Palifermin reduces diarrhea and increases survival following irinotecan treatment in tumor-bearing DA rats

Authors

  • Rachel J. Gibson,

    Corresponding author
    1. Department of Medical Oncology, Royal Adelaide Hospital, Adelaide, Australia
    2. Department of Medicine, University of Adelaide, Adelaide, Australia
    • Department of Medical Oncology, Royal Adelaide Hospital, North Terrace, Adelaide 5000, South Australia, Australia
    Search for more papers by this author
    • Fax: +61-8-8222-3217.

  • Joanne M. Bowen,

    1. Department of Medical Oncology, Royal Adelaide Hospital, Adelaide, Australia
    2. Department of Medicine, University of Adelaide, Adelaide, Australia
    Search for more papers by this author
  • Dorothy M.K. Keefe

    1. Department of Medical Oncology, Royal Adelaide Hospital, Adelaide, Australia
    2. Department of Medicine, University of Adelaide, Adelaide, Australia
    Search for more papers by this author

Abstract

Mucositis is a common side effect of cancer chemotherapy for which there is currently no treatment. Irinotecan is a commonly used effective chemotherapeutic agent, causing severe gastrointestinal mucositis and diarrhea. Previous research suggests that palifermin is potentially antimucotoxic. The primary aim of this study was to determine whether palifermin was effective in ameliorating irinotecan-induced gastrointestinal mucositis. We also determined the protective effects of single large and multiple small doses of palifermin. Tumor-bearing DA rats were treated with a single large (10 mg/kg) dose of palifermin 3 days prior to, or multiple small (3 mg/kg day for 3 days) doses of palifermin or vehicle control prior to, receiving 2 doses of 150 mg/kg irinotecan. Animals were killed at 6, 24, 48, 72, 96, 120, or 144 hr after treatment. The primary endpoints were diarrhea and mortality. Gastrointestinal morphometry, histopathology and apoptosis were assessed. Tumor weights and mitoses were measured to ensure palifermin did not promote tumor growth. Data were analyzed using Peritz' F-test, Student's t-test and Tukey-Kramer multiple comparison test. Animals pretreated with palifermin tolerated irinotecan treatment better than control animals with less severe diarrhea (5% in animals receiving 10 mg/kg palifermin, 11% in animals receiving 3 × 3 mg/kg palifermin and 28% in animals receiving irinotecan only) and reduced mortality (2% in animals receiving 10 mg/kg palifermin, 11% in animals receiving 3 × 3 mg/kg palifermin and 28% in animals receiving irinotecan only). Small and large intestinal weights were maintained. Intestinal morphometry was not maintained in palifermin-pretreated rats despite being increased prior to irinotecan treatment. Palifermin pretreatment did not prevent apoptosis that peaked at 6 hr in the jejunum or colon, but prevented apoptosis at 96 hr in the small intestine. Palifermin pretreatment in both treatment regimens significantly reduces diarrhea and mortality following irinotecan administration without adversely affecting tumor growth. This positive response warrants further investigation, particularly in humans. © 2005 Wiley-Liss, Inc.

Mucositis is a major oncologic problem caused by the cytotoxic effects of cancer chemotherapy. The condition affects the entire gastrointestinal tract from the mouth to the anus and causes pain and ulceration in both the mouth and small and large intestines. In addition, it causes abdominal bloating, vomiting and diarrhea.1, 2, 3 Irinotecan hydrochloride is a common and effective chemotherapeutic drug used in the treatment of gastrointestinal cancers. Administration of this drug, however, causes 2 forms of diarrhea, an early cholinergic diarrhea and a severe delayed diarrhea, both of which are distressing to patients and can cause disruptions in treatment.4, 5 Previous research has proposed varying mechanisms for the delayed-onset diarrhea,6, 7, 8, 9 but the definitive mechanism behind this remains unknown. It has been previously shown that irinotecan differs from most cytotoxic agents in that it causes severe gastrointestinal damage with increased apoptosis in both the small and large intestines, changes in intestinal morphometry and excessive mucus secretion in the colon.8

Keratinocyte growth factor (palifermin) has been the subject of recent animal and clinical studies within the gastrointestinal tract. Many of these studies have indicated that this growth factor has antimucotoxic potential.10, 11, 12, 13, 14, 15, 16, 17, 18, 19 Farrell et al.12 have shown that palifermin is protective in some animal models of gastrointestinal mucositis, in ameliorating weight loss and improving crypt survival. Furthermore, Dorr et al.14, 15, 20, 21, 22 have shown that palifermin offers significant protection to the oral mucosa of mice following radiation therapy. In the clinical setting, reports from recent trials (phases 1, 2 and 3) have shown that palifermin is effective in reducing oral mucositis.16, 17, 18, 19 These reports document for the first time an effective antimucotoxic agent that is capable of reducing the side effects of cancer therapy.16, 17, 18, 19 All of the previous studies investigating the effect of palifermin have used multiple small doses; however, it has recently been suggested that palifermin may be more effective given as a single large dose prior to chemotherapy treatment (Amgen, Thousand Oaks, CA; personal communication). Therefore, the aims of this study were to investigate whether palifermin given as a single large dose or multiple smaller doses was able to prevent the severe diarrhea and mortality seen following the administration of irinotecan in rats with breast cancer.

Material and methods

Ethics

The study described here was approved by the Animal Ethics Committees of the Institute of Medical and Veterinary Sciences and of the University of Adelaide and complied with the National Health and Medical Research Council (Australia) Code of Practice for Animal Care in Research and Training (1997). Due to the nature of the diarrhea induced by irinotecan, animals were monitored 4 × daily and if any animal showed certain criteria (as defined by the Animal Ethics Committee), they were euthanized. These criteria included a dull ruffled coat with accompanying dull and sunken eyes, cold to touch with no spontaneous movement and a hunched appearance. Mortality in this study therefore was defined as an animal that had to be euthanized or died prior to its scheduled time point.

Passage of tumor

The mammary adenocarcinoma used in this study has been used by our group for a number of years and its passage has been described elsewhere.8, 23, 24 This tumor is specific to the DA rat. Tumor-bearing animals were used in this study to ensure that palifermin did not promote tumor growth or interfere with the cytotoxicity of irinotecan.

Preparation of tumor inoculum

Mammary adenocarcinoma tumors were diced, homogenized and then filtered through sterile gauze. The resulting tumor cell suspension was spun 3 times at 1,100g for 3 min each time, with the resultant pellet resuspended in fresh PBS prior to the next spin. A viable cell count was conducted using 0.5% w/v Trypan blue.

Experimental plan

The experiments described below were conducted in female DA rats weighing approximately 150 g. All animals received breast cancer inoculum as described above. Irinotecan was used as the chemotherapeutic agent.

Two hundred and eleven DA rats were randomly assigned to 1 of 7 groups (Table I). Palifermin was reconstituted immediately prior to injection, with a specific reconstitution solution provided by Amgen. Rats received subcutaneous pretreatment with either a single large dose (10 mg/kg) of palifermin administered 3 days prior to chemotherapy treatment, multiple smaller doses (3 × 3 mg/kg) administered daily for 3 days prior to chemotherapy treatment, or vehicle control [0.9% isotonic saline (NaCl)] administered for 3 days prior to treatment.

Table I. Treatment Groups and Kill Days
 0 hr6 hr24 hr48 hr72 hr92 hr120 hr144 hr
  1. Control rats were killed at time designated 0 time with rats from each treatment regimen killed 6, 24, 48, 72, 92, 120, and 144 hr after chemotherapy.

Single-dose (1 × 10 mg/kg) palifermin onlyX       
Multiple-dose (3 × 3 mg/kg) palifermin onlyX       
Saline onlyX       
Single-dose (1 × 10 mg/kg) palifermin/irinotecan XXXXXXX
Multiple-dose (3 × 3 mg/kg) palifermin/irinotecan XXXXXXX
Irinotecan only XXXXXXX
Multiple-dose (3 × 3 mg/kg) palifermin/buffer XXXXXXX
Saline/buffer XXXXXXX

Three groups of rats were subsequently killed at time designated as 0 hr to confirm that palifermin was biologically active and also to determine the effect on the gastrointestinal tract without the confounding effects of chemotherapy. All remaining groups received irinotecan or buffer control at a dose of 150 mg/kg on 2 consecutive days, which has been previously described to cause reproducible gastrointestinal mucositis.8 Rats received 0.01 mg/kg s.c. atropine (to reduce any cholinergic reaction to irinotecan) immediately prior to the administration of the i.p. irinotecan or buffer control. Irinotecan (Pharmacia, Kalamazoo, MI) was administered in a sorbitol-lactic acid buffer (45 mg/mL sorbitol, 0.9 mg/mL lactic acid, pH 3.4), required for activation of the drug on 2 consecutive days. Control rats received buffer only.

Groups of rats from each treatment group were killed by CO2 asphyxiation followed by cervical dislocation at 6, 24, 48, 72, 96, 120 and 144 hr after treatment, with tissues being collected for further analysis. The entire gastrointestinal tract from the pyloric sphincter to the rectum was dissected out and flushed with chilled isotonic saline (0.9% w/v) to remove contents, then the wet weight of small and large intestines were recorded. Samples (1 cm in length) of the small intestine (taken at 25% of the length of the small intestine from the pylorus) and the colon (taken at mid colon position) were collected and placed into Clarke's fixative (60% w/v ethanol, 40% w/v acetic acid) for intestinal morphometry and in 10% neutral buffered formalin for histopathologic and apoptotic assessment. The tumors were removed and weighed. Small cross-sections of nonnecrotic tumor were dissected and placed into 10% neutral buffered formalin for histologic examination. In addition, the liver and spleen were removed, weighed and fixed in 10% neutral buffered formalin for histologic examination.

Diarrhea assessment

All animals were assessed 4 times daily and diarrhea was recorded according to previous gradings.8 Briefly, there were 4 grades: 0, no diarrhea; 1, mild diarrhea (staining of anus); 2, moderate diarrhea (staining over the top of the legs and lower abdomen); 3, severe diarrhea (staining over the legs and higher abdomen, often associated with continual oozing). All diarrhea assessments were conducted in a blinded fashion by 2 investigators (R.J.G. and J.M.B.).

Histopathologic examination

Samples of jejunum, colon, liver and spleen were taken, weighed and fixed in 10% neutral buffered formalin for routine histologic examination as previously described.8 Expert histopathologic examination of all sections was performed by specialist veterinary pathologist, Dr. John Finnie, from the Institute of Medical and Veterinary Science (Adelaide, Australia). This examination was performed in a blinded fashion.

Intestinal morphometry

Sections (1 cm in length) of small and large intestine (jejunum, corresponding to 25% of the jejunoileum; colon, corresponding to 50% of the colon length) were opened onto cardboard, fixed in Clarke's fixative for 24 hr and stored in 70% ethanol at room temperature prior to use. Tissue was rehydrated through a graded series of ethanols prior to being hydrolyzed in 1 M HCI for 7 min at 60°C. Following 2 washes in double-distilled water, the tissue was subsequently stained with Schiff's reagent for 30 min prior to being microdissected with a cataract knife and stereomicroscope. Microdissected tissue was mounted in 45% (w/v) acetic acid and measurements were taken using a calibrated graticule of the villus length, apical and basal widths of 15 villi (jejunum) and the lengths of 15 crypts (jejunum and colon).2, 8, 23, 24, 25 Villus area was calculated using a trapezoid approximation with villus area and crypt length correlating with villus and crypt epithelial cell populations, respectively.2, 8, 23, 24, 25 All assessments were conducted in a blinded fashion by 2 investigators (R.J.G. and J.M.B.).

Apoptosis assessment

Sections (1 cm in length) of small and large intestine (jejunum, corresponding to 25% of the jejunoileum; colon, corresponding to 50% of the colon length) were fixed for 24 hr in 10% neutral buffered formalin before being dehydrated and embedded in paraffin wax. Sections 4 μm in size were cut using a rotary microtome and mounted onto saline-treated slides for later analysis. The method used to stain for apoptosis was performed using the In Situ Cell Death Detection Kit AP (Roche, Mannheim, Germany) and has been described elsewhere.2, 8, 23, 24, 25 Briefly, slides were dewaxed and rehydrated through graded ethanols prior to immersion in a 0.1% TX-100 in 0.1% (w/v) sodium citrate buffer solution for 8 min at room temperature. After 2 washes in PBS, slides were placed in a TUNEL buffer solution (150 mM Tris, 0.7M NaCaCo, 10 mM CoCL2, 10% BSA, sterile H20) for 10 min at room temperature. Slides were placed into the reaction mixture and incubated in a humid chamber for 3 hr at 37°C. After 3 rinses in PBS, slides were incubated with converter-AP for 60 min at 37°C. Slides were rinsed in 2 changes of PBS prior to fast red chromagen being applied for 15 min. Following counterstaining with hematoxylin, slides were mounted and apoptotic bodies counted per crypt per 4 μm section.2, 8, 23, 24, 25 All assessments were conducted in a blinded fashion by one investigator (J.M.B.).

Tumor cell proliferation

Cross-sections of breast adenocarcinoma were fixed in 10% formalin before being processed and embedded in paraffin wax. Sections of tumor were cut (4 μm), stained with hematoxylin and eosin and the number of mitotic figures were counted. These were expressed as the number of mitotic figures per mm2. To avoid necrosis, counts were taken from the periphery of the tumor. All assessments were conducted in a blinded fashion by one investigator (R.J.G.).

Statistical analysis

Group means were compared using the Peritz' F-test, which is a robust multiple comparison test of group means where group size is not necessarily equal.26 Student's t-test and the Tukey-Kramer multiple comparison test were also used.

Results

Biologic efficacy of palifermin

Palifermin is a known stimulator of hepatocytes.13 We therefore recorded the weight of the liver following pretreatment with both doses of palifermin in order to confirm biologic activity. Rats pretreated with palifermin had significantly increased liver weights (single-dose palifermin, 8.29 ± 1.05 g; multiple-dose palifermin, 9.89 ± 1.00 g) compared to saline control (7.11 ± 0.58 g; p < 0.022 and 0.0002, respectively). In addition, pretreatment with palifermin caused an increase in small intestinal weight (single-dose palifermin, 5.53 ± 0.67 g; multiple-dose palifermin, 6.38 ± 0.89 g; control, 5.08 ± 0.54 g) without increasing overall body weight or adversely increasing tumor weights (single-dose palifermin, 2.48 ± 2.07 g; multiple-dose palifermin, 1.79 ± 1.25 g; control, 2.66 ± 1.73 g; p = NS).

Response to treatment

No control rat, including those that were pretreated with palifermin, developed diarrhea. Approximately 25% of rats that received irinotecan alone had mild diarrhea immediately following the first injection, with a further 2% developing moderate diarrhea after the first injection (Fig. 1). By 24 hr after irinotecan treatment, 4% of all rats had developed severe diarrhea, which increased to 18% at 48 hr, 40% at 72 hr, before decreasing to 16% at 96 hr. No further severe diarrhea was recorded in this treatment group after 96 hr. The peak in severe diarrhea at 72 hr posttreatment directly corresponded to the mortality peak (Fig. 2).

Figure 1.

Incidence of diarrhea in rats pretreated with a single large (10 mg/kg) dose of palifermin, multiple small (3 × 3 mg/kg) doses of palifermin, or irinotecan alone. Both regimens of palifermin pretreatment were effective in delaying the onset and reducing the severity of diarrhea compared to irinotecan alone.

Figure 2.

Graph showing mortality over time in rats pretreated with a single large (10 mg/kg) dose of palifermin, multiple small (3 × 3 mg/kg) doses of palifermin, irinotecan alone and control rats. Mortality was significantly reduced in both regimens of palifermin pretreatment. Rats that received a single large (10 mg/kg) dose had improved mortality over rats that received multiple small (3 × 3 mg/kg) doses.

Animals pretreated with multiple doses (3 × 3 mg/kg) of palifermin also developed mild diarrhea immediately following irinotecan treatment. However, this diarrhea did not increase in severity until 72 hr after irinotecan treatment, when approximately 33% recorded severe, 33% recorded moderate and 33% recorded mild diarrhea (Fig. 1). The diarrhea improved over subsequent days and only mild diarrhea was recorded at 144 hr. Again, the peak in severe diarrhea at 72 hr corresponded with peak mortality. (Fig. 2).

Rats pretreated with a single large dose (1 × 10 mg/kg) of palifermin prior to chemotherapy did not develop severe diarrhea until 72 hr after treatment, when this grading was noted in approximately 10% of animals (Fig. 1). However, the severity of the diarrhea quickly resolved with no further severe diarrhea recorded at any other time point. The severe diarrhea with this treatment group did not correspond to peaks in mortality, with no animal dying at this time point. At 96 hr after treatment, one death was recorded, this being the sole mortality for this treatment group (Fig. 2).

Autopsies were performed on all animals that died prematurely and, as in our previous studies,8 the cause of death was duodena perforation, leading to peritonitis, with the underlying pathology being paralytic ileus. Abdominal cavities were fluid-filled, and areas of perforation with accompanying adherent peritoneum or omentum were observed.8

Body and organ weights

At autopsy, final body weights, small and large intestinal weights, liver and spleen weights were recorded. Body weights of palifermin-pretreated groups were maintained until 72 hr after irinotecan treatment. At this time point, rats pretreated with multiple doses of palifermin had similar body weights to rats receiving irinotecan only (154.7 ± 13.6 vs. 149.8 ± 9.8; Fig. 3). In contrast, rats pretreated with a single large dose of palifermin were considerably heavier than rats receiving irinotecan only, although this did not reach statistical significance (168.4 ± 12.3 vs. 149.8 ± 9.8). Furthermore, across all other time points measured, this group of rats maintained body weights (Fig. 3). In addition, following treatment with irinotecan, both single and multiple doses of palifermin were able to maintain small and large intestinal weights when compared to irinotecan alone at 72 hr (data not shown).

Figure 3.

Graph showing bodyweights over time of rats pretreated with a single large (10 mg/kg) dose of palifermin, multiple small (3 × 3 mg/kg) doses of palifermin, irinotecan alone and control rats. Control rats gradually put on weight over the course of the experiment. Rats that were pretreated with both regimens of palifermin maintained a higher bodyweight than rats receiving irinotecan alone.

Histology

Pathological changes throughout the gastrointestinal tract were examined. Results were highly variable in severity but marked crypt dilatation was observed with crypt lumens containing mucinous material (which was sometimes copius) and degenerate polymorphonucleocytes (PMNs). The enterocytes were attenuated. In addition, goblet cell metaplasia and enterocyte hyperplasia were observed. Lamina propria edema was seen. No increase was observed in the normal mixed inflammatory cell population within the lamina. In addition to the gastrointestinal change, a number of definite pathologies were recorded and these included mesenteric hemorrhage, perirenal hematoma, chronic interstitial pneumonia, chronic active mesentery inflammation with PMN, transitional epithelial hyperplasia and chronic active dermatitis.

Intestinal morphometry

Jejunum.

Pretreatment with single or multiple doses of palifermin alone caused a variable response. Rats receiving a single large-dose of palifermin did not have significantly longer crypts than rats that received saline only (p = NS). However, rats that received multiple small doses of palifermin had significantly longer crypts (p < 0.034) than controls, with no differences in villus area and crypt cell counts (data not shown). Rats pretreated with single or multiple doses of palifermin did not have improved crypt lengths, villus area, or crypt cell counts at any time point following irinotecan treatment (data not shown).

Colon.

Pretreatment with single and multiple doses of palifermin alone caused a similar response in the colon as in the jejunum. There was no effect seen in rats pretreated with single large-dose palifermin (data not shown). However, multiple small doses caused increases in crypt length and crypt cell counts. Following treatment with irinotecan, there was no change in colonic length between any groups, including those pretreated with palifermin. However, from 72 hr onward, there was a rapid colonic hyperplasia in all treatment groups, with all lengths increasing substantially (data not shown).

Intestinal apoptosis

Pretreatment with either single or multiple doses of palifermin did not alter the rates of apoptosis in either the jejunum or colon prior to irinotecan treatment (Fig. 4). Following chemotherapy, palifermin pretreatment was unable to protect against the early peak (6 hr posttreatment) of irinotecan-induced apoptosis, with similar levels observed for irinotecan alone and both single and multiple doses of palifermin pretreatment (Fig. 4). However, single and multiple doses of palifermin were able to offer protection from the late peak of apoptosis (96 hr posttreatment; single-dose palifermin/irinotecan vs. irinotecan only, p < 0.01; multiple-dose palifermin/irinotecan vs. irinotecan only, p < 0.001; Tukey-Kramer). In the colon, pretreatment with either single or multiple doses of palifermin offered no protection from the 6-hr posttreatment irinotecan-induced apoptosis (single-dose palifermin/irinotecan vs. irinotecan only, p = NS; multiple-dose palifermin/irinotecan vs. irinotecan only, p = NS; Tukey-Kramer). Unlike the jejunum, no second peak of apoptosis at 96 hr was observed with any group (Fig. 4).

Figure 4.

Graph showing presence of crypt cell apoptosis in both the small and large intestines of rats pretreated with a single large (10 mg/kg) dose of palifermin, multiple small (3 × 3 mg/kg) doses of palifermin, irinotecan alone and control rats at autopsy. The duration of the apoptotic response was reduced in rats that were pretreated with both regimens of palifermin. Control rats showed little apoptosis over the time course of the experiment.

Tumor proliferation

To ensure that palifermin pretreatment did not affect the efficacy of irinotecan, we measured tumor weights at all time points. There was no significant difference between any of the groups at any time point measured (data not shown). Pretreatment with single or multiple doses of palifermin did not cause increased tumor cell proliferation (single-dose palifermin, 17.97; multiple-dose palifermin only, 19.09; saline only, 20.41 mitoses per mm2 of tumor). It also did not interfere with the cytotoxicity of irinotecan (single-dose palifermin/irinotecan, 11.71; multiple-dose palifermin/irinotecan, 10.89; irinotecan only, 8.04 mitoses per mm2 of tumor at 24 hr posttreatment; p = NS; data not shown).

Discussion

Oral and gastrointestinal mucositis (OM and GIM, respectively) are both severe toxicities of anticancer treatments. They share some mechanisms and thus may well share some treatment responses. Palifermin (Amgen) has been shown to reduce OM following bone marrow transplantation for hematologic malignancy in a phase 3 clinical trial.19 It has also been shown to reduce OM in colon cancer patients treated with 5-FU.16 However, it has not been previously tested in GIM. Irinotecan is an effective cytotoxic agent in certain tumors. However, its use is limited by the severity of the diarrhea it causes, making an otherwise effective drug very difficult to use clinically. Oral loperamide works to slow diarrhea in many,27 but there is still a significant incidence of grade 4 diarrhea and occasional mortality, leading to cessation of treatment.4, 5, 28 Previous research has shown that palifermin is an effective antimucotoxic, reducing severity and duration of mucositis in both animal models12, 15, 20, 21, 22, 29, 30 and the clinical setting.16, 17, 18, 19 A drug that would reduce irinotecan-induced diarrhea would improve the clinical options, particularly in colon cancer, but also in other tumors. The aim of this study was to assess the effect of palifermin in ameliorating gastrointestinal mucositis (specifically irinotecan-induced diarrhea) and to see whether once-per-cycle dosing was as effective as the standard triple-dose regimen.

This study confirmed that pretreatment with either a single large dose or multiple small doses of palifermin was biologically active in the rat model by causing increased liver weight. It also confirmed that when given alone, palifermin did not cause increased gastrointestinal apoptosis or tumor cell proliferation. These results suggest that although many epithelial tumors, such as the breast adenocarcinoma used in this study, may express the palifermin receptor,31, 32 when palifermin is administered in therapeutic doses, there is no increase in tumor cell proliferation. This is an important finding for future clinical trials.

This study demonstrated that rats that were given a single large dose of palifermin prior to irinotecan administration were able to maintain their bodyweights and had delayed onset and reduced the severity of diarrhea compared to all other treatment groups over all time points. Only 5% of animals recorded severe diarrhea compared to 11% in the multiple dose of palifermin (p < 0.05) and 28% in irinotecan-alone groups (p < 0.05). In addition, animals receiving single-dose palifermin had improved mortality with only 2% dying prematurely, compared to 11% in the multiple dose of palifermin and 28% in the irinotecan-alone groups. Thus, a single large dose of palifermin was highly effective in maintaining bodyweights, ameliorating the severe diarrhea and improving the mortality associated with irinotecan administration.

Traditional measures such as intestinal morphometry and intestinal apoptosis have been the hallmarks of measuring response to chemotherapy and gastrointestinal mucositis,2, 8, 23, 24 therefore these were employed in the present study. Despite pretreatment with multiple smaller doses of palifermin causing increases in crypt length, villus area (jejunum only) and crypt cell counts, these were not maintained following chemotherapy treatment even though diarrhea was decreased. Similarly, pretreatment with palifermin did not cause changes in gastrointestinal apoptosis and was unable to protect the gastrointestinal tract from the early peak of apoptosis. However, in the jejunum, palifermin pretreatment was able to protect from the second peak of apoptosis, although the mechanisms behind this are unclear.

From this study, it seems that the traditional ways of measuring the gastrointestinal response to mucositis is not sufficient, as there was no correlation between gastrointestinal histomorphometry and apoptosis and ultimate response to treatment. Therefore, it would seem that more advanced and novel techniques such as changes in cytokine expression and gene profile changes need to be utilized. Recent reviews33, 34 have suggested that the whole tissue is involved in the response to cytotoxic damage and not just the epithelium; therefore, the damage of the underlying submucosa and its subsequent relationship to the development of mucositis need to be investigated.

In conclusion, this study has shown that palifermin is an effective antimucotoxic for irinotecan-induced gastrointestinal toxicity. It reduces onset and severity of diarrhea and improves survival. This has important implications for the use of palifermin in gastrointestinal mucositis in humans and should be studied further.

Acknowledgements

The authors gratefully acknowledge Amgen for the supply of the palifermin and Pharmacia (Pfizer) for the supply of irinotecan. In addition, the authors acknowledge Dr. John Finnie for expert histopathologic advice as well as Professor A. Woods, Mr. C. Gordon, Mrs. D. Lazauskas and Mrs. R. Pradham for assistance with histologic preparation. Supported by an Amgen Unrestricted Educational Grant (to D.M.K.K.).

Ancillary