was determined as a function of the YFP quenching rate following the iodide influx in the cells stimulated with FSK alone (20 µM; light gray) or with FSK plus ivacaftor (1 µM; dark gray). The data are means ± SD (n = 3). The symbols indicate the statistical significance: # p < 0.05 vs. WT-CFTR protein; ## p < 0.01 vs. WT-CFTR protein; * p < 0.05, ** p < 0.01 vs. DMSO-treated, FSK-stimulated variant protein; §, p < 0.05, §§, p < 0.01 vs. FSK-stimulated variant protein (upon the same chronic treatment). ^ this type of variant is expected to result in little or no CFTR protein (see CFTR2 database at https://cftr2.org, accessed on 3 November 2022).