The invention relates to the treatment of antibiotic as¬ sociated colitis, typically due to Clostridium difficile usin Platelet Activating Factor antagonists, such as WEB 2170, SR 27417 or BN 52021, or the cyclooxygenase antagonists, such as indomethacin. The PAF antagonists BN 52021 and SR 27417 and the cyclooxygenase antagonist indomethacin were effective in inhibiting the secretory effects caused by C. difficle Toxin and even by Cholera toxin. Background of the Invention

Clostridium difficile is one of the most frequently recog¬ nized bacterial causes of diarrheal disease in hospitalized adults in industrialized countries. The microorganism can be acquired nosocomially and is present in environmental sources Antibiotic associated colitis and pseudomembranous colitis ar frequently associated with cytotoxigenic Clostridium difficil The frequency of Clostridium difficile toxin associated with antibiotic associated colitis is 50-80% and with pseudo¬ membranous colitis is 90-100%. Despite available treatment f antibiotic associated colitis and pseudomembranous colitis, relapses occur in 20-25% of patients. Vancomycin and metronidazole can be effective, but are subject to relapse after the use of the drugs may occur.

Clostridium difficile produces two toxins, A (enterotoxic) and B (cytopathic) , the former (Toxin A) being implicated in the pathogenesis of pseudomembranous colitis. Toxin A causes

hemorrhagic fluid accumulation associated with mucosal damage and a cytopathic effect in tissue culture cells. In experimen¬ tal animals such as rabbits, toxin A causes hemorrhagic fluid secretion and cell damage in ligated intestinal segments of loops, and is considered to be the cause of antibiotic- associated colitis in experimental hamsters and in people.

Toxin B is a more potent cytotoxin than toxin A, but it has no enterotoxic activity. It has been reported that the toxin A is important in the pathogenesis of Clostridium difficile dis¬ ease.

In animal models of the disease, Clostridium difficile culture supernatant filtrate or purified toxin A causes mucosal lesions characteristic of antibiotic-associated colitis. In human dis¬ ease, as well as in animal models, a spectrum of mucosal changes ranging from nonspecific colitis (with or without diar¬ rhea) to severe disease with complete mucosal necrosis has been described in prior literature.

Antibiotics, including vancomycin or metronidazole, directed against C. difficile, or anion exchange resins that bind the toxins of C. difficile, including cholestyramine, have been used to treat C. difficile diarrhea. However, the antibiotics prolong carriage of C. difficile and relapses are common and resins are not always well tolerated.

PAF inhibitors have been have been found to be effective in treating Inflammatory Bowel Disease (IBD) when administered 4 - 7 days after the induction of colitis. The PAF inhibitor BN52021, however, shows no effect in the first four days. Chemical Abstracts 110:190302k (1989). IBD is distinct from

the specific Clostridium difficile toxin-induced antibiotic a sociated acute colitis. IBD, including Crohn's disease, is a chronic, recurring, granulo atous inflammatory process that i largely related to host susceptibility. C. difficile diarrhe is a specific, acute, toxin-mediated diarrhea that can occur any individual and does not recur over many years as is chara teristic of IBD. Clostridium difficile is defined as "a speci that is part of the normal colon flora in human infants and sometimes in adults. It sometimes produces a toxin that caus pseudomembranous enterocolitis in patients receiving antibiot therapy" Dorland's Illustrated Medical Dictionary. W. B. Saunders Co., Philadelphia, Pa., 27th Edition, 1988, p. 348. Pseudomembranous enterocolitis is defined as "an acute inflam tion of the bowel muscosa with the formation of pseudo¬ membranous plaques overlying an area of superficial ulceratio and the passage of the pseudomembranous material in the feces; it may result from shock and ischemia or be associated with a tibiotic therapy. Dorland s Illustrated Medical Dictionary, supra, p. 560. Crohn's disease is defined as "a chronic granulomatous inflammatory disease of unknown etiology, invol ing any art of the gastrointestinal tract from mouth to anus, but commonly involving the terminal ileu with scarring and thickening of the bowel wall; it frequently leads to intestin obstruction and fistula and abscess formation and has a high rate of recurrence after treatment." Dorland's Illustrated Medical Dictionary, supra, p. 484. IBD is, to an extent, susc tible to control by diet, fitness and stress reduction. Eati

Right for a Bad Gut: The Complete Nutritional Guide to Ileitis. Colitis. Crohn's Disease & Inflammatory Bowel Disease, Dr. James Scala, Penguin.

SUMMARY OF THE INVENTION

It has now been found that the hemorrhagic fluid secretion caused by Clostridium difficile toxin A in ligated intestinal segments can be significantly blocked by administering an effe tive amount of either a Platelet Activating Factor (PAF) in¬ hibitor, or by a cyclooxygenase inhibitor, or by the combina¬ tion of inhibitors. The PAF inhibitors are from the group con sisting of BN52021, WEB 2170 or SR27417. The phospholipase ₂ inhibitor is from the group consisting of aristolochic acid, bromophenacyl bromide and quinacrine. The cyclooxygenase in¬ hibitor is indomethacin.

The fluid secretion caused by Cholera toxin in ligated inte tinal segments can be blocked by administering an effective amount of a Platelet Activating Factor inhibitor selected from the group consisting of BN52021 and SR27417 or the cyclooxygenase inhibitor indomethacin. The combination of the cyclooxygenase inhibitor indomethacin and at least one Platele Activating Factor antagonist from the group consisting of SR 27417 and BN 52021 also blocks the hemorrhagic fluid secretion caused by Cholera toxin.

BRIEF DESCRIPTION OF THE DRAWINGS

The specification and disclosed invention will be better un derstood when read in reference to the drawings in which:

FIGURE 1 is a graph illustrating the effects of Aristolochi Acid on C. difficile toxin A in rabbit ileum;

FIGURE 2 is a chart illustrating the effects of Aristoloch Acid on C. difficile toxin A in rabbit ileal ligated loops;

FIGURE 3 is a graph illustrating the effects of PAF in¬ hibitors (BN, WEB, SR) and indomethacin on C. difficile toxin in rabbit ileum;

FIGURE 4 is a chart illustrating the effects of PAF in¬ hibitors (BN, WEB, SR) or indomethacin on C. difficile toxin in ligated rabbit ileal loops;

FIGURE 5 is a graph illustrating the effects of Quinacrine on C. difficile toxin A in rabbit ileum;

FIGURE 6 is a chart illustrating the effects of Quinacrine on C. difficile toxin A in ligated rabbit ileal loops;

FIGURE 7 is a chart illustrating the effects of Quinacrine on C. difficile toxin A in T84 cell monolayers;

FIGURE 8 is a graph illustrating the effects of Quinacrine on C. difficile toxin A in T84 cell monolayers;

FIGURE 9 is a graph illustrating the effects of PAF in¬ hibitors (BN and SR) and indomethacin on cholera toxin in rab bit loops; and,

FIGURE 10 is a graph illustrating the effects of bromophenacyl bromide on C. difficile toxin A in rabbit ileum. DETAIL DESCRIPTION OF THE PREFERRED EMBODIMENTS

The hemorrhagic fluid secretion caused by Clostridium dif¬ ficile toxin A in ligated intestinal segments can be sig¬ nificantly blocked by either Platelet Activating Factor (PAF) inhibitors BN52021, WEB 2170, SR27417; phospholipase A ₂ in-

-6-

hibitors, Aristolochic Acid, Quinacrine, and bromophenacyl bromide or by cyclooxygenase inhibitor indomethacin, or by th combination of BN 52021 or WEB 2170 and indomethacin.

CRC Handbook of PAF and PAF Antagonists. Edited by Braquet CRC Press,-Boston Mass., identifies WEB 2170 on page 51. The ginkgolide BN 52021 (IHB nomemclature) is illustrated on page 52 of the CRC Handbook, the subject matter of the CRC Handboo being incorporated herein by reference as though completely reproduced.

The structural formula for BN 52021, with R ¹ and R ² equal and R- ³ equals H, is as follows:

The structural formula for WEB 2170 is as follows:

The availability of highly purified Clostridium difficile toxin A, has allowed accurate, reproducable study of the specific effects involved in early mucosal damage and its cor relation with the intestinal response. The preparation of toxin A and the use of toxin A in studies in ligated small in testinal segments and other studies, is disclosed in Laborato Investigation. Vol. 61, No. 4, 1989, pages 419 through 425, Clostridium difficile Toxin A, Lima et al., the subject matte of which is incorporated herein by reference, as though com¬ pletely reproduced.

The instant disclosure conducted tests using: (1) platelet activating factor (PAF) inhibitors, BN52021 (BN) , WEB2170 (WE and SR27417 (SR) ; (2) the phospholipase A ₂ inhibitors, aris¬ tolochic acid (AA) , Quinacrine (QN) and bromophenacyl bromide (BPB) ; and (3) cyclooxygenase inhibitor, indomethacin (INDO) . These were tested both in vivo in the rabbit ileal ligated lo model and in vitro in monolayers of T-84 cells in tissue cul¬ ture (a colonic epithelial carcinoma cell line) to determine the effects of these inhibitors on the cytotoxicity and secretory responses due to C. difficile toxin A.

EXPERIMENT I

Clostridium difficile toxin A alone and in combination wit the inhibitors were placed in 4-6 cm ligated Ileal segments i 1.5-2.5 kg, male, New Zealand white rabbits. The rabbits were fasted overnight and provided water ad libitum. The animals were anesthetized with an initial intramuscular dose of ketamine (35mg/kg body weight) and xylazine (5mg/kg) . Ligate rabbit small intestinal segments of approximately 5 cm were

made with double ties of umbilical tape. These segments were injected with the following combinations and then returned to the abdominal cavity without separation from the remaining in¬ testine or interruption of the blood supply. After six (6) hours the animals were sacrificed and the fluid volume ( illiliters) and length (centimeters) for each loop was measured and a volume to length ratio calculated. Toxin A 10 μg/segment in 1 ml consistently caused severe hemorrhagic in¬ flammatory fluid secretion at 6-8 hours (V/L=0.64± 0.13 ml/cm segment, n=6) . The molarity (M) indicates the drug concentra¬ tion in the 1 ml containing the lOμg/ml Toxin A. Phosphate buf fered saline (PBS) is used as a control.

Each rabbit was injected with one of the following: lml toxin A in PBS (10 μg/ml) lml PBS (as a control) , lml toxin A and BN52021 (10 ^"4 M and 10 ^"5 ) [BN] lml toxin A and WEB2170 (10 ^"5 M and 10 ^"6 M) [WEB] lml toxin A and SR27417 (10~ ⁵ and 10 ^~6 M) [SR] lml toxin A and Aristolochic acid (10~ ⁵ M) [AA] lml toxin A and a combination of quinacrine (10~ ⁵ M) [QN] , indomethacin (10 ^"5 M) [INDO] lml toxin A and bromophenacyl bromide(10~ ⁵ ) [BPB]

0.5ml toxin A (20μg/ml) plus 0.5ml of BN(2xlθ~ ⁴ M and 2xlθ ^"5 )

0.5ml toxin A and WEB(2xlθ ^"5 M and 2xl0~ ⁶ ) 0.5ml toxin A and SR(2xlO~ ⁵ M and 2xl0 ^~6 ) 0.5ml toxin A and AA(2xlθ ^"5 M) 0.5ml toxin A and QN(2xlθ ^"5 M)

0.5ml toxin A and INDO(2xlθ ^"5 M)

The toxin A 10 μg/segment hemorrhagic inflammatory fluid secretion was significantly inhibited by 10 ^~5 M AA, a phos¬ pholipase A ₂ inhibitor. AA provided an 87% inhibition, (p<0.0001, n=6) . The secretory effect of toxin A was also in¬ hibited in ligated segments adjacent to those with AA (89% in¬ hibition, p≤O.0001, n=6) . The results of these tests are il¬ lustrated in the graph of Figure 1 and the chart of Figure 2.

The PAF inhibitors, BN (10 ^"4 M and 10 ^"5 M) , WEB (>10 ⁶ M) or th cyclooxygenase inhibitor, INDO (10 ⁵ M) , both significantly in¬ hibited the secretory effects of toxin A in ligated but not ad jacent rabbit intestinal segments (71% inhibition for BN; 84% for WEB; 79% for INDO; all p<0.0001) as noted in Figures 3 and 4. illustrate the test results using p factors of <0.001 and <0. SR gave a reduction of 49% (p<0.001) in the same loops an a reduction of 48% (p<0.01) in the adjacent loops.

The secretory effects of C. difficile toxin A with ligated ileal segments were significantly inhibited by phospholipase A inhibitor, QN (p<0.05), in both the same or the adjacent loops with toxin A. The test results where p<0.015 and p<0.008 are illustrated in Figures 5 and 6.

The phospholipase A ₂ pathway is involved in the secretory responses to C. difficile toxin A, via the subsequent genera¬ tion of prostaglandin and/or PAF. The prostaglandin is decreased by the phospholipase The generation of prostaglandi is reduced by the PLA ₂ or cyclooxygenase inhibitors.

The effects of BPB on the D. difficile toxin A are il¬ lustrated in Figure 10, where p<0.01 and n=4.

The numeric results of the experiments with BN, INDO and WE

The hemorrhagic fluid secretion caused by Clostridium dif¬ ficile is significantly blocked by two Platelet Activating Fac tor (PAF) inhibitions, BN 52021 (10 ⁴ M) or WEB 2170 (<10 ⁶ M) or by the cyclooxygenase inhibitor indomethacin (10~ ⁵ M) ; inhibi¬ tion was 62% with BN52021, 74% for WEB 2170, 65% for in¬ domethacin, all p<0.05). PAF inhibitors, alone or in combina tion with cyclooxygenase inhibitors, provide a means to block the common and often severe antibiotic-associated diarrhea an colitis caused by Clostridium difficile.

EXPERIMENT II

T84 cells, a human intestinal epithelial cell line, were grown, passaged, then grown on the 4-well LAB-TEK tissue cul¬ ture chamber slides (Nunc, Inc., Naperville, II) to confluency C. difficile toxin A 0.7 μg/ well (1 ml, final concentration 0.7 μg/ l) alone was placed onto a T84 cell monolayer. Addi¬ tional cell monolayers were applied with equal concentration o toxin A in combination with:

1 ml BN(10 ^~4 M) ,

1 ml WEB(10~ ⁴ M)

1 ml AA(5xlO ^"4 M)

1 ml QN(5xlO ^"4 M)

1 ml BPB(10 ^~4 M)

1 ml INDO(10 ^"4 M)

Cell monolayers were prepared with toxin A 0.7 μg (0.5 ml o 1.4 μg/ml). Additional cells were prepared with the same toxi A concentration in combination with:

0.5 ml BN(2X10 ^"4 M)

0.5 ml WEB(2xlO ^"4 M) or

0.5 ml AA(10 ^"4 M)

0.5 ml QN(10 ^~3 M)

0.5 ml BPB(2X10 ^"4 M)

0.5 ml INDO(2xlO~ ⁴ M)

After eight (8) hours incubation, the cell-coated chamber slides were rinsed in PBS and fixed for 15 minutes at room tem perature in 3.7% formaldehyde. The slides were rinsed again i PBS, per eabilized for five (5) minutes in 20°C acetone, air dried, stained with rhodamine-labeled phalloidin for 30 minute at room temperature in the dark, rinsed twice in PBS and kept

in the dark at 4°C. Cells were then examined and photographed using an inverted Olympus fluorescence microscope. To compare the pattern of labeling between toxin-treated and PAF in¬ hibitors or phospholipase A ₂ inhibitors or cyclooxygenase in- . hibitor plus toxin-treated mono-layers, photomicrographs were taken with Kodak 400, daylight slides and exposures were varied to optimize the resulting images.

The destruction by toxin A (0.7 μg/ml) of rhodamine-labeled phalloidin-stained F-actin at the tight junctions of T-84 cell monolayers was inhibited by 5xlθ ^"4 M AA or 10~ ⁴ M BPB (n=2-4 experiments each) . BN, INDO, WEB, and QN also inhibited the reduction in F-acting staining by toxin A.

EXPERIMENT III

T84 cells, a human intestinal epithelial cell line, were grown, passaged and mounted on collagen-coated microfilters. Transepithelial resistances generated by the ring-mounted monolayers used for these studies ranged from 2600 to 4000Ω

When the T84 cells grown on the microfilters reached con¬ fluence as determined by phase microscopy and development of a high, stable resistance, toxin A 0.02μg(10μl, final concentra¬ tion 0.07μg/ml), or 36μl QN(2xlθ ^"4 M, -final concentration 2x10 ^{" 5} M) , or toxin A 0.02 μg plus 36 μl QN(2xlO~ ⁴ M) were added into 300 μl tissue culture medium. All above reagents were applied only to the apical (mucosal) surface of the monolayers. Resis¬ tance was measured every hour during the first eight hours and then at 10, 12 and 24 hours.

The results of the QN resistance are shown in Figures 7 an 8. Simultaneous administration of QN (p<0.05) additionally s nificantly inhibited tissue resistance in the T84 cell monolayers caused by toxin A at 5-6 hours exposure.

The foregoing PAF inhibitors, BN and SR and Indomethacin, have been found to significantly reduce the enterotoxin Chole toxin (choleragen) produced by Vibrio cholerae, as well as C. difficile. Choleragen acts on the epithelial cells in the small intestine to cause secretion of large quantities of isotonic fluid from the mucosal surface. Choleragen is af¬ fected by the PAF inhibitors similarly to the toxin A created by Clostridium difficile. The Cholera related diarrhea resul in massive gastrointestinal fluid loss and saline depletion, acidosis and shock. The ability to control the effects of th diarrhea would allow the body to absorb badly needed fluids, nutrients and medication, and eliminate the further depletion of the body's resources and resulting deficiencies.

EXPERIMENT IV

Cholera toxin, alone and in combination with the inhibitor BN, SR and indomethacin, were placed in 4-6 cm ligated Ileal segments in 1.5-2.5 kg, male, New Zealand white rabbits. The rabbits were fasted overnight and provided water ad libitum. The animals were anesthetized with an initial intramuscular dose of ketamine (35mg/kg body weight) and xylazine (5mg/kg) . Ligated rabbit small intestinal segments of approximately 5 c were made with double ties of umbilical tape. These segments were injected with the following combinations and then return to the abdominal cavity without separation from the remaining

intestine or interruption of the blood supply. After six (6) hours the animals were sacrificed and the fluid volume (milliliters) and length (centimeters) for each loop was measured and a volume to length ratio calculated. Choleragen 10 μg/segment in 1 ml consistently caused severe hemorrhagic inflammatory fluid secretion at 6-8 hours (V/L=0.64+0.13 ml/cm segment, n=6) . The molarity (M) indicates the drug concentra¬ tion in the 1 ml containing the lOμg/ml choleragen.

Each rabbit was injected with one of the following: lml choleragen in PBS (lμg/ l) lml PBS (as a control) , lml choleragen and BN52021 (~ ⁴ M and 10 ^"5 M) [BN] lml choleragen and SR27417 (10~ ⁵ and 10 ^"6 M) [SR] lml choleragen and a combination of BN52021 (10 ^~5 M) [BN] , indomethacin (10 ^~5 M) [INDO]

0.5ml choleragen and SR(2xlO~ ⁵ M and 2xl0~ ⁶ )

0.5ml choleragen and BN(2xlO ^~5 M)

0.5ml choleragen and INDO(2xlO~ ⁵ M)

The choleragen Iμg/segment hemorrhagic inflammatory fluid secretion was significantly inhibited by 10 ₅ - 10 ⁶ M SR27417. SR provided an 86% inhibition when placed in the loop with the toxin, as well as a slightly lesser inhibition when in the loo adjacent to the toxin, (both p<0.011), as illustrated in Figur 9.

The PAF inhibitors, BN (10 ^"4 M and 10 ^"5 M) , and INDO (10 ⁵ M) , both significantly inhibited the secretory effects of toxin A in ligated intestinal segments (44% inhibition for BN; 38% for INDO, both p<0.01) .