Activation of mast cells by bile acids

GASTROENTEROLOGY

LIVER,

PANCREAS,

AND BILIARY

1991;101:446-456

TRACT

Activation of Mast Cells by Bile Acids RICHARD G. QUIST, HUONG-THU TON-NU, ALAN F. HOFMANN, and KIM E. BARRETT

JAN LILLIENAU,

From the Divisions of Gastroenterology and Allergic and Immunologic Medicine, University of California, San Diego, La Jolla, California

To test whether bile acids interact with mast cells, dilute, aqueous solutions of five pure unconjugated natural bile acids and their corresponding glycine or taurine conjugates were incubated with murine PT-18 cells (a mast cell line functionally and cy tochemically similar to mucosal mast cells) or with freshly isolated rat peritoneal mast cells. Bile acid solutions ranged in concentration from 0.3 to 10 mmol/L; histamine release was assessed by a fluorimetric assay, and cell lysis by cytosolic enzyme (lactate dehydrogenase) release. Lipophilic, dihydroxy bile acids (chenodeoxycholic acid and deoxycholic acid as well as their glycine and taurine conjugates) caused histamine release in a doserelated manner; cholic acid and its conjugates caused much less or no histamine release. Two hydrophilic bile acids (ursodeoxycholic acid and ursocholic acid and their conjugates) were virtually devoid of activity. Histamine release, which was independent of extracellular Ca’+, occurred at 0.3 mmol/L, well below the critical micellization concentration. For a given concentration, unconjugated bile acids and glycine-conjugated bile acids induced more histamine release than taurine-conjugated bile acids; maximal release was observed at 3 mmol/L for lipophilic, dihydroxy bile acids. To test whether bile acids could also cause histamine release from cutaneous mast cells in vivo, rats were injected intradermally with bile acid solutions and histamine release assessed by capillary leakage of Evan’s blue dye. Cutaneous blueing was greater with cytotoxic bile acids, chenodeoxycholylglycine or deoxycholylglytine, than with ursodeoxycholylglycine and was inhibited by prior antihistamine treatment. Histamine release correlated highly and positively with lipophilicity and with bile acid surface activity. It was concluded that lipophilic but not hydrophilic bile acids possess concentration-dependent cytotoxicity toward mast cells causing histamine release, that unconjugated and glycine-conjugated bile acids are more potent than taurine-conjugated bile acids,

Diseases, Department

of

and that mast cell histamine release is highly correlated with lipophilicity of bile acids as well as their surface activity.

M

ast cells are widely distributed in the body and have been implicated in the pathogenesis of inflammation, because they contain a variety of mediators, including histamine, eosinophil and neutrophil chemotactic factors, platelet-activating factor, proteoglycans such as heparin or chondroitin sulfate, and specific proteases (2). Release of these mediators from mast cells is induced by immunologic stimulation as well as by a variety of nonimmunologic activators (2). A role for mast cells has been proposed in the inflammatory response seen in several gastrointestinal diseases (3,4). Two such diseases, celiac disease (5) and inflammatory bowel disease (6), are accompanied by increased epithelial permeability. In such conditions, mast cells present in the subepithelium of the intestinal mucosa might become exposed to potential secretagogues in the lumen. Among secretagogues present at relatively high concentrations (l-10 mmol/L) in the small intestine are bile acids. In this paper, we report studies in vitro on the effect of pure conjugated and unconjugated bile acids on cultured mast cells and on freshly isolated rat peritoneal mast cells. These studies were complemented by in vivo studies aimed at testing whether bile acids can cause histamine release from cutaneous mast cells. Lastly, to gain some insight into the mechanism by which bile acids cause mediator release from mast

Abbreviations used in this paper: [BA],,,,, 50% histamine release; BSA, bovine serum albumin; CDC, chenodeoxycholic acid; C, cholic acid; CMC, critical micellar concentration: DC, deoxycholic acid; GLY, glycine conjugate; IL-3, interleukin 3; HPLC, high-performance liquid chromatography; LDH, lactate dehydrogenase; TAU, taurine conjugate; UC, ursocholic acid; UDC, ursodeoxycholic acid. o 1991 by the American Gastroenterological Association OOl&5085/91/$3.00

August

EFFECT OF BILE ACIDS ON MAST CELLS

1991

cells, the relationship physicochemical examined.

between histamine release and properties of the bile acids used was

447

peritoneal mast cells always exceeded 95%, with a typical yield of 1.4 x 10~ mast cells per rat.

Histamine Release Studies Materials

and Methods

Materials Chenodeoxycholic acid (CDC) and ursodeoxycholic acid (UDC) were received as gifts from Diamalt AG (Raubling, Germany), deoxycholic acid (DC) was purchased from Aldrich Chemical Co. (Milwaukee, WI), and deoxycholylglytine (DC-GLY) and cholic acid (C) were purchased from Calbiochem Inc. (La Jolla, CA). Ursocholic acid (UC) was a gift of Gipharmex S.P.A., Milan, Italy. All other conjugated bile acids tested were synthesized in the laboratory of A.F.H. using methods that have been reported in detail elsewhere (7). All bile acids tested possessed a purity of at least 98%, as demonstrated by thin-layer and highperformance liquid chromatography (8). Diphenhydramine in the in vivo studies was from Sigma Chemical Co. (St. Louis, MO). All other chemicals used were of reagent grade and were purchased commercially.

Mast Cells A line of interleukin 3 (IL-3)-dependent murine mast cells (PT-18) was used for in vitro studies. These cells, which were originally derived from the culture of murine spleen cells with IL-3-containing conditioned medium, are functionally and cytochemically similar to the mast cell type present in the intestinal mucosa (9). The cells were maintained in culture in a humidified atmosphere of 95% sir/5% CO,. The culture medium was RPMI-1640 (40%), to which were added penicillin (50 LJ/mL), streptomycin (50 mg/mL), nonessential amino acids (100 kmol/L), 2-mercaptoethanol (50 kmol/L), sodium pyruvate (1 mmol/L), and L-glutamine (4 mmol/L). The medium was also supplemented with conditioned medium from the WEHI-3B cell line [a source of IL-3, 50%), and fetal calf serum (10%). The cultures were divided in half and replenished with fresh medium every 4-5 days. To obtain rat peritoneal mast cells, male Sprague-Dawley rats weighing between 250 and 300 g were killed by CO, asphyxiation. The peritoneal cavity was then injected with 20 mL of calcium and magnesium-free Tyrode’s buffer containing 10 U/mL of heparin [Abbott Laboratories, North Chicago, IL). The peritoneal cavity was then briefly massaged manually then drained through an excision in the abdominal wall. The peritoneal cells were collected and immediately placed on ice, then centrifuged (150 x g, 5 minutes, 4”C), washed twice in calcium and magnesiumfree Tyrode’s buffer, and resuspended in a final volume of 2 mL. For separation of mast cells from other cells in the lavage fluid, the 2 mL suspension was layered onto a 38% bovine serum albumin (BSA) gradient and centrifuged (250 X g, 20 minutes, 4°C). The BSA gradient was then aspirated, and the pellet was washed three times in full Tyrode’s buffer (100 x g,5 minutes) and resuspended in an appropriate volume for the experiment. The purity of rat

PT-18 cells or rat peritoneal mast cells were washed once in HEPES-buffered Tyrode’s solution (pH 7.4) containing (mmol/L): NaCl, 137; glucose, 5.6; HEPES, 10; KCl, 2.7; NaH,PO,, 0.4; CaCl,, 1; and MgCl,, 1. Aliquots of the cell suspension (1 mL) were prewarmed to 37°C for 10 minutes, then various concentrations of bile acids were added in a volume of 10 or 20 ~_LLand incubated for a further 10 minutes. The reaction was arrested by addition of 2.0 mL ice-cold Tyrode’s buffer. Samples were centrifuged (100 x g, 10 minutes, 25”C), the supernatants removed to fresh tubes, and the pellets resuspended in 3 mL Tyrode’s buffer. The cell pellets were boiled for 10 minutes or sonicated to release residual histamine. The histamine content of all samples was then assayed using an o-phthaldialdehyde-based fluorimetric method as described by Shore et al. (10) (without the extraction steps). Release of histamine into the supernatant was expressed as a percentage of histamine initially present in the cells and was corrected for the spontaneous release of histamine (usually < 15% for PT-18 cells and < 5% for rat peritoneal mast cells) occurring in the absence of stimulation. For in vitro experiments aimed at defining the dependency of histamine release on extracellular Ca2+ concentration, calcium and magnesium were omitted from the Tyrode’s buffer, and EGTA (1 mmol/L) was added.

Lactate Dehydrogenase

Studies

To investigate whether bile acids influenced the viability of mast cells, release of the cytosolic enzyme lactate dehydrogenase (LDH) was measured and compared with histamine release. Cultured PT-18 mast cells were incubated with bile acids in Tyrode’s buffer as described in the histamine release studies. After separation of cellular and supernatant fractions, samples were sonicated and LDH activity was determined spectrophotometrically using an LDH assay kit (Sigma). Results are expressed as the percentage of total LDH that was released into the supernatant fraction minus the release occurring in the absence of bile acids (usually < 10%).

In Viva Studies In vivo studies were performed to assess whether bile acids cause histamine release from cutaneous mast cells and if the bile acid structure-potency relationships observed in vitro were present in vivo. Male SpragueDawley rats (200-350 g) were lightly anesthetized with sodium pentobarbital (50 mgikg) and were then injected IV with 0.5 mL Evan’s blue dye (1.8% wtivol). After 1 hour, the rats were first shaved and then injected intradermally [into the dorsal skin) with 50 uL of a bile acid solution; positive control sites were injected with histamine (1 mg/mL; 9 mmol/L) and negative control sites with saline. After 10 minutes, the rats were killed by sodium pentobarbital

448

QUIST ET AL.

GASTROENTEROLOGY Vol. 101, No. 2

overdose, and the dorsal skin was removed and inverted. Vascular leakage of Evan’s blue was quantitated as the sum of the crossed diameters (in millimeters) of blueing at the intradermal injection sites. To confirm mast cell activation in vivo, some rats received diphenhydramine (5 mg/kg IV) 60 minutes before the bile acid injection. The effect of bile acids on cutaneous mast cells was also assessed histologically by examining tissue sections that had been intradermally injected with bile acids. Male Sprague-Dawley rats (300 g) were lightly anesthetized with sodium pentobarbital (50 mg/kg), and the dorsal skin was shaved. After intradermal injections of bile acids, 6-mm punch biopsy samples were obtained and placed in Carnoy’s solution for 18 hours, then placed in absolute ethanol. The tissues were embedded in paraffin, sectioned at 5 km, and stained with 0.5% toluidine blue. The number of mast cells in the sections were assessed by light microscopy by an observer blinded to the experimental treatment. Results are expressed as the number of mast cells per 100 high-power fields of 0.0289 mm2 in area.

Physicochemical

Properties

Lipophilicity. Lipophilicity, or hydrophobicity, of bile acids has been hypothesized to be a predictor of their cytotoxicity (11). Several investigators have proposed that lipophilicity is indicated by the relative retention time during reversed-phase partition chromatography (12,13). We have used the values published by Heuman et al. (13), which were obtained using a C,, reversed-phase column at pH 9. Under these conditions, lipophilicity is not influenced by the degree of ionization of bile acids. Swface pressure. Because bile acids are surface active and because surface activity is a measure of the affinity of compounds for a hydrophobic interface such as the lipid domains of the cell membrane, we hypothesized that surface activity should correlate with histamine release. Surface activity may be expressed as surface pressure, i.e., the amount by which the surface tension of a solution is lowered by the presence of a surfactant (in dynes per centimeter) (14). The maximal surface pressure occurs at a concentration close to the critical micellar concentration (CMC), because at concentrations above the CMC there is little increase in surface pressure. The surface pressure values and values for the solutions of the bile acids used in these studies were determined using the maximum bubble pressure method, the principles of which have been fully described (15,16) or were taken from data published by Roda et al. (17).

All data are expressed as means 2 SE, and statistical comparisons were made using Student’s t test. Results The trivial name, abbreviation, nuclear substituents, and pK, values for the bile acids studied are 1. Table

Nuclear substituents

R, Trivial name” Unconjugated bile acids (R, = 0-); pK, = 5.0b Chenodeoxycholate Ursodeoxycholate Deoxycholate Cholate Ursocholate Glycine-conjugated bile acids [R, = NH(CH,)CO,-1; pK, = 3.7’ Chenodeoxycholylglycine Ursodeoxycholylglycine Deoxycholylglycine Cholylglycine Ursocholylglycine Taurine-conjugated bile acids [R, = NH(CH,),SO,-1; pK, < 1.0 Chenodeoxycholyltaurine Ursodeoxycholyltaurine Deoxycholyltaurine Cholyltaurine Ursocholyltaurine

Abbreviation

R,

R,

3-OH 7-OH Z-OH

CDC UDC DC C UC

(Y (Y (Y (Y cx

0. P a P

(Y CX a

CDC-GLY UDC-GLY DC-GLY C-GLY UC-GLY

(Y cx ci CX (Y

(Y p (Y P

(Y cl a

CDC-TAU UDC-TAU DC-TAU C-TAU UC-TAU

(Y (Y (Y (Y ci

; a P

(Y (Y a

“The United States Adopted Names (USAN) for chenodeoxycholate and ursodeoxycholate are chenodiol and ursodiol, respectively. An alternative trivial name for a conjugated bile acid such as cholylglytine is glycocholate. bThe values for pK, have been published by Fini and Roda (28).

histamine released at 3 mmol/L bile acid concentration, as well as the concentration of bile acid causing 50% histamine release. Table 2 also gives the physicochemical properties (CMC, lipophilicity index, and surface activity index). Ejject of Structure of Steroid Moiety on Mast Cell Activation

Data Analysis

given in Table

Table I. Bile Acids Studied

2 shows the percentage

of

Unconjugated bile acids. The unconjugated dihydroxy bile acids, CDC and DC, caused histamine release from mast cells in a dose-related manner (Figure 1). Both CDC and DC initiated histamine release at 0.3 mmol/L and caused maximal histamine release, approximately 80%, at 3 mmol/L. At 1 mmol/L, a concentration well below the CMC value, CDC and DC caused histamine release of 51.5% ? 6.2% and

August

EFFECT

1991

OF BILE ACIDS ON MAST CELLS

449

Table 2. Comparison of Histamine Release With Physicochemical Properties of Bile Acids Cytotoxicity Histamine % at 3 mmol/L

Bile acids Unconjugated bile acids CDC UDC DC C UC Glycine-conjugated bile acids CDC-GLY UDC-GLY DC-GLY C-GLY UC-GLY Taurine-conjugated bile acids CDC-TAU UDC-TAU DC-TAU C-TAU UC-TAU

-c 4.9

Physicochemical

release

WI,,,,

CMC (mmol/L.)b

Lipophilicity

properties Surface

index

activity

1.0 >lO 1.3 9.4 >lO

2.7 8.4 2.4 11 39

+0.59

11.4 k 7.4 80.4 -c 3.3 5.6 + 6.0 4.1 * 3.2

-0.31 +0.72 f0.13 -0.74

24.7 20.4 25.1 20.0 16.5

77.6 + 7.7 2 80.0 2 11.3 e 3.3 5

2.0 >lO 2.3 9.4 >lO

1.8 4.4 2.6 8.9 30

f0.51 -0.43 +0.65 +0.07 -0.86

25.1 20.4 24.5 20.8 18.1

4.7 >lO 3.6 210 >lO

1.8 4.3 1.5 10.3 40

+0.46 -0.47 +0.59 0.00 -0.94

23.3 18.9 22.9 18.5 16.0

79.4

1.6 4.6 0.9 2.5 2.9

14.7 2 4.6 4.1 ? 3.8 44.6 k 5.4 13.9 * 1.9 13.0 -t 10.9

index’

“As measured in vitro from PT-18 mast cells (n = 3-10). bThe CMC value is in the presence of 0.15 mol/L Na+. Although the CMC is given as a single concentration, bile acid molecules self-associate over a concentration range: the more hydrophilic the bile acid, the greater the concentration range. The CMC is about the midpoint of the concentration range over which micelle formation occurs (40). Values for C, UC, UC-GLY, and UC-TAU were obtained from previously published data of Roda et al. using the maximum bubble pressure method (15-17). Values for other bile acids were obtained by J. Lillienau, Y. Peng, K. J. Mysels, and A. F. Hofmann [unpublished data). ‘.The surface activity index is defined as the decrease in surface tension [surface pressure) at the CMC.

33.8% + 8.0% (n = 4). In contrast to CDC and DC, UDC, a less lipophilic dihydroxy bile acid caused insignificant amounts of histamine release even at concentrations up to 10 mmol/L. Cholic acid, a trihydroxy bile acid, was found to cause intermediate amounts of histamine release, with 53.4% k 10.8% (n = 3) histamine release occurring at 10 mmol/L. Ursocholic acid, a less lipophilic 7p hydroxy epimer,

caused insignificant amounts all concentrations tested.

of histamine

release at

Glycine-conjugated bile acids. Chenodeoxycholylglycine and DC-GLY caused histamine release at concentrations as low as 0.3 mmol/L, a concentration well below their respective CMC values. The two hydroxy-conjugated bile acids caused maximal release at 3 mmol/L, as shown in Figure 2. In contrast,

0 0.3

3 Conlcentratbn (mM)

10

Figure 1. Effect of unconjugated bile acid concentration on histamine release from cultured PT-18 mast cells in vitro. Each point represents net histamine release from the cells after a lo-minute incubation with the bile acid tested and is expressed as the mean f SE for 3-10 experiments. Histamine release occurring in the absence of stimulation was 9.8% -t 2.7% (n = 4). 0, CDC; H, DC; 0, UDC; 0, C; A, UC.

0.3

1 Concentration

3

10

(mM)

Figure 2. Effect of glycine-conjugated bile acid concentration on histamine release from cultured &T-l&l mast cells in vitro. Each point represents net histamine release from the cells after a lominute incubation with the bile acid tested and is expressed as the mean ? SE for 3-10 experiments. Histamine release occurring in the absence of stimulation was 16.5% -C 2.4% (n = 7). Symbols as given in Figure I (for the corresponding unconjugated derivative).

450 QUISTETAL.

GASTROENTEROLOGYVol. lOl,No.2

UDC-GLY caused no histamine release at 0.3mmol/L; even at 10 mmol/L (well above its CMC), UDC-GLY caused only slight amounts of histamine release (20.1% + 8.4% n = 8). Cholic acid glycine conjugate, a trihydroxy-conjugated bile acid, induced considerable histamine release (54.4% f 9.7% n = 3) at a concentration of 10 mmol/L. Ursocholic acid glycine conjugate, its less lipophilic epimer, was inactive at all concentrations tested. Taurine-conjugated bile acids. The effect of taurine-conjugated bile acids is shown in Figure 3. Chenodeoxycholic acid taurine conjugate and DCTAU both initiated histamine release at approximately 1 mmol/L (3.9% + 0.7% and 13.4% k 6.5%, respectively, n = 3) and caused maximal release of approximately 80% at 10 mmol/L. Ursodeoxycholic acid taurine conjugate and C-TAU both caused low levels of histamine release, because 20.5% k 13.9% and 22.0% + 5.0% (n = 3) release occurred, respectively, at 10 mmol/L. Ursodeoxycholic acid taurine conjugate caused only slight amounts of histamine release up to 10 mmol/L. Effect of amidation among bile acids causing histamine release. In the in vitro studies, histamine

release was clearly dependent on bile acid class and species. For a given steroid moiety with both CDC and DC, histamine release was significantly higher for the unconjugated bile acids at submaximal concentrations (1 mmol/L) than for their corresponding glycine and taurine conjugates (P < 0.05), as shown in Figure 4. Lactate dehydrogenase release studies. To investigate whether the histamine releasing effect of bile acids on mast cells reflected cytotoxicity, cultured mast cells were treated with various concentrations of

0.3

1

Concentration

3

10

(mM)

Figure 3. Effect of taurine-conjugated bile acid concentration on histamine release from cultured PT-18 mast cells in vitro. Each point represents net histamine release from the cells after a lo-minute incubation with the bile acid tested and is expressed as the mean 2 SE for three experiments. Histamine release occurring in the absence of stimulation was 16.4% k 2.6% (n = 5). Symbols as given in Figure 1 [for the corresponding unconjugated derivative).

60

T

Figure 4. Comparison of histamine release induced by CDC and DC (open bars) or their glycine (stippled bars) or taurine (hatched bars) conjugates from cultured FT-18mast cells in vitro. Values represent histamine released from the cells after a to-minute incubation with the bile acid tested and are expressed as the mean +- SE for 10 experiments. Values for conjugated bile acids denoted by asterisks differ significantly from the corresponding values with unconjugated bile acids, P < 0.05.

bile acids, and the amount of LDH released was compared with histamine release. Comparisons of LDH release with histamine release are shown in Figure 5A-C for the bile acids CDC, DC, and UDC, respectively. For CDC and DC, percent histamine release and percent LDH release were consistently similar. For example, with 3 mmol/L CDC, histamine release was 79.2% + 4.3% and LDH release was 71.7% + 6.6%. For 3 mmol/L DC, histamine release was 77.1% + 2.9% and LDH release was 79% k 0.8%. In contrast, for UDC, concentrations < 3 mmol/L induced significant histamine release without an associated release of LDH. Kinetics of histamine release. To investigate whether addition of low, noncytotoxic concentrations of bile acids to mast cells caused histamine release when incubated for prolonged time periods, PTl8 cells were incubated with 0.3 mmol/L of CDC, UDC, or DC for ZO-, 40-, and 60-minute intervals, and histamine release was measured as stated in the histamine release studies. As shown in Figure 6, increases in histamine release occurred after 40 minutes of incubation with CDC or DC. At 60 minutes, histamine release caused by 0.3 mmol/L CDC was 12.8 +- 1.5 and was significantly higher (P < 0.05) than release occurring at 10 minutes (3.5% k 2.9%). For 0.3 mmol/L DC, histamine release at 10 minutes was 1.6% 2 1.4% and was significantly increased at 60 minutes to 14.0% + 1.3% (P < 0.01). However, for 0.3 mmol/L UDC, there

EFFECT OF BILE ACIDS ON MAST CELLS 451

August1991

80

80

CDC (mM)

DC(

100

C

80

20

0 -10

I

I

I

10.0

0.3 GC

(rn$O

were no significant increases in histamine occurring up to 60 minutes of incubation. Ej’fect of bile acids on rat peritoneal

release

mast cells.

Because mast cells may display considerable variations in their functional properties that are both tissue and species specific, experiments were performed to determine whether bile acids induce histamine from another mast cell type. Rat peritoneal mast cells, which are cytochemically and functionally similar to the mast cell type present in cutaneous tissue, were isolated as described in Materials and Methods and incubated with various concentrations of CDC, UDC, DC, C, or UC. Results shown in Figure 7 show that the histamine release occurred in a dose-related manner for all five unconjugated bile acids tested. For CDC and DC, the percent release that occurred was quite similar to that from the PT-18-cultured mast cells, because the threshold concentration for release was

Figure 5. Comparison of the ability of unconjugated bile acids, CDC (A), DC (B), and UDC (C), to release histamine (open bars) or LDH (hatched bars) from cultured PT-18 mast cells after 10 minutes of incubation. Values are means + SE for three experiments. Values for LDH release denoted by asterisks (C) differ significantly from the percentage histamine release induced by the same concentration of bile acid (P < 0.05).

0.3 mmol/L (2.1% + 2.1% and 0.4% 2 0.5%, n = 3) and maximal release occurred at 2.5 mmol/L (75.1% + 8.4% and 80.8% 2 0.9% n = 3) for both bile acids. However, for UDC, the rat peritoneal mast cells were found to be more sensitive to the bile acids than the PT-18 cultured mast cells. Histamine release at 10 mmol/L UDC for the peritoneal mast cells was 67.0% -+ 7.3% and was thus significantly higher (P < 0.01) than release occurring with the PT-18cultured mast cells (3.2% +- 3.0%). Effect of extracellular calcium on histamine release in vitro. Because bile acids have been claimed

to induce increases in cytosolic calcium by either acting as calcium ionophores (18-20) or by mobilizing intracellular Ca’+ stores (Zl), we tested whether bile acid-induced histamine release required calcium in the incubation solution. The effect of unconjugated and glycine-conjugated bile acids on histamine re-

452

QUIST

GASTROENTEROLOGY

ET AL.

I

I 10

-loo

I 20

I 30

I 40

I 50

I 60

I

70

Time (min) Figure 6. Time course of histamine release induced by 0.3 mmol/L CDC, UDC, or DC from cultured IT-18 mast cells. Values are means k SE for three experiments. 0, CDC; 0, UDC; a, DC.

lease was measured in the presence and absence of extracellular calcium or magnesium. Histamine release occurred in a dose-related fashion in the absence of calcium or magnesium, similar to the release observed in the presence of extracellular calcium, regardless of the dose of bile acid used. For example, histamine release in the presence of calcium caused by 1 mmol/L CDC was 51.5% + 6.2% and in the absence of calcium was 68.7% + 1.9%, n = 3. Likewise for 1 mmol/L DC-GLY, histamine release in the presence of calcium was 33.3% + 8.0% and did

100

Vol. 101, No. 2

not differ from release occurring in the absence of extracellular calcium, which was 30.8% ? 10.8%. At higher concentrations of bile acids (3 or 10 mmol/L), histamine release occurring in the absence of calcium also did not differ significantly from release occurring in the presence of calcium or magnesium. Thus, histamine release occurring by bile acids in vitro was independent of extracellular calcium or magnesium. Effect of bile acids on mast cells in vivo. Intradermal injection of bile acids caused an increase in capillary permeability. Data for unconjugated bile acids are shown in Table 3. Chenodeoxycholic acid, DC, and UDC caused dose-dependent blueing, but CDC and DC were more potent than UDC. We also examined the effect of glycine-conjugated bile acids at 3 mmol/L. Chenodeoxycholic acid glycine conjugate and DC-GLY induced significant blueing, whereas the blueing induced by 3 mmol/L UDC-GLY did not differ significantly from that induced by the saline negative control (Figure 8). Thus in vivo structure activity relationships mirrored those observed in vitro. Capillary leakage induced by bile acids was decreased by pretreatment with diphenhydramine (Figure 8). Blueing induced by 3 mmol/L DC-GLY was decreased by 91.7% ? 6.9% (n = 7) and that induced by CDC-GLY injection was decreased by 77.3% ? 7.0% (P < 0.05) (n = 7). In control sites injected with histamine, the same dose of diphenhydramine caused a significant decrease in blueing (60.4% + 10.2%, P < 0.05) (n = 7). To confirm histologically that bile acids could have effects on mast cells in vivo, punch biopsy specimens from cutaneous rat tissue injected with bile acids were stained for mast cells. For tissue injected with saline alone, there were 80 mast cells/100 fields. In the sections injected with 0.3 mmol/L CDC or DC, there were essentially no changes in the number of mast cells (82 cells/100 fields and 77.7 cells/100 fields, respectively). However, when higher doses of bile acid were injected, there was a sizable decrease in the number of stainable mast cells (32.5 cells/100 fields, 3 Table 3. Dose Dependence of the Effect of Unconjugated Bile Acids on Cutaneous Vascular Permeability in the Rat Diameter

0.1

1 IO 0.3 Concentration (mM)

20

Figure 7. Histamine release induced from freshly isolated rat peritoneal mast cells by unconjugated bile acids. Values show net histamine release from the cells after a lo-minute incubation with the bile acid tested and are means k SE for three to four experiments. Histamine release occurring in the absence of stimulation was 3.0 f 0.3 (n = 5). 0, CDC; 0, UDC; W, DC; 0, C; A, UC.

of blueing

(mm)“

Bile acid concentration

Bile acid

0.3 mmol/L

CDC DC UDC

13.7 2 2.4 17.3 f 1.8 13.3 f 0.9

1 mmoliL 25.0 -t 2.1* 24.7 2 1.2b 16.0 2 1.7

3 mmol/L 30.3 2 1.7b 27.7 2 l.ab 20.0 2 1.7"

NOTE. Values are expressed as mean 2 SE. “Values for three rats are expressed as the sum of crossed spot diameters. Responses that differ significantly from that of the saline control;*P < O.Ol;"P < 0.05.

EFFECT OF BILE ACIDS ON MAST CELLS

August 1991

453

30

0

Control

m f

DPH (ontihlatamina

20

s z ‘f B ??

ij

10

.

0

cot-CLY

L

DC-CLY

:

UDC-CLY

Figure 8. Effect of intradermally injected bile acids, CDC-GLY, DC-GLY, or UDC-GLY (3 mmol/L), on cutaneous capillary leakage in the rat. Open bars represent the mean sum of blue spot diameters (in millimeters) caused by leakage of Evan’s blue dye from plasma at sites injected intradermally with bile acids in rats pretreated with placebo. Hatched bars represent the sum of blue spot diameters caused by intradermal bile acid injection when animals were pretreated 1 hour previously with diphenhy (50 m&g). Cutaneous blueing was assessed at 10 dramine minutes after intradermal injections. Values are means f SE for seven experiments. Asterisks denote responses in the presence of diphenhydramine that were significantly less than those occurring in its absence (P < 0.05 by Student’s t test).

mmol/L sections 0.3, 1, cells/100 alone.

CDC; 48.1 cells/100 fields, 3 mmol/L DC). In treated with UDC, sites injected with either or 3 mmol/L had approximately 70 mast fields, similar to those injected with saline

Correlation of histamine-releasing physicochemicalproperties of bile acids.

effects with

The relationship between lipophilicity, assessed by high-performance liquid chromatography (HPLC) retention times, and the concentration of bile acids causing 50% histamine release ([BA],,,,) from PT-18 mast cells is shown in Figure 9. Bile acids with a lipophilicity index 5 0.0 did not induce 50% histamine release at the highest concentration tested (i.e., [BA],,,, > 10 mmol/L). For bile acids with lipophilicity indices > 0.0, as the lipophilicity increased, there was a progressive increase in ability to release histamine (Figure 9). For these active bile acids (C, CDC, DC, C-GLY, CDC-GLY, DC-GLY, CDC-TAU, and DC-TAU), there was a statistically significant linear correlation (r = -0.95; P < 0.001) between the lipophilicity index of the bile acid and the concentration at which it caused 50% histamine release. Histamine release was also correlated with the surface activity of bile acids. Figure 10 shows [BA],,,, for those bile acids where this parameter was < 10 mmol/L, plotted against surface pressure at the CMC, which was termed the surface activity index. The

0.0

0.1

0.2

0.4

0.3

0.5

Lipophiliclty

0.6

0.7

0.8

Index

Figure 9. Comparison of the concentration of bile acid causing 50% histamine release ([BA],,,,) for bile acids causing >50% histamine release below 10 mmol/L, with the lipophilicity index as reported by Heuman et al. (15). The line was fitted to the points by linear regression analysis.

surface activity index also correlated highly with the cytotoxic activity of bile acids (r = 0.97; P < 0.001). However, the surface activity index was not 100% specific in discriminating between cytotoxic and noncytotoxic bile acids, because UDC-GLY was noncytotoxic and had a surface activity index of 20.4, whereas C was cytotoxic with a surface activity index of 20.0. Discussion These studies show that three natural bile acids, CDC, DC, and C, and their corresponding glycine conjugates are potent releasers of histamine 10

1

1

I

I

I

I

I

1

I

20

21

22

23

24

25

1

I

I

8-

6-

4-

2-

01 19

Surface

activity

I

29

indrx

Figure 10. Comparison of [BA],,,, [for bile acids causing > 50% histamine release below 10 mmol/L) with the surface affinity index (see text for definition). The Iine was fitted to the points by linear regression analysis.

454

QUIST ET AL.

from mast cells in vitro and in vivo. Chenodeoxycholic acid glycine conjugate, C-GLY, and DC-GLY are the major bile acids present in the gallbladder and small intestine in humans (22). Two less lipophilic bile acids (UDC and UC) caused little histamine release. The histamine-releasing effect was dependent not only on the steroid moiety of the particular bile acid but also on the mode of conjugation. The order of potency for histamine release for CDC and DC was unconjugated > glycine conjugated > taurine conjugated, as depicted in Figure 4. For C and its conjugates, C and C-GLY had equivalent histaminereleasing activity, while C-TAU was not active at the concentrations used. Bile acids are known to partition into membrane bilayers at the concentrations used in these studies (2 3-26) I Lipophilic unconjugated dihydroxy bile acids have also been shown to flip-flop rapidly across artificial lipid bilayers (27), so that even at low concentrations unconjugated bile acids should have entered the cytosol of the mast cells. The rate of membrane permeation of glycine-conjugated bile acids is not known but could well be greater than that of taurine-conjugated bile acids, because the pK, of glycine-conjugated bile acids (3.7) is only 1.3 units lower than that of unconjugated bile acids (5.0) (28) whereas taurine-conjugated bile acids are extremely strong acids. In the hepatocyte, cytotoxic bile acids have been shown to mobilize Ca2+ from intracellular stores (2129); such is likely to have occurred in the present experiments. At concentrations close to their CMC values, bile acids are known to solubilize membranes (30); also, our observations of LDH release suggest that such membrane solubilization is largely responsible for the observed effect of histamine release in this study. Histamine release correlated with lipophilicity, as inferred from HPLC behavior; hydrophilic bile acids caused little, if any, histamine release, whereas lipophilic bile acids induced histamine release in direct relation to their degree of lipophilicity. Histamine release also correlated with the affinity of bile acids for the air/water interface. It has been known for many years that common anionic surface active agents (surfactants) are potent releasers of histamine from rat peritoneal mast cells (3 1). The cytotoxic effects of dihydroxy bile acids, such as CDC and DC, have been shown previously for hepatocytes (32,333 and for intestinal epithelial cells in perfusion experiments (34). Whenever examined, CDC or DC was more cytotoxic than UDC (32,33,35), in agreement with the findings reported here. Cytotoxicity appeared at concentrations similar to those observed in our studies. The concentration at which DC and CDC caused

GASTROENTEROLOGY

Vol. 101, No. 2

significant histamine release was about 1 mmol/L , which is likely to be in the range of monomeric concentration of bile acids in intestinal content and bile (36,37). Presumably, the apical membranes of biliary and gastrointestinal tract epithelial cells are highly resistant to the cytotoxic properties of bile acids. However, were the epithelial surface of the intestine to be disrupted by, for example, ischemia or a chemical agent such as ethanol, mast cells in the subepithelial mucosa might be exposed to relatively high concentrations of bile acids. Bile acids are also present in high concentrations in hepatic and gallbladder bile (38), in close proximity to mast cells present under the biliary epithelium (39). When bile leaks into the peritoneal cavity, as may occur with injuries to the bile duct or gallbladder, mast cells would be exposed to high bile acid concentrations; this might contribute to bile peritonitis, This study makes it highly unlikely that the effect of bile acids on mast cells is specific for any particular mast cell phenotype. Histamine release by bile acids occurred in a dose-related manner from not only isolated rat peritoneal mast cells, which are of the connective tissue phenotype, but also from PT-18 cultured mast cells, which are cytochemically similar to mast ceIIs present in the gastrointestina1 mucosa. Further, data presented in the in vivo studies demonstrated that bile acids caused histamine release from connective tissue mast cells in rat skin, in support of this view. Thus, these studies suggest that bile acids could cause mast cell cytolysis in any body site, if present in sufficient concentration. Prottey and Ferguson (31) have reported that, among inflammatory cell types, mast cells are especially sensitive to the membrane-damaging effects of common anionic surfactants; however, studies examining the susceptibility of other types of inflammatory cells to bile acids have not been performed. We conclude that the ability of bile acids to cause histamine release from mast cells is most likely explained by nonspecific cytotoxicity. Nonetheless, the combination of the susceptibility of mast cells to lysis, together with the wide range of mediators stored in this cell type, raises the possibility that bile acid-induced mast cell cytolysis could play a role in the pathogenesis of a number of hepatobiliary and digestive disease processes. Experiments to test whether mediator release from mast cells by bile acids does occur in disease and whether it is important in the progression of disease processes appear to be of interest.

References 1. Quist R, Hofmann AF, Barrett KE. Activation of mast cells by bile acids (abstr). Gastroenterology 1989;96:A691.

August

1991

EFFECT

OF BILE ACIDS ON MAST

CELLS

455

2. Barrett KE, Metcalfe DD. Mucosal mast cells and IgE. In: Jones AL, Heyworth MF, eds. Immunology of the gastrointestinal tract and liver. New York: Raven, 198865-92. 3. Rao SN. Mast cells as a component of the granuloma in Crohn’s disease. J Path01 1973;109:79-82. 4. Kirsner JB, Shorter RG. Recent developments in “non-specific” inflammatory bowel disease. Part I. N Engl J Med 1982;306:775785. 5. Juby LB, Rothwell J, Axon ATR. Cellobiose/mannitol sugar test-a sensitive tubeless test for coeliac disease: results on 1010 unselected patients. Gut 1988;30:476-480.

Hanson RF, Hersh T, Hightower NC Jr, Marks JW, Mekhjian H, Shaefer RA, Soloway RD, Thistle JL, Thomas FB, Tyor MP, the National Cooperative Gallstone Study Group. Pretreatment biliary lipid composition in white patients with radiolucent gallstones in the National Cooperative Gallstone Study. Gastroenterology 1982;83:738-752. 23. Schurtenberger P, Mazer N, Kaenzig W. Micelle to vesicle transition in aqueous solutions of bile salt and lecithin. J Phys

6. Ukabam SO, Clamp JR, Cooper BT. Abnormal intestinal permeability to sugars in patients with Crohn’s disease of the terminal ileum and colon. Digestion 1982;27:70-74. 7. Huijghebaert SM, Hofmann AF. Influence of the amino acid moiety on deconjugation of bile acid amidates by cholylglycine hydrolase or human fecal cultures. J Lipid Res 1986;27:742752. 8. Rossi SS, Converse JL, Hofmann AF. High pressure liquid chromatographic analysis of conjugated bile acids in human bile: simultaneous resolution of sulfated and unsulfated lithocholyl amidates and the common conjugated bile acids. J Lipid Res 1987;28:589-595.

vesicles by dilution of phosphatidylcholine-sodium cholate mixed micelles. Biochemistry 1986;25:2597-2605. 25. Schubert R, Schmidt KH. Structural changes in vesicle membranes and mixed micelles of various lipid compositions after binding of different salts. Biochem 1988;27:8787-8794. 26. Saito H, Sugimoto Y, Tabeta R, Suzuki S, Izumi G, Kodama M, Toyoshima S, Nagata C. Incorporation of bile acid of low

9. Barrett KE, Pluznik DH, Metcalfe DD. Histamine release from the cultured mouse mast cell line PT-18 in response to immunologic and nonimmunologic stimuli. Agents Actions 1984;14:488-493. 10. Shore PA, Burkhalter A, Coh,n VH. A method for the fluorimetric assay of histamine in tissues. J Pharmacol Exp Ther 1959;127:182-185. 11. Attili AF, Angelic0 M, Cantafora A, Alvaro D, Capocaccia L. Bile acid induced liver toxicity: relation to the hydrophobichydrophilic balance of bile acids. Med Hypotheses 1986;19:5769. 12. Armstrong MJ, Carey MC. The hydrophobic-hydrophilic balance of bile salts. Inverse correlation between reverse-phase high performance liquid chromatographic mobilities and micellar cholesterol-solubilizing capacities. J Lipid Res 1982;23:7080. 13. Heuman DM. Quantitative estimation of the hydrophilichydrophobic balance of mixed bile salt solutions. J Lipid Res 1989;30:719-730. 14. Adamson AN. Physical chemistry of surfaces. 3rd ed. New York: Wiley, 1976. 15. Mysels KJ. Improvements in maximum bubble pressure measuring surface tension. Langmuir 1986;2:428-432. 16. Mysels KJ. Some limitations in the interpretations of the time dependence of surface tension measured by the maximum bubble pressure method. Langmuir 1989;5:442-447. 17. Roda A, Hofmann AF, Mysels KJ. The influence of bile salt structure on self-association in aqueous solutions. J Biol Chem 1983;258:6362-6370. 18. Oelberg DG, Dubinsky WP, Sackman JW, Wang LB, Adcock EW, Lester R. Bile salts induce calcium uptake in vitro by human erythrocytes. Hepatology 1987;7:245-252. 19. Maenz DD, Forsyth GW. Calcium ionophore activity of intestinal secretory compounds. An in vitro porcine model for the effects of bile acids, hydroxy-fatty acids and diotyl sulfosuccinate. Digestion 1984;30:138-150. 20. Anwer MS. Hepatotoxic bile acids increase cytosolic calcium activity of isolated rat hepatocytes. Hepatology 1988;8:887891. 21. Combettes L, Dumont M, Berthon D, Erlinger S, Claret M. Release of calcium from the endoplasmic reticulum by bile acids in rat liver cells. J Biol Chem 1988;263:2299-2303. 22. Hofmann AF, Grundy SM, Lachin JM, Lan S-P, Baum RA,

Chem 1985;89:1042-1049. 24. Almog S, Kushnir T, Nir S, Lichtenberg D. Kinetic and structural aspects of reconstitution of phosphatidylcholine

27.

28.

29.

30.

concentration into model and biological membranes studied by deuterium and phosphorus-31 NMR. J Biochem 1983;94:18771887. Cabral DJ, Small DM, Lilly HS, Hamilton JA. Transbilayer movement of bile acids in model membranes. Biochem 1987;26: 1801-1804. Fini A, Roda A. Chemical properties of bile acids. IV. Acidity constants of glycine-conjugated bile acids. J Lipid Res 1987;28: 755-759. Combettes L, Berthon B, Doucet E, Erlinger S, Claret M. Characteristics of bile acid-mediated calcium release from permeabilized liver cells and liver microsomes. J Biol Chem 1989;264:156-167, Lichtenberg D, Robson RJ, and Dennis EA. Solubilization of phospholipids by detergents. Biochim Biophys Acta 1983;737:

285-304. 31. Prottey C, Ferguson TFM. The effect of surfactants upon rat mast cells in vitro. Food Cosmet Toxic01 1976;14:425-430. 32. Schoelmerich J, Becher M-S, Schmidt K, Schubert R, Kremer B, Feldhaus S, Gerok W. Influence of hydroxylation and conjugation of bile salts on their membrane-damaging propertiesstudies on isolated hepatocytes and lipid membrane vesicles. Hepatology 1984;4:661-666. 33. Miyazaki K, Nakayama F, Koga A. Effect of chenodeoxycholic and ursodeoxycholic acids on isolated adult human hepatocytes. Dig Dis Sci 1984;29:1123-1130. 34. Chadwick VS, Phillips SF, Hofmann AF. Measurements of intestinal permeability using low molecular weight polyethylene glycols [PEG 400). II. Application to normal and abnormal permeability states in man and animals. Gastroenterology 1977;73:247-251. 35. Chadwick VS, Gaginella TS, Carlson GL, Debongnie JC, Phillips SF, Hofmann AF. Effect of molecular structure on bile acid-induced alterations in absorptive function, permeability and morphology in the perfused rabbit colon. J Lab Clin Med 1979;94:661-674. 36. Duane WC. The intermicellar bile salt concentration in equilibrium with the mixed micelles of human bile. Biochim Biophys Acta 1975;398:275-286. 37. Higuchi WI, Arakawa M, Lee PH, Noro S. Simple micellemixed micelle coexistence equilibria for the taurocholate-, taurochenodeoxycholate-, and tauroursodeoxycholate-lecithin systems. J Colloid Interface Sci 1987:119:30-37. 38. Carey MC, Small DM. The physical chemistry of cholesterol solubility in bile. J Clin Invest 1978;61:998-1026. 39. Gilloteaux J, Pomerants B, Kelly TR. Human gallbladder mucosa ultrastructure: evidence of intraepithelial nerve structures. Am JAnatomy 1989;184:321-333.

456

QUIST ET AL.

40. Mukerjee P, Moroi Y, Murata M, Yang AYS. Bile salts as atypical surfactants and solubilizers. Hepatology 1984;4:61S65s.

Received December 12, 1989. Accepted October 5,199O. Address requests for reprints to: Kim E. Barrett, Ph.D., University of California, San Diego Medical Center, H-81lG, 225 Dickinson Street, San Diego, California 92103. This study was supported in part by grants from the National Institutes of Health (DK 21506, DK 32130, and A124992) as well as grants-in-aid from the Falk Foundation e.V., Freiburg, and Diamalt AG, Raubling, Germany; Medstone International, Inc., Irvine, CA, and Ciba-Geigy, Summit, NJ. This study was presented in part at the annual meeting of the AmericanGastroenterological Association, Washington, DC., 1989,

GASTROENTEROLOGY Vol. 101, No. 2

and has been published in abstract form (Gastroenterology 1989;96: A691). Dr. Quist was a recipient of a summer fellowship from the American Gastroenterological Association. His present address is: School of Medicine, George Washington University, 2150 Pennsylvania Avenue Northwest, Washington, DC. 20037. For questions regarding mast cells, contact Kim E. Barrett, Ph.D.; for questions relating to bile acids, contact Alan F. Hofmann, M.D., Ph.D., Division of Gastroenterology, 0813, Department of Medicine, University of California, San Diego, La Jolla, CA 92093. The authors acknowledge the assistance of Dr. Claudio Schteingart for his thought-provoking advice and Vicky Huebner for her assistance in preparing the manuscript. The authors also acknowledge the anonymous reviewers whose criticisms have greatly improved their thinking and this paper.