Pancreatic Enzyme Extract Improves Survival in Murine Pancreatic Cancer

ORIGINAL ARTICLE

Pancreatic Enzyme Extract Improves Survival in MurinePancreatic Cancer

Murat Saruc, MD,* Silke Standop,* Jens Standop, MD,* Fumiaki Nozawa, MD, PhD,*Atsushi Itami, MD, PhD,* Krishan K. Pandey, PhD,* Surinder K. Batra, PhD,*

Nicholas J. Gonzalez, MD, Pierre Guesry, MD, and Parviz M. Pour, MD*

Objectives: The disappointing current therapeutic approaches forpancreatic cancer (PC) represent an urgent need for the developmentof novel methods to control the disease. Based on a recent report onthe effectiveness of pancreatic enzyme therapy, we examined the ef-fect of porcine pancreatic enzyme extracts (PPE) on human PC xeno-grafts in nude mice.

Methods: The malignant human PC cell line AsPC1 was trans-planted into the pancreas of male beige XID nude mice that weretreated or not with PPE in drinking water. The survival, size, andvolume of tumors, plasma pancreatic enzyme levels, fecal fat, andurine were examined as were the expression of transforming growthfactor , insulinlike growth factor-I, epidermal growth factor, epider-mal growth factor receptor, apoptosis, and proliferation rate of tumorcells.

Results: PPE-treated mice survived significantly longer than thecontrol group (P < 0.002). Tumors in the PPE-treated group weresignificantly smaller than in the control group. All mice in the controlgroup showed steatorrhea, hyperglucosuria, hyperbilirubinuria, andketonuria at early stages of tumor growth, whereas only a few in thetreated group showed some of these abnormalities at the final stage.There were no differences in the expression of growth factors, epider-mal growth factor receptor, or the apoptotic rate between the tumorsof treated and control mice.

Conclusions: The treatment with PPE significantly prolongs thesurvival of mice with human PC xenografts and slows the tumorgrowth. The data indicate that the beneficial effect of PPE on survivalis primarily related to the nutritional advantage of the treated mice.

Key Words: pancreatic cancer, xenograft, nude mice, pancreatic en-zyme, survival, therapy

(Pancreas 2004;28:401412)

Despite advances in the clinical and biologic areas of pan-creatic cancer (PC), the disease has maintained its deadlycourse. It is still the fifth leading cause of cancer death in bothmen and women, accounting for more than 27,000 deaths an-nually in the United States.1 The most common symptoms ofthe disease, jaundice, abdominal pain, and weight loss togetherwith other presenting factors, are nonspecific for this disease.Thus, the diagnosis of PC at an early stage is difficult at bestand requires an extensive diagnostic workup, including ex-ploratory surgery.2 In the majority of patients who are referredto the hospital, the tumor is generally in an advanced stage, hasinvaded the surrounding tissues, or has metastasized.

The only effective therapy, surgery, is still limited toabout 25% of the patients, and even in these patients, cancerrecurrence has remained inescapable.3 Although the operativemortality over the years has decreased, the postoperative sur-vival has remained below 20% at 5 years.47 These featureshighlight the urgent need for the establishment of effectivetherapeutic modalities because the current conventionaltherapy has proven ineffective.

Among the new chemotherapeutic drugs, gemcitabinehas been shown to provide a clinical benefit in the treatment ofPC compared with 5-fluorouracil (5-FU).2 This benefit hasbeen limited mainly to the palliation of disease symptoms in-cluding pain intensity, analgesic consumption, Karnofsky per-formance status, and weight gain.8,9 In a study, the only benefitfor gemcitabine-treated patients was an approximate 1-monthincrease in median survival time compared with the survival ofpatients treated with 5-FU.9 The probability of surviving 1year for those receiving gemcitabine was 18% compared with2% for those receiving 5-FU.9 These survival numbers are,however, not significantly different. Also, the combination ofchemotherapy and radiation therapy for PC has remained a ma-jor toxicity problem.10

Because of the limited scientific design and the lack ofparticularly efficacious therapy, the role of adjuvant therapy in

Received for publication August 6, 2003; accepted October 3, 2003.From the *Eppley Institute for Research in Cancer and Allied Diseases, De-

partments of Biochemistry and Molecular Biology and Pathology andMicrobiology, University of Nebraska Medical Center, Omaha, Nebraska;private practice, New York, New York, Nestle Research Center,Lausanne, Switzerland.

This work was supported by the Nestle Company and, in part, by an AmericanCancer Society Special Institutional grant. Dr. Saruc is the recipient of ascholarship from the Turkish Gastroenterology Foundation.

Reprints: Parviz M. Pour, MD, The Eppley Institute for Research in Cancerand Allied Diseases, 986805 Nebraska Medical Center, Omaha, NE68198-6805 (e-mail: [email protected]).

Copyright 2004 by Lippincott Williams & Wilkins

Pancreas Volume 28, Number 4, May 2004 401

PC remains largely unestablished.1113 Although neoadjuvanttherapy has been advocated to improve curative resection ratesand survival, it carries its own morbidity. As a result, the prog-nosis of patients with PC has remained extremely poor.3 Thepitfalls of therapy in experimental models also highlight theinherent resistance of PC cells to any conventional therapy.

Alternative medicine has been employed in hard-to-control cancers with some success. Recently, a significant in-crease in survival rates has been reported in PC patients whowere given high doses of porcine lyophilized pancreas, con-taining 3080 USP units of proteolytic activity per milligramand 1540 U of lipolytic activity per milligram.14 In this study,81% of the patients survived 1 year, 45% survived 2 years, and36% survived 3 years.14 These results are significantly abovethe 25% survival rate at 1 year and 10% survival rate at 2 yearsfor all stages of pancreatic adenocarcinoma reported in the Na-tional Cancer Database.15 The data were generated from 11patients in a private clinic and no controls were used. In thepresent studies, we examined the therapeutic efficacy of thesame enzyme used in the clinical study in the nude mousemodel with human PC cell xenografts.

MATERIALS AND METHODS

AnimalsThirty 5- to 6-week-old, male beige XID nude mice from

the National Cancer Institute were used. They were housed inplastic cages on autoclaved corncob bedding (Anderson 1/8-in. bed corncob autoclavable bedding) in groups of 5 and weremaintained in a controlled, germ-free environment (tempera-ture, 2226C; humidity, 44%64%; 12-hour light/darkcycles) in sterilized plastic containers covered with polyesterbacterial air filters. All mice had free access to commercialfood (Teklad 8064 Sterilizable Rodent Chow; Harland Teklad,Madison, WI) and autoclaved water.

MethodsTwo separate experiments were performed. The first

study was planned as a survival study followed by a compara-tive study on the cancer growth rate and some laboratory data.

Orthotopic PC Cell InoculationUnder Nembutal anesthesia, PC cells (AsPC1), which

were originally derived from the ascites of the patient with PCand exhibit an extremely malignant behavior in vitro (we haveshown that it is resistant to the apoptotic effect of TRAIL),16

were injected into the splenic lobe of the pancreas by a methoddeveloped in our laboratories.17 Briefly, the abdomen wasopened at the right upper quadrant at the projection of the du-odenum instead of the left upper quadrant. This techniqueavoids the adherence of the growing tumor in the splenic lobeto the laparotomy wound at the projection of that lobe. Then,the pancreas was pulled out and 1 106 tumor cells were in-

jected into the splenic lobe with a 12-gauge syringe. The in-jection was carefully performed to avoid spillage of cells intothe peritoneal cavity or the liver. The pancreas and attachedspleen were replaced to the original positions, and the woundwas closed.

Study 1 (Survival Study)Thirty male mice were divided into 2 equal groups. Both

groups received human PC cells, AsPC1, injected into theirpancreas as described above. One week after the injection oftumor cells, when tumor growth was evident by palpation ofthe abdomen at the projection of the transplantation area, thetreatment group received PPE in water at a dose of 400 mg/kgbody weight. The dose of PPE corresponded to that used inpatients.14 The control group was given only autoclaved tapwater. Water intake was measured daily and body weightweekly. The concentration of PPE solution was adjustedweekly according to their water intake and body weightchanges. PPE intake per day was calculated for each mouse.Mice were closely observed during the study and examinedperiodically for tumor growth by palpating the abdomen at theprojection of the tumor inoculation site.

Animals were allowed to die spontaneously, except forthe last 2 animals from the treatment group, which were killedon day 62, when no control mice were alive. A complete nec-ropsy was performed in each mouse, and all abdominal andthoracic tissues were examined carefully for abnormalities andmetastases. The pancreas was removed, and the dimensions,volume, and weight of tumors (isolated and freed from the sur-rounding tissues) were measured by previously reported meth-ods.18 The liver, kidneys, and lungs were sliced in 2-mm thickslices for the detection of possible metastases. All tissues werefixed in buffered formalin, and the pancreas, duodenum, liver,kidneys, lungs, and any tissues with abnormalities were pro-cessed for histology by conventional methods. The histologicevaluation of the tissues was performed by 2 pathologists in ablinded fashion.

Study 2 (Tumor Growth Study)The same study, using the same number of mice and pro-

cedures, was carried out 2 months after the termination of thefirst experiment. Tumors were carefully isolated from the sur-rounding pancreatic tissue, if recognizable, as well as fromconnective tissue as much as possible. The tumor size andweight were determined in 2 mice from each group on day 52as well as 2 mice from the control group and 5 mice from thetreatment group on day 60. In these mice, plasma amylase andlipase levels were also examined. The fecal fat content and theurinary ketone, glucose, and bilirubin of all mice were exam-ined weekly.

In all mice, a complete necropsy was performed and tis-sues were examined as in the first study. All tumors from micethat were euthanized at intervals were examined by immuno-

Saruc et al Pancreas Volume 28, Number 4, May 2004

402 2004 Lippincott Williams & Wilkins

histochemistry and polymerase chain reaction (PCR) for theexpression of transforming growth factor (TGF)-, insulinlikegrowth factor (IGF)-I, epidermal growth factor receptor(EGFR), apoptotic rate, and cell proliferation markers.

Cell CultureAsPC1 cells were grown in RPMI 1640 supplemented

with 10% heat-inactivated fetal bovine serum (FBS), 100U/mL penicillin and 100 g/mL streptomycin. Cells weremaintained at 37C in a humid chamber with 5% CO2 condi-tions. At the time of their use, the viability of these cells, asdetermined by the trypan blue staining method, was over 98%.RPMI 1640, FBS and penicillin-streptomycin were purchasedfrom Gibco BRL (Grand Island, NY).

Porcine Pancreatic EnzymeThe PPE preparation (Nutritional Services, Inc., CA)

was used clinically in patients.14 Specifically, porcine lyophi-lized pancreas containing 3080 USP units of proteolytic ac-tivity per milligram and 1540 U of lipolytic activity per mil-ligram was used. PPE differs from the conventional commer-cial enzyme preparation in that it is not alcohol extracted andvacuum distilled, a procedure that produces a very purifiedprotein extract, removing some yet unknown active factors,cofactors, protective factors, enhancers, and so on. PPE in theless activated form is far more stable in the gut, and with thecomponent of unknown as well as known cofactors, theproduct is more efficient in controlling the digestion. Accord-ing to the supplier, each lot of the enzyme preparation had beenassayed qualitatively and bacteriologically. Comparison of theactivity of PPE with the commercial preparation Pancreatin issummarized in Table 1. The comparative assay performed atthe Laboratories of Nestle in Lausanne, Switzerland, haveshown that Pancreastatin has more proteolytic activities (tryp-sin, -chymotrypsin, and elastase) than PPE. On the contrary,PPE has a higher -amylase, aminopeptidase (leucyl-, valyl-AP), acid phosphatase, and phosphoamidase activity than Pan-creatin. It is assumed that the differences are due to the condi-tions of tissue extraction and drying. The rapid freeze-dryingassociated with PPE would favor release, activation, and pres-ervation of lysosomal enzymes (eg, acid phosphatase and gly-cosidases) and could better preserve the activities of pancreaticantiprotease inhibitors.

The fresh batch, which was kept at 4C immediately af-ter delivery, was replaced every 3 months.

Because feeding of the PPE via serial gavage proceduresseveral times daily (consistent with the clinical therapy)14 re-sulted in morbidity and mortality, the enzyme was supple-mented in drinking water. We determined the activity of PPEin water in various conditions to optimally preserve the en-zyme activity over 24 hours in water. PPE, which was deliv-ered in flake form and not soluble in water, was ground by atissue tearer at low speed. One gram of the ground PPE could

then be dissolved in water at 1:10 mL in 4C water and kept atroom temperature. Samples were taken for the determinationof amylase, lipase, and trypsin at 0, 4, 12, and 24 hours. Theprocedure was repeated in every batch to check enzyme activ-ity. Cold water was used for the enzyme solution to avoid pos-sible enzyme degradation at room temperature. For compari-son, samples were also taken from a solution where PPE wasdissolved in water at room temperature. The PPE solution wasprepared freshly every day by dissolving the extract in auto-claved water.

Histology and ImmunohistochemistryTumor tissues derived from nude mice were embedded

in paraffin and cut into 4-m serial sections. One set of sectionswas stained with hematoxylin and eosin for the determinationof morphologic type, invasion, extent of necrosis, and metas-tases. For adequate screening, in the immunohistochemicallyprocessed slides, depending on the size of the tumors, the sameareas of tumors (as many as possible) were marked with a redcircle on the back of each of the serial sections. In each section,prepared by different staining methods, the identical areaswere subjected to evaluation. In each area, the cells positive forthe staining were counted (at least 1000 cells per slide). Twopathologists independently counted the cells, and the average

TABLE 1. Comparative Activity of Enzymes BetweenCommercial Preparation (Pancreatin) and PPE

Enzyme Pancreatin PPE

Control 0.00 0.00Alkaline phosphatase 1.25 1.00Esterase (C4) 1.75 2.00Esterase Lipase (C8) 2.25 3.00Lipase (C14) 1.00 1.50Leucin arylamidase 0.50 5.00Valine arylamidase 0.50 2.50Cystine arylamidase 1.00 1.00Trypsin 5.00 3.00-Chymotrypsin 5.00 2.50Acid phosphatase 0.50 4.50Phosphoamidase 0.75 4.50-Galactosidase 0.00 0.00-Galactosidase 4.00 4.50-Glucuronidase 1.75 1.00-Glucosidase 1.00 2.00-Glucosidase 0.50 1.25N-acetyl--glucosaminidase 0.00 1.00-Mannosidase 0.50 1.50-Fucosidase 0.00 0.75

Activities were assayed by the Api Zym system (Bio Merieux SA, France),which is a semiquantitive microassay system.

Pancreas Volume 28, Number 4, May 2004 Pancreatic Cancer Treatment by Pancreatic Enzymes

2004 Lippincott Williams & Wilkins 403

number of cells counted was regarded as the representativenumber. When differences in the results of these pathologistsexceeded 20%, the slides were reevaluated.

All immunohistochemical examinations were carriedout using an avidin-biotin-peroxidase complex (ABC)method.19 Antibodies used in this study are summarized inTable 2.

Apoptosis AssayApoptosis was measured on paraffin-embedded tissue

sections using the TUNEL staining. TUNEL staining was per-formed using the DeadEnd TUNEL System (Promega, Madi-son, WI), according to the manufacturers instructions. Fromformalin-fixed and paraffin-embedded specimens, 4-m thicksections were prepared on glass slides. The specimens weredeparaffinized in xylene for 5 minutes 2 times and were thenhydrated in 100, 95, 85, 70, and 50% ethanol, washed with0.85% NaCl, and finally in phosphate-buffered saline (PBS)(Gibco BRL). These specimens were fixed in 4% paraformal-dehyde solution and washed with PBS. These were treatedwith 20 g/mL proteinase K (Promega) for 20 minutes andrefixed by 4% paraformaldehyde solution. After washing withPBS, equilibration buffer was added. These specimens weretreated with 100 L of labeling reagent at 37C for 60 minutes.This reaction was terminated by 2 SSC. After washing byPBS, we blocked the endogenous peroxidases by 0.3% hydro-gen peroxidase for 5 minutes at room temperature. Thesespecimens were incubated with streptavidinhorseradish per-oxidase solution at room temperature for 30 minutes. Afterwashing in PBS at room temperature, specimens were visual-ized with DAB substrate, DAB chromogen, and hydrogen per-oxidase solutions. After rinsing in deionized water, thesespecimens were counterstained with hematoxylin and thenmounted. The results were observed with an optical micro-scope. For the evaluation of the positive cell ratios, a total of atleast 300 cells in 1 optical field was counted 3 times, and themean value was taken as the apoptotic ratio.

Reverse Transcriptase PCR (RT-PCR)Total RNA was isolated from 2 AsPC1 xenografts in

treated mice, from 4 untreated mice (control), and from the

native cultured AsPC1 cells. Thin sections of tumor tissuewere cut out and lysed in guanidinium isothiocyanate (GITC)for 5 hours with shaking at room temperature. The lysate wasoverlaid on cesium chloride and centrifuged for 20 hours at32,000 rpm at room temperature. The RNA pellet was resus-pended in DEPC water, and the RNA was precipitated withethanol and sodium acetate.

Two micrograms of RNA were reverse transcribed usingSuperscript II RNase H-reverse transcriptase (Invitrogen,Carlsbad, CA). The cDNA was used in PCR to check theexpression of EGF, EGFR, and TGF- using a gene specificprimer (Table 3). The integrity of the RNA was con-firmed by amplifying a constitutively expressed riboso-mal gene, RPL13A. PCR cycling parameters included aninitial denaturation at 94C for 4 minutes followed by 3540cycles of denaturation at 94C for 45 seconds, annealingat 60C for 45 seconds, polymerization at 72C for 1 minute,and a final extension of 20 minutes at 72C. The PCRproducts were separated by 1% agarose gel electrophoresisanalysis.

Enzyme AssaysLipase

The Ortho-Clinical Diagnostics, Inc. Vitros ClinicalChemistry Slide (LIPA) is a dry, multilayered, analytical ele-ment coated on a polyester support. A 10-L drop of patientsample is deposited on the slide and evenly distributed by thespreading layer to the underlying layers. The method incorpo-rates colipase, which facilitates the adsorption of lipase to thesubstrate micelles in the presence of bile salts. Lipase thencatalyzes the hydrolysis of water-insoluble triacetylglycerolesters. The method uses the enzyme diacetinase to convert thesubstrate 2,3-diacylglycerol to glycerol. Glycerol kinase con-verts the glycerol to L-glycerophosphate to generate hydrogenperoxide. Peroxidase oxidizes a leuco dye to produce a coloreddye. The resulting change in reflection density is measured af-ter 3.85 and 5 minutes at a wavelength of 540 nm. The rate ofchange in reflection density is proportional to the activity oflipase present in the sample.

TABLE 2. Antibodies Used for Immunohistochemistry

Antibody Pretreatment Dilution Supplier

PCNA Microwave heat 1:100 Novocastra Laboratories (Newcastle, UK)TGF- 0.05% Saponin 1:20 Oncogene Research Products (Cambridge, MA)IGF-I Microwave heat 1:2000 Cymbus Bioscience (Chandlers Ford, UK)IGF-IR Microwave heat 1:200 Santa Cruz Biotechnology, Inc. (Santa Cruz, CA)Cyclin D1 Microwave heat Ready to use Novocastra Laboratories

PCNA, proliferating cell nuclear antibody; TGF-, transforming growth factor , IGF-1, insulinlike growth factor-I; IGF-IR, IGF-I receptor.



Amylase

Using the Ortho-Clinical Diagnostics, Inc. Vitros Clini-cal Chemistry Slide (AMYL), a 10-L drop of patient sampleis deposited on the slide and is evenly distributed by thespreading layer to the underlying layers. The spreading layercontains the dyed starch substrate (dye covalently linked toamylopectin) for the reaction. The amylase in the sample cata-lyzes the hydrolysis of this dyed starch into smaller dyed sac-charides and then diffuses into an underlying reagent layer thatcontains a cationic hydrophobic polymeric mordant, whichbinds the released anionic dye fragments and removes themfrom solution. The reflection density of the dyed saccharides inthe reagent layer is measured by reflectance spectrophotom-etry at 2.3 and 5 minutes. The change in the slide reflectiondensity between the 2 readings is proportional to sample amy-lase activity.

Trypsin

Trypsin was assayed by a radioimmunoassay method atAssociated Regional and University Pathologists, Inc. (SaltLake City, UT).

Fecal Fat

Feces were collected and stained for fat according to themethods of Drummey et al20 and Fine and Ogunji21 with minormodifications. Briefly, a small amount of stool was placed ona glass slide and 2 drops of 36% acetic acid were added. An-other glass slide was placed on top immediately, and the con-tent was homogenized by grinding between the 2 glass slides,after which 2 drops of 1% Sudan III (Rowley Biochemical Inc.,Danvers, MA) was added. The slides were held, by hand, overa hot plate until bubbles appeared, then they were quickly re-moved and reheated 2 more times. Microscopy was performedimmediately. The assay was performed weekly in all mice. Thefeces of 3 untreated mice were used as control.

Urine Analysis

Reagent strips of Bayer Multistix 10 SG (Bayer Corpo-ration, Elkhart, IN) were used to determine the urine levels ofglucose, bilirubin, and ketone (acetoacetic acid), which wascollected weekly from all mice. The urine of 3 untreated micewas used as control.

Statistical EvaluationThe survival rates of the 2 groups were evaluated by us-

ing the Kaplan-Meier survival curves and compared by thelog-rank test. The Mann-Whitney and Wilcoxon tests wereused to evaluate the statistical significance of tumor size, vol-ume, weight, and body weight.

RESULTS

Stability Testing of the PPESamples taken from the PPE solution at 0, 4, 12, and 24

hours were assayed for amylase, lipase, and trypsinogen. Theresults are given in Table 4. Although the level of amylase andtrypsin was not different in the samples, the level of the lipasewas higher at 12 and 24 hours. Similar results were obtainedwhen PPE was dissolved in water at room temperature. Hence,all enzyme activities were found to be stable in the solution atroom temperature for at least 24 hours (Table 4).

Study 1 (Survival Study)During the tumor cell injection procedure, 1 mouse from

the treatment group and 2 from the control group died. As aresult, treatment and control groups consisted of 14 and 13mice, respectively.

All mice in the treatment group showed normal physicalactivities and no signs of the disease, whereas all mice in thecontrol group showed reduced activity. No differences werefound in the eating and drinking habits between the groupsexcept that from day 21 on, treated mice consumed more waterthan those in the control group (P < 0.0001, Fig. 1), althoughinitially, from days 0 to 7, treated mice drank less water (pos-sibly due to the taste of the PPE) than did the control mice (P 0.05). Weekly food intake per mousewas 29.0 3.7 g in the treated group and 29.5 2.6 g in thecontrol group. There was no difference in the total amount ofenzyme taken by each mouse (Table 6). The average PPE in-take was 11.2 0.01 mg/day.

Survival RatesBecause of the experimental design, an adequate sur-

vival comparison could not be made. Nevertheless, based on afew mice that were allowed to die spontaneously, treated micesurvived considerably longer (58 days) than the control mice(44 days). At day 49, 9 of 14 control mice (64%) in the treat-ment group were alive compared with only 4 of 14 (29%) in thecontrol group (Table 7, Fig. 4).

FIGURE 1. Daily water intake in the treated and controlgroups. The small circle and square indicate the mean, the linein the box indicates the median (50th percentile), and thebottom and the top of the box indicate the 25th and 75thpercentiles of the distribution, respectively.

FIGURE 2. Survival curves of both groups in study 1. The es-timated survival of the 2 groups of animals using Kaplan-Meiersurvival curves and comparing them using the log-rank test.The difference between the survival experience of the 2 groupsis statistically significant (P = 0.009).



Tumor Size and WeightThe treatment group had significantly smaller tumors

than the control group (Fig. 5). The mean tumor weight was 1.2g in the control group and 0.75 g in the control group (P =0.001). Tumor volumes were smaller in the treated mice (0.42cm3) than in the control group (0.91 cm3; P = 0.02) (Fig. 6).The difference in tumor size was more obvious when tumorweight and size were compared between the animals that werekilled at the same time (P = 0.0001). Figure 7 shows the dis-tribution of tumor volume between the 2 groups.

Histology and ImmunohistochemistryAll animals in the treatment and control groups devel-

oped poorly differentiated adenocarcinoma. Most of the pan-

creatic tissues were destroyed by the carcinoma. Little or nointact pancreatic tissue remained in most of the cases from bothgroups. The pancreatic destruction was more significant, espe-cially in the mice with a longer survival. The average tumorsize, taken from the largest diameter of the tumor, was 420mm3 in the treatment group and 1060 mm3 in the control group(P = 0.001). Invasion was found in 2 of 3 mice in the controls(66.6%) and 2 of 4 in the treatment group (50%) (P > 0.05).Necrosis was present in 50% of the tumor tissue of the treat-ment group and 40% in the control group (P > 0.05). No distantmetastasis was found in any mice from both groups.

Table 8 shows the results of immunohistochemicalstaining of the tumor tissue from the mice in the treated andcontrol groups. Proliferating cell nuclear antigen (PCNA)

FIGURE 3. Tumor weights of the groups. This figure displaysboxplots indicating the distribution of tumor weight betweenthe 2 groups. The small circle indicates the mean, the line inthe box indicates the median (50th percentile), and the bot-tom and the top of the box indicate the 25th and 75th per-centiles of the distribution, respectively. The distribution oftumor weight at the time of death was higher for the treat-ment group, and this difference was statistically significant (P =0.04, Wilcoxon test).

TABLE 6. Amount of PPE Consumed by Each Mouse

No. of Mice in a Cage

1 2 3 4 5

Study 1Case 1*

Surv/PPE 38/393 41/419 48/492 53/558Cage 2

Surv/PPE 32/322 38/391 46/488 47/492 53/580Cage 3

Surv/PPE 22/232 32/338 36/386 53/589 53/589Study 2

Cage 1Surv/PPE 60/671 37/414 60/671 51/570 52/581

Cage 2Surv/PPE 60/671 52/581 58/648 60/671 60/671

Cage 3Surv/PPE 47/525 45/503 42/470 43/481

*In each cage, there were 4 or 5 mice.Survival (in days)/PPE intake in milligrams (relationship between lon-

gevity and the amount of PPF consumed). For example, 1 mouse in 1 cagesurvived 38 days and consumed a total of 393 mg PPE). On average, eachmouse from either group consumed about 1011 mg/day.

TABLE 5. Estimated Survival Functions of the Groups in Study 1

Group

Survivalat

T = 35

95% ConfidenceInterval for

Survival at T = 35MedianSurvival

95% ConfidenceInterval for

Median Survival

Treated 79% (57%, 99%) 43.5 (36%, 55%)Control 38% (12%, 65%) 35 (16%, 43%)

We estimated the survival of the 2 groups of animals using Kaplan-Meier survival curves and compared themusing the log-rank test. Some parameter estimates are shown. The difference between the survival experience of the2 groups is statistically significant (P = 0.009).



staining was positive in 67.8% of the cells in the control groupand 42.4% in the treatment group; however, the difference wasnot statistically significant (P > 0.05). Tumor tissue was notstained for IGF-I and IGF-I receptor, whereas normal pancreatictissue was stained strongly positive. Tumor tissue was also notstained for TGF-. The apoptotic rate and cyclin D1 staining didnot show any difference between the groups (P > 0.05).

Expression Analysis of Growth FactorsRT-PCR was used to detect specific mRNA for EGF,

EGFR, and TGF- in control and treated tumor samples. Thepancreatic tumor cell line HPAF was used as a positive controlfor this expression analysis. Ribosomal protein RPL13A wasused as an internal control to show the quality of mRNA.22 Noexpression could be detected in both control and treated samples(Fig. 8). There was no expression of mRNA for EGF, EGFR,

or TGF- in the control or in the treated samples. In PC cellline AsPC1 cells, however, a weak expression of these geneswas detected.

Plasma Amylase and Lipase LevelsAlthough the plasma amylase levels were higher in the

control mice (1545 793 U/L) than in the treated group (1286 604 U/L), the difference was not statistically significant.Also, the level of lipase did not vary between the control group(445.8 100 U/L) and the treatment group (475.6 86 U/L).

Fecal Fat and Urine AnalysisFecal fat was examined in 14 mice in the treatment group

and in 14 mice in the control group. Thirteen of 14

FIGURE 4. Survival curves of both groups in study 2. We estimated the survival of the 2 groups of animals using Kaplan-Meiersurvival curves and compared them using the log-rank test. The survival between the survival experience of the 2 groups is notstatistically significant (P = 0.13).

TABLE 7. Estimated Survival Functions of the Groups in Study 2

GroupSurvival at

T = 4995% Confidence Interval

for Survival at T = 49MedianSurvival

95% Confidence Intervalfor Median Survival

Treated 64% (39%, 89%) 58 (45%, )Control 29% (5%, 52%) 44.5 (42%, )

We estimated the survival of the 2 groups of animals using Kaplan-Meier survival curves and compared them using the log-rank test. Some parameter estimatesare shown. The means that the upper limit of the confidence interval for the median is not defined. The difference between the survival experience of the 2 groupsis not statistically significant (P = 0.13).



mice in the control group and 2 of 14 mice in the treatmentgroup (P < 0.001) had fecal fat (Fig. 9). In the control group,fecal fat appeared as early as week 5. In the treatment group,1 animal had fecal fat at week 6 and the other at week 8. Sixof 9 mice in the control group had fecal fat at week 6, 3 of 4mice at week 7, and both mice at week 8 had fecal fat.

Urinalysis showed that 10 of 14 mice in the controlgroup had ketonuria compared with 4 of 14 mice in the treat-

ment group (P = 0.057) (Fig. 9). Glucosuria developed in 6 of14 in the control group and in 5 of 14 mice in the treatmentgroup. For hyperbilirubinuria, the figures were 5 of 14 and 4 of14, respectively. All values for the untreated, age-matchedmale beige XID nude mice were 0.

DISCUSSIONNo effective therapy guidelines have been established

for PC patients.23 Although neoadjuvant and adjuvant therapyhave support, the respective studies have flaws in design andhave been generally ineffective. Therapy for measurable re-sidual or unresectable PC is attractive in concept, although ef-fective regimens have yet to be established.

PPE has been used in different countries for many yearsin the therapy for a variety of cancers.14 The limited amount ofpublished data suggests that PPE causes tumor regressions oreven remission in terminally ill cancer patients.24,25

Recently, significantly improved survival rates havebeen reported in inoperable PC patients who received PPEorally.14 The article was the case reports of 11 PC patients withstage IV disease. Nine of 11 treated patients (81%) survivedfor 1 year, 5 of 11 patients (45%) survived for 2 years, and 4 of11 (36%) survived for 3 years.14 Two were still alive and doing

FIGURE 5. Xenografts of human PC cells, AsPC1 in nudemice. a: At day 52, tumor size was significantly larger in thecontrol group (left) than in the treatment group (right). b:The same differences were found at day 60. c: At the latestage, tumors in the control group appeared as multiple nod-ules invading the surrounding tissues and peritoneal and pel-vic cavities.

FIGURE 6. Distribution of tumor weights (gram) of the groupsin study 2. Boxplots indicate the distribution of tumor weightbetween the 2 groups. The small circle indicates the mean, theline in the box indicates the median (50th percentile), and thebottom and the top of the box indicate the 25th and 75thpercentiles of the distribution, respectively. The values of tu-mor weight were smaller for the animals in the treatmentgroup, and this difference was statistically significant (P =0.001, Wilcoxon test). Control group mean, 1.20; treatmentgroup mean, 0.75.



well when the article was published (1 at 3 years, the other at 4years). Overall, the median survival was 17 months and themean survival was 25.2 months. These survival rates are insharp contrast to the data of the National Cancer Database Re-port on Pancreatic Cancer from 199515 showing 26% and 6%survival rates for stage IV PC patients after 1 and 2 years, re-spectively. In the statistics from a trial of gemcitabine, the sur-vival rate was 5.7 months; only 18% of patients lived 1 year,and none of them survived beyond 19 months.9 These resultsand the results of the study in which alternative treatment wasused were far from comparable. It must be noted that the diag-nosis was confirmed by a core biopsy of the pancreas in only 4of the 11 patients undergoing PPE therapy, whereas in the re-maining patients, the diagnosis was based on histopathologicexaminations of metastatic foci strongly suggesting (but withno proof) the pancreas as the primary site.

Our intensive search to develop a therapeutic modalityfor PC in an animal model during the past 30 years has re-mained unsuccessful and indicated PCs exceptional resis-tance to therapy. The results of the clinical PPE studiesprompted us to confirm the efficacy of this treatment in anexperimental model. We used the mouse model, which hasgained popularity as a model to study human tumor biology2629

because tumor transplants in nude mice normally and gener-ally maintain the phenotype and biology of the original tu-mor30,31 and respond to therapeutic agents.32 We used AsPC1cells, derived from nude mouse xenografts initiated with cellsfrom the ascites of a patient with PC because this is one of themost malignant human PC cell lines.16

The initial study, which was merely a survival study,showed that the PPE treatment significantly prolonged sur-vival. The larger tumor size in the treated mice survivinglonger indicated that the treatment apparently does not affect

FIGURE 7. Distribution of tumor volume (cm3) of the groups instudy 2. Boxplots indicate the distribution of tumor volumebetween the 2 groups. The small circle and indicates themean, the line in the box indicates the median (50th percen-tile), and the bottom and the top of the box indicate the 25thand 75th percentiles of the distribution, respectively. The val-ues of tumor volume were smaller for the animals in the treat-ment group, and this difference was statistically significant (P =0.02, Wilcoxon test). Control group mean, 0.91; treatmentgroup mean, 0.42.

FIGURE 8. Expression of EGF, EGFR, and TGF- in control andenzyme-treated mice pancreas. RPLA13, a constituent of ribo-somal protein, was used as an internal control. Lane M is a kbladder (Promega Corp). Lanes 14 contained control samples;5 and 6 were PPE-treated samples. CD18-10% was used as apositive control for all these genes. Since these tumors weregenerated in mice by injecting the AsPC I cell line, the expres-sion of these genes was also checked in this cell line.

TABLE 8. Results of Immunohistochemical Staining inBoth Groups

Treated Control P Value

Tumor size (cm3) 0.42 1.06 0.001Necrosis 50% 40% >0.05Invasion 50% 66.6% >0.05PCNA 42.4% 67.8% >0.05TGF No staining No staining >0.05IGF-I No staining No staining >0.05IGF-IR No staining No staining >0.05Apoptosis 15.5% 18.1% >0.05Cyclin D1 50.2% 56% >0.05

PCNA, proliferating cell nuclear antibody; TGF-, transforming growthfactor ; IGF-I, insulin-like growth factor I; IGF-IR, IGF-I receptor.



tumor growth. Because treated mice with large tumors showedphysical activity and behavior comparable with those of thehealthy mice, it was thought that PPE has a beneficial effect onnutrition. The earlier death and the poor health condition of thecontrol mice could well be due to the destruction of the exo-crine pancreas by cancer cells leading to permanent pancreaticenzyme deficiency and, hence, malnutrition. This nutritionalinsufficiency was not reflected by the body weight, possiblybecause of the presence of large tumors and ascites. Hence, thelonger survival of treated mice, whose cancer showed the sameinvasive patterns as those in the control group, could be due tothe compensatory effect of PPE on pancreatic enzyme insuffi-ciency. To clarify these issues, in the second experiment, wecompared some nutritional parameters during the tumorgrowth. The results confirmed the role of PPE in the nutritionalstatus of the mice because control mice showed a significantdecrease of serum pancreatic enzyme levels, steatorrhea, keto-nuria, hyperglucosuria, and hyperbilirubinuria, which oc-curred early and in a severe form. Only a few treated mice atthe late stage presented these abnormalities and in a less severedegree. This comparative study also showed that tumors in thetreated mice grow significantly slower and are less invasivethan those in the matched control mice. The reduced cell pro-liferation index was consistent with this finding. Our hypoth-esis that the PPE treatment, directly or indirectly, affects thetumor growth by altering the expression of the growth factors

known to be overexpressed in human PC cells, however, wasnot validated. There was also no correlation between the sizeof tumors, cyclin D1 expression, and apoptosis rate. Themechanism involved in the suppression of tumor growth is notclear. Theoretically, the enzymes could alter the release ofCCK, a growth-promoting factor, which has receptors in bothendocrine and exocrine cells. Alteration of CCK may alter therelease of insulin, the potent growth hormone. It is also pos-sible that some other ingredients of the PPE act as transcrip-tional factors in inhibiting the release of growth factors.Mechanistic studies are required to understand the actualevent.

Confirming clinical studies, we did not observe any sideeffect that may be associated with PPE treatment. Pancreaticenzyme preparations have been used in several human diseaseswith a wide range of safety. In the literature, the only side ef-fect that may be related to the use of pancreatic enzyme prepa-ration has been observed in children with cystic fibrosis. Inthese patients, taking megadoses of pancreatic enzymes (theequivalent of 10,000 U lipase per kilogram of body weight perday), a distinctive form of fibrosing colonopathy has been re-ported.3335 We could not confirm this alteration in the micetreated with PPE.

In summary, PPE is the first experimentally and clini-cally proven agent for the effective treatment of PC. The sig-nificant advantages of PPE over any other currently availabletherapeutic modalities include its effects on physical condi-tion, nutrition, and lack of toxicity.

ACKNOWLEDGMENTThe authors thank Dr. Linda Lee Isaacs for providing the

porcine pancreatic extract.

REFERENCES1. Ahmedin J, Thomas A, Murray T, et al. Cancer statistics, 2002. CA Can-

cer J Clin. 2002;52:2342.2. Cardillo TM, Blumenthal R, Ying Z, et al. Combined gemcitabine and

radioimmunotherapy for the treatment of PC. Int J Cancer. 2002;97:386392.

3. Cooperman AM. Pancreatic cancer: the bigger picture. Surg Clin NorthAm. 2001;81:557574.

4. Ahmad NA, Lewis JD, Ginsberg GG, et al. Long term survival after pan-creatic resection for pancreatic adenocarcinoma. Am J Gastroenterol.2001;96:26092615.

5. Benassai G, Mastrorilli M, Quarto G, et al. Factors influencing survivalafter resection for ductal adenocarcinoma of the head of the pancreas. JSurg Oncol. 2000;73:212218.

6. Huguier M. Survival after duodenopancreatectomy with mesenteroportalresection for cancer of the head of the pancreas. Chirurgie. 1998;123:421422.

7. Sperti C, Pasquali C, Piccoli A, et al. Survival after resection for ductaladenocarcinoma of the pancreas. Br J Surg. 1996;83:625631.

8. Burris H, Storniolo AM. Assessing clinical benefit in the treatment ofpancreas cancer: gemcitabine compared to 5-fluorouracil. Eur J Cancer.1997;33(Suppl 1):S18S22.

9. Burris HA 3rd, Moore MJ, Andersen J, et al. Improvements in survivaland clinical benefit with gemcitabine as first-line therapy for patients withadvanced pancreas cancer: a randomized trial. J Clin Oncol. 1997;15:24032413.

FIGURE 9. Urine ketone, glucose, and bilirubin and fecal fat.Ketonuria, glucosuria, bilirubinuria, and fecal fat were exam-ined in 14 mice in the treatment group and in 14 mice in thecontrol group. Data were shown as a positive ratio (percentageof total at the end of the week).



10. Talamonti MS, Catalano PJ, Vaughn DJ, et al. Eastern Cooperative On-cology Group Phase I trial of protracted venous infusion fluorouracil plusweekly gemcitabine with concurrent radiation therapy in patients withlocally advanced pancreas cancer: a regimen with unexpected early tox-icity. J Clin Oncol. 2000;18:33843389.

11. Kalser MH, Ellenberg SS. Pancreatic cancer. Adjuvant combined radia-tion and chemotherapy following curative resection. Arch Surg. 1985;120:899903.

12. Gastrointestinal Tumor Study Group. Further evidence of effective adju-vant combined radiation and chemotherapy following curative resectionof PC. Gastrointestinal Tumor Study Group. Cancer. 1987;59:20062010.

13. Klinkenbijl JH, Jeekel J, Sahmoud T, et al. Adjuvant radiotherapy and5-fluorouracil after curative resection of cancer of the pancreas and peri-ampullary region: phase III trial of the EORTC gastrointestinal tract can-cer cooperative group. Ann Surg. 1999;230:776784.

14. Gonzalez NJ, Isaacs LL. Evaluation of pancreatic proteolytic enzymetreatment of adenocarcinoma of the pancreas, with nutrition and detoxi-fication support. Nutr Cancer. 1999;33:117124.

15. Niederhuber JE, Brennan MF, Menck HR. The National Cancer DataBase report on PC. Cancer. 1995;76:16711677.

16. Matsuzaki H, Schmied BM, Ulrich A, et al. Combination of tumor necro-sis factor-related apoptosis-inducing ligand (TRAIL) and actinomycin Dinduces apoptosis even in TRAIL-resistant human PC cells. Clin CancerRes. 2001;7:407414.

17. Egami H, Tomioka T, Tempero M, et al. Development of intrapancreatictransplantable model of pancreatic duct adenocarcinoma in Syrian goldenhamsters. Am J Pathol. 1991;138:557561.

18. Corbert T, Valeriote F, LoRusso P, et al. In vivo methods for screeningand preclinical testing. Use of rodent solid tumors for drug discovery. In:Teicher B, ed. Anticancer Drug Development Guide: Preclinical Screen-ing, Clinical Trials, and Approval. Totowa, NJ: Humana Press; 1997:7599.

19. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex(ABC) in immunoperoxidase techniques: a comparison between ABC andunlabeled antibody (PAP) procedures. J Histochem Cytochem. 1981;29:577580.

20. Drummey GD, Benson JAJ. Microscopical examination of the stool forsteatorrhea. N Engl J Med. 1961;264:823835.

21. Fine KD, Ogunji F. A new method of quantitative fecal fat microscopy

and its correlation with chemically measured fecal fat output. Am J ClinPathol. 2000;113:528534.

22. Jesnowski R, Backhaus C, Ringel J, et al. Ribosomal highly basic 23-kDaprotein as a reliable standard for gene expression analysis. Pancreatology.2002;2:421424.

23. Rosemurgy AS, Serafini FM. New directions in systemic therapy of PC.Cancer Control. 2000;7:437451.

24. Shively FL. Multiple proteolytic enzyme therapy of cancer. Dayton:Johnson-Watson; 1969.

25. Sakalova A, Bock PR, Dedik L, et al. Retrospective cohort study of anadditive therapy with an oral enzyme preparation in patients with multiplemyeloma. Cancer Chemother Pharmacol. 2001;47(Suppl):S38S44.

26. Flanagan SP. Nude, a new hairless gene with pleiotropic effects in themouse. Genet Res. 1966;8:295309.

27. Marincola FM, Drucker BJ, Siao DY, et al. The nude mouse as a model forthe study of human PC. J Surg Res. 1989;47:520529.

28. Dexter DL, Matook GM, Meitner PA, et al. Establishment and character-ization of two human PC cell lines tumorigenic in athymic mice. CancerRes. 1982;42:27052714.

29. Kyriazis AP, McCombs WB 3rd, Sandberg AA, et al. Establishment andcharacterization of human pancreatic adenocarcinoma cell line SW-1990in tissue culture and the nude mouse. Cancer Res. 1983;43:43934401.

30. Bettan-Renaud L, Bayle C, Teyssier JR, et al. Stability of phenotypic andgenotypic traits during the establishment of a human neuroblastoma cellline, IGR-N-835. Int J Cancer. 1989;44:460466.

31. Satyaswaroop PG, Zaino R, Clarke CL, et al. Nude mouse system in thestudy of tumor biology, treatment strategies and progesterone receptorphysiology in human endometrial carcinoma. J Steroid Biochem. 1987;27:431438.

32. Standop J, Schneider MB, Ulrich A, et al. Experimental animal models inpancreatic carcinogenesis: lessons for human PC. Dig Dis. 2001;19:2431.

33. Lowdon J, Goodchild MC, Ryley HC, et al. Maintenance of growth incystic fibrosis despite reduction in pancreatic enzyme supplementation.Arch Dis Child. 1998;78:377378.

34. Smyth RL, Ashby D, OHea U, et al. Fibrosing colonopathy in cysticfibrosis: results of a case-control study. Lancet. 1995;346:12471251.

35. FitzSimmons SC, Burkhart GA, Borowitz D, et al. High-dose pancreatic-enzyme supplements and fibrosing colonopathy in children with cysticfibrosis. N Engl J Med. 1997;336:12831289.



Documents

Pancreatic Enzyme Extract Improves Survival in Murine Pancreatic Cancer