Cholesterol secretion in bile

The elucidation of the exact events leading to lipid secretion in bile is still being investigated. In an elegant ultrastructural study, Crawford et al 17 have proved that biliary phospholipid molecules are secreted by hepatocytes into the bile canalicular lumina as unilamellar vesicles 63–67 nm in diameter. This process occurs rapidly and seems to be facilitated by the detergent action of luminal bile salts at the level of the exoplasmic part of the canalicular membrane. Cholesterol is also secreted into hepatic bile and is taken up by biliary lipid vesicles in which the cholesterol:phospholipid ratio is normally 0.34–0.38. In lithogenic bile, however, this ratio is larger.18-21 As mixed micelle formation requires the solubilisation of more phospholipid than cholesterol, the cholesterol:phospholipid ratio in the remaining vesicles progressively increases to >1 (about 2:1).22These cholesterol-rich vesicles tend to aggregate or to fuse into multilamellar vesicles of increasing size. As emphasised by Cohenet al,23 cholesterol solubilisation and transport in bile is explained by the classic paradigm of the simple micelle-mixed micelle-unilamellar vesicle, with multilamellar unstable vesicles heralding the initiation of the lithogenic process. Classic triclinic cholesterol crystals may precipitate from multilamellar vesicles, as observed by video enhanced time lapse microscopy.24 The process of mixed micelle formation starts in the biliary tree but is strongly promoted in the gall bladder11 as the gall bladder absorbs electrolytes actively and water passively. The concentration of biliary lipids then increases and high bile salt concentrations promote mixed micelle formation with growth of cholesterol-rich vesicles. Thus, vesicular cholesterol is important for cholesterol crystal formation.25 Recent studies in cholesterol fed prairie dogs, however, have suggested that supersaturated mixed micelles may also have a role in cholesterol crystal precipitation.26

Biliary proteins and cholesterol crystallisation

Bile from patients with cholesterol stones cannot be discriminated from those of control patients simply on the basis of lipid composition.2 ,27 ,28 Thus, other factors, such as biliary proteins, have been suggested to play a part in the pathogenesis of cholesterol crystallisation and stone formation.29Although several proteins with cholesterol crystallisation inhibiting as well as promoting properties have been isolated, the active factor has not been identified as yet. Apolipoprotein A-1 and A-230 and a 120 kDa glycoprotein isolated from normal human gall bladder bile31 inhibit cholesterol crystallisation.

Mucin and non-mucin glycoproteins might have a role as promoting agents of cholesterol crystallisation. In fact, gall bladder mucin promotes crystallisation of cholesterol18 ,32 ,33 and mucin hypersecretion precedes cholesterol crystal precipitation and stone formation both in animals34 ,35 and humans (such as in obese people with weight reduction and gallstone formation).36-38 Mucin provides a microenvironment and might act as a nidus for cholesterol crystal aggregation. In this way sludge is formed, a viscous gel of mucin and entrapped cholesterol crystals or calcium bilirubinate precipitates.39 ,40Sludge may disappear spontaneously41 but progressive cholesterol crystal precipitation and subsequent aggregation into macroscopic gallstones may also occur.33 Groenet al 12 isolated from T-tube bile of patients with cholesterol gallstones a concanavalin-A binding non-mucin glycoprotein (molecular weight 130 kDa) that stimulates cholesterol crystallisation in model bile. As it is found in hepatic bile, this factor is likely to be of hepatic origin. Other non-mucin factors have crystallisation promoting effects in vitro: a pronase resistant low density lipoprotein particle (±90% lipid),42 aminopeptidase N (a 130 kDa biliary canalicular glycoprotein),43 ,44 phospholipase C,45immunoglobulins of the IgM, IgA and IgG classes,46-48fibronectin,49 a biliary anionic polypeptide fraction (APF),50 acute phase proteins, such as a 42 kDa α₁-acid glycoprotein,51-54α₁-antichymotrypsin,⁵⁵ and haptoglobin.⁵⁶ Although the liver secretes promoting and inhibiting factors simultaneously, the ultimate pathophysiological importance of these proteins in human bile remains controversial.51 There is also evidence for the existence of pronase resistant pronucleating proteins as protein hydrolysis does not influence rapid42 ,57 crystallisation in human bile.55 ,58

Cholesterol crystallisation in bile

Precipitation of crystalline cholesterol in bile is likely to occur when the cholesterol concentration exeeds the micellar solubility limit and fusion or aggregation of cholesterol-rich vesicles occurs.24 ,59 The “classic” crystallisation pathway was believed to result in the appearance of typical monohydrate cholesterol plates in both human2 ,60 and model bile.60 ,61 Although early studies showed that hypovitaminotic C guinea-pigs fed a high cholesterol diet developed randomly arranged, needle shaped (possibly anhydrous cholesterol) crystals in the gall bladder62 and that needle shaped crystals existed both in model and human bile,63-65 the possibility that crystal forms other than plate-like structures could develop in bile was only recently tackled. In an important study Konikoff et al 7 described filamentous crystals that appeared within two to four hours of dilution of model bile and within one to three days of ex vivo incubation of human gall bladder bile (preliminary report: in six of eight patients with cholesterol gallstones but in none of those with pigmented stones). Over a few days filaments evolved through needle, helical, and tubular structures to end as typical plate-like cholesterol monohydrate crystals. These findings were confirmed in a subsequent study from a taurocholate-rich dilute bile salt-rich model bile.6 In this system cholesterol low density (1.030 g/ml) filamentous crystals reached a maximal concentration at 24 hours and disappeared gradually after 156 hours of incubation, whereas high density (1.045 g/ml) plate-like crystals increased reciprocally. The authors used several analytical procedures which pointed to the mainly anhydrous nature of early cholesterol filaments,6 with subsequent hydration as water becomes incorporated within the crystals. In an initial report66 two non-perturbing techniques (that is, cryotransmission electron microscopy and video enhanced light microscopy) were used for direct imaging of cholesterol crystallisation. This study confirmed the dynamic evolution of lipid microstructure which is similar in model and human bile. According to a recent hypothesis,5 the two putative pathways of cholesterol nucleation in bile might involve (i) precipitation of classic plate-like cholesterol crystals from multilamellar vesicles24 and (ii) nucleation of anhydrous cholesterol occurring internally as a liquid core within unilamellar vesicles (that is, “internal nucleation”).5 ,57

Studies on the different habits of cholesterol crystals have been extended in model bile.4 ,67 Our group has studied this in model68 ,69 and human bile.8 ,70 ,71 We believe that the complete process of cholesterol crystallisation in bile is still poorly understood and that the pathophysiological importance of the pathway through cholesterol microcrystalline intermediates needs to be carefully investigated. In this respect, a major contribution comes from the recent work of Wang and Carey4 who used complementary physical–chemical techniques, and have now carefully defined five different crystallisation pathways which are based on the main forms of liquid and solid cholesterol crystals in pathophysiologically relevant model bile. As shown by the analysis of crystallisation sequences as functions of time and increasing lecithin content within the phase diagram for the mixed bile salt-egg yolk lecithin-cholesterol system, at equilibrium five regions were identified above the one-phase micellar zone and named A to E (moving from left to right across the phase diagram). Crystal observation times for plate-like crystals were progressively retarded between pathwaysA and D, whereas plates never appeared in pathway E. Studies from the same group suggest that the events in model biles correlate with crystallisation pathways in humans72 and in inbred mice on a lithogenic diet.35 Accordingly, it is anticipated that factors in bile such as total lipid concentration, phospholipids and bile salt hydrophobicity can modulate crystal growth by shifting the relative composition of bile to different regions of the phase diagram, as discussed in the following paragraphs.

The finding that cholesterol crystallisation is longer in dilute supersaturated gall bladder11 and model73bile would be explained by a shift delution of the crystallisation pathway towards the E region (liquid crystals) of the phase diagram. The opposite would apply to concentrated bile.4

Phospholipid molecular species and class can influence cholesterol crystallisation in model and human bile.74-76 Saturated lecithins delay whereas polyunsaturated phospolipids accelerate the crystal observation time in model bile.7 ,74 Although the surface layer of phosphatidylcholine, which covers the surface of cholesterol filamentous crystals, might interfere directly with crystal growth,7 a valid alternative explanation is that modulation of phospholipid species can shift phase boundaries towards pathways where the plate-like crystal observation time is either shorter (a leftward shift) or longer (a rightward shift) within the phase diagram.4 Possible strategies for cholesterol crystallisation and gallstone prevention might theoretically imply increased biliary secretion of phospholipids75 as dietary factors can influence phospholipid content.76 Further studies are warranted on this topic.

Several studies have focused on the relation between cholesterol crystal shape in model bile and the hydrophobic–hydrophilic balance of bile salts, and have concluded that crystallisation of several forms is increased in the presence of more hydrophobic bile salts.4 ,21 ,67-69 At least seven different crystal forms have been identified, including helical, thread-like and ribbon-like structures, the growth of which decreased as bile salt hydrophobicity decreased (that is, deoxycholate > chenodeoxycholate > cholate > taurocholate > ursodeoxycholic acid (UDCA)).^{68 77} Van Erpecum et al 69 found that the most striking crystallisation of all cholesterol crystal forms, including large plate aggregates, was seen in the very hydrophobic system in taurodeoxycholate-rich model bile and that bile salt hydrophobicity influenced both crystal shape and extent of crystallisation. Similar studies with an accurate description of cholesterol crystal shapes have been carried out by Juste et al 67 and by Wang and Carey4 in supersaturated model biles with a number of bile salts at different hydrophobicities. It is possible that hydrophobic bile salts which exhibit lower critical micellar concentrations induce rapid formation of mixed micelles with increased formation of cholesterol-rich vesicles and rapid crystal precipitation.21 A better explanation is provided by the study of the equilibrium phase diagram4: crystallisation pathways shift with decreasing bile salt hydrophobicity and the right two-phase area (that is, region E, where only liquid cholesterol crystals occur) becomes progressively larger. This mechanism would explain why crystallisation of all forms is greatly inhibited or prevented by very hydrophilic bile salt species such as UDCA, tauroursodeoxycholate, taurohyodeoxycholate, etc.4 ,67-69

Recently, we studied cholesterol crystal forms in both ultrafiltered (isotropic) bile and native bile of patients with cholesterol or pigment gallstones.8 Daily polarising microscopy was used to quantify crystal observation time and to construct crystal-time growing curves of the different forms of solid crystals such as arcs, needles, spirals, tubules, plates, and their aggregates over three weeks. Figure 1 shows most of these crystal forms. Pigment stone biles never crystallised; cholesterol gallstone bile, in contrast, had a rich assortment of crystals forms. Apparently, arcs and needles act as substrates for the development of progressively thicker forms giving rise to young and mature (stable) plates. This evolution suggests that slow hydration of filamentous anhydrous crystals occurs with time.6 It has been suggested that arcs/needles are transient forms.6 ,7 Indeed, arcs and needles are seen infrequently in freshly sampled bile of patients with cholesterol stones.8 However, we also found that in vitro these filamentous forms of cholesterol increased progressively to a maximum and persisted throughout the observation time.8 Whether this behaviour is relevant in vivo during gallstone formation is still unclear. Both tubules and spirals were transient forms usually seen within the first week of observation and with a tendency to decrease and/or to convert into mature plates. The presence of tubules or spirals was associated with a high concentration of the hydrophobic deoxycholate and rapid growth of other crystal forms. As anticipated by a number of studies,10-12 all crystal forms had a high rate/extent of crystallisation in patients with multiple cholesterol stones, compared with those with solitary stones. Overall, the two most important factors independently associated with fast crystallisation of plates were the number of gallstones and the presence of tubules/spirals in ultrafiltered bile.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

: Light microscopy of ultrafiltered bile from a patient with cholesterol gallstones showing classic plate-like monohydrate cholesterol crystals (P) and other cholesterol crystal shapes such as needles (N), arcs (A), spirals (S), and tubules (T). Original magnification ×100.

We found that treatment with UDCA (which results in bile enriched with glycine and taurine conjugates of UDCA) led to some interesting effects on crystallisation of solid cholesterol. UDCA, in fact, totally prevented the precipitation of all crystal forms in isotropic bile of patients with gallstones treated with UDCA8 and was associated—in native bile—with the appearance of rather thickened atypical cholesterol crystals with multilobate edges which, in some cases, tended to decrease in size and even disappeared after prolonged observation.8 ,71 This has been described previously in in vitro dissolution studies in UDCA containing systems.78 As anticipated4 from in vitro studies the effect of UDCA on crystallisation might be due to the expansion of the liquid crystal region of the equilibrium phase diagram by reducing bile salt hydrophobicity. We also found recently71 that treatment with UDCA leads to a decrease in the concentrations of various crystallisation promoting proteins (IgG, IgA, haptoglobin, and α₁-acid glycoprotein, aminopeptidase N, and total proteins) and reduces the crystallisation promoting activity of concanavalin A eluate in patients with cholesterol gallstones. This was confirmed by two independent techniques including a nephelometric assay79and microscopy.8

Conclusions

Rapid crystallisation of cholesterol in bile is as a key factor in cholesterol gallstone formation. The original observation that solid cholesterol precipitates only as characteristic plates of monohydrate cholesterol has been recently revised as filamentous, helical and tubular forms of non-hydrated cholesterol can grow together with plates and aggregates of plates in both isotropic and native bile. Several factors such as biliary proteins and lipids might act as potential cofactors of cholesterol crystallisation. The most active crystallisation of cholesterol in bile is found in patients with multiple cholesterol gallstones. Gall bladder bile enriched with the more hydrophilic bile salts, such as glycine and taurine conjugates of UDCA, inhibits cholesterol crystallisation and decreases the concentrations of crystallisation promoting biliary proteins. Exciting studies are awaited to clarify whether prevention of cholesterol gallstone formation in at-risk subjects can be effectively achieved by manipulating biliary factors which have pronucleating or antinucleating effects.