The role of essential fatty acids in cystic fibrosis and normalizing effect of fenretinide

Abstract

Cystic fibrosis (CF) is the most common autosomal-recessive disease in Caucasians caused by mutations in the CF transmembrane regulator (CFTR) gene. Patients are usually diagnosed in infancy and are burdened with extensive medical treatments throughout their lives. One of the first documented biochemical defects in CF, which predates the cloning of CFTR gene for almost three decades, is an imbalance in the levels of polyunsaturated fatty acids (PUFAs). The principal hallmarks of this imbalance are increased levels of arachidonic acid and decreased levels of docosahexaenoic acids (DHA) in CF. This pro-inflammatory profile of PUFAs is an important component of sterile inflammation in CF, which is known to be detrimental, rather than protective for the patients. Despite decades of intensive research, the mechanistic basis of this phenomenon remains unclear. In this review we summarized the current knowledge on the biochemistry of PUFAs, with a focus on the metabolism of AA and DHA in CF. Finally, a synthetic retinoid called fenretinide (N-(4-hydroxy-phenyl) retinamide) was shown to be able to correct the pro-inflammatory imbalance of PUFAs in CF. Therefore, its pharmacological actions and clinical potential are briefly discussed as well.

Polyunsaturated fatty acids and their metabolism in CF

Metabolism of PUFAs

Unsaturated fatty acids are carboxylic acids that contain one or more double bonds and more than four carbons in the acyl chain. Monounsaturated fatty acids contain one double bond, whereas polyunsaturated fatty acids (PUFAs) contain two or more double bonds. The most common naturally occurring PUFAs are ω-3 and ω-6 PUFAs. The designation ω refers to the position of the most distal double bond counted from the last carbon of the acyl chain (e.g., ω-3 PUFAs have its first double-bond three carbons away from the most distal methyl group). Alternatively, the n letter instead of the ω can be used with the same meaning and fatty acids can be designated as n-3 and n-6. For example, docosahexaenoic acid is a ω-3 fatty acid that has 22 carbons and 6 alternating double bonds counted from the most distal carbon atom and thus can be categorized as 22:6n-3. Similarly, arachidonic acid is a ω-6 PUFA that has 20 carbons and 4 alternating double bonds counted from the 20th carbon and, thus, can be labeled as 20:4n-6.

Mammals do not have the ability to introduce double bonds in fatty acids beyond carbon 9 and carbon 10 (counted from the carboxyl carbon), and therefore, they must obtain some PUFAs from their food [1]. The two essential fatty acids for humans are linoleic acid (18:2n-6) and α-linolenic acid (18:3n-3), which are metabolized through the pathways shown in Fig. 1. Therefore, linoleic acid is the precursor for AA, and α-linolenic acid is the precursor for DHA. Importantly, both AA and DHA can also be introduced in the metabolism directly from food. Indeed, AA and its precursors are found in plant-derived oils (corn, peanut, and sunflower) and red meat, while DHA is found in fish oil and algae [2], as illustrated with the coloured boxes in Fig. 1. Some earlier studies expressed concerns that rodents may not be good models to study the metabolism of DHA with relevance to humans, because microsomal desaturase activity seemed much more prominent in rats than in humans and, consequently, conversion of α-linolenic acid would occur at a much faster rate in rats than in humans [3, 4]. However, the methods used to estimate the rate of DHA production from α-linolenic acid in these two studies were not the same. To address this issue, Domenichiello et al. [5] orally administered radioactively labeled ²H₅-α-linolenic acid to rats and sampled blood over a 6 h experiment to measure the amount of ²H₅-DHA, and they applied calculations used previously in human studies with a single oral bolus of labeled α-linolenic acid [5]. The percentage of α-linolenic acid that was converted to DHA in this study ranged from 0.12 to 0.64%, which is not higher than the previously estimated α-linolenic acid to DHA ratio in humans [6–8]. Based on these results, rodents may represent good models for the study of unsaturated fatty acids modulation. However, the same group conducted a study with 12 week DHA supplementation in humans, and despite a 130% increase in plasma EPA concentrations, there was no change in the plasma δ13 C-EPA levels suggesting that there was no or very little retroconversion of DHA to EPA. Plasma δ13 C-EPA remained similar to plasma δ13 C-ALA, which means that, virtually, all of the increase in plasma EPA with DHA supplementation was derived from ALA, which supports the slower turnover of EPA responsible for this phenomenon [9]. Therefore, even though retroconversion of DHA to EPA appears to be a real phenomenon in both rodents and humans, the increases in plasma and tissue EPA levels following DHA supplementation are not the result of increased flux through the retroconversion pathway, but are largely due to the slowed metabolism and accumulation of EPA.

Fig. 1.

Metabolism of ω-6 and ω-3 fatty acids starting with essential fatty acids obtained from food. The food sources that are rich in specific fatty acids are shown in the boxes placed next to the name of the corresponding fatty acid

Both AA and DHA are transported in the bloodstream either in the esterified form (incorporated into triglycerides or cholesteryl esters) or as “free fatty acids” bound to albumin and/or lipoproteins, but once inside the cell, they then get incorporated into phospholipids, and other complex lipid moieties which are principal lipid constituents of cellular membranes. Upon activation of various signaling pathways initiated either by endogenous (hormones and cytokines) or exogenous factors (bacterial lipopolysaccharide), AA and DHA are released from phospholipids through the action of enzymes called phospholipases. There are three major categories of phospholipases: cytosolic phospholipases, secreted phospholipases (sPLA₂), which are Ca²⁺-dependent and Ca²⁺-independent phospholipases (iPLA₂), and intracellular phospholipases, cPLA₂ and iPLA_2.

Each of these categories of enzymes is coded by multiple paralogous genes. For example, in the human genome, there are six genes for cytosolic phospholipases (PLA2G4A, PLA2G4B, PLA2G4C, PLA2G4D, PLA2G4E, and PLA2G4F coding for cPLA2-α, cPLA2-β, cPLA2-γ, cPLA2-δ, cPLA2-ε, and cPLA2-ζ proteins respectively), 11 genes for secreted phospholipases (PLA2G2A, PLA2G2D, PLA2G2F, PLA2G3, PLA2G5, PLA2G10, PLA2G1B, PLA2G2C, PLA2G2E, PLA2G12A, and PLA2G12B coding for group IIA sPLA2, group IID sPLA2, group IIF sPLA2, group III sPLA2, group V sPLA2, group X sPLA2, sPLA2-IB, sPLA2-IIC, sPLA2-IIE, sPLA2-XIIA, and sPLA2-XIIB proteins, respectively), and six genes for iPLA2 (PNPLA9, PNPLA8, PNPLA6, PNPLA3, PNPLA2, and PNPLA4 coding for iPLA2β, iPLA2γ, iPLA2δ, iPLA2ε, iPLA2ζ and iPLA2η proteins, respectively) [10, 11].

AA is released from the phospholipids mostly through the actions of cPLA2 and sPLA2 enzymes [10, 11] while DHA can be released through the action of iPLA2 enzymes [12, 13] and sPLA2 [14, 15] (Fig. 2). Interestingly, once released from the membrane, DHA can inhibit cPLA2 and thus limit the release of AA [16].

Fig. 2.

Subcellular metabolism of ω-6 and ω-3 fatty acids starting with essential fatty acids obtained from food (LA and ALA) and from the complex phospholipid membrane (AA and DHA)

Once released from the membrane, AA and DHA are metabolized through a series of enzymatic steps into molecules with potent pro- and anti-inflammatory properties.

AA is the substrate for cyclooxygenases: constitutively expressed COX-1 and inflammation induced COX-2 are both well-known targets of aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) [17]. These two enzymes initially convert AA into an unstable intermediate called prostaglandin H₂, which is converted to diverse biologically active prostanoids depending on the presence of different tissue-specific isomerases. These are prostaglandins (PGE₂, PGD₂, and PGF2α), prostacyclin (PGI₂), and thromboxanes (TXA₂ and TXB₂) [18]. AA is also the substrate for lipooxygenases (5-LOX, 12-LOX and 15-LOX), which convert it into leukotrienes, lipoxins, and hepoxilins [19]. 5-LOX produces leukotriene A4 (LTA₄), which leads to cysteinyl-leukotrienes (LTC₄, LTD₄, and LTE₄), leukotriene B4 (LTB₄), and lipoxins (LXA₄ and LXB₄). 12-LOX converts AA into 12-hydroxyeicosatetraenoic acid (12-HETE), and 15-LOX converts AA into LXA₄ and LXB₄ [20]. Indeed, LXA4 exerts potent anti-inflammatory activities most notably by counteracting LTB4-induced neutrophil migration [21–23] and mice treated with LXA4 analogs resolved P. aeruginosa lung infection more efficiently than placebo-treated controls [24]. Importantly, the LXA4/LTB4 ratio is strongly decreased in the lower airways of children with CF [25] (Figs. 3 and 4)

Fig. 3.

Lipid mediators derived from arachidonic acid. Most of the lipid mediators derived from arachidonic acid have pro-inflammatory properties (red), with the notable exception of lipoxins which are anti-inflammatory (green). In one branch of the pathway, arachidonic acid is a substrate for cyclooxygenases and resulting mediators are prostacyclin, prostaglandins, and thromboxanes, all of which have pro-inflammatory properties. Alternatively, arachidonic acid serves as the substrate for 5-lipooxygenase, which leads to the synthesis of leukotrienes and lipoxins. Lipoxins are the only known anti-inflammatory lipid mediators derived from arachidonic acid

Fig. 4.

Lipid mediators derived from Ω-3 fatty acids. All known lipid mediators derived from Ω-3 fatty acids have anti-inflammatory properties. EPA-derived mediators are resolvins of E series, whereas resolvins of D series, maresins, and protectins are derived from DPA and DHA through the series of intermediates. All of these intermediary compounds are short-living, unstable species, and their physiological functions, if any, remain unknown

All secondary metabolites produced from AA, with the notable exception of lipoxins, are potent pro-inflammatory molecules. The activation of numerous G-protein coupled receptors (GPCRs) through which these derivatives of AA act leads to the well-known hallmarks of inflammation, such as synthesis of pyrogenic cytokines which raise the body’s temperature, increased chemotaxis of neutrophils, and degranulation of platelets at their site of production [26].

On the other hand, all known metabolites of DHA have anti-inflammatory properties [27]. Once released from the membrane by the action of phospholipases, DHA serves as the substrate for 15-LOX and 12-LOX enzymes. 15-LOX converts DHA into the intermediate 17S-hydroperoxy-DHA which is further converted by neutrophil-derived 5-LOX to the D series of resolvins (RvD1–RvD4) [28, 29]. DHA can also be converted by 15-LOX into 16, 7-epoxy-protectin intermediate and then to protectin D1 [28, 29]. Finally, 12-LOX in macrophages converts DHA into maresins (Macrophage Mediators in Resolving Inflammation) [30, 31]. Resolvins, protectins, and maresins are potent anti-inflammatory molecules that also act through specific GPCRs and activate multiple signaling pathways that lead to the resolution of inflammation [27]. These include reduction of neutrophils chemotaxis, stimulation of phagocytosis of apoptotic neutrophils, clearance of allergens, and many others [27]. A detailed discussion of the physiological actions and biosynthetic pathways of these metabolites can be found in the other recent excellent reviews on this topic [32–36].

PUFAs in CF

The first association of abnormal metabolism of PUFAs with CF was made back in 1962 when Kuo et al. [37] described abnormally low levels of linoleic acid (18:2n-6) in the plasma and tissues of children with CF. The next two important milestones were the discoveries that the proportion of DHA to AA is decreased in different tissues of CF patients examined post mortem [38], whereas the proportion of AA to DHA is increased in leukocytes [39–41]. These historical findings were later confirmed in both CF patients [42] and CFTR-KO mice [43, 44]. Taking into consideration the previously discussed pro- and anti-inflammatory roles of AA and DHA, respectively, these findings collectively defined what is nowadays known as a pro-inflammatory imbalance of fatty acids in CF, which is also frequently illustrated as a high AA/DHA or low DHA/AA ratio.

However, with the cloning of the CFTR gene in 1989, which sparked the hope that the genetic defect itself could be corrected by gene therapy, the initial description of pro-inflammatory imbalance of fatty acids fell into obscurity. New interest in this field was rekindled in the early 2000s with the finding that the dietary supplementation with DHA of cftr −/− mice ameliorated many of the CF-related pathologic phenotypes, including reduction of villi hypertrophy, reversal of pancreatic duct dilation, and most importantly, restrained LPS-stimulated pulmonary inflammation [43, 45]. Although an improvement of liver function in mice was observed [43, 45–47], the supplementation of DHA alone and other n-3 PUFAs alone did not confer a significant clinical benefit to patients with CF [48–50]. Nevertheless, it was shown that the pharmacological correction of the pro-inflammatory AA/DHA ratio by fenretinide that leads to the simultaneous decrease in AA and increase in DHA (which may not be achieved with DHA treatment alone) resulted in the significantly improved ability of cftr −/− mice to resolve pulmonary infection with P. aeruginosa [44].

For a long time, it was thought that the pro-inflammatory imbalance of fatty acids observed in patients with CF is the consequence of pancreatic insufficiency and the resultant lipid malabsorption [51–53]. However, several independent lines of evidence undoubtedly disproved the malabsorption hypothesis. First, fatty acid imbalance persists in CF patients in whom pancreatic insufficiency has been successfully corrected with enzyme replacement and nutritional support [54, 55]. Second, the pro-inflammatory imbalance of fatty acids is seen in cftr −/− mice [43, 44], mostly in tissues that in littermate, wt (cftr + / +) mice express high levels of CFTR protein such as the lungs, pancreas, and intestine, but not in the brain, kidney, or heart [43, 56]. Finally, the presence of the pro-inflammatory imbalance of fatty acids in cultured cells derived from the lungs of CF patients demonstrates that this phenomenon is not secondary to pancreatic insufficiency or pathogen-induced lung inflammation, but is indeed causally linked to lack of functional CFTR protein and is, therefore, an intrinsic defect of CF [57, 58].

Several non-conflicting hypotheses were proposed to explain the pro-inflammatory imbalance of fatty acids in CF, but none of them provides a mechanistic connection between the absence of functional CFTR protein and high AA/DHA ratio. Historically, the first described abnormality of fatty acids in CF, i.e., the LA deficiency, as well as the increased level of AA could simply be explained by the rapid turnover of LA into AA [59]. Indeed, increased conversion of radioactively labeled LA into AA was demonstrated in CF cells compared to wild-type controls [60]. These changes parallel with the increased expression of Δ5- and Δ6-desaturases (coded by FADS1 and FADS2 genes, respectively) and could be induced by the treatment of cells with a small molecule inhibitor of CFTR (CFTR_inh-172). However, these results do not fully explain the low levels of DHA observed in CF.

Interestingly, DHA is known to suppress the expression of Δ5- and Δ6-desaturases in cultured cells [61] and mice [43, 56], but the causal connection between CFTR deficiency and low DHA levels remains unclear. It was demonstrated that the metabolism of EPA to DHA is decreased in cultured CF cells [60], although in that study, the levels of 22:5n-3 (DPA), an immediate downstream product of EPA, were not decreased. On the other hand, retroconversion of DHA to EPA which occurs through the modified β-oxidation in peroxisomes [62, 63] was also described in CF [61]. Abnormal metabolism of phosphatidylcholine (PC) was proposed as a possible explanation for lower levels of DHA in CF. PC formed de novo through the methylation of phosphatidylethanolamine (PE) is known to have higher DHA content than PC formed using choline derived from the diet [64, 65]. CF cells appear to have a defect in the metabolism of methyl groups; therefore, the de novo synthesis of PC is favored, thereby depleting DHA levels [66, 67]. Whatever the mechanism of DHA deficiency in CF is, it is clear that this abnormality may result in a lower level of anti-inflammatory lipid mediators derived from DHA and EPA, as it was indeed reported for Resolvin E1 in CF patients, impairing their ability to resolve inflammatory responses induced by frequent lung infections [68]. Consistent with the anti-inflammatory role not only of DHA, but also of other metabolites from the n-3 pathway, lipidomic analysis of plasma samples revealed that the higher levels of docosapentaenoic acid (DPA) showed a very strong positive correlation with higher FEV1 in patients with CF, whereas lower levels of α-linolenic acid were associated with chronic infection with P. aeruginosa [69].

Finally, numerous studies have convincingly demonstrated that the increased release of AA in CF cells from the cell membrane is mediated by cytosolic phospholipases (cPLA₂) [39, 70–73]. In this scenario, the increased conversion of LA to AA could be merely a compensatory response to this change, which inadvertently contributes to even higher levels of AA. The association between these two pathways, if any, remains to be explored.

Nevertheless, it has never been established which one out of the six cytosolic phospholipases contributes the most to the increased levels of AA. We examined the mRNA levels of all six cytosolic phospholipases in CFBE 41o-cells derived from a ΔF508/ΔF508 patient and we could not observe any difference in mRNA between parental cells and those transfected with wt-CFTR protein [58]. However, this leaves an open possibility that the protein levels and/or activities of cPLA₂ enzymes are affected by posttranscriptional modifications. Indeed, the treatment of CFBE41o- parental cells with fenretinide, an orphan drug which was previously reported to decrease the level of AA and increase the level of DHA [44], was found to decrease the total cPLA2 activity assayed by colorimetric assay [58]. The mechanism of this change remains to be further explored.

Finally, an important role for anti-inflammatory metabolites of DHA in CF is starting to emerge. Indeed, resolvin D1 (RvD1) has been detected in the plasma and sputum of CF patients, and an increased RvD1/IL-8 ratio is associated with a better pulmonary function and higher FEV₁ in CF patients [74]. Interestingly, RvD1 can increase airway surface liquid height in CF bronchial epithelial cells, restoring the nasal trans-epithelial potential difference in CF mice by decreasing the amiloride-sensitive sodium absorption and by stimulating CFTR-independent chloride secretion [75]. RvD1 decreased TNFα-induced IL-8 secretion and enhanced the phagocytic and bacterial killing capacity of alveolar macrophages from CF patients [75].

Taken together, these studies underline the importance of restoring the normal levels of DHA and AA in patients with CF and suggest that CF patients would benefit from such a biochemical correction.

Pro-inflammatory fatty acid imbalance in CF and its normalization by treatment with fenretinide

Understanding the mechanisms of pro-inflammatory fatty acid imbalance in CF is of direct relevance to dietary therapy [76, 77]. Unfortunately, the typical modern Western diet contains a high ratio of ω-6 to ω-3 PUFAs [78, 79], and, therefore, may exacerbate the already present intrinsic lipid imbalance resulting from high AA and low DHA production in CF patients. This, in turn, may result in an increased production of pro-inflammatory lipid mediators either triggered directly by AA-mediated phosphorylation of signaling molecules such as phospho-ERK or via decreased production of anti-inflammatory lipid mediators for which DHA serves as the starting substrate, such as D-class resolvins. Unfortunately, supplementation of DHA alone and other n-3 PUFAs alone did not confer a significant clinical benefit to patients with CF [48–50]. The outcomes of these trials raise many questions regarding the physiological role of not only AA and DHA but also of their precursors and downstream metabolites. For example, dihommo-γ-linolenic acid (DGLA) is a known substrate for cyclooxygenases, giving rise to PGE₁ and for 15-LOX, increasing 15-hydroxy-eicosatrienoic acid levels, both of which are anti-inflammatory eicosanoids [80]. Additionally, the supplementation of γ-linolenic acid, the immediate upstream precursor of dihommo-γ-linolenic acid, was shown to decrease the production of pro-inflammatory LTB₄ by neutrophils ex vivo [81–83]. Several studies involving an increased intake of GLA reported an increment of DGLA in PMBCs and neutrophils as expected, but surprisingly found no effect on the levels of AA [81, 82, 84–86]. An epidemiological study assessing 1123 subjects in Italy found no correlation between the concentrations of LA in plasma and eight markers of inflammation (IL-6, soluble IL-6R, IL-1β, IL-1 receptor antagonist, TNF-α, IL-10, TGF-β, and CRP) [87]. Although subjects in the lowest quartile of plasma LA concentrations had the highest concentrations of pro-inflammatory markers (IL-6 and CRP) and the lowest concentrations of anti-inflammatory cytokines (IL-10 and TGF-β) [87]. Finally, even though AA is a known downstream metabolite of LA, increased levels of LA are not necessarily associated with high levels of AA [88–91] and it is still unknown why. This clearly implies that our linear view of PUFA metabolism was oversimplified thus far.

Fenretinide [4-hydroxy(phenyl)retinamide; 4-HPR] is a synthetic analog of retinoic acid which has a phenyl moiety instead of carboxylic group linked through an amide bond (Fig. 5).

Fig. 5.

Structures of fenretinide and retinoic acid

We have previously reported that 5–10 mg/kg doses of fenretinide correct pro-inflammatory lipid imbalance in CFTR-KO mice by decreasing the levels of AA while simultaneously upregulating the level of DHA [44]. Fenretinide was used in the long-term chemopreventive trial (5 years daily) at a 100 mg per day dosage in patients with breast cancers [92] and relatively high doses (several times higher) for treatment of patients with neuroblastoma [93]. In both cases, fenretinide was shown to be a relatively safe agent and the most common adverse effects reported were minor and reversible visual disturbances such as diminished dark adaptation. This adverse effect is most likely due to the ability of fenretinide to decrease endogenous retinol concentrations by competing for the same transport protein RBP4 [94]. Fenretinide binds strongly to RBP4, but this complex does not bind to transthyretin [95]. However, fenretinide–RBP4 complex is eliminated through renal filtration, thereby depleting the level of RBP4 necessary to transport retinol, which eventually results in lower retinol levels in fenretinide-treated patients [95]. However, a few days of interruption of treatment after every 4 weeks of treatment to recuperate the RBP4 levels were sufficient to resolve this issue.

Similarly to retinoic acid, which is a physiologically active metabolite of retinol, fenretinide binds to retinoid receptors (RARs), which belong to the family of nuclear receptors. Upon binding of their ligand (which may either be all-trans-retinoic acid or 9-cis retinoic acid), several isoforms of RARs, known as RAR-α, -β, and -γ, bind to the DNA sequence known as retinoic acid receptor elements (RAREs) and regulate transcription of hundreds of genes [96]. On the other hand, retinoid X receptors (RXR-α, -β, and -γ) bind only 9-cis retinoic acid [97]. Fenretinide was found to bind strongly to RARγ, weakly to RARβ [98, 99] and not at all to RXRα [99]. Additionally, fenretinide activates transcription by a nuclear receptor called PPARγ [100, 101], although there is no evidence for the direct interaction with PPARγ. The consequences of these actions are certainly cell-type dependent and result in the down-regulation or up-regulation of the transcription of the genes normally controlled by these nuclear receptors. Importantly, PPARγ receptor has long been known to be reduced in CF [102–104] and pharmacological activation of PPARγ by rosiglitazone was shown to confer some therapeutic benefits in CF mice [105]. However, rosiglitazone has been recalled from the market due to serious side effects such as elevated risk of heart attacks in diabetic patients [106]. Fenretinide was found to prevent the activation of the NF–κB pathway by inhibiting the phosphorylation of IκBα kinase [107] and to inhibit phosphorylation of Erk kinase [108], thereby inhibiting the pathways generated from multiple Tyr-phosphorylated growth factor receptors.

Due to its pleiotropic effects, outlined above, it is challenging to pinpoint the exact mechanism whereby fenretinide affects the activity of cytosolic phospholipases. Additionally, fenretinide was found to increase the level of anti-inflammatory very-long-chain ceramides while decreasing the levels of pro-inflammatory long-chain ceramides and this effect could be attributed to the down-regulation of Cers5 enzyme [109].

Even though the exact mechanism of fenretinide action on different lipid species is still not fully elucidated, fenretinide is a safe agent for the pharmacological correction of aberrant lipid metabolism in patients with CF. Long-term safety of fenretinide was well established in the population of breast cancer patients [110, 111]. The safety profile of fenretinide in CF patients was already evaluated in a Phase Ib clinical trial with a novel orally bioavailable formulation of fenretinide (LAU-7B). The reassuring safety profile and excellent pharmacokinetics of the LAU-7B formulation obtained in a dose escalation (100, 200, and 300 mg per day) Phase Ib trial, led to the Phase II (APPLAUD) trial (NCT03265288). This is a currently ongoing international Phase II, double-blind, randomized, placebo-controlled study to confirm the safety of long-term (6 months) treatment using 300 mg once-daily dosage and to evaluate the efficacy of LAU-7b in CF patients.

Finally, since the signaling through retinoid receptors and NF–κB pathway is known to be essential for constitutive and pathogen-induced expression of mucin genes [112–114], we examined how fenretinide treatment affects the expression of MU5AC and MUC5B genes in a goblet cell line derived from rat lungs. Interestingly, our recent results demonstrate that the low dose of fenretinide (1.25 µM) selectively prevents MUC5AC overexpression, without significantly affecting MUC5B expression [58, 115]. These results are especially important considering significant biological differences between MUC5AC and MUC5B [116, 117]. It was recently demonstrated that MUC5B is required for mucociliary clearance and immunological defense of the lungs, whereas MUC5AC is dispensable and is, in fact, responsible for pathological plugging and airway obstruction [117]. Nevertheless, the mechanistic connection between pro-inflammatory lipid imbalance and the regulation of mucin gene expression remains unexplored. Most likely, such a connection would be indirect, through the metabolites of fatty acids such as eicosanoids [118] or 12(R)-HETE [119].

Conclusion

In conclusion, the pro-inflammatory imbalance of PUFAs reflected in a high AA/DHA ratio is an important component of sterile inflammation which is detrimental for the lung defense of CF patients. This underexplored biochemical aspect of CF may lead to the development of novel therapeutic interventions in the future. To this end, fenretinide holds promise as a safe pharmacological agent for the correction of abnormal lipid metabolism and as a novel mucoregulatory agent.

Author contributions

DG and DR wrote the manuscript. DCD created the figures and edited the text. JS and JBdS edited the text, performed literature search, and provided valuable comments.

Compliance with ethical standards

Conflict of interest

Dr. Danuta Radzioch is a scientific advisor and a shareholder in Laurent Pharmaceuticals Inc., which is currently running Phase II (APPLAUD) multicenter clinical trial aimed to assess the efficacy of fenretinide in patients with CF.