Exogenous alkaline phosphatase for the treatment of patients with moderate to severe ulcerative colitis

Abstract

Background:

Increased activity of intestinal alkaline phosphatase (AP) occurs locally in patients with ulcerative colitis (UC), aimed at repairing inflammatory tissue damage. We evaluated the safety and preliminary efficacy of exogenous AP administered to patients with UC in an open-label, first-in-patient exploratory trial, conducted in the Internal Medicine and Gastroenterology hospital departments in the Czech Republic and Italy.

Methods:

Twenty-one patients were enrolled (13 females), age 23–54 years, with steroid- and/or immunosuppressant-refractory, moderate/severe UC (Mayo score 6–11). Oral AP enzyme 30,000 U was administered daily for 7 days, intraduodenally. Efficacy outcomes were changes in Mayo score at Day 21 posttreatment; changes in Modified Truelove–Witts Severity index (MTWSI) at Days 21, 63; C-reactive protein and stool calprotectin levels at Days 7, 21, 63. Safety evaluations were adverse events and laboratory abnormalities reported up to Day 63 posttreatment.

Results:

No clinically relevant adverse events causing withdrawal or considered serious, or laboratory abnormalities or antibody formation against AP were observed. Mayo scores were significantly decreased at Day 21, and MTWSI at Days 21 and 63. C-reactive protein and stool calprotectin levels were decreased at Days 21 and 63. Clinical response on the Mayo score after a single 7-day AP course was 48% at Day 21.

Conclusions:

In this uncontrolled trial, administration of exogenous AP enzyme daily over a 7-day course to patients with UC was associated with short-term improvement in disease activity scores, with clinical effects being observed within 21 days and associated with reductions in C-reactive protein and stool calprotectin. AP enzyme treatment was well tolerated and nonimmunogenic. (Inflamm Bowel Dis 2009;)

Alkaline phosphatase (AP) is present in several human cells and organs (e.g., intestines, kidney, placenta, liver, bone, neutrophils),1 as in nature AP is 1 of the most highly prevalent enzymes,2 from bacteria3 to marine life4 and mammals.5 Yet beyond its general dephosphorylating properties,6, 7 the role of this enzyme is largely unknown. Recently, AP has been shown to act in host defense and inflammatory reactions through dephosphorylation of extracellular adenosine triphosphate (ATP)8, 9 and lipopolysaccharide endotoxin (LPS) released by bacteria leading to detoxification.10, 11

It has been amply demonstrated that AP dephosphorylates LPS,12–16 resulting in the formation of a nontoxic lipid A group within the LPS molecule. In general, the lipid A group of LPS contains 2 phosphate groups that are responsible for the toxicity of LPS and AP removes at least 1 of these phosphate groups. This enzyme is abundantly present along the microvilli in the small intestine of all species,17 indicating a possible role in the protection of the host against LPS in colitis.18

Intravenous administration of exogenous AP to healthy volunteers and patients has been shown to be well tolerated,19, 20 which confirmed literature reports that the effects of elevated levels of endogenous enzyme denote tissue effort to repair damage while elevated enzyme by itself is neither toxic nor damaging.21, 22

In ulcerative colitis (UC), damaged intestinal mucosa facilitates the influx of LPS or bacterial gut translocation, resulting in activation or worsening of intestinal inflammation.23–25 Additionally, circulating levels of LPS have been documented in patients with UC,23, 24 with increased local activity of intestinal AP also occurring in these patients,18 while locally available intestinal AP has been shown to protect mice intestinal tissue from inflammatory injury in a chronic colitis model.26, 27

The aim of the present open-label exploratory trial was to assess the safety and preliminary efficacy of exogenous intestinal AP enzyme administered intraduodenally to patients with UC and thus provide a basis for future studies of AP as a potential biologic approach in the treatment of inflammatory bowel disease (IBD).

MATERIALS AND METHODS

Protocol

This was an exploratory, open-label study, as defined by the International Conference on Harmonization's General Considerations for Clinical Trials,28 study conducted at 6 departments of General Medicine and Gastroenterology, from May to October 2006.

Patients were screened for eligibility within 5 days prior to enrolment. The inclusion criteria were: age between 18–60 years, ability to provide written informed consent, confirmed diagnosis of active UC, classified according to the Montreal IBD classification29 with Mayo score 6–11 and Modified Truelove–Witts Severity index (MTWSI) score 7–15, and 1) inability to reduce daily corticosteroid dose below 40 mg oral prednisone-equivalent, unchanged for a minimum 2 weeks preentry; or 2) clinical inability to decrease or stop the course of oral steroid medication for a minimum of 12 weeks pretreatment; or 3) moderate to severe relapse (Mayo score 6–11 and MTWSI score 7–15) despite treatment with a stable dosage of azathioprine for a minimum of 12 weeks pretreatment, with or without corticosteroids. The following protocol-stopping rules applied: either by the patient's own request with or without giving reason, or by the investigator in the case of a serious adverse event or worsening symptoms that were considered by the investigator to preclude continued participation in the study. We used both Mayo and MTWSI because this was a small uncontrolled study in which we wished to obtain more detailed information on clinical symptoms provided by MTWSI as well as sigmoidoscopy data provided by the Mayo score.

The following therapies were allowed: azathioprine or 6-mercaptopurine, if started ≥12 weeks before entry and if the dose remained constant for at least 4 weeks prior to entry; 5-aminosalicylates in stable doses for at least 4 weeks prior to entry. There was no mandatory steroid-tapering regimen; records of daily doses were implemented to ascertain whether any improvement/deterioration in disease activity might be related to increase/decrease in doses.

The exclusion criteria were: UC requiring immediate surgical intervention; massive hemorrhage, perforation, sepsis, suppurative complications (intraabdominal or perianal abscess), toxic megacolon; previous large bowel surgery; positive stool culture including Clostridium difficile; clinically significant other major organ dysfunction; pregnancy or breast feeding, preentry positive pregnancy test (compulsory for fecund females) or inadequate contraception. The use of probiotics, antibiotics, methotrexate, cyclosporine, infliximab, tacrolimus, or any experimental therapy was not allowed within 8 weeks prior to entry.

Study Medication

Exogenous AP enzyme for oral administration was supplied by AM-Pharma (Netherlands) as calf intestinal AP in glycerol solution (CAS Registry: 9001-78-9) in 20-mL vials containing AP 14,500 U/mL (specific activity: 1000 U/mg protein), compliant with European Union's guideline on animal-derived products in medicines.30 Since this was the first administration of exogenous AP enzyme enterally to patients, the dose was derived from previous safety data with intravenous administration in human volunteers19 and from preclinical data on systemic absorption of oral AP in a piglet model.31 The dose was set at 30,000 U/24h (428 U/kg) assuming a 5-fold (piglet/human in U/kg) safety margin. A Phase I ascending dose study was conducted in volunteers (unpubl. data) to ascertain safety of the current dose prior to our patient trial. Prior to administration patients underwent placement of a naso-duodenal tube with radiological control for appropriate positioning. Delivery of medication into the naso-duodenal tube was through a fixed-rate (20–30 drops per minute for 24 hours) infusion pump for 7 days.

Study Design and Statistics

The present study was designed to assess the safety and preliminary efficacy of enteral administration of exogenous AP enzyme. The sample size was determined as a minimum of 20 patients to allow safety evaluations, while also allowing evaluation of Mayo score changes. Thus, the sample size was primarily aimed to assess safety, rather than response/remission rates. We based our efficacy assumptions on recent results in moderate-severe UC,32 whereby a median difference of 3 points on the Mayo score (mean difference >2.5) would be considered clinically relevant, while a standard deviation in the range 2.5–2.8 would allow for variability associated with the small sample size. With these assumptions, for 2-tailed testing a sample size of n = 20 would provide a power >90%, if the assumptions were correct. Since this was an exploratory study, there were no alpha-adjustments for multiple testing. The protocol also stipulated that although estimation of response and remission rates could not be robustly estimated after a single treatment course, it would be conducted according to standard definitions.33 All patients were admitted to the hospital 24–48 hours prior to treatment start, remained hospitalized for the duration of the 7-day treatment period, and were followed-up to Day 63 as outpatients. Treatment efficacy was assessed using the following activity scores: 1) the Mayo score,32, 34 measured at baseline (average of 3 days pretreatment) and Day 21; 2) MTWSI (Lichtiger score; 21-point based)33, 35 measured at baseline, Days 21 and 63.

Response was defined as a decrease in Mayo score of at least 3 points with a decrease of at least 1 point in rectal bleeding, thus resulting in a final 0–1 bleeding score. Remission was defined as a total score ≤2 points with no individual subscore item exceeding 1 point.32

Endoscopic investigations (proctosigmoidoscopy) were performed at baseline and Day 21 by the attending physicians. To minimize interobserver variability, all procedures were recorded photographically and later reassessed centrally by a single gastroenterologist (M.L.) as per Mayo Score Flexible Proctosigmoidoscopy Assessment.35 To confirm disease features, histopathology samples obtained during endoscopic examination were also evaluated according to established criteria36 centrally by a single pathologist (Dr. I. Vitkova).

Blood samples were taken to determine levels of C-reactive protein (baseline, Days 7, 21, and 63) and serum cytokines (Days 1 and 7): tumor necrosis factor (TNF)-α, interleukin (IL)-1, IL-1β, IL-4, IL-6, IL-8, and IL-10 (BioSource enzyme-linked immunosorbent assay [ELISA] kits; Invitrogen, Carlsbad, CA). Stool calprotectin levels were measured at baseline, Days 7, 21, and 63. Except for C-reactive protein, all other parameters were analyzed centrally (TNO Laboratories, Zeist, Netherlands). Serum AP activity was evaluated by kinetic assay as described by Beumer et al.15

Safety and tolerability evaluations were based on adverse events recorded from the start of treatment up to Day 63 and on laboratory tests (hematology, biochemistry, qualitative urinalysis, chest and abdominal x-rays, 12-lead electrocardiogram) performed before, during, and after treatment cessation (on Day 8). Antibodies against exogenous AP enzyme were measured at baseline and at 3 timepoints (Days 7, 14, and 63) using validated ELISA methods for IgA, IgE, and IgG-anti-AP by a central laboratory (TNO Laboratories).

Patients were assigned treatment sequentially after providing consent and undergoing screening including preentry tests. Although this was the first ever study in patients, the protocol stipulated that all patients who received any study medication would be evaluated on an intent-to-treat (ITT) basis. A per-protocol analysis was also conducted. The protocol planned a minimum of 20 patients to be enrolled for the evaluation of safety and preliminary estimation of efficacy. All data were externally analyzed by CRM Biometrics (Rheinbach, Germany) using SAS (v. 8.02 for Windows, Cary, ND). Means, standard deviations, medians, ranges, and upper and lower quartiles were calculated. Absolute and relative frequencies were determined for qualitative data. Parameters were listed by subject, summarized and evaluated using descriptive statistics. Changes from baseline for efficacy variables were assessed by means of t-tests for paired data in an exploratory manner without α-adjustments.

Ethical Considerations

The protocol was approved by the Ethics Committees of all participating hospitals. All patients provided written informed consent prior to enrolment into the study. The study was conducted in accordance with current Guidelines on Good Clinical Practice,37 and centrally registered (http://clinicaltrials.gov/; NCT00727324).

RESULTS

Patient Characteristics and Progress Through Study

A total of 22 patients were screened. One patient failed the preentry screening because of imminent surgery and disallowed medication, thus 21 patients were included (Table 1). Two patients were subsequently found to have violated the protocol (1 had received cyclosporine outside the allowed time interval and 1 was found to have positive stool culture for C/difficile). However, both completed the study treatment course and were included in ITT analysis. As per entry criteria, at baseline 9 patients were on: 1) oral prednisone equivalent ≥40 mg daily systemic steroids for ≥2 weeks (mean: 59.4 mg; range: 40–95 mg); 2) 5 patients were on <40 mg steroids daily unable to taper for ≥12 weeks (mean: 25.0 mg; range: 15–30 mg); and 3) 7 patients had been on stable azathioprine (AZA) ≥12 weeks (mean: 107.1 mg) and 20.0 mg (steroid); ranges: 100–125 mg (AZA) and 10–50 mg (steroid). One patient defined as criterion (3) also fulfilled criterion (1). All patients also had the required Mayo scores and MTWSI for entry. Patient characteristics at entry are summarized in Table 1. There were no treatment interruptions or requirements for dose decreases during the study.

Table 1. Patient Enrolment and Characteristics at Entry

ITT, intent-to-treat; BMI, body mass index; MTWSI, modified Truelove-Witts severity index; Mean (SD), arithmetic mean (standard deviation); Q1, Q3, 25%, 75% quartiles.
Screened: n	22
Enrolled (safety/efficacy: ITT): n	21
Females/males: n	13/8
Age: mean years (SD)	35 (9)
Q1–Q3	27–43
BMI: mean Kg/m^2 (SD)	21 (2)
Q1–Q3	19–23
Current smoker: n (%)	3 (14)
Extent of bowel involvement: n (%)
Pancolitis:	15 (71)
Distal colitis:	5 (24)
Proctitis:	1 (5)
Baseline Mayo scores: mean (SD)	7.8 (1.3)
Q1–Q3	7–9
Baseline MTWSI scores: mean (SD)	10.6 (2.9)
Q1–Q3	8–13
Concomitant medications: n
Mesalazine	21
Corticosteroids	20
Azathioprine	7

Safety

Thirteen of 21 patients reported 20 adverse events of mild to moderate severity (Table 2). No serious adverse events or events leading to treatment withdrawal were reported. The events of hypoglycemia (in Table 2) were investigated and were found to have been present before the study started (at baseline), the values remaining either below or within normal range during treatment and follow-up. No other clinically relevant deviations in laboratory parameters, radiological investigations, or electrocardiography were recorded. Evaluation of antigenicity of exogenous AP enzyme did not reveal any antibody formation. Mean serum levels of AP remained stable, within the normal range, with minimal variations (within 5 U/L in daily means) measured at 9 timepoints between baseline and Day 63.

Table 2. Safety and Tolerability Results (n = 21)

	n	%
* Includes multiple events in single patients, regardless of relationship to treatment.
Patients reporting adverse events	13	61.9
Total reported adverse events*	20
Ulcerative colitis symptoms	3	14.3
Flatulence	3	14.3
Hypoglycemia	3	14.3
Arthralgia	2	9.5
Epigastric pain	1	4.8
Odynophagia	1	4.8
Vomiting	1	4.8
Pyrexia	1	4.8
Oral candidiasis	1	4.8
Pharyngitis	1	4.8
Body weight loss	1	4.8
Headache	1	4.8
Insomnia	1	4.8

Efficacy

A significant decrease in Mayo score was observed at Day 21; similarly, a significant reduction in MTWSI was found at Day 21 and sustained up to Day 63 (Fig. 1).

Figure 1. Changes in UC disease activity (n = 21).

*: Arithmetic means (Standard deviation). MTWSI: modified Truelove-Witts severity index. p-value: t-test of change from baseline. Medians (Q1-Q3): Baseline-Mayo = 8.0 (7.0–9.0); Baseline-MTWSI = 11.0 (8.0–13.0); Day 21-Mayo = 5.0 (4.0–7.0); Day 21-MTWSI = 6.0 (5.0–8.0); Day 63-MTWSI = 6.0 (2.0–7.0). Mayo score range: 0-12. Higher scores indicate more severe disease. Each patient serves as his/her own control to establish degree of abnormality of stool frequency. Daily bleeding score represents most severe bleeding of the day. Physician's global assessment acknowledges the three other criteria, the patient's daily recollection of abdominal discomfort and general sense of wellbeing, and other observations, such as physical findings and the patient s performance status. MTWSI range: 0–21. Higher scores indicate more severe disease. Individual items: diarrhea; nocturnal diarrhea; visible blood in stool; fecal incontinence; abdominal pain/cramping; general well-being; abdominal tenderness; need for antidiarrheal drugs. Each symptom is calculated a three-day average covering the period immediately preceding each assessment.

At Day 21, 48% (10/21) of patients had a clinical response, while 19% (4/21) were in clinical remission and 33% (7/21) did not respond, as per total Mayo score prospectively defined. The ITT and per-protocol analyses yielded similar results. Likewise, using local or central rating of proctosigmoidoscopy images did not alter the results. Daily corticosteroid requirement decreased gradually during the study up to Day 63 (Table 3). Histopathology samples (pre- and posttreatment, evaluated by a single pathologist and used primarily to reduce bias due to the uncontrolled design) were available for 13 patients and were in agreement with individual response to treatment and sigmoidoscopies, particularly in relation to inactive/quiescent ulcerative colitis without active mucosal inflammatory cell infiltration and crypt abscesses or ulcerations. Evaluation of inflammatory parameters showed reductions in C-reactive protein levels and stool calprotectin at Days 21 and 63, but not at Day 7 (Table 4). Cytokines were measured only up to Day 7 and did not show changes in this time period, except for a trend toward a decrease in IL-4.

Table 3. Oral Prednisolone-equivalent Daily Requirements (n = 21)

Dose (mg)	Baseline	Day 7	Day 21	Day 63
Mean	33	29	24	19
Median	32	27	24	18

Table 4. Inflammatory Parameters

	Baseline	Day 7	Day 21	Day 63
Individual values below limit of detection (LOD) were set to equal to LOD for calculations.
C-reactive protein (mg/L)
Median (range)	6.3 (1.0; 36.8)	7.0 (0.4; 26.2)	5.8 (0.5; 136.9)	2.0 (0.2; 31.5)
n	20	19	17	20
Stool calprotectin (mg/kg)
Median (range)	1029.9 (23.7; 3644.0)	1134.2 (89.7; 9972.6)	592.4 (19.5; 3120.2)	810.6 (19.5; 9929.8)
n	19	17	15	15
IL-1ß (pg/mL)
Median (range)	43.0 (43.0; 46.3)	43.0 (43.0; 63.3)
n	15	14
IL-4 (pg/mL)
Median (range)	0.8 (0.4; 2.9)	0.4 (0.4; 4.2)
n	21	18
IL-6 (pg/mL)
Median (range)	19.0 (19.0; 19.0)	19.0 (19.0; 83.7)
n	13	8
IL-8 (pg/mL)
Median (range)	15.0 (15.0; 56.8)	15.0 (15.0; 62.2)
n	18	16
IL-10 (pg/mL)
Median (range)	12.3 (12.3; 49.0)	12.3 (12.3; 68.4)
n	15	13
TNF-α (pg/mL)
Median (range)	14.7 (14.7; 26.4)	14.7 (14.7; 174.1)
n	21	19

DISCUSSION

Mammalian AP enzyme is membrane-bound and in humans 4 AP isoenzymes exist, of which 3 are tissue-specific (placental, germ cell, and intestinal) and 1 is tissue nonspecific (TNAP) expressed virtually in all tissues (e.g., bone, liver, and kidney). This hydrolase removes phosphate groups from proinflammatory molecules,8–10 dephosphorylation being a key mechanism in host defense against inflammation, whether due to exogenous toxins such as bacterial LPS present in the gut, or to endogenous extracellular ATP,8, 38 the latter having a potent proinflammatory activity.9 Therefore, AP administration may enhance the dephosphorylating capacity when the need is greatest, such as during an acute inflammatory response, to reduce inflammation, local tissue damage, and to retain organ function. Animal models of IBD have shown exogenous AP to attenuate the inflammatory response and to improve tissue damage,18 while endogenous intestinal AP has an antiinflammatory role in experimental chronic colitis in mice.26 Recently, exogenous intravenous AP given to patients with severe sepsis was reported to exert renal protection and improve outcomes,20 while the clinical pharmacology of exogenous enzyme has been shown to be linear, dose-proportional, and not affected by the systemically impaired metabolic status of sepsis.19

Since this was the first enteral administration of exogenous AP enzyme to patients, we were very cautious about the dose, which was a fraction (in U/kg) of the dose tested in animal models, and we limited the treatment period to a single 7-day course. However, our safety results suggest that higher doses and/or longer treatment periods with multiple courses are warranted. A good safety profile of exogenous AP administration has also been reported in previous studies with intravenous AP enzyme.19, 20 The anticipated absence of systemic absorption from the gut, due to the large molecular size of enzyme protein (140 kDa) was also confirmed by the stable serum activity levels of AP enzyme recorded during our study.

Regarding efficacy of oral AP enzyme in UC, our results derive from an uncontrolled exploratory trial and, as such, require confirmation. However, despite the short treatment period limited to a single short course, the observed decrease in disease activity scores was significant and clinically relevant. Thus, our findings strongly suggest that future research of AP in the treatment of UC is warranted, particularly due to its reported good tolerability, also observed in our data. Future controlled studies with bigger cohorts should confirm, or otherwise, our findings.

There were several limitations to conducting this first patient study with a new oral biologic treatment, other than the usual uncertainty about any new potential therapy. The need for a naso-duodenal tube (with radiological confirmation of positioning) for administration by infusion pump with constant delivery of drug for 7 days required hospitalization. These were key factors precluding the addition of a placebo-control group and for repeating courses. However, to our knowledge, a gastro-resistant tablet is being developed that will allow future larger trials with repeated treatment courses and placebo-control (Personal communication from the manufacturer, AM-Pharma BV). Despite the challenges, we made every attempt to enhance the information obtained from our patients by collecting images from all proctosigmoidoscopies for central evaluation by using a single pathologist to review biopsies, as well as by using single laboratories for evaluation of key inflammatory parameters. Therefore, we strove to adhere to strictest methodology to enhance the quality of the information within the limitations of a first trial in patients.

Our results also suggest that the efficacy of AP treatment can be evaluated relatively quickly, within 21 days of starting treatment, and may be sustained for at least 8 weeks after treatment cessation. In the absence of a placebo group, one cannot estimate the relevance of response/remission rates as estimated by large trials in this setting,32 as we were restricted to a single treatment course in a limited patient sample. Future trials should assess repeated courses of AP after 4–8 weeks in order to potentially obtain higher therapeutic efficacy. This is supported by the finding that C-reactive protein levels and stool calprotectin, although not affected during the first week, were both decreased at Day 21 and remained lower at Day 63 relative to baseline. Similarly, no changes in cytokine levels were observed during the first 7 days. However, since we do not have data beyond this timepoint a similar trend as seen in C-reactive protein and calprotectin cannot be excluded, and should be investigated for longer periods in future studies.

Finally, no antibodies against AP enzyme were detected. Although no such effect would be expected from orally administered bovine protein, since most humans consume bovine proteins daily in their diet, the origin of the enzyme raises a number of questions about reliable sources of safe materials. We have been informed by the manufacturers that a human recombinant form of the enzyme is being produced for future trials, which will resolve this issue.

In conclusion, enteral administration of 30,000 U AP enzyme daily for a single 7-day course to patients with moderate/severe UC was associated with short-term improvement in Mayo and MTWSI scores. Clinical response was observed within 21 days and was associated with a decrease in C-reactive protein and stool calprotectin. Enterally administered AP enzyme was well tolerated and nonimmunogenic. A solid form of gastro-resistant, oral AP enzyme is required in order to evaluate the efficacy of higher doses and repeated courses of AP in future trials.

Acknowledgements

The authors thank Jennifer Schwamborn (CRM Biometrics GmbH), Luigi Jonk, and Willem Raaben (AM-Pharma BV) for their help in compiling the data, and Dr. Ivana Vitkova (Medical Pathology Dept., Prague Univ. Hospital) for reviewing all biopsies. Competing interests: None declared. The interpretation of the results was the prerogative of the attending/participating physicians. Monitoring and auditing of the hospitals' study data were conducted externally by Sintesi Research S.r.l. Milan, Italy. Statistical analyses were conducted externally by CRM Biometrics GmbH, Rheinbach, Germany.