Discussion
CFTR mutations are one of the factors that determine the severity of CF.20 However, there are many studies performed with patients bearing the same mutation but with different clinical severity.13,14,15,18,19 Research regarding clinical heterogeneity in twins and siblings with same CFTR mutation suggests that modifier genes could play a role in disease severity. Studies which use transcriptomic approaches determined that clinical severity is associated with many molecular mechanisms such as CFTR protein folding, function, and trafficking, inflammation, immune response, epithelial cell differentiation, ion transport and oxidative stress.
In a study performed with nasal epithelial cells of F508del homozygotes patients and healthy controls, the expression of genes involved in cell proliferation (MMP1, ADM, AREG, GJA1, RUNX2) was increased and the expression of genes involved in cilia formation (DNAH9, DNAH12, DNAI1, DNAI2 and DNAAF1) was decreased in patients.13 In another transcriptomic study, nasal and bronchial samples obtained from same patients have been used, and as a result the primary affected molecular pathways in nasal and bronchial epithelial cells were found to be amino acid metabolism and inflammatory response, respectively.19 In another recent study conducted by Kormann et al. it was found that genes involved in type-1 interferon response and ribosomal stalk proteins that interact with rRNA have an effect on disease severity.18 Similar to this study, Wright et al. have shown that genes involved in lipid metabolism, protein ubiquitination mechanism, protein metabolism and immune response are related to disease severity.15
So far, there is no transcriptomic study that investigates phenotypic variability between mild and severe siblings with the same mutation. In our study we aimed to explain clinical heterogeneity by using targeted transcriptome analysis of nasal epithelial cells to investigate differentially expressed genes and associated pathways of between mild and severe siblings. For this purpose, we determined three families that include siblings with different clinical severity determined by recurrent lung infection, pulmonary involvement and comorbidity. The mutations of Family 1 (F508del/G85E) and Family 2 (F508del/F508del) belonged to Class II group whereas Family 3 (I1234V/I1234V) had mutations of Class IV.
As a result of comparison between mild and severe siblings with Class II mutation, seven genes (TNFRSF11A, KCNE1, STX1A, SLC9A3R2, CXCL1, CFTR, CXCL2) were found to be differentially expressed. TNFRSF11A, KCNE1, STX1A and SLC9A3R2 genes were up regulated while CXCL1, CFTR and CXCL2 genes were down regulated in severe patients. TNFRSF11A also known as NF kappa B activator, is a member of the TNF receptor family and triggers the transcriptional activation of NF kappa B pathway mediated proinflammatory cytokines and chemokines such as CXCL1-2-8 and IL1B.21 Similarly, in our study the reason for up regulation of TNFRSF11A (4,03 fold), may be to increase the inflammatory response to infection in severe patients with Class II mutation. The mRNA expression of the SLC9A3R2 (2,02 fold) and SCNN1G (2,77 fold) genes in the aldosterone regulated sodium reabsorption pathway were also increased in severe patients. Studies have shown that the ion transporters, negative CFTR regulators SCL9A3R2 (SLC9A3 regulator) and SCNN1G (ENaC subunit) are associated with decreased CFTR expression.22,23 STX1A was determined in SNARE interactions in vesicular transport pathway. Up regulated STX1A, have been associated with CFTR expression and its conductance.24,25 The decreased expression of CXCL1 (-5,03 fold), CXCL2 (-2,65 fold), IL1B (-3,22 fold), CXCL8 (-10,59 fold), HSP90AA1 (-2,09) and PTGS2 (-3,61 fold) in inflammation related pathways (IL-17 signaling, NF-kappa B signaling, cytokine-cytokine receptor interaction, TNF signaling, NOD-like receptor signaling and, chemokine signaling pathways) suggested that the inflammatory response could be suppressed in severe patients with Class II mutation. Some studies have shown increased expression of chemokines and cytokines detected in CF patients compared to controls.13,19,26But some reported conflicting results, similar to our findings. Ideozu et al. showed that the expression of CXCL1 and IL1B were decreased in CF patients compared to healthy controls.27 Similar to this study, Kormann et al. and Wright et al. showed that CXCL8 expression was decreased in severe CF patients.15,18In our study, the decreased chemokine expression detected in severe patients with Class II mutation could also be related to endoplasmic reticulum stress and chemokine secretion.28 Pathway enrichment analysis showed that NOD-like receptor signaling pathway was among the most prominent ones. Significantly down regulated CXCL1 and CXCL2 genes were found to be functional in NOD-like receptor signaling pathway in severe patients. In a study conducted on primary bronchial epithelial cell culture and monocytes of CF patients, it was detected that as a result of decreased CFTR protein expression, ENaC expression was increased. This caused stimulation of NLRP3 inflammasome and the release of IL18 and IL1B.29 This finding supported the formation of the inflamasomme mechanism in nasal epithelial cells.29,30 The decrease in the expression of genes in this pathway in severe patients could be elucidated by the decreased CFTR expression and ER stress.
In Class IV group, 24 genes (Up regulated 3 genes, down regulated 21 genes) were found to be differentially expressed in severe sibling. Since Class IV group consisted of only one family (3 siblings), the statistical significance of differentially expressed genes could not be evaluated and the pathways were interpreted by means of biological significance. As a result of pathway analysis, different from the Class II group, specific to the Class IV group, AGE-RAGE signaling pathways in diabetic complications and TLR signaling pathways were found. Especially, the AGE-RAGE signaling pathway in diabetic complications was the most important pathway in the severe patient. AGE-RAGE interaction triggers the activation of inflammation-related pathways such as NF-kappa B signaling.31 CCL2, CXCL8, TNF and IL1B were included in this pathway and were found to be down regulated in the severe patient (F3-P3). NF kappa B signaling pathway associated inflammation has suggested that in this patient the inflammatory response may be suppressed. Also, NOS3 gene was identified in the AGE-RAGE signaling pathway and played a role in caspase-3 activation.
In order to identify common differentially expressed genes between severe patients with Class II mutation and severe patients with Class IV mutation Venn diagram was created. 13 genes (KCNE1, ACE, CFTR, SFTPB, CXCR2, CXCL1, CXCL8, IL1B, PLA2G5, HSP90AA1, DUSP1, PTGS2, MBL2) which were found to be common between Class II and Class IV severe patients show that their expression change independent of mutation type (Figure S1). Gene Ontology (GO) analysis showed that the common genes were involved in neutrophil chemotaxis and regulation, inflammation response, chemokine signaling pathway and immune response-related processes.
As a result of heat map interpretation, it was an interesting finding that controls demonstrated similar gene expression profiles with mild Class IV group with I1234V mutation. F3-P1, which is the mildest sibling in the Class IV group, has a close relationship with C3. The common up and down regulated genes between intermediate F3-P2 and severe F3-P3 siblings with Class IV mutation, were associated with hepatic involvement. In addition, 17 genes (KIT, DEFB1, SNAP23, HSPA4, CLU, AHSA1, LCN2, HSP90AA1, EZR, HSPH1, CCL2, TCF7L2, FAS, KCNE1, PPP2R4, SERPINA1 and CXCL2) which showed distinct expression pattern between F3-P2 and F3-P3 were responsible for cell chemotaxis, inflammation, immune response, protein folding, stabilization, TNF, chemokine and IL-17 signaling pathways. Also, it can be concluded that inflammation and protein processing mechanisms may explain the clinical severity between intermediate F3-P2 and severe F3-P3 siblings with Class IV mutation. Another significant finding in the heat map was the close correlation of mild F1-P1 and severe F1-P2 in Family 1. Genes that showed a similar expression profile in Family 1, were the ones responsible for LPS response, cytokine-chemokine activity and protein folding processing mechanism. On the other hand, 32 genes which were differentially expressed between mild and severe siblings in Family 1 were associated with hepatic involvement and may have a role in determining clinical severity.
In order to investigate whether there was an association between clinical parameters (recurrent lung infection, FEV1%, hepatic involvement, CFRD) and differentially expressed genes in all CF patients, correlation analysis was performed. The increase in expression of 29 genes was found to have a significant correlation with an increase in severe hepatic involvement and liver inflammation. Among the significant pathways, AGE-RAGE signaling pathways in diabetic complications (PRKCE, CCL2, TNF, NOS3, EDN1 and TGFB1) was determined as the most outstanding one and may be regarded as a prognostic biomarker for CFLD which is characterized by focal biliary cirrhosis and occurs in 5-10% of patients with CF.32 A recent study performed by Paranjapye et al. has determined ABCA4, GSTP1, MBL2 and SERPINA1 as modifier genes for CFLD.33 Also, Debray et al. has confirmed SERPINA1 as a modifier gene for severe CFLD.34 Also they concluded that ACE, TGFB1 and CXCL8 could be associated with CFLD severity. However, in our study, we did not find an association between ABCA4, GSTP1, MBL2 and SERPINA1 genes and hepatic involvement in CF patients. We also showed that ACE and TGFB1 had a positive correlation with hepatic involvement.