Paul Hadvary

Hydrolysis of dietary fat by pancreatic lipase stimulates cholecystokinin release

BACKGROUND & AIMS The hypothesis that cholecystokinin release requires adequate dietary fat digestion in the small intestine was investigated in 10 healthy volunteers, and the consequences of reduced fat hydrolysis on pancreaticobiliary secretions were assessed. METHODS Fat hydrolysis was inhibited by intraduodenal perfusion of tetrahydrolipstatin, an irreversible lipase inhibitor. An oil emulsion containing 0, 30, 60, or 120 mg tetrahydrolipstatin was perfused. After a 40-minute basal period, a test meal was eaten to stimulate cholecystokinin release and pancreaticobiliary responses. RESULTS In the control without tetrahydrolipstatin, lipase output increased threefold with meal ingestion and remained doubled for 4 hours. At the ligament of Treitz, free fatty acid concentration averaged 60% of total fatty acids. Increasing doses of tetrahydrolipstatin induced a dose-dependent inhibition of duodenal lipase activity (P < 0.01); 120 mg tetrahydrolipstatin eliminated the postprandial lipase peak activity, free fatty acid levels decreased to < 5% of total fatty acids, and plasma cholecystokinin levels were suppressed by 77% (P < 0.01). Amylase and trypsin outputs were reduced by 77% and 59%, respectively, and bilirubin secretion was virtually abolished (P < 0.01). CONCLUSIONS These findings show that tetrahydrolipstatin prevents triglyceride hydrolysis and that plasma cholecystokinin release, gallbladder emptying, and pancreatic enzyme secretion require adequate triglyceride digestion. These data also support the concept of negative feedback regulation of cholecystokinin secretion.

Identification of the active‐site serine in human pancreatic lipase by chemical modification with tetrahydrolipstatin

A chemical modification approach was used in this study to identify the active site serine residue of human pancreatic lipase. Purified human pancreatic lipase was covalently modified by incubation with [3H], [14C]tetrahydrolipstatin (THL), a potent inhibitor of pancreatic lipase. The radiolabeled lipase was digested with thermolysin, and the peptides were separated by HPLC. A single THL‐peptide‐adduct was obtained which was identical to that obtained earlier from porcine pancreatic lipase. This pentapeptide with the sequence VIGHS is covalently bound to a THL molecule via the side chain hydroxyl group of the serine unit corresponding to Ser‐152 of the lipase. The selective cleavage of the THL‐serine bond by mild acid treatment resulted in the formation of the δ‐lactone Ro 40–444] in high yield and clearly proves that THL is attached via an ester bond and with retention of stereochemistry at all chiral centers to the side chain hydroxyl group of Ser‐152 of the lipase. The results obtained for human pancreatic lipase corroborate the inhibition mechanism of THL found on the porcine enzyme, and are in full agreement with the identification of the Ser‐152…His‐263…Asp‐176 catalytic triad in the X‐ray structure of human pancreatic lipase.

[10] Covalent inactivation of lipases

Publisher Summary This chapter presents and discusses results chiefly concerning the covalent inhibition of gastric and pancreatic lipases. Lipolysis is catalyzed by preduodenal and pancreatic lipases. In humans, the hydrolysis of alimentary triacylglycerols begins in the stomach and is catalyzed by human gastric lipase (HGL) that is able to hydrolyze short- and long-chain triacylglycerols at comparable rates. Under acidic pH conditions, HGL has been shown to be remarkably stable and active, whereas pancreatic lipase irreversibly loses its lipolytic capacity. Achieving specific and covalent inhibition of lipolytic enzymes is a difficult task, because of nonmutually exclusive processes such as interfacial denaturation, changes in “interfacial quality,” and surface dilution phenomena. Furthermore, the interfacial enzyme binding and/or the catalytic turnover can be diversely affected by the presence of potential amphipathic inhibitors. The chapter provides selected experimental data illustrating the specific problems encountered during the covalent inhibition of digestive lipases.

Surface behaviour of bile salts and tetrahydrolipstatin at air/water and oil/water interfaces

The surface behaviour of two bile salts, sodium deoxycholate (NaDC) and sodium taurodeoxycholate (NaTDC), as well as that of tetrahydrolipstatin (THL), a potent gastrointestinal lipase inhibitor, was studied at air/water and oil/water interfaces, using interfacial tensiometry methods. The surface behaviour of NaDC and NaTDC was comparable at both oil/water and air/water interfaces. A fairly compact interfacial monolayer of bile salts is formed well below the critical micellar concentration (CMC) and can help to explain the well-known effects of bile salts on the kinetic behaviour of pancreatic lipases. Using the Wilhelmy plate technique, the surface pressure-molecular area curves recorded with THL at the air/water interface showed a collapse point at a surface pressure of 24.5 mN.m(-1), corresponding to a molecular area of 70 A(2). Surprisingly, using the oil drop method, a limiting molecular area of 160 A(2) was found to exist at the oil/water interface. On the basis of the above data, space-filling models were proposed for bile salts and THL at air/water and oil/water interfaces.

Activation of human and rabbit blood platelets by synthetic structural analogs of platelet activating factor

The proaggregatory potencies of synthetic analogs of platelet activating factor (PAF-acether: 1-0-alkyl 2-0-acetyl sn-glycero-3-phosphorylcholine) were tested in rabbit and human platelet rich plasma (PRP). For activation of human compared to rabbit platelets about 30 to 100 times higher concentrations of the compounds were needed, but the relative potencies of the analogs tested were quite similar in both species. The absolute stereospecificity required for platelet activation could be confirmed and extended also to the pair of enantiomers with a cis-9 double bond in the 0-alkyl chain and to the enantiomers with the 0-alkyl chain at C-2 of glycerol. The propionyl ester at C-2 of glycerol proved to be as potent as the acetyl ester but replacement by the formyl as well as by the butyryl ester reduced the activity to less than three per cent. Variations at the 3-phosphorylcholine substituent all abolished the activity. It appears that the correct stereochemistry and a phosphorylcholine group in a strictly defined position towards the asymmetric C-2 of the glycerol backbone are absolute requirements for activity whereas variations in the alkyl and the short chain acyl groups have not as dramatic consequences.

Role of Free Fatty Acids in Regulating Gastric Emptying and Gallbladder Contraction

The limited effectiveness of orlistat, an inhibitor of gastrointestinal lipases, in inhibiting fat digestion is not completely understood. Therefore we studied the effect of orally and duodenally administered orlistat on gastric emptying, cholecystokinin (CCK) secretion, and gallbladder contraction. In healthy males, gastric emptying of solids and fat were quantified scintigraphically, gallbladder contraction by ultrasound and CCK release by radioimmunoassay. Three studies were performed: (1) oral and (2) duodenal orlistat with a fat-containing meal, and (3) duodenal orlistat with a fat-free meal. Gastric emptying rates of solids and fat (T50% accelerated by 16 and by 22%, p< 0.05, respectively) were significantly faster after duodenal perfusion of orlistat; gallbladder contraction and CCK release were reduced under these conditions (p < 0.005, respectively). With oral orlistat no significant effect was documented on these parameters. We conclude that fat hydrolysis is essential in the regulation of fat-induced gastric emptying and gallbladder contraction.

Inhibitor Binding to Thrombin: X-Ray Crystallographic Studies

It is a great pleasure to see all the excellent structures presented by Wolfram Bode. I am grateful to him for agreeing to leave me something new to present, namely our work on the binding of the Mitsubishi compound MD-805 (Argatroban) to human thrombin. Before I proceed to the details of that, I would first like to give a short overview of our work at Roche, Basle in this field. Our interest in thrombin inhibitors as clinical candidates led us to purify human thrombin and set about crystallization trials. As others had had poor success with uninhibited thrombin, (presumably because of autolysis), the first efforts were concentrated on the binary complex of thrombin with hirudin extracted from whole leeches. When I joined the project five years ago Fritz Winkler and Allan D’Arcy had found a hexagonal crystal form HEX1 (see Table 1) which, however, only diffracted to about 3.5A. We managed to grow some large crystals of this complex but, for some unknown reason, could never collect good native data and were also unable to find heavy atom derivatives. We looked for alternatives, and reading the work of Stuart Stone and Jan Hofsteenge decided to try the ternary complex of PPACK-thrombin — hirudin, hoping (Figure 1) to gain information about binding in the P1′, P2′, etc. sites as well as the active site and the anion binding exosite. We found after some time an orthorhombic crystal form (OR1, Table 1) which was difficult to reproduce and optimized instead a second orthorhombic form, OR2, which was used for heavy atom trials with no success. We later found one very large crystal of form OR2 and collected from it data to 2.35A. As a third attempt we decided to use, instead of hirudin, just a hirudin C terminal peptide, hoping to prevent autolysis and to give us different crystals. This worked well and we solved the structure of the human thrombin-PPACK- hirudin peptide complex in the tetragonal Tl form using the heavy-atom method. The two hirudin containing forms were then solved by molecular replacement (with the assistance of Wolfgang Janes and Christian Oefner, whom we thank). To our disappointment we have not been able to interpret the weak electron density for the hirudin N-terminal domain in either crystal form, and have concluded that in the presence of PPACK this domain is bound unspecifically. The well ordered hirudin density corresponds to residues 55–65, i.e. the same peptide used to obtain the Ti form, and both structures turn out to be similar. There is, however, a difference in that the natural hirudin is sulphated on Tyr 63. The S03-group makes two very good hydrogen bonds to thrombin, consistent with the tighter binding of sulphated peptides. The hirudin complexes both present Trp148 towards the active site, although the two conformations of the loops (Lys145 — Lys149E, chymotrypsin numbering) are not strictly identical. In contrast, this loop swings away from the active site in the Tl crystal form and is stabilised by a crystal contact. Otherwise the three PPACK ternary complexes are remarkably similar, and so the refined 2.35A hirudin complex was most useful in helping to give a good model for complexes in the Tl crystal form, where it is not possible to measure data beyond about 3A resolution. As far as we can tell, our structures, determined independently, are entirely consistent with those described at this meeting and the preceding Amsterdam meeting by Wolfram Bode and Al Tulinsky.

Interference of the Paf antagonist Ro 19–3704 with Paf and antigen‐induced bronchoconstriction in the guinea‐pig

1 In vitro, Ro 19–3704, a structurally related antagonist of platelet‐activating factor (Paf) inhibited selectively rabbit platelet aggregation. In vivo, administered intravenously, it inhibited bronchoconstriction, leukopenia, thrombocytopenia and the accompanying accumulation of platelet aggregates in guinea‐pig lung microvessels induced by i.v. Paf. Administered by aerosol, Ro 19–3704 failed to inhibit bronchoconstriction, thrombocytopenia or leukopenia due to i.v. Paf. 2 Bronchoconstriction induced by Paf, in aerosol form, was blocked by Ro 19–3704 administered by the i.v. or aerosol route, which suggests that it interacts with pulmonary cells responsible for bronchoconstriction. 3 Ro 19–3704 has free radical scavenging properties, since it inhibited the production of superoxide anions by macrophages stimulated by Paf and by N‐formyl‐methionyl‐leucyl‐phenylalanine (FMLP). Ro 18–7715, another Paf antagonist and analogue of Ro 19–3704, failed to inhibit the production of superoxide anions by macrophages stimulated by FMLP at concentrations which were effective against Paf. 4 Administered intravenously, Ro 19–3704 failed to block bronchoconstriction induced by an i.v. injection of ovalbumin to guinea‐pigs passively sensitized with anti‐ovalbumin antiserum. Passive pulmonary anaphylaxis due to an aerosol of ovalbumin was blocked by i.v. Ro 19–3704.

Transfer of orlistat through oil–water interfaces

The transfer of radiolabelled orlistat ([14C]orlistat), a potent gastrointestinal lipase inhibitor, through an oil-water interface from a single oil droplet to an aqueous phase was investigated, using an oil drop tensiometer. The absolute transfer fluxes were found to be very low, even in the presence of micellar concentrations of bile salts, which increased their values from 0.2 to 2.5 and 6.5 pmol cm(-2) min(-1) in the presence of 0, 4 and 15 mM NaTDC, respectively. Adding either a lipid emulsion or pure human pancreatic lipase (HPL) or human serum albumin or beta-lactoglobulin had no effect on the flux of transfer of orlistat. The presence of colipase or a mixture of colipase and HPL was found, however, to reduce the flux of orlistat transfer, probably because it partly covered the single oil drop surface, even in the presence of bile salts. Using a finely emulsified system, we investigated the partitioning of orlistat between the aqueous and oil phases, in the absence or presence of bile salts above their CMC (4 mM NaTDC, final concentration). Under these emulsified conditions, orlistat was found to be mostly associated with the oil phase, since more than 98.8% of the total radioactivity was recovered after decantation with the oil phase. The low transfer rates of orlistat, as well as its partitioning coefficient between the oil and the aqueous phases, should help us to better understand the inhibitory effects of orlistat on lipid digestion in humans.

Tetrafibricin: A nonpeptidic fibrinogen receptor inhibitor from Streptomyces neyagawaensis (I) its GPIIB/IIIA blockage on solid phase binding assay

Tetrafibricin is a novel nonpeptidic fibrinogen receptor inhibitor isolated from Streptomyces neyagawaensis NR0577. Its competitive and selective fibrinogen receptor blockage was demonstrated in this study. Tetrafibricin competitively inhibited (Ki = 9.9 nM) the binding of biotinylated fibrinogen to purified active glycoprotein (GP) IIb/IIIa immobilized on plastic plate. When RGDS and tetrafibricin were added in combination, the inhibition was additive. The binding of other RGD-containing proteins, fibronectin and von Willebrand factor, to active GPIIb/IIIa were also completely inhibited by tetrafibricin. The fact that tetrafibricin did not inhibit the binding of von Willebrand factor to GPIb/IX indicates the specific blockage of tetrafibricin for GPIIb/IIIa. Fibrinogen receptor inhibition of tetrafibricin was also confirmed by its ability to inhibit 125I-fibrinogen binding to platelets stimulated with ADP. Because of its competitiveness and specificity, tetrafibricin can be used in a new structural model for the design of fibrinogen receptor inhibitors.