Overexpression of claudin-3 and claudin-4 receptors in uterine serous papillary carcinoma

Overexpression of Claudin-3 and Claudin-4Receptors in Uterine Serous Papillary CarcinomaNovel Targets for a Type-Specific Therapy Using Clostridiumperfringens Enterotoxin (CPE)

Alessandro D. Santin, MD1

Stefania Bellone, PhD1

Moira Marizzoni, PhD1

Michela Palmieri, MS1

Eric R. Siegel, MS2

Jesse K. McKenney, MD3

Leah Hennings, PhD3

Fabrizio Comper, PhD1

Elisabetta Bandiera, PhD4

Sergio Pecorelli, MD4

1 Department of Obstetrics and Gynecology, Divisionof Gynecologic Oncology, University of Arkansas forMedical Sciences, Little Rock, Arkansas.

2 Department of Biostatistics, University of Arkan-sas for Medical Sciences, Little Rock, Arkansas.

3 Department of Pathology, University of Arkansasfor Medical Sciences, Little Rock, Arkansas.

4 Division of Gynecologic Oncology, University ofBrescia, Brescia, Italy.

BACKGROUND. Uterine serous papillary carcinoma (USPC) represents a highly

aggressive variant of endometrial cancer. Using gene expression profiling, we

recently identified high expression of the claudin-3 and claudin-4 receptors in a

limited set of USPC. These tight junction proteins represent the low- and high-

affinity receptors, respectively, for the cytotoxic Clostridium perfringens entero-

toxin (CPE) and are sufficient to mediate CPE binding and trigger subsequent

toxin-mediated cytolysis. The potential for targeting this pathway in the treat-

ment of USPC was explored.

METHODS. Claudin-3 and claudin-4 receptor expression was analyzed at the

mRNA and protein levels in flash-frozen and formalin-fixed, paraffin-embedded

tissue from 20 consecutive USPC patients. The potential of recombinant CPE as

a novel therapy against primary, metastatic, and chemotherapy-resistant USPC

cell lines was also investigated in vitro. Finally, the in vivo therapeutic effect of

sublethal doses of CPE was studied in SCID mouse xenografts harboring subcuta-

neous and intraperitoneal USPC that expressed claudin-3 and claudin-4.

RESULTS. In all, 100% (20 out of 20) of the primary flash-frozen USPC tested

overexpressed 1 or both CPE receptors by quantitative reverse-transcriptase poly-

merase chain reaction (RT-PCR). Membranous immunoreactivity for claudin-4

protein expression was documented in the majority of USPC specimens tested by

immunohistochemistry, whereas only a low level of membranous staining was

found in normal endometrial control tissue samples. When primary and meta-

static short-term USPC cell lines were incubated with different concentrations of

CPE in vitro, a dose-dependent cytotoxic effect was demonstrated. In vivo, intra-

tumoral injections of well-tolerated doses of CPE in large subcutaneous USPC

xenografts led to large areas of tumor cell necrosis and tumor disappearance in

all the treated animals, whereas sublethal intraperitoneal injections of CPE had a

significant inhibitory effect on tumor progression, with extended survival of ani-

mals harboring chemotherapy-resistant intra-abdominal USPC carcinomatosis.

CONCLUSIONS. Claudin-3 and claudin-4 receptors may offer promising targets for

the use of CPE as a novel type-specific therapy against this highly aggressive and

chemotherapy-resistant variant of endometrial cancer. Cancer 2007;109:1312–22.

� 2007 American Cancer Society.

KEYWORDS: uterine serous carcinoma, Clostridium perfringens enterotoxin, clau-din-3 and -4.

C ancer of the uterine corpus represents the most prevalent gyne-

cologic tumor in women, with an estimated 40,880 cases and

7310 deaths in the US in 2005.1 Two subtypes of endometrial carci-

noma, namely, Type I and Type II tumors, have been described,

Supported in part by grants from the Mary KayAsh Foundation; the Angelo Nocivelli; the CamilloGolgi and the Berlucchi foundations, Brescia,Italy; and the Istituto Superiore di Sanita’ (ItalianNational Institute of Health, Programma Oncotec-nologico), Rome, Italy.

Address for reprints: Alessandro D. Santin, MD,UAMS Medical Center, Department of Obstetrics& Gynecology, 4301 W. Markham, Slot 518, LittleRock, AR, 72205-7199; Fax: (501) 686-8091.E-mail: [email protected]

Received October 23, 2006; revision receivedDecember 15, 2006; accepted December 21, 2006.

ª 2007 American Cancer SocietyDOI 10.1002/cncr.22536Published online 26 February 2007 in Wiley InterScience (www.interscience.wiley.com).

1312

based on both clinical and histopathologic features.2

Type I endometrial cancers, which account for the

majority of cases, are usually well differentiated and

endometrioid in histology. These neoplasms are fre-

quently diagnosed in younger women and are asso-

ciated with a history of hyperestrogenism as the

main risk factor, and typically have a favorable prog-

nosis with appropriate therapy. In contrast, Type II

endometrial cancers are poorly differentiated tumors,

often with serous papillary or clear cell histology.

Although Type II tumors account for only a minority

of endometrial cancers, about 50% of all recurrences

occur in this group of patients.3

Our group recently used microarray technology

to analyze the genetic fingerprints of uterine serous

papillary carcinomas (USPC), the biologically more

aggressive variant of Type II endometrial cancer.4

Among the several candidate target genes identified,

genes encoding tight junction (TJ) proteins claudin-3

and claudin-4 were consistently found as 2 of the

most highly up-regulated genes in USPC. Impor-

tantly, claudin-3 and claudin-4 proteins have recently

been shown to represent the natural receptors for

Clostridium perfringens enterotoxin (CPE), and to be

the only family members of the transmembrane tis-

sue-specific claudin proteins capable of mediating

CPE binding and cytolysis.5,6 CPE is a single polypep-

tide of 35 kDa comprised of 319 amino acids that is

associated with C. perfringens type A food poisoning,

the second most commonly reported food-borne ill-

ness in the US.7 CPE triggers lysis of epithelial cells

through interaction with claudin-3 and claudin-4,

with resultant initiation of massive permeability

changes, osmotic cell ballooning, and lysis.5–7 Mam-

malian cells that do not express either claudin-3

and/or claudin-4 fail to bind CPE and are not sus-

ceptible to CPE cytotoxicity 7. The CPE structure/

function relation has been extensively investigated,

mainly by characterizing the functional properties of

enterotoxin fragments and point mutants.8–10 The

CPE290-319 COOH-terminus fragment is sufficient

for high-affinity binding to target cell receptor and

small complex formation, although this fragment is

incapable of initiating large complex formation and

cytolysis. Furthermore, the CPE290-319 COOH-termi-

nus fragment inhibits cytolysis of susceptible target

cells by full-length CPE. Residues 45-116 of CPE are

essential for large complex formation and cytotoxic-

ity, whereas deletion of the NH2 terminus generates

a CPE45-319 fragment with enhanced large mem-

brane complex formation and cytotoxic activity.

In this study we quantified the expression levels

of claudin-3 and claudin-4 receptors by real-time

polymerase chain reaction (PCR) in freshly explanted

USPC. In addition, we analyzed claudin-4 receptor

expression at the protein level by immunohistochem-

istry in a set of formalin-fixed, paraffin-embedded

USPC. Finally, we tested the ability of recombinant

CPE to kill several primary, metastatic, and chemo-

therapy-resistant USPC cell lines in vitro as well as in

vivo. Here we report that USPC highly overexpress

the claudin-3 and claudin-4 receptors and that these

biologically aggressive tumors, regardless of their re-

sistance to chemotherapy, are highly sensitive to CPE

treatment in vitro as well as in vivo. CPE-mediated

therapy may thus represent a novel, potentially

highly effective strategy for the treatment of USPC

refractory to standard treatment modalities.

MATERIALS AND METHODSCloning and Purification of NH2 Terminus His-Tagged CPEClostridium perfringens strain 12917, obtained from

American Type Culture Collection (ATCC, Manassas,

Va), was grown from a single colony and used to pre-

pare bacterial DNA with the InstaGene kit according

to manufacturer’s directions (Bio-Rad, Hercules,

Calif). The bacterial DNA fragment encoding the full-

length CPE gene (GenBank AJ000766) was PCR-

amplified (primer 1, 50-AGA TGT TAA TGG ATC CAT

GCT TAG TAA CAA TTT AAA TCC-30; primer 2, 50-AAAGCT TTT AAA ATT TTT GAA ATA ATA TTG AAT AAG

GG-30). The PCR products were digested with the

restriction enzymes NdeI/BamHI and cloned into an

NdeI/BamHI-digested pET-16b expression vector

(Novagen, Madison, Wis) to generate an in-frame

NH2-terminus His-tagged CPE expression plasmid,

pET-16b-10xHIS-CPE. His-tagged CPE toxin was pre-

pared from pET-16b-10xHIS-CPE transformed Esche-

richia coli M15. Transformed bacteria were grown at

378C to 0.3-0.4 optical density at 600 nm, after which

CPE protein expression was induced overnight with

1 mM isopropyl b-D-thio-galactoside and the cells

harvested, resuspended in 150 mM NaH2PO4, 25 mM

Tris-HCL, and 8 M urea pH 8.0 buffer, and lysed by

centrifugation at 10,000 rpm for 30 minutes. The

fusion protein was isolated from the supernatant on

a Poly-Prep Chromatography column (Bio-Rad). His-

tagged CPE was washed with 300 mM NaH2PO4, 25

mM Tris-HCl, and 10 M urea pH 6.0, and eluted

from the column with 200 mM NaH2PO4, 25 mM

Tris-HCl, and 8 M urea pH 6.0. In order to reduce

the level of endotoxin from His-tagged CPE protein,

10 washings with ice-cold phosphate-buffered saline

(PBS) with Triton X-114 (from 1% to 0.1%) and 10

washings with ice-cold PBS alone were performed.

Dialysis (Mr 3500 cutoff dialysis tubing) against PBS

was performed overnight. Purified CPE protein was

CPE-Mediated Therapy for USPC/Santin et al. 1313

then sterilized by 0.2 mm filtration and frozen in ali-

quots at �708C.

Flash-Frozen Biopsies and Primary Cell LinesTwenty flash-frozen biopsy specimens containing

>75% neoplastic cells were obtained from USPC

patients at the time of primary surgery and 6 pri-

mary USPC cell lines (3 generated from tumor sam-

ples obtained from the primary uterine site and 3

from metastatic intra-abdominal disease or ascites)

were evaluated for claudin-3 and claudin-4 expres-

sion by real-time PCR. Patient characteristics from

which primary specimens were obtained are depicted

in Table 1. Among the 20 USPC cases, 16 were pure

forms, whereas 4 were admixed with endometrioid

or clear cell histology. Specimens containing less

than 25% tumor cells with serous papillary features

were considered mixed USPC. Two of the 6 USPC

specimens established as short-term cultures (ie,

USPC-4 and USPC-6) were confirmed to be highly

resistant to multiple chemotherapeutic agents when

measured as percentage cell inhibition (PCI) by in

vitro Extreme Drug Resistance (EDR) assay (Onco-

tech, Irvine, Calif).11 Specifically, USPC-4 and USPC-

6 were found highly resistant by EDR to the following

drugs: taxol, cisplatin, carboplatin, doxorubicin, and

the combination cisplatin þ gemcitabine). Control

cell lines evaluated in the CPE assays included Vero

cells (purchased from ATCC), normal endometrial

cells (NEC), normal cervical keratinocytes, peripheral

blood lymphocytes (PBL), Epstein-Barr-transformed

B lymphocytes (LCL), and human fibroblasts. With

the exception of normal cervical keratinocytes, which

were cultured in serum-free keratinocyte medium,

supplemented with 5 ng/mL epidermal growth factor

and 35 to 50 mg/mL bovine pituitary extract (Invi-

trogen, Grand Island, NY) at 378C, 5% CO2, all other

fresh specimens were cultured in RPMI 1640 me-

dium (Invitrogen) containing 10% fetal bovine serum

(FBS; Gemini Bio-products, Calabasas, Calif), 200 U/mL

penicillin, and 200 mg/mL streptomycin. All sam-

ples were obtained with appropriate consent accord-

ing to Institutional Review Board (IRB) guidelines.

Tumors were staged according to the FIGO operative

staging system. Tumor debulking included a total

abdominal hysterectomy with bilateral salpingo-

oophorectomy and bilateral pelvic lymphadenectomy

in all USPC patients, whereas normal tissues were

obtained from consenting donors undergoing surgery

for benign pathology. Tumors and normal primary

control tissues were established after sterile process-

ing of the samples from surgical biopsies as pre-

viously described.4,12 Briefly, viable tumor tissue was

mechanically minced in RPMI 1640 to portions no

larger than 1-3 mm3 and washed twice with RPMI

1640. The portions of minced tumor were then

placed into 250-mL flasks containing 30 mL of

enzyme solution (0.14% collagenase Type I and

0.01% DNAse 2000 KU/mg; Sigma, St. Louis, Mo) in

RPMI 1640, and incubated on a magnetic stirring ap-

paratus overnight at 48C. Enzymatically dissociated

tumor was then filtered through 150-mm nylon mesh

to generate a single cell suspension. The resultant

cell suspension was then washed twice in RPMI 1640

plus 10% FBS. The epithelial nature and the purity of

epithelial tumor cultures was verified by immunohis-

tochemical staining and flow cytometric analysis

with antibodies against cytokeratin as previously

described.4,12 In USPC cell lines RNA extraction was

performed at a tumor cell confluence of 50% to 80%

after a minimum of 2 to a maximum of 10 passages

in vitro. Only primary cultures which had at least

90% viability and contained >99% epithelial cells

were used to test sensitivity to CPE in vitro.

RNA Extraction and Quantitative Real-Time PCRRNA isolation from flash-frozen and primary cell

lines was performed using TRIzol Reagent (Invi-

trogen) according to the manufacturer’s instructions.

Quantitative (q-PCR) was performed with an ABI

TABLE 1Clinical and Pathological Characteristics of the Patients

USPC Patients Age Race Stage Histology

1 62* Black IV USPC Pure

2 63* Black III USPC Pure

3 83* White IV USPC Pure

4 64y Black IV USPC Pure

5 43y White III USPC Pure

6 75y White III USPC Pure

7 65 White IV USPC Pure

8 61 Black III USPC Pure

9 78 White III USPC Mixed

10 75 Black IV USPC Pure

11 84 White IV USPC Mixed

12 61 Black III USPC Pure

13 48 White IB USPC Mixed

14 79 White IV USPC Pure

15 84 White IV USPC Mixed

16 60 White III USPC Pure

17 47 Black IB USPC Pure

18 84 White III USPC Pure

19 73 White III USPC Pure

20 71 White III USPC Pure

USPC indicates uterine serous papillary carcinoma.

* Patients from which primary USPC short-term cell lines were established to be used for in vitro

and in vivo Clostridium perfringens enterotoxin (CPE) sensitivity studies.y Patients from which metastatic USPC short-term cell lines were established to be used for in vitro

and in vivo CPE sensitivity studies.

1314 CANCER April 1, 2007 / Volume 109 / Number 7

Prism 7000 Sequence Analyzer using the manufac-

turer’s recommended protocol (Applied Biosystems,

Foster City, Calif) to evaluate expression of claudin-3

and claudin-4 in all the samples. Each reaction was

run in triplicate. Briefly, 1 mg of total RNA from each

sample was reverse-transcribed using SuperScript III

first-strand cDNA synthesis (Invitrogen). Five mL of

reverse-transcribed RNA samples (from 500 mL of

total volume) were amplified by using the TaqMan

Universal PCR Master Mix (Applied Biosystems) to

produce PCR products specific for claudin-3 and

claudin-4. The primers for claudin-3 and claudin-4

were obtained from Applied Biosystems as Assay-on-

Demand products. Assay IDs were Hs00265816_s1

(claudin-3) and Hs00533616_s1 (claudin-4). The

comparative threshold cycle (CT) method (PE

Applied Biosystems) was used to determine gene

expression in each sample relative to the value

observed in the lowest nonmalignant endometrial

epithelial cell sample, using GAPDH (Assay-on-

Demand Hs99999905_m1) RNA as internal controls.

Claudin-4 Immunostaining of Formalin-FixedTumor TissuesUSPC were evaluated by standard immunohisto-

chemical staining on formalin-fixed tumor tissue for

claudin-4 surface expression. Study blocks were

selected after histopathologic review by a surgical pa-

thologist. The most representative hematoxylin and

eosin (H&E)-stained block sections were used for

each specimen. Briefly, immunohistochemical stains

were performed on 4-mm-thick sections of formalin-

fixed, paraffin-embedded tissue. After pretreatment

with 10 mM citrate buffer at pH 6.0 using a steamer,

they were incubated with mouse anti-claudin-4 anti-

bodies (Zymed Laboratories, San Francisco, Calif).

Antigen-bound primary antibody was detected using

standard avidin-biotin immunoperoxidase complex

(Dako, Carpinteria, Calif). Cases with less than 10%

membranous staining in tumor cells were considered

negative for claudin expression. The intensity of

membranous immunoreactivity for claudin-4 in

tumor cells was subjectively scored as follows: a)

1þ, weak staining; b) 2þ, medium staining; and c)

3þ, intense staining. Negative controls, in which the

primary antibodies were not added, were processed

in parallel.

CPE Treatment of Cell Lines and TrypanBlue Exclusion TestTumor samples and normal control cells were seeded

at a concentration of 1 3 105 cells/well into 6-well

culture plates (Costar, Cambridge, Mass) with the

appropriate medium. Adherent tumor samples, fibro-

blasts, and normal epithelial control cell lines were

grown to 80% confluence. After washing and renewal

of the medium, CPE was added to final concentra-

tions ranging from 0.03 to 3.3 mg/mL. After incuba-

tion for 60 minutes to 24 hours at 378C, 5% CO2,

floating cells were removed and stored and attached

cells were trypsinized and pooled with the floating

cells. After staining with Trypan blue, viability was

determined by counting the number of Trypan blue-

positive cells and the total cell number.

SCID Mouse Tumor Xenografts and CPE TreatmentC.B-17/SCID female mice 5–7 weeks old were

obtained from Harlan Sprague-Dawley (Indianapolis,

Ind) and housed in a pathogen-free environment at

the University of Arkansas for Medical Sciences

(UAMS). They were given commercial basal diet and

water ad libitum. The experimental protocol for the

use of these animals for these studies was approved

by the UAMS Institutional Animal Care and Use

Committee. Animals were used to generate subcuta-

neous and intraperitoneal USPC xenografts. Briefly,

for subcutaneous xenografts 10 3 106 USPC-6 cells

in 0.1 mL normal saline were injected subcuta-

neously into the left flank in groups of 3 to 5 mice.

Eight to 10 weeks after USPC-6 injection, when tu-

mor size was approximately 2 cm in diameter, 1 mgCPE dissolved in 100 mL 0.9% NaCl or saline alone

was injected intratumorally daily in the large tumor

area for 5 consecutive days for 2 weeks. On Day 15

some mice were euthanized to determine tumor size,

whereas others were followed to evaluate survival

and possible tumor reappearance. Explanted tumors

were stored in 2% formaldehyde and stained with

H&E for histologic examination. For intraperitoneal

(i.p.) xenografts USPC-6 cells were injected at a dose

of 15 3 106 into C.B-17/SCID mice in groups of 5.

One week after i.p. tumor injection mice were

injected i.p. with 6.0 mg of CPE dissolved in 1 mL of

sterile saline at 96-hour intervals for a total of 5

injections. All animals were observed twice daily and

weighed weekly and survival was monitored. In ad-

dition, groups of mice injected i.p. at a dose of 15

3 106 USPC-6 tumor cells were sacrificed at 1, 2, 3,

and 4 weeks for necropsy and pathologic analysis.

The remaining animals were euthanized and exam-

ined when they became moribund due to i.p. carci-

nomatosis and malignant ascites.

StatisticsClaudin-3 and claudin-4 expression between the

treatment and control groups were compared using

the unpaired t-test. For the USPC-6 animal model,

survivals were plotted using Kaplan-Meier methods

CPE-Mediated Therapy for USPC/Santin et al. 1315

and compared using the log-rank test. A P value less

than .05 was used for statistical significance.

RESULTSClaudin-3 and Claudin-4 Transcript Levels in USPCWe used q-RT-PCR assays to get highly sensitive

measurements of claudin-3 and claudin-4 expression

in 20 USPC flash-frozen specimens, 6 short-term

USPC cultures, and several flash-frozen normal tis-

sues and cell lines. Both claudin-3 and claudin-4

genes were highly expressed in USPC when com-

pared with NEC, normal cervical keratinocytes, EBV-

transformed B cells (LCL), and human fibroblasts

(Fig. 1A). With no exception, all USPC samples (20

out of 20 ¼ 100%) were found positive for claudin-3

and claudin-4 expression by RT-PCR (Fig. 1B). In

contrast, only low levels of claudin-3 and claudin-4

gene expression were found in the NEC control cul-

tures tested (Fig. 1A,B).

When claudin-3 and claudin-4 receptor expres-

sion was compared by RT-PCR between primary cell

lines established from tumors obtained from the en-

dometrial cavity (USCP-1, -2, and -3) vs primary cell

lines established from metastatic sites of disease (ie,

USPC-4, -5, and -6), metastatic USPC were found to

express significantly higher levels of claudin-3 and

claudin-4 receptors (Fig. 1C; P < .05).

Claudin-4 Expression by Immunohistochemistryon USPC and NEC Tissue BlocksTo determine whether the high expression of the

claudin-4 gene detected by q-RT-PCR assays in pri-

mary USPC cell lines is the result of a selection of a

FIGURE 1. (A) Quantitative reverse-transcriptase polymerase chain reac-tion (RT-PCR) analysis of claudin-3 and claudin-4 expression. Y-axis, mRNA

copy number relative to normal endometrial cell (NEC) expression. X-axis,

each sample tested for claudin-3 (upper panel) and claudin-4 (lower panel).

1: Normal cervical keratinocytes (mean � SEM of 3 samples); 2: normal

human fibroblasts (mean � SEM of 3 samples); 3: Epstein-Barr-transformed

B lymphocytes (LCL) (mean � SEM of 3 samples); NEC: normal endometrial

epithelium (mean � SEM of 3 samples); USPC: uterine serous papillary car-

cinomas (mean � SEM of 20 samples). (B) Quantitative RT-PCR analysis of

claudin-3 and claudin-4 mRNA copy number in individual NEC controls and

USPC samples. The first 2 bars represent normal endometrial epithelium.

The following 20 bars represent primary uterine serous papillary carcinoma

samples.3�22 (C) Quantitative RT-PCR analysis of claudin-3 and claudin-4

expression in primary vs metastatic USPC. The Y-axis represents the fold

induction relative to normal endometrial cell expression. The X-axis repre-

sents each sample tested for claudin-3 and claudin-4. 1 (primary uterine

cancers ¼ 3 USPC samples, mean � SEM); 2 (metastatic uterine

cancer ¼ 3 USPC samples, mean � SEM) (P < .05).

3

1316 CANCER April 1, 2007 / Volume 109 / Number 7

subpopulation of cancer cells present in the original

tumor, or whether in vitro expansion conditions may

have modified gene expression, we performed immu-

nohistochemical analysis of claudin-4 protein expres-

sion on formalin-fixed tumor tissue from the

uncultured primary surgical specimens from which

fresh USPC cancers were derived as well as a sepa-

rate set of USPC. As shown in Table 2 and represen-

tatively in the lower panel of Figure 2, moderate to

heavy membranous staining for claudin-4 protein

expression was noted in the cancer specimens that

overexpressed the claudin-4 transcript as well as in

the separate set of USPC tested. In contrast, a much

lower intensity staining pattern was found in the

NEC specimen tested by immunohistochemistry

(Table 2, Fig. 2). Normal mesothelial cells from the peri-

toneal cavity were nonreactive for claudin-4 (Fig. 2).

Effects of CPE on Primary and Metastatic USPC andNormal Control CellsOn the basis of the high expression of claudin-3

and/or claudin-4 on primary USPC cell lines, it was

expected that USPC tumors expressing either clau-

din-3 or claudin-4 would be sensitive to CPE-

mediated lysis. However, it was important to demon-

strate this directly on fresh human USPC carcinoma

cells, particularly in a clinically relevant setting of

USPC disease for which current salvage therapies are

ineffective (ie, metastatic/chemotherapy-resistant

disease). For this reason we examined and tested the

sensitivity of short-term in vitro cultures of USPC

obtained either from the primary endometrial site

(USPC-1 to -3) or from metastatic locations (USPC-4

to -6) to scalar doses of CPE. The sensitivity of these

primary USPC tumor cultures to CPE-mediated cyto-

lysis was tested along with an appropriate claudin-3

and claudin-4-expressing positive control (ie, Vero

cells), and negative controls which do not express

detectable levels of either claudin-3 or claudin-4. As

representatively shown in Figure 3 for USPC-6, a

metastatic tumor cell line found highly resistant to

chemotherapy in vivo as well as in vitro by EDR

(data not shown), dramatic morphological alterations

of tumor cells including rapid detachment and ‘‘bleb

balloon’’ formation were evident after 5 minutes expo-

sure to concentrations of CPE as low as 0.8 mg/mL.

Importantly, all USPC tested, regardless of their

primary or metastatic origin or their resistance to

chemotherapy, were highly sensitive to CPE-me-

diated cytolysis (Fig. 4). The cytotoxic effect was

dose-dependent and correlated positively with the

levels of either claudin-3 or claudin-4 expression as

tested by RT-PCR in tumor samples. With the single

exception of the USPC-3 cell line, which was partially

resistant to CPE (ie, 58% tumor killing at 24 hours

CPE exposure), all remaining USPC cultures includ-

ing those derived from metastatic/chemotherapy re-

sistant disease (USPC-4 to -6) were highly sensitive

to the exposure to CPE (Fig. 4). In contrast, all nor-

mal control cells tested including NEC, cervical kera-

tinocytes, human fibroblasts, and mononuclear cells

lacking claudin-3 or claudin-4 expression were not

affected by CPE (Fig. 4).

Effect of CPE on Xenografted Chemotherapy-ResistantUSPC Cells In VivoFor in vivo confirmation of our in vitro data, we

established subcutaneous and intraperitoneal USPC

xenografts in SCID mice by injection of claudin-3

and claudin-4 expressing USPC-6, a metastatic and

highly chemotherapy-resistant USPC cell line. In sub-

cutaneous experiments, after the formation of large

subcutaneous tumors (ie, 2 cm in diameter), animals

were treated with a total of 10 injections of CPE or

saline directly into the tumor over a 2-week period.

After 15 days from the beginning of CPE therapy,

some animals were killed and the tumors were exam-

ined histologically. Tumor size was compared by

measuring the tumor volume before treatment and

upon death. As representatively shown in Figure 5,

subcutaneous CPE-treated USPC-6 tumors showed a

TABLE 2Claudin-4 Staining

Patient Claudin-4 Positivity

NEC 1 1þNEC 2 1þNEC 3 1þUSPC 1 3þUSPC 2 3þUSPC 3 3þUSPC 4 2þUSPC 5 3þUSPC 6 2þUSPC 7 3þUSPC 8 1þUSPC 9 3þUSPC 10 3þUSPC 11 3þUSPC 12 3þUSPC 13 1þUSPC 14 2þUSPC 15 3þUSPC 16 2þUSPC 17 3þUSPC 18 3þUSPC 19 2þUSPC 20 3þ

USPC indicates uterine serous papillary carcinoma; NEC, normal endometrial cells.

CPE-Mediated Therapy for USPC/Santin et al. 1317

dramatic reduction in tumor size, with clinical disap-

pearance of the disease in all CPE-treated animals.

In contrast, animals treated with saline control injec-

tions developed rapidly progressing disease and had

to be euthanized 12 to 14 weeks after tumor injec-

tion. All animals examined for histologic confirma-

tion of regression after CPE treatment showed large

necrotic areas comprising up to 90% of the histologic

sections (Fig. 6). In contrast, we found no necrosis in

the histologic sections of all control animals injected

with saline (not shown).

Because USPC, like ovarian serous papillary

tumors, may rapidly spread to the abdominal cavity

and induce a lethal peritoneal carcinomatosis in

humans, we established a further xenograft tumor

model in SCID mice by i.p. injection with the meta-

static chemotherapy-resistant USPC-6 cell line. Pri-

mary USPC-6 tumor cells grew progressively as

numerous serosal nodules adherent to virtually all

intra-abdominal organs (peritoneum, omentum, dia-

phragm, bowel, liver, pancreas, spleen) and exhibited

the capacity for local tissue invasion and formation

of malignant ascites after 2 to 3 weeks from injec-

tion. Tumors first appeared grossly by the second

week as small nodules on the omentum and continu-

ously grew to form a confluent omental mass by the

time the animals died (Fig. 7). Necropsies revealed

massive hemorrhagic ascites and numerous tumor

nodules, measuring 1 to 15 mm in diameter, stud-

ding the entire peritoneal surface and implanting the

serosa of virtually all intra-abdominal organs (Fig. 7).

Because in a previous report we found 6 mg of CPE

administered i.p. in 1 mL of saline to be a well-toler-

ated and safe dose in 100% of the animals (ie,

16.5 � 1.0 g female SCID mice),13 mice harboring

USPC-6 (a week after tumor injection with 15 3 106

cells) were treated with 6 mg of CPE administered i.p.

in 1 mL of saline every 96 hours for a total of 5 injec-

FIGURE 2. Representative immunohistochemical staining for (A) claudin-4 on normal endometrial cell (NEC) 1 paraffin-embedded specimen, (B) human meso-thelial cells, and (C,D) 2 USPC specimens. NEC 1 and mesothelial cells showed light to negligible membrane staining for claudin-4, whereas both USPC showed

strong cytoplasmic and membranous reactivity for claudin-4. Original magnification 3400.

1318 CANCER April 1, 2007 / Volume 109 / Number 7

tions. CPE injections were well tolerated and no

adverse events were observed throughout the com-

plete treatment protocol either in control mice

receiving CPE alone or CPE-treated mice harboring

tumors. Importantly, whereas mice harboring i.p.

USPC-6 treated with saline all died within 9 weeks of

tumor injection (Fig. 8), animals treated with 5 CPE

injections survived significantly longer than control

animals did (P < .002, Fig. 8).

DISCUSSIONOur group has recently evaluated the genetic finger-

print of USPC,4 the most biologically aggressive and

chemotherapy-resistant variant of endometrial can-

cer. Claudin-3 and claudin-4, the natural receptors

for CPE, were found among the highest differentially

expressed genes in USPC. Because USPC are histolo-

gically similar to high-grade ovarian serous tumors

in their ability to rapidly spread to the abdominal

cavity and, in addition, are notorious for their high

resistance to chemotherapy,4,14–16 these findings

imply that USPC refractory to standard treatment

modalities may be susceptible to CPE-based thera-

peutic approaches. In this study we carefully evalu-

ated the expression of CPE receptors at both the

RNA and protein levels in multiple flash-frozen

USPC biopsies and primary USPC cell lines. In addi-

tion, we studied the sensitivity of primary, metastatic,

and chemotherapy-resistant USPC cell lines to CPE

treatment in vitro. Finally, we tested the in vivo effi-

cacy of local/regional CPE administrations as novel

therapy in 2 SCID mouse xenograft models harboring

established USPC refractory to chemotherapy.

Our studies demonstrated that 100% (20 out of

20) of the USPC tested for claudin-3 and claudin-4

expression by quantitative RT-PCR overexpress both

the high-affinity CPE receptor (claudin-4) as well as

the low-affinity CPE receptor (claudin-3). Impor-

tantly, USPC cell lines established from metastatic

foci of disease were found to express significantly

higher levels of claudin-3 and claudin-4 receptors

when compared with USPC cell lines established

from biopsies obtained from the primary tumor site

FIGURE 3. Clostridium perfringens enterotoxin (CPE)-mediated cytotoxicityof uterine serous papillary carcinoma (USPC)-6 chemotherapy-resistant tumor

cells. Upper panel: Typical phenotype of primary monolayer USPC-6 tumor

cells in vitro before CPE exposure. Lower panel: USPC-6 cell line after 5-mi-

nute exposure to 0.8 mg/mL of CPE. Note the dramatic morphological altera-

tions of USPC-6 tumor cells, which are now partially detached, nonviable,

and characterized by a ‘‘bleb balloon’’ formation.

FIGURE 4. Representative dose dependent Clostridium perfringens entero-toxin (CPE)-mediated cytotoxicity of uterine serous papillary carcinoma

(USPC) cell lines compared with positive control Vero cells or negative con-

trols (ie, normal cells) after 24 hours exposure to CPE. Vero, positive control

cells. USPC-1 to USPC-6 uterine serous tumors. NEC, normal endometrial

cells; norm CX, normal cervix keratinocytes; fibroblast, normal human fibro-

blasts; LCL, Epstein-Barr-transformed B lymphocytes; PBL, normal peripheral

blood lymphocytes.

CPE-Mediated Therapy for USPC/Santin et al. 1319

(ie, endometrial cavity). Although the biological rea-

sons for these differences in claudin-3 and claudin-4

receptor expression between primary and metastatic

sites of disease are not well understood, in high-

grade serous papillary ovarian cancer (OSPC), a tu-

mor histologically indistinguishable from USPC,

overexpression of claudin-3 and claudin-4 has

recently been associated with an increase in tumor

cell motility, invasion capability, and tumor cell sur-

vival.17 These data are consistent with our results in

USPC and suggest that high expression of claudin-3

and claudin-4 receptors in serous papillary uterine

and ovarian tumors may represent a marker of bio-

logical aggressiveness that correlates with the meta-

static process. Importantly, the extremely high

expression of claudin-3 and claudin-4 detected in all

metastatic USPC cell lines evaluated so far, including

USPC-6, a highly chemotherapy-resistant tumor cell

line, suggests that biologically aggressive USPC dis-

ease may be particularly susceptible to CPE-

mediated killing in vitro as well as in vivo. In agree-

ment with this view, with a single exception (ie,

USPC-3), no USPC cell line was found viable after 24

hours exposure to CPE at the concentration of 3.3

mg/mL, a dose well tolerated when administered i.p.

in animal models harboring chemotherapy-resistant

USPC xenografts overexpressing claudin-3 and clau-

din-4 receptors. Importantly, in 5 out of 6 USPC cell

lines we found that exposure to CPE doses as low as

0.8 mg/mL was able to induce massive ‘‘bleb balloon-

ing’’ and tumor cell death throughout USPC cultures

within minutes, demonstrating the dramatic sensitiv-

ity of these biologically aggressive tumors to CPE.

These results were in strong contrast with the lack of

sensitivity of normal endometrial epithelium as well

as other normal control cells to CPE-mediated cyto-

lysis. These findings are likely explained by a limited

expression of claudin-3 and claudin-4 in normal epi-

thelia compared with uterine serous tumor cells.

In vivo, multiple injections of well-tolerated

doses of CPE administered intratumorally led to

massive tumor necrosis and dramatically inhibited

subcutaneous tumor growth in claudin-3 and clau-

din-4-expressing USPC-6 SCID mouse xenografts.

Importantly, these results were obtained by challen-

ging mice harboring extremely large tumor xeno-

grafts of USPC (ie, 10-week established tumors).

Furthermore, 5 i.p. injections of sublethal doses of

CPE had a significant inhibitory effect on tumor pro-

FIGURE 5. Gross appearance of subcutaneous uterine serous papillarycarcinoma (USPC)-6 xenografts in SCID mice (A) before and (B) after Clos-

tridium perfringens enterotoxin (CPE) treatment. Note the almost complete

disappearance of the large tumor nodule after multiple intratumoral CPE

injections.

FIGURE 6. Intratumoral injections of Clostridium perfringens enterotoxin

(CPE) in uterine serous papillary carcinoma (USPC)-6 subcutaneous xeno-

grafts. After H&E staining, large necrotic areas comprising up to 90% of the

histologic section were detected in CPE-treated animals harboring chemo-

therapy-resistant disease (3200 magnification).

1320 CANCER April 1, 2007 / Volume 109 / Number 7

gression, with extended survival of animals harboring

established chemotherapy-resistant USPC in the ab-

dominal cavity. It is worth noting, however, that

because claudin-3 and/or claudin-4 may be

expressed in some normal human tissues, including

the gut, lungs, and kidney,18 the potential high toxic-

ity of CPE at doses used for systemic cancer therapy

in animal models may ultimately limit its use in

humans to regional applications. In this regard,

unlike previous toxin-based anticancer approaches,

such as antibody-based targeting strategies to deliver

activated forms of protein synthesis directed toxins

(for example, Pseudomonas aeruginosa exotoxin),19

CPE is a membrane integrity-directed toxin with a

very different intrinsic mechanism of sensitivity and

toxicity.6–8,18 Taken together, our results suggest that

the local/regional administration of well-tolerated

doses of CPE may have great potential as a novel

treatment modality in patients harboring surgically

unresectable chemotherapy-resistant USPC. The

future design and implementation of Phase I clinical

trials in patients harboring chemotherapy-resistant

USPC will determine the feasibility and validity of

this novel therapeutic approach.

REFERENCES1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005.

CA Cancer J Clin. 2005;55:10–30.

2. Bohkman JV. Two pathogenetic types of endometrial carci-

noma. Gynecol Oncol. 1983;15:10–17.

3. Santin AD, Bellone S, O’Brien TJ, Pecorelli S, Cannon MJ,

Roman JJ. Current treatment options for endometrial can-

cer. Exp Rev Anticancer Ther. 2004;4:679–689.

FIGURE 7. Typical necropsy specimen from C.B-17/SCID mice after 8 to 9 weeks from the intraperitoneal injection of 15 3 106 viable uterine serous papil-

lary carcinoma (USPC)-6 cells. Note the large omental and pelvic tumor masses and the presence of superficial liver implants.

FIGURE 8. Survival of C.B-17/SCID Mice after intraperitoneal (i.p.) injectionof 15 3 106 viable uterine serous papillary carcinoma (USPC)-6 tumor cells.

Animals harboring 1 week established USPC-6 xenografts were injected i.p.

with a total dose of 6 mg of Clostridium perfringens enterotoxin (CPE) in

1 mL of sterile saline solution as described in Materials and Methods. CPE

was administered i.p. every 96 hours for a total of 5 times. Mice were eval-

uated on a daily basis and sacrificed when moribund. The log-rank test

yielded P < .002 for the differences in survival.

CPE-Mediated Therapy for USPC/Santin et al. 1321

4. Santin AD, Zhan F, Bellone S, et al. Discrimination be-

tween uterine serous papillary carcinomas and ovarian se-

rous papillary tumours by gene expression profiling. Br J

Cancer. 2004;90:1814–1824.

5. Katahira J, Inoue N, Horiguchi ,Y, Matsuda M, Sugimoto N.

Molecular cloning and functional characterization of the

receptor for Clostridium perfringens enterotoxin. J Cell

Biol. 1977;136:1239–1247.

6. Katahira J, Sugiyama H, Inoue N, Horiguchi Y, Matsuda M,

Sugimoto N. Clostridium perfringens enterotoxin utilizes

two structurally related membrane proteins as functional

receptors in vivo. J Biol Chem. 1997;272:26652–26658.

7. McClane BA. An overview of Clostridium perfringens en-

terotoxin. Toxicon. 1996;34:1335–1343.

8. Kokai-Kun JF, McClane BA. Evidence that a region(s) of the

Clostridium perfringens enterotoxin molecule remains ex-

posed on the external surface of the mammalian plasma

membrane when the toxin is sequestered in small or large

complexes. Infect Immun. 1996;64:1020–1025.

9. Kokai-Kun JF, McClane BA. Determination of functional

regions of Clostridium perfringens enterotoxin through de-

letion analysis. Clin Infect Dis. 1977;25(Suppl 2):S165-167.

10. Kokai-Kun JF, McClane BA. Deletion analysis of the Clos-

tridium perfringens enterotoxin. Infect Immun. 1997;65:1014–

1022.

11. Holloway RW, Mehta RS, Finkler NJ, et al. Association

between in vitro platinum resistance in the EDR assay and

clinical outcomes for ovarian cancer patients. Gynecol

Oncol. 2002;87:8–16.

12. Santin AD, Diamandis EP, Bellone S, et al. Human kallik-

rein 6: a new potential serum biomarker for uterine serous

papillary cancer. Clin Cancer Res. 2005;11:3320–3325.

13. Santin AD, Cane S, Bellone S, et al. Treatment of chemo-

therapy-resistant human ovarian cancer xenografts in C.B-

17/SCID mice by intraperitoneal administration of Clos-

tridium perfringens enterotoxin. Cancer Res. 2005;65:4334–

4342.

14. Levenback C, Burke TW, Silva E, Gershenson D. Uterine

papillary serous carcinoma (USPC) treated with cisplatin,

doxorubicin, and cyclophosphamide (PAC). Gynecol Oncol.

1992;46:317–32.

15. Nicklin JL, Copeland LJ. Endometrial papillary serous car-

cinoma: pattern of spread and treatment. Clin Obstet Gyne-

col. 1996;39:686–695.

16. Trope C, Kristensen GB, Abeler VM. Clear-cell and papillary

serous cancer: treatment options. Best Pract Res Clin Obstet

Gynaecol. 2001;15:433–446.

17. Agarwal R, D’Souza T, Morin PJ. Claudin-3 and claudin-4

expression in ovarian epithelial cells enhances invasion

and is associated with increased matrix metalloproteinase-

2 activity. Cancer Res. 2005;65:7378–7385.

18. Michl P, Buchholz M, Rolke M, et al. Claudin-4: a new tar-

get for pancreatic cancer treatment using Clostridium per-

fringens enterotoxin. Gastroenterology. 2001;121:678–684.

19. Mansfield E, Amlot P, Pastan I, FitzGerald DJ. Recombinant

RFB4 immunotoxins exhibit potent cytotoxic activity

for CD22-bearing cells and tumors. Blood. 1997;90:2020–

2026.

1322 CANCER April 1, 2007 / Volume 109 / Number 7