Outside the critical micellar concentration range, the amount of lipid solubilized is significantly reduced. Consequently, steatorrhea appears 138 and FA absorption decreases 139. This situation is found in CF patients where total bile salt secretion is impaired 140. About 36% of CF children showed reduced BA secretion into the duodenum 141. Despite this low BA secretion, impaired water secretion leads to high concentrations of bile salts in hepatobiliary secretions 142, which may contribute to the cholelithiasis reported in CF patients 94143.

EFAs regulate BA excretion: PUFAs are reported to induce the excretion of BA in rats 144 and humans 145, but this effect depends on the EFA family, since n-3 FA increases bile flow more than n-6 FA 146. The EFAD effect on bile flow and BA secretion varies among animal species: decreasing in rats 147148, having no effect in mice 114 and increasing in hamsters 114. In humans, the impact of EFAD on BA secretion and composition remains unclear. However, prostaglandins, which are AA metabolites, alter hepatic bile flow 149 indicating that disturbances of AA status may affect choleretic response in CF patients.

The nature of BAs influences their ability to solubilize lipids. Cholesterol absorption by enterocytes is greater with cholyl-taurine than with chenodeoxycholyl-taurine 150151152 and BAs with conjugated taurine are more effective than glycine-conjugated BAs at solubilizing lipids. The effect of EFA on BA composition variation and fat absorption remains enigmatic in humans, since there are no available data.

In CF, oral taurine load appears to be normal, but excessive fecal loss 153 increases the ratio of glycine/taurine-conjugated BA 94140141. Oral taurine supplementation in CF children is effective in decreasing the glycine/taurine ratio in duodenal fluid 153, but its ability to improve fat malabsorption is controversial 154155156157. Apparently, taurine supplementation improves BA malabsorption, mainly in patients with a high degree of steatorrhea 158159160.

Studies on animals show that EFAD leads to impaired biliary excretion of taurocholate 161 and to reduced EFA content (linoleic and arachidonic acid) in biliary PC that is essential for micelle formation 162. However, BA and bile lipid composition vary across species. In rats and hamsters with EFAD, bile composition is markedly impaired 147161. On the other hand, BA composition appears to be similar in EFA-deficient and EFA-sufficient mice 114.

The relationship between BA and EFA was investigated in CF-EFAD children and it showed that ursodeoxycholic acid supplementation improves the hepatic metabolism of EFA. After 1 year, ursodeoxycholic acid supplementation led to an improvement of EFA status (reduction of triene/tetraene FA ratio) 163.

BAs are recycled through the enterohepatic cycle with remarkable efficiency (95% reabsorbed) 164165. The enterohepatic cycle is often interrupted in CF because of excessive fecal losses of BA, 95166167168169170, which diminish the BA pool 171172. Excessive BA losses may be attributed to: 1) BA irreversibly bound to unhydrolyzed TG and PL within the intestinal lumen 173 ; 2) BA precipitation due to acidic pH in the duodenum 174; and 3) intestinal bacterial overgrowth, which is present in 40% of CF patients and results in BA deconjugation and dehydratation 175. BA loss may also result from a primary cell defect in the active absorption of BA in the ileum. In vitro studies using brush border membrane vesicles from CF patients have shown that total ileal BA uptake is diminished 176177. Surprisingly, a study of Ileal Biliary Acid Transporter (IBAT) in CFTR knock-out mice shows a BA uptake rate four-fold that of wild-type mice 178. The increase in IBAT protein and BA uptake can be interpreted as an up-regulation in response to a low BA rate. This result should lead to a renewed interest in intraluminal events, especially for the implication of the thick mucus barrier in CF pathophysiology.

Fat digestion begins in the stomach with the action of acid lipase. Following hydrolysis by gastric lipase, medium-chain FAs are partly absorbed by the stomach 179. Thus, the stomach plays an essential role in fat digestion, especially in pancreatic insufficiency 147.

In CF patients with pancreatic insufficiency, altered motility with an increase in gastric emptying 180 and small bowel transit time has been described 181182183184. It has been reported that slow gastric emptying reduces the success of pancreatic enzyme replacement therapy in improving TG hydrolysis, which could explain in part the variation in pancreatic enzyme replacement therapy efficiency 180185. However, other studies have shown that gastric emptying time is similar in both CF patients and healthy controls 90182.

In EFAD rats, AA and linoleic acid are emptied from the stomach at similar rates and these rates do not differ from controls 186. However, EFA-deficient mice show that the motility of epithelial cells is increased in the jejunum 187.

In the majority of CF patients, histological brush border studies reveal normal morphology 188. However, in some cases, abnormalities are reported, such as ileal hypertrophy 189 or partial villous atrophy in the small intestine 190191. This atrophic mucosa, which also occurs in knock-out mice 189, may result from acidic aggression by unbuffered stomach chyme, chronic inflammation or denutrition. Furthermore, a thick mucinous layer covering the brush border and large areas of the microvilli is revealed by electron microscopic examination of biopsies from CF patients 188, which may contribute to malabsorption in CF patients. The viscosity of the intact CF glycoprotein is almost two-fold that of normal glycoprotein 192. Elevated viscosity may be caused by defective chloride transport, mucin hyperglycosylation or a high level of disulfide-linked peptides 193194. Obviously, the highly acidic properties of surface components related to oversulfatation could modify the micro acidic climate on the intestinal epithelium surface and influence interaction between micelles and enterocytes, impairing the protonation of FAs.

In EFAD, histological and biochemical alterations of the intestinal mucosa are described 195. In rats and mice with EFAD, the height of villi is decreased leading to a diminished absorption area. Cellular differentiation was also found to be impaired 196, highlighting the role of EFA in the formation of new tissues, such as the maintenance of tissue and cell structure 197. Furthermore, the abnormal structure of mitochondria and microvilli is correlated to decreased fat absorption 196. It seems difficult to reconcile jejunal hypertrophy in CF patients and microvilli in EFAD animals. However, EFAD seems to induce histological abnormalities in the CFTR function. In CFTR knock-out mice, jejunal hypertrophy is corrected with an oral administration of high doses of DHA 198, which are associated with an inhibitory proliferative cell effect 199. Then, more than in typical EFAD, an abnormal EFA imbalance could lead to hypertrophic villosities in CF patients. Hypertrophy is only localized in the jejunal portion. Accordingly, the FA composition of jejunum mucosa, the main segment for optimal lipid absorption, is markedly different from ileal and colonic mucosa 195. Furthermore, it is interesting to note that dietary influences are tissue specific, since serum or red blood cell membranes do not reflect local changes in any of the different intestinal segments 195. At present, no studies have correlated intestinal morphology in CF patients with their EFAD status.

Recently, studies have suggested that pathological modifications in EFAD may be the consequence of cell adhesion disorders 200.

In CF patients, the dual sugar permeability test 201 demonstrates that the paracellular pathway is more permeable to large molecules, while the passive transcellular uptake of small molecules is normal 182. Intestinal permeability in CF is related to patient genotype: patients homozygous or heterozygous for ΔF508 mutation exhibit significantly increased intestinal permeability compared with patients with unidentified genotypes or controls 202. Moreover, abnormalities in the tight junction, an essential structure for the control of intestinal permeability, have been reported in the intestinal epithelium of fetuses with CF 203.

EFAs are able to modify cells ultrastructurally and to alter intestinal permeability. In the culture of enterocytes from EFA-deficient CF, the lateral surfaces between cells are fairly straight, a consequence of the absence of complex interdigitations that play an essential intercellular cohesive role 196. Moreover, a reduction in the number of desmosomes has been reported in the intestinal tract of EFAD rats 59204. Studies in endothelial cells suggest a possible mechanism of EFA-modulating cell adhesion. However, this EFA function has to be demonstrated in intestinal cells. In endothelial cells, some PUFAs such as γ-linolenic acid or EPA increase transepithelial electrical resistance and reduce paracellular permeability. Studies involving CF patients are definitely needed to assay intestinal cell adhesion molecules and their relation to FA status.

EFAs may be absorbed by enterocytes, mostly in the form of EFA and MG. Until recently, it was assumed that these lipids diffused passively through the enterocyte brush border membrane 205. Indeed, earlier studies reported that the uptake of FAs occurred at 0°C, implying that the process is strictly passive. In particular, the intestinal absorption mechanism of linoleate was noted to depend on its intraluminal concentration, showing a passive diffusion at high concentrations. However, a transporter was required at weaker concentrations 206. Accordingly, the absorption of FFA was found to be a saturable phenomenon that can be inhibited through competition with long-chain PUFA 207. These observations suggest that FA uptake is a concentration-dependent dual transport mechanism involving both passive diffusion and a carrier-dependent process. Recently, several membrane transport systems have been identified and they seem to be involved in the enterocyte absorption of lipids: membrane FA binding protein (FABPm), capable of facilitating transmembrane passage mainly of FAs, but other lipids as well 208209210; FA translocase/cluster determinant 36 (FAT/CD36) implicated in long-chain FA transport. 211212213; scavenger Receptor class B type I (SR-BI) that plays a role especially in cholesterol movement (absorption and/or efflux) at the enterocyte level 214; ATP Binding Cassette transporter family that provides several cholesterol carriers: ABCG5 and ABCG8 cooperate to limit sterol intestinal absorption, rather facilitating cholesterol efflux toward the intestinal lumen and their mutations predispose to sitostrolemia 215216; and ABCA1 expressed in the enterocyte, which could partially control cholesterol efflux toward the intestinal lumen 217, although its exact role in the brush border membrane remains controversial 218219. Recently, a new protein called Nieman Pick C1-Like1 was identified in the small intestine. It seems closely involved in intestinal cholesterol absorption, a pathway sensitive to sterol absorption inhibitors such as ezetimibe 220221.

No study to our knowledge has investigated the interaction that may exist between the CFTR and these lipid transporters. Yet, the CFTR is known to modulate the activity of other carriers as well as certain ionic channels, for instance 37222. The CFTR regulation of other intestinal ionic transporters is effectively diminished in CF patients 223. Furthermore, a microarray study on pulmonary tissue from knock-out CFTR mice shows that membrane transporters specific for ligands as different as glutamate, hormones or neurotransmitters have their expression influenced by CFTR 224. The completion of the same type of study on intestinal tissue would likely offer interesting tracks to target lipid transport proteins capable of being influenced by the CFTR. The large diversity of transporters interacting with the CFTR could lead to impairment in enterocyte lipid uptake and trafficking in CF, which would represent another cause for nutrient malabsorption.

The existence of anomalies in the enterocyte uptake to EFA is controversial. In effect, EFA intestinal absorption in patients does not always appear to be impaired. Some studies report that the rate of linoleic acid absorption is normal when pancreatic enzyme supplementation is given at sufficient doses 225. Apparently, even the presence of steatorrhea was not accompanied by diminished EFA absorption 186 and no correlation has been established between steatorrhea and EFAD in preadolescents with CF 226. On the other hand, a recent study has shown that patients undergoing pancreatic enzyme treatment display a reduction in the intestinal uptake of long-chain FAs 9. The differences between these studies may indicate either that the intestinal malabsorption does not in itself explain the EFA deficit or that the severity of CFTR mutation could influence the enterocyte absorption of lipids. Unfortunately, genotyping analysis has not been carried out in most of these investigations. Future developments will get to the bottom of these major unsolved questions by tracing the defects in enterocyte lipid uptake, the status of lipid transporters, the relationship with CFTR in CF patients and fat malabsorption.

Early studies on EFAD did not notice an anomaly in the enterocyte transport of EFA 186196227. However, these data have not been confirmed. The studies are somewhat outdated and the sensitivity of the techniques used may be called into question. Furthermore, several elements suggest that EFAD may be implicated, at least to some extent, in lipid intestinal malabsorption. Indeed, EFAD can affect the lipid composition of the enterocyte membrane and modify its fluidity, which may directly disturb the functioning of the transporters that can be found there. Moreover, some transporters like the SR-BI or CD36 could act as transporters, not directly, but by creating a special micro-environment in the neighbouring membrane lipids. This local change favors the transfer of their ligands 228. In this model, it appears very likely that a modification of the physiochemical properties of the membrane, as is the case in EFAD, may impact on the transporter lipid transfer abilities. EFAD could also influence these lipid transporters more directly. In effect, most of these transporters are regulated by long-chain FAs: the expression of the FABP and CD36 genes is increased by long-chain FAs 229; the SR-BI protein is also regulated according to the type of long-chain FAs (unpublished personal data); and finally, the polyunsaturated FAs trigger a decrease in the ABCA1 protein, thereby reducing the basolateral efflux of cholesterol in human Caco-2 cells 230. This regulation could be drilled through a reduction in ABCA1 expression 231 or an increase in the degradation of the protein 232.

FAs can also act directly on the level of expression of the CFTR protein. In this way, a short-chain FA, the butyrate, increases CFTR expression significantly in animal epithelium cell cultures 233. It is interesting to note that this same FA modulates lipid synthesis, the biogenesis of apolipoproteins (apos) and the assembly of LP in the enterocyte 234. To our knowledge, there are no similar studies with long-chain FAs or EFAs. The potential role for the CFTR in all the enterocyte lipid synthesis steps underlines the need for new investigations, which may lead to new therapeutic strategies.

After crossing the brush border membrane, the lipids must be processed by cytosolic proteins for their intracellular trafficking toward various compartments, including the endoplasmic reticulum (ER), where their reesterification may take place. However, the precise mechanisms behind this transport remain unclear. Certain transport proteins have been identified, but their roles often remain hypothetical: the Sterol Carrier Protein (SCP-2) and two cytosolic FABP capable of linking FA in particular, but also PL. According to our unpublished and preliminary data, intestinal-FABP (I-FABP), which is restricted to the intestine, could direct the FAs to membranes for lipid cycling or to a degradation pathway (peroxisome or mitochondria), whereas the liver-FABP (L-FABP), also found in the liver, could guide them to the ER to be assembled into LP. I-FABP and L-FABP bind differently according to lipid class, but both with a greater affinity for unsaturated FAs than saturated FAs.

No study has examined the interactions that can exist between the various proteins of the cytoplasmic transport of lipids and the CFTR. However, it has been clearly demonstrated that anomalies in CF exist in the intracellular movement of the CFTR itself. Various physiochemical conditions, such as low temperatures 233 or chemical agents 235, are able to increase the stability of the mutated CFTR protein and avoid its abnormal retention in the ER 236237. The latter would, in part, be attributable to a defective interaction between the CFTR and certain PLs that play the role of lipid chaperones. Obviously, the mutation of ΔF508 causes the CFTR protein to lose its ability to bind preferentially with phosphatidylserine rather than with PC. The replacement of PC with non charged analogues in mutated cell cultures increases CFTR expression and the quantity of its mature form. Therefore, certain PLs through their lipid chaperone function seem important for the intracellular trafficking of the CFTR 238. Moreover, the invalidation of the CFTR initiates changes in gene expression as well as protein degradation via ubiquitin-dependent proteasome, which may modify several transport proteins 224. Overall, these studies support the concept that CFTR is modulated by PLs and indicate potential relationships between CFTR and other local transporters. Unfortunately, it remains unknown how EFAs contained in PLs directly alter the activity of CFTR and whether a "partnership" exists between CFTR and intracellular lipid transporters in the enterocyte. Forthcoming investigations will highlight new links between the CFTR, EFAs and cellular processes, which may identify important factors that play a role in networks of lipid signalling and transport.

The assembly of microtubules is critical for intracellular trafficking and chylomicron transport 239240. Early reports had indicated that the administration of microtubule inhibitors led to a decrease in the conveyance of radioactive lipids in rat enterocytes 241. Certain EFAs, such as γ-linoleic acid or AA, regulate the microtubule polymerization 242243. Thus, EFAD could hinder intraenterocyte trafficking through an alteration of microtubule polymerization, synthesis or function.

The products of lipolysis once absorbed and transported to the ER are resesterified to form TG, PL and CE through specific enzymatic pathways. This reesterification involves several enzymes monoacylglycerol acyltransferase (MGAT), diacylglycerol acyltransferase (DGAT), glycerophosphate acyltransferase, phosphatidate phosphodiesterase, lyso-PC acyltransferase, CE and acyl-coenzyme A:cholesterol acyltransferase (ACAT). After their synthesis, hydrophobic lipids must be associated with proteins or apolipoproteins, in order to allow their solubilization in the blood circulation, thereby forming complexes known as LP. The intestine is capable of secreting most lipoprotein classes (chylomicrons, VLDL, HDL), but chylomicrons represent the specific and most abundant class in the enterocyte.

Note that apo B is a component that is essential for LP assembly and secretion by the intestine. It exists in two forms: apo B-100, present particularly in the liver, but to a limited extent in the intestine, and apo B-48, which is specific to the enterocyte that results from the posttranscriptional modification of the apo B-100 mRNA, or editing, involving an enzymatic complex called APOBEC-I (apo B mRNA-editing catalytic subunit-1) 244. Other apos are produced in the enterocyte, mainly apo A-I and apo A-IV that is exclusively of intestinal origin in humans 245. During its synthesis in the ER, apo B must undergo lipidation that protects it from degradation by the proteasome 246. Lipid transfer from the endoplasmic reticulum to apo B requires microsomal triglyceride transfer protein (MTP) intervention. This step is crucial for chylomicron assembly as noted in abetalipoproteinemia, an illness brought on by the mutation of the MTP gene, where there is defective lipoprotein secretion 247.

Pre-chylomicrons are exported to the Golgi apparatus where they undergo their final maturation (glycosylation of the apos, modification of certain PLs, etc.) before being secreted through the basolateral membrane into the lymphatic milieu. It is important to note the role of cargo proteins in the secretion step, which has been highlighted in chylomicron retention disease: the mutation of the Sar-1 GTPase protein prevents the secretion of chylomicrons through the dependent COPII vesicles 248. TG transfer from the ER to the Golgi apparatus seems to be a limiting step in fat absorption 249. After being secreted into the lymphatic capillaries, intestinal LPs are discharged into the systemic blood circulation through the thoracic canal.

In CF patients, lipid composition, concentration and size are irregular 250. At present, the role played by CFTR anomalies is unknown. To our knowledge, no study has focused on the possible interactions of the CFTR with the implicated enzymes in lipid reesterification, apo B biosynthesis, MTP activity or the relationship with Sar-1 GTPase protein expression.

It is during this LP secretion step that interactions with the CFTR may be the most likely. In fact, CFTR mutations alter the secretion not only of electrolytes, but also of substances as different as BAs by the hepatocyte 172, γ-light chain antibodies by lymphocytes 251, INFγ by monocytes 252 or neurotransmitters by pulmonary neuroendocrine cells 253. Since the enterocyte represents a key cell in the physiopathology of CF and given the numerous lipid aberrations observed in this disease, it appears essential to study the secretion abilities of lipids by epithelial cells instead of focusing only on the digestive mechanism in CF as is presently the case. For example, the recent analysis of RNA, influenced by CFTR knock-out, identified a number of proteins involved in LP metabolism, including proteasome 26S 224. The latter subunit is the major proteolytic component of the ubiquitin-dependent proteasome. As mentioned before, the ER-localized ubiquitin-proteasome pathway is primarily involved in the intracellular degradation of apo B. Its alteration in the absence of CFTR 224 indicates potential relationships between CFTR and the apo B recovery/degradation pathway, thereby determining LP assembly and secretion. Furthermore, there is increasing evidence that CFTR regulates endosomal fusion and vesicular trafficking 254, indicating potential relationships between the CFTR and ADP-ribosylation factor 254, which has a central role in VLDL assembly 255. Finally, the Sar-1/COPII complex is necessary when apo B-100 exits the hepatic ER 256 and apo B-48 containing chylomicrons are exported from the enterocyte 248. However, the Sar-1/COPII complex is also implicated in the exiting of CFTR from the ER 257 for its entry into the proteasome degradation pathway 258. Overall, these observations demonstrate the existence of possible interactions among apo B elaboration, chylomicron packaging and CFTR function. Hence, CFTR mutations may significantly affect lipid transport and obviously studies are needed to clarify this relationship.

It is important to underline that the anomalies in the reported LP during CF are most marked in patients with EFAD 250. These defects may either be the direct consequence of EFAD on LP synthesis or may only reflect an association between the type of CFTR mutation and lipid metabolism. A phenotypic classification of LP profiles according to genotype would allow us to elucidate this question. In fact, EFAD could cause interference at several levels within the process of LP synthesis and secretion.

TG reesterification rates are diminished in rats during EFAD 259260. Lipid membrane modifications, notably long-chain FAs, are known to change the functioning of reesterification key enzymes such as MGAT 261. Similarly, intake rich in (n-3) FA lowers the ratio (n-6)/(n-3) and leads to an activation of DGAT and ACAT 262. Therefore, an unbalancing of the EFA status could affect the lipid reesterification step.

Biosynthesis through the enterocytes of several apos, such as apo B or apo A-IV, is regulated in a specific way by certain EFAs 263264265. Experimental studies confirm investigations carried out on humans 266. Thus, EFAD is possibly deleterious for the synthesis of the main intestinal LP. Accordingly, enterocyte secretion of synthesized lipids is impaired during EFAD in mice and rats, which translates into an accumulation of large lipid droplets in the intercellular space 196. Notably, in EFAD, the balance between the various EFAs rather than directly the absence of EFAs could impair the exocytosis mechanism. In effect, in enterocyte cultures, certain EFAs, such as EPA, decrease TG esterification and PL transfer from the ER to the Golgi and the mechanisms responsible for these processes have not yet been specified [267]. Similarly, EFAD may affect microtubules [242-268], as we previously mentioned, which could impact on the assembly and secretion of chylomicrons 239240242.