Design optimization and characterization of Her2/neu-targeted immunotoxins: comparative in vitro and in vivo efficacy studies.

Targeted therapeutics are potential therapeutic agents because of their selectivity and efficacy against tumors resistant to conventional therapy. The goal of this study was to determine the comparative activity of monovalent, engineered anti-Her2/neu immunotoxins fused to recombinant gelonin (rGel) to the activity of bivalent IgG-containing immunoconjugates. Utilizing Herceptin and its derived humanized single-chain antibody (single-chain fragment variable, designated 4D5), we generated bivalent chemical Herceptin/rGel conjugate, and the corresponding monovalent recombinant immunotoxins in two orientations, 4D5/rGel and rGel/4D5. All the constructs showed similar affinity to Her2/neu-overexpressing cancer cells, but significantly different antitumor activities. The rGel/4D5 orientation construct and Herceptin/rGel conjugate were superior to 4D5/rGel construct in in vitro and in vivo efficacy. The enhanced activity was attributed to improved intracellular toxin uptake into target cells and efficient downregulation of Her2/neu-related signaling pathways. The Her2/neu-targeted immunotoxins effectively targeted cells with Her2/neu expression level >1.5 × 10(5) sites per cell. Cells resistant to Herceptin or chemotherapeutic agents were not cross-resistant to rGel-based immunotoxins. Against SK-OV-3 tumor xenografts, the rGel/4D5 construct with excellent tumor penetration showed impressive tumor inhibition. Although Herceptin/rGel conjugate demonstrated comparatively longer serum half-life, the in vivo efficacy of the conjugate was similar to the rGel/4D5 fusion. These comparative studies demonstrate that the monovalent, engineered rGel/4D5 construct displayed comparable in vitro and in vivo antitumor efficacy as bivalent Herceptin/rGel conjugate. Immunotoxin orientation can significantly impact the overall functionality and performance of these agents. The recombinant rGel/4D5 construct with excellent tumor penetration and rapid blood clearance may reduce the unwanted toxicity when administrating to patients, and warrants consideration for further clinical evaluation.

Estrogen replacement therapy and gallbladder disease in postmenopausal women.

OBJECTIVE: To examine the role of estrogen replacement therapy on the development of gallbladder disease in postmenopausal women. DESIGN: Systematic review of the English literature was conducted. All studies that addressed the association between hormone replacement therapy and gallbladder disease published from 1970 to the present were reviewed. RESULTS: Seven observational studies, two clinical trials, two case series, and one nonrandomized and three randomized investigations were reviewed. The results of each study were reported and analyzed. CONCLUSIONS: Estrogen replacement therapy in postmenopausal women increased the chances for gallstone formation.

Comparative cytotoxicity and pharmacokinetics of antimelanoma immunotoxins containing either natural or recombinant gelonin.

Immunotoxins are a class of targeted therapeutic agents under development by various research groups. The murine monoclonal antibody designated ZME-018 recognizes a high molecular weight glycoprotein present on most human melanoma cells and biopsy specimens and has been utilized for clinical imaging studies in patients with melanoma. The plant toxin gelonin is a ribosome-inactivating protein (RIP) with n-glycosidase activity similar to that of ricin A chain. In previous studies by our group, the gelonin toxin was sequenced, cloned and expressed in E. coli. The purified recombinant gelonin (RG) was found to have identical protein synthesis inhibitory activity to that of natural gelonin (NG). For comparative purposes, chemical conjugates of antibody ZME and either RG or NG were produced using the heterobifunctional crosslinking reagents SPDP and SMPT. The ZME-NG and ZME-RG immunotoxins were found to be 10(4)- to 10(5)-fold more cytotoxic to antigen-positive human melanoma cells than free toxin. NG toxin alone was cytotoxic to intact cells (IC(50) = 100 nM) while RG was nontoxic to cells at doses up to 1 microM. Both ZME-NG and ZME-RG immunoconjugates were nontoxic to antigen-negative (Me-180) cells. ZME-RG immunotoxins constructed with the more stable SMPT reagent were slightly more effective in culture than conjugates made with SPDP. Tissue distribution studies in tumor-bearing nude mice demonstrated that tumor uptake of the ZME-RG immunotoxin was similar to that of the intact ZME antibody with reduced distribution to normal organs compared to an immunoconjugate produced with NG. Pharmacokinetic studies showed that the terminal-phase plasma half-life of ZME-RG was similar to that of ZME itself (42 h vs 50 h) and almost threefold higher than that of ZME-NG (11.5 h). The area under the concentration curve (Cxt) for ZME-RG was 50% lower than that for ZME due to an increased apparent volume of distribution (Vd(a)) but was almost tenfold higher than the Cxt for ZME-NG. These studies suggest that immunoconjugates comprising RG demonstrate identical in vitro cytotoxic effects to immunoconjugates produced with NG and immunotoxins with RG display improved in vivo pharmacodynamics and tissue distribution compared to immunotoxins containing NG.

Recombinant immunotoxins directed against the c-erb-2/HER2/neu oncogene product: in vitro cytotoxicity, pharmacokinetics, and in vivo efficacy studies in xenograft models.

TAB-250 and BACH-250 are murine and human chimeric antibodies directed at the extracellular domain of the gp185c-erb-2 (HER2/neu) growth factor receptor overexpressed in a variety of tumor types, including ovarian and breast carcinoma. The ribosome-inhibiting plant toxin gelonin (rGel) was chemically coupled to both antibodies, and the resulting immunotoxins were purified and tested in vitro against human tumor cells expressing various levels of HER-2/neu and in vivo against human tumor xenograft models. The binding of both BACH-250 and BACH-250/rGel conjugate to target cells was essentially equivalent. Against SKOV-3 cells, the IC50 of BACH-250/rGel was 97 pM (17 ng/ml), whereas BACH-250 and rGel alone showed no cytotoxic effects. There was a clear correlation between expression levels of HER-2/neu and cytoimmunotoxin. Tissue distribution studies showed that the antibody and immunotoxin both concentrate 2-10-fold higher in tumors than in normal tissues, with optimal tumor uptake occurring 48-96 h after administration. Plasma clearance curves for BACH-250 and BACH-250/rGel showed terminal-phase half-lives of 26 and 72 h, respectively. In athymic mice bearing s.c. or i.p. SKOV-3 tumors, immunotoxin treatment slowed tumor growth by 99 and 94 % at days 35 and 49 after implantation, respectively, and lengthened the median survival by 40% (from 30 to 50 days) in mice bearing lethal i.p. tumors. We conclude that clinical development of BACH-250/rGel may be warranted in patients with HER2/neu-expressing malignancies.

High-dose preoperative radiation and the challenge of sphincter-preservation surgery for cancer of the distal 2 cm of the rectum.

PURPOSE: Sphincter-preserving surgery for the management of distal rectal cancer is gaining recognition as an alternative to abdominoperineal resection and loss of anal function. The use of high-dose preoperative radiation appears to enhance the options for sphincter preservation, even in the most distal segments of the rectum. MATERIALS AND METHODS: Seventy patients with tumors located in the distal 2 cm of the rectum received a minimum dose of 40 to 45 Gy over 4 1/2 weeks at 1.8 to 2.5 Gy per fraction. Patients with unfavorable tumors were given an additional boost of 10 to 15 Gy. Surgery was performed 5 to 10 weeks following completion of radiation. Radical surgical resection was performed in 48 patients and full thickness local excision in 22. Follow-up ranged from a minimum of 1 year to a maximum of 10 years, with a median of 4 years. RESULTS: There was one perioperative mortality. Two patients did not have their colostomy closed because of complications. Late diversion was required in 4 patients, primarily for recurrent disease. Sixty patients (86%) maintained long-term satisfactory sphincter function. Local recurrence was observed in 9 patients (13%) and distant metastases in 12 patients (17%). The overall five-year actuarial survival rate was 82%. The 5-year survival and local recurrence for postradiation pathological stage of disease was: T0, T1, T2, N0--95% and 8%, T3, T4, N0--91% and 4%, T(any) N+--50% and 41%, respectively. CONCLUSION: High-dose preoperative radiation, in properly selected patients with rectal cancers of the distal 2 cm, offers opportunities for sphincter-preserving surgical resection with excellent local control, survival, and enhanced quality of life.

Comparison of the impact of transdermal versus oral estrogens on biliary markers of gallstone formation in postmenopausal women.

This prospective, randomized, double blind, parallel study was undertaken to elucidate further the potential mechanisms through which estrogens could promote the formation of cholesterol gallstones and to compare the impact of nonoral (transdermal) and oral estrogens on serum, hepatic, and biliary markers of estrogen action. Ninety-seven postmenopausal women were randomized to receive either transdermal estradiol (E2; 0.1 mg every 3.5 days; n = 48) or oral conjugated equine estrogens (1.25 mg every day; n = 49) for 8 weeks. Blood samples were drawn, and bile samples were obtained by cholecystokinin-stimulated duodenal drainage before and after 8 weeks of estrogen administration. The main outcome measures included serum FSH, LH, E2, estrone, estrone sulfate, sex hormone-binding globulin, lipid profiles, biliary cholesterol saturation index, cholesterol nucleation time, presence of cholesterol crystals in bile, as well as biliary arachidonate, PGE2, and mucous glycoproteins. Estrogens administered by both routes increased circulating estrogens and resulted in similar suppression of both gonadotropins. Sex hormone-binding globulin was clearly increased, and the changes in serum lipids were more pronounced with oral conjugated equine estrogens than with transdermal E2. The biliary cholesterol saturation index was significantly increased compared to the baseline values with both transdermal E2 (1.08 +/- 0.04 vs. 1.00 +/- 0.03; mean change, 8%) and oral conjugated equine estrogens (1.04 +/- 0.03 vs. 0.99 +/- 0.03; mean change, 6%); however, there was no difference between the treatments. The number of patients with cholesterol crystals detected in bile was similar after both estrogen regimens. Transdermal and oral estrogens decreased nucleation time in vitro, increased arachidonate and PGE2 levels, and minimally raised total glycoprotein concentrations. In conclusion, transdermal and oral estrogens exerted comparable nonhepatic effects, as evidenced by similar reductions of gonadotropin levels, but oral therapy exhibited substantially greater actions on hepatic markers of estrogen action. Both transdermal E2 and oral conjugated equine estrogens significantly elevated the biliary cholesterol saturation index and reduced the nucleation time. These results suggest that estrogens at the doses studied could promote gallstone formation by alteration of biliary lipids and cholesterol nucleation time that have been incriminated in this process.

Nucleation of biliary cholesterol, arachidonate, prostaglandin E2, and glycoproteins in postmenopausal women.

BACKGROUND & AIMS: Among dieting obese patients, cholesterol gallstone formation is preceded by increases in levels of biliary cholesterol saturation, arachidonate, prostaglandin E2, total glycoproteins, and rapid nucleation of cholesterol. The aim of this study was to determine if similar increases occur among postmenopausal women with cholesterol crystals in their bile. METHODS: In 101 postmenopausal women without gallstones, gallbladder bile was sampled via nasoduodenal tube and analyzed. RESULTS: Nineteen of the women had saturated bile and crystals. Levels of cholesterol saturation, arachidonate, prostaglandin E2, and total glycoprotein were highest among women with cholesterol-saturated bile and cholesterol crystals and lowest among women with unsaturated bile. Levels were intermediate among women with saturated bile but no crystals. CONCLUSIONS: Among postmenopausal women, increases in levels of biliary cholesterol saturation, arachidonate, prostaglandin E2, and total glycoproteins may be important pathophysiologically in the rapid nucleation of cholesterol crystals.

Effects of ursodiol or ibuprofen on contraction of gallbladder and bile among obese patients during weight loss.

Rapid loss of weight in obese patients is associated with increased saturation of bile with cholesterol, increased nucleation and growth of cholesterol crystals, and gallstones. The aims of this study were to determine the effects of rapid weight loss on contraction of the gallbladder and to evaluate the effects of ursodiol and ibuprofen on saturation, nucleation and growth, and contraction. Forty-seven obese patients entering a very low calorie dietary program were randomized to receive ursodiol, 1200 mg/day, ibuprofen, 1600 mg/day, or placebo for 12 weeks. Contraction of the gallbladder to a liquid meal was evaluated by ultrasonography, and duodenal bile was collected initially and after six and 12 weeks. Diet caused reduced contraction of the gallbladder, increased cholesterol saturation, and increased nucleation and growth of crystals. Ursodiol reduced saturation and prevented increases in nucleation and growth and contraction. Ibuprofen prevented the increase in saturation and the reduction in contraction with a trend opposing the increase in nucleation and growth. In conclusion, during dieting, contractility of the gallbladder to meals is reduced. The effectiveness of ursodiol in preventing gallstones may be explained partially by effects on contraction. Ibuprofen deserves further study because of its effects on saturation, nucleation and growth, and contraction.

Natural history and treatment with ursodiol of gallstones formed during rapid loss of weight in man.

Cholesterol gallstones form frequently among obese patients during rapid loss of weight. The aims of the present study were to determine the short-term natural history of these gallstones and the efficacy of ursodiol for their dissolution. Twenty-two patients whose gallstones had formed during rapid loss of weight were randomized in double-masked fashion to either ursodiol, 1200 mg/day, or placebo for nine months. Ultrasonography of the gallbladder was performed after three and nine months of treatment. All patients without disappearance of their gallstones after nine months received open-label ursodiol for an additional nine months with ultrasonography after three and nine months. Among the patients completing three months of masked treatment, disappearance of gallstones was seen in five of 11 patients who received placebo and four of seven patients who received ursodiol. Only one additional patient of six continuing placebo for nine months experienced disappearance. Neither of two patients continuing ursodiol for nine months had disappearance of gallstones. None of the five patients treated with open-label ursodiol for nine months had disappearance of gallstones. Thus, half of the gallstones that form during rapid loss of weight disappear rapidly once loss of weight ceases; ursodiol may not increase the frequency or rapidity of their disappearance.

Oral and contact dissolution of gallstones.

The appropriate selection of patients for treatment with oral ursodeoxycholic acid (UDCA)--a drug that has virtually no side effects--results in about 50% of patients experiencing safe dissolution of gallstones within 2 years. Eligible patients have small (less than 20 mm in diameter) radiolucent gallstones in a gallbladder visualized by oral cholecystography (OCG); ideal candidates are thin women who have gallstones that are less than 15 mm in diameter, floating when observed by OCG, or of low density on computed tomographic (CT) scanning. Contact dissolution with methyl tert-butyl ether (MTBE) is rapid, effective more often than UDCA, and safe but requires the expertise of an interventional radiologist. Any size and number of cholesterol gallstones that are not CT-dense will be dissolved by MTBE, leaving at most only insoluble debris that is clinically innocuous. Although gallstones recur after dissolution by UDCA or MTBE in 50% of patients within 5 years, recurrent gallstones are usually asymptomatic and/or can probably be dissolved. We conclude that oral or contact dissolution provides an alternative treatment to cholecystectomy for about 30% of patients with symptomatic gallstones.

A gelonin-containing immunotoxin directed against human breast carcinoma.

Toxins may be specifically directed to tumor cells and the toxins' potency greatly increased by covalent conjugation to monoclonal antibodies recognizing tumor-associated antigens. Antibody 15A8, an immunoglobulin G1 (IgG1) subclass anti-human breast carcinoma murine monoclonal antibody and gelonin, a plant toxin, were covalently modified with N-succimindyl 3-(2-pyridyldithio) proprionate and iminothiolane, respectively, and allowed to cross-link. 15A8-gelonin conjugates were purified from unreacted antibody and free gelonin by gel filtration and blue sepharose chromatography. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed that the final product contained two bands corresponding to antibody:gelonin conjugates of 1:1 (predominant) and 1:2. There were no contaminating amounts of free antibody or free toxin in the preparation. The yield of the final purified 15A8-gelonin conjugate was approximately 20% based on the amount of starting antibody. The protein synthesis inhibitory activity of the immunoconjugate was assessed by in vitro rabbit reticulocyte translation assay. This functional activity was normalized to that of unmodified gelonin for use in in vitro antiproliferative assays against antigen-negative (Hs294t human melanoma) and antigen-positive (ME-180 human cervical carcinoma) cell lines. Antigen-negative Hs294t cells incubated for 72 hours with 15A8-gelonin immunotoxin showed no increased cytotoxicity compared with HS294t cells exposed to free gelonin alone. However, the immunotoxin was preferentially toxic to antigen-positive ME-180 cells; over 5 logs greater cell kill was observed after 72 hours exposure to 15A8-gelonin than after the same exposure to gelonin alone. Various lysosomotropic agents augmented 15A8-gelonin cytotoxicity; the most effective potentiating agent appeared to be monensin. In addition, the chemotherapeutic agents L-phenylalanine mustard (L-PAM), 5-fluorouracil, vincristine, and bleomycin, and the biological response modifiers interferon-alpha and tumor necrosis factor-alpha were shown to augment 15A8-gelonin cytotoxicity. Should in vivo pharmacology and therapeutic studies confirm these in vitro findings, 15A8-gelonin conjugate may be a potent agent for therapy of cancer in man.

The sequence of biliary events preceding the formation of gallstones in humans.

The aim of the present study was to determine the sequence of events leading to formation of gallstones among obese patients predisposed to cholesterol gallstones by a very low-calorie diet. Nine obese patients beginning a 520-kcal diet had gallbladder bile collected from the duodenum before beginning the diet and seven times during the first 56 days of the diet. Biliary cholesterol saturation index and levels of arachidonate, prostaglandin E2, and glycoprotein increased significantly; nucleation time decreased; and total lipid concentration did not change. Decreases in nucleation time preceded the appearance of cholesterol crystals. Significant (P less than 0.05) increases in prostaglandin E2 level were preceded by significant increases in arachidonate level and followed by significant increases in glycoprotein level. These observations support the hypotheses that in obese patients predisposed to gallstones by very low-calorie diets (a) decreases in nucleation time are necessary before cholesterol crystals form in the gallbladder; (b) biliary arachidonate, through its conversion to prostaglandins, promotes biliary synthesis and secretion of glycoprotein; (c) biliary glycoprotein promotes nucleation; and (d) increases in the concentration of gallbladder bile are not necessary for cholesterol nucleation to occur in vivo.

Risk factors for gallstone formation during rapid loss of weight.

Risk factors for the development of gallstones during rapid weight loss were assessed in 457 subjects who entered a weight control program (520 kcal/day). Absence of gallstones in these subjects was documented by ultrasonography prior to entry into the study. Ultrasonography was performed again at 16 weeks on the subjects who remained in the study (N = 248). The incidence of gallstones by 16 weeks of rapid weight loss was 10.9% (27/248). Most factors associated with gallstones in the general population, eg, older age, female gender, parity, positive family history, etc, were not associated with gallstones in this population. The risk factors for developing gallstones included increased initial body mass index [weight (kg)/height (m)2], amount of body mass index loss, and serum triglyceride levels. The positive predictive value of these risk factors was 75%, but the sensitivity was only 12%. These observations indicate that risk factors for the development of gallstones during rapid weight loss are probably different from those in the general population. The factors identified by this study are useful in predicting patients at high risk for gallstones. However, since only a minority of gallstones that form can be predicted, further study is needed to identify additional factors that will improve our ability to predict gallstone formation.

Ulcer recurrence following duodenal ulcer healing with omeprazole, ranitidine, or placebo: a double-blind, multicenter, 6-month study. The Omeprazole Duodenal Ulcer Study Group.

The present study determined whether the rate of relapse of duodenal ulcer was reduced after ulcer healing with omeprazole compared with ranitidine or placebo. It was made up of a double-blind, randomized, controlled multiple-center trial set within the United States. Patients were candidates if their duodenal or pyloric channel ulcer successfully healed in one of two large multicenter U.S. trials; one compared omeprazole, 20 mg once daily, before breakfast with ranitidine, 150 mg twice daily, and the other compared the same dose of omeprazole with placebo. Two hundred forty (73.8%) of the 325 patients with complete ulcer healing within 4 weeks of starting therapy who were eligible to enter the follow-up study were enrolled. There was no intervention. Endoscopic assessment of ulcer status was performed at 2, 4, and 6 months and whenever patients had symptoms thought to represent return of an ulcer. The lifetable relapse rates for duodenal ulcer according to initial ulcer therapy with omeprazole, ranitidine, or placebo were 76.7% [95% confidence interval (CI), 64%-89.3%], 59.8% (95% CI, 47.8-71.7%), and 50.4% (95% CI, 15.7%-85.2%), respectively. These rates were not statistically significantly different. Seventeen percent of recurrent ulcers occurred at a site different from that of the original ulcer. It is concluded that despite the more rapid rate of duodenal ulcer healing with omeprazole therapy, the rate of ulcer relapse appears similar and independent of whether ulcer healing was accelerated with omeprazole or ranitidine.

Factors affecting the measurement of cholesterol nucleation in human gallbladder and duodenal bile.

The study of cholesterol gallstone disease would be facilitated if the nucleation time of cholesterol crystals could be measured in duodenal bile and was correlated with nucleation occurring in vivo. Therefore, our aims were to determine (a) if nucleation time could be measured in duodenal bile, (b) the effect of bacteria, phospholipase, protease, and dilution on the measurement of nucleation time, and (c) the ability of nucleation time of duodenal bile to reflect changes occurring in vivo that promote the formation of gallstones and, therefore, the potential usefulness of nucleation time in predicting and studying the formation of gallstones. Gallbladder bile was obtained from 27 patients undergoing elective cholecystectomy and 19 patients undergoing diagnostic duodenal biliary drainage. Among the 14 bile samples collected by drainage that nucleated within 21 days, mean nucleation time was 6.3 +/- 2.8 days. The addition of inhibitors of phospholipase or protease prolonged nucleation time slightly. Bacteria were cultured from one bile sample at the time of collection and five samples at the time of nucleation. The addition of antibiotics had no effect on nucleation time. Dilution of bile collected at cholecystectomy to the concentration of duodenal bile prolonged nucleation time. In 4 of 5 obese patients receiving a very low calorie diet and predisposed to gallstones, the nucleation time in duodenal bile shortened, and the shortest nucleation times were associated with the formation of cholesterol crystals in vivo. Thus, measurement of nucleation time in duodenal bile may be useful in predicting and studying the formation of cholesterol gallstones.

Roles of deoxycholate and arachidonate in pathogenesis of cholesterol gallstones in obese patients during rapid loss of weight.

Our aim was to examine the relationship between biliary deoxycholate and arachidonate in obese patients and the relationship of deoxycholate and arachidonate to the stimulation of biliary mucous glycoprotein among obese patients predisposed to cholesterol gallstones. Thirty-four obese patients predisposed to cholesterol gallstones by a weight-reducing diet (520 kcal/day) received placebo, ursodiol (1200 mg/day), or aspirin (1300 mg/day). Duodenal bile was collected prior to beginning the diet and at four weeks. There was no correlation between deoxycholate and arachidonate among the 34 patients before beginning the diet. With placebo, deoxycholate decreased while arachidonate and glycoprotein increased. With ursodiol, deoxycholate decreased while arachidonate decreased and glycoprotein did not change. With aspirin, there was no change in deoxycholate but a decrease in arachidonate and no change in glycoprotein. Our data do not support a role for biliary deoxycholate in the regulation of biliary arachidonate. Our data do support a role for arachidonate, but not deoxycholate, in the regulation of biliary glycoprotein during the formation of cholesterol gallstones.

Gallstones: an update.

Gallstones are a major health problem in the United States. More than 20 million Americans have gallstones, and a million new cases are discovered each year. A resurgence of interest in this field has been stimulated by advances in understanding the pathogenesis of cholesterol gallstones and in nonsurgical treatments. The purpose of this symposium is to update these advances.

Effects of ursodeoxycholic acid and aspirin on the formation of lithogenic bile and gallstones during loss of weight.

We attempted to determine whether the administration of aspirin or ursodeoxycholic acid during loss of weight could prevent the development of lithogenic changes in bile and the formation of gallstones. Sixty-eight obese subjects without gallstones who were entered in a program (520 kcal per day) to lose weight were randomly assigned to receive ursodeoxycholic acid (1200 mg per day), aspirin (1300 mg per day), or placebo in double-blind fashion for up to 16 weeks. At entry, at four weeks of treatment, and at three weeks after the completion of treatment, the subjects underwent ultrasonography to detect gallstones and duodenal drainage of bile to detect cholesterol crystals and to determine the bile saturation index and glycoprotein concentration. No gallstones or cholesterol crystals formed in the patients treated with ursodeoxycholic acid. Among the patients given placebo, gallstones formed in five (P less than 0.05 vs. ursodeoxycholic acid) and cholesterol crystals in six (P less than 0.001 vs. ursodeoxycholic acid); among those given aspirin, gallstones formed in two and cholesterol crystals in one (no significant difference from ursodeoxycholic acid treatment). By the fourth week, the bile saturation index increased in the placebo group (from 1.07 +/- 0.26 to 1.29 +/- 0.27; P less than 0.001), decreased in the ursodeoxycholic acid group (from 1.11 +/- 0.34 to 0.91 +/- 0.24; P less than 0.001), and did not change significantly in the aspirin group. The concentration of glycoprotein in bile increased in the placebo group (27.9 +/- 14.5 percent; P less than 0.001) but did not change significantly in the groups treated with ursodeoxycholic acid or aspirin. We conclude that ursodeoxycholic acid prevents lithogenic changes in bile and the formation of gallstones in obese subjects during loss of weight.