Characterization of 5-Oxo-L-prolinase in Normal and Tumor ... · enhanced in vivo tumor response to...

Vol. 4, 131-138, Janua�’ /998 Clinical Cancer Research 131

Characterization of 5-Oxo-L-prolinase in Normal and Tumor Tissues

of Humans and Rats: A Potential New Target for Biochemical

Modulation of Glutathione1

Xiang Chen, Robyn L. Schecter,

Owen W. Griffith, Michael A. Hayward,

Lesley C. Alpert, and Gerald Batist2

McGill Center for Translational Research in Cancer, Lady DavisInstitute for Medical Research, Sir Mortimer B. Davis-Jewish GeneralHospital, McGill University, Montreal, Quebec, H3T 1E2 Canada

[X. C., R. L. S., L. C. A., G. B.], and Department of Biochemistry,Medical College of Wisconsin, Milwaukee, Wisconsin 53226[O.W.G.. M.A.H.]

ABSTRACT

5-Oxo-L-prolinase (5-OPase) is an enzyme of the y-glu-

tamyl cycle involved in the synthesis and metabolism ofglutathione (GSH), which is known to protect cells from thecytotoxic effects of chemotherapy and radiation. Previous

studies on rats have shown that administration of the cys-teine prodrug L-2-oxothiazolidine-4-carboxylate, a 5-oxo-L-proline analogue that is metabolized by 5-OPase, preferen-flabby increases the GSH content of normal tissues whileparadoxically decreasing it in the tumor and results in anenhanced in vivo tumor response to the anticancer drug

melphalan. These observations initiated the present study of5-OPase in experimental models and clinical specimens to

investigate the potential role of this enzyme in the selectivemodulation of GSH in normal and tumor tissues. First,

5-OPase activity was measured in tissues of tumor-bearing

rats, in the peripheral mononuclear cells of normal humansubjects, and in surgically resected tumor and the adjacentnormal tissues from patients. We found that the activity of5-OPase in human kidney, liver, and lung is significantly

lower than that found in rats. Second, we have raised apolyclonal IgG anti-5-OPase antibody by immunizing rab-bits with purified 5-OPase from rat kidney. This antibody

has very high affinity (shown by immunoprecipitation) and

specificity (shown by Western blot) and cross-reacts withhuman 5-OPase (shown by Western blot and immunohisto-chemistry). It was then used to examine the distribution of

Received 4/10/97; revised 9/16/97; accepted 10/10/97.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to

indicate this fact.

I Supported by grants from the Cancer Research Society, Inc., Canadaand the National Cancer Institute of Canada.2 To whom requests for reprints should be addressed, at McGill Centerfor Translational Research in Cancer, Sir Mortimer B. Davis-JewishGeneral Hospital, 3755 Chemin de Ia Cote-Ste-Catherine, Suite D-127,Montreal, Quebec, H3T lE2 Canada. Phone: (514) 340-8248; Fax:(514) 340-8302.

5-OPase in paired normal and neoplastic human specimens

using Western blot and immunohistochemistry. Examina-

tion of paired normal and neoplastic tissues of stomach and

lung revealed a significantly lower level of 5-OPase in tumortissues than in the paired normal tissues. In colon tissues,

there is no significant difference in 5-OPase level betweenthe normal and tumor tissues. These findings could haveimplications for both carcinogenesis and therapy.

INTRODUCTION

5-OPase3 is one of the five enzymes involved in the -y-glu-

tamyl cycle, an interrelated series of reactions involved in the

synthesis and metabolism of GSH. 5-OPase catalyzes the ATP-

dependent hydrolysis of 5-oxo-L-proline to L-glutamate.

Whereas 5-oxo-L-proline is the metabolite of the -y-glutamyl

residue in GSH, L-glutamate is one of the three amino acid

substrates for GSH synthesis. Therefore, 5-OPase links the

reactions involved in GSH metabolism with those involved in

GSH synthesis in the ‘y-glutamyb cycle (1-3).

Because GSH plays a critical role in protecting cells from

toxic oxygen species and various xenobiotics, including many

anticancer drugs and carcinogens, numerous studies have been

undertaken to modulate the cellular GSH levels to improve the

effect of chemotherapy on cancer or produce chemoprotection

against carcinogens. Most of the enzymes involved in the -y-glu-

tamyl cycle have been investigated thoroughly. However, nei-

ther 5-OPase itself nor its role in GSH modulation has been as

extensively studied (4-8). The main reason is the limited char-

acterization of the enzyme and lack of probes such as antibodies

for its study. Another reason may be the observation that under

normal conditions, cells do not necessarily depend on S-OPase

for GSH synthesis. because 5-oxo-L-proline has no significant

feedback inhibition on OSH synthesis, and glutamate from

many other sources is usually available and sufficient for this

purpose (9). Rare cases of an inborn deficiency of 5-OPase have

been reported that are associated with a variety of clinical

manifestations but do not necessarily lead to GSH deficiency (9.

10). However, several studies have shown that OTZ, an ana-

bogue of S-oxo-L-probine, can significantly increase the cellular

GSH level when cells are under oxidative stress ( 1 1-1 3). Be-

cause 5-OPase intracelbularly converts OTZ to L-cysteine. a

rate-limiting substrate for GSH synthesis, these studies suggest

that in certain circumstances, the role of 5-OPase in maintaining

cellular GSH content seems to be more important than under

normal conditions.

3 The abbreviations used are: 5-OPase, 5-oxo-t-prolinase: GSH, gluta-

thione; OTZ, L-2-oxothiazolidine-4-carboxylate; ECL. enhanced chemi-

luminescence.

Research. on March 17, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

132 5-OPase in Cancer Chemotherapy

A number of studies have identified another potentially

important robe of 5-OPase in GSH modulation in the context of

cancer chemotherapy. It was observed that tumor cell lines

usually have lower 5-OPase content than normal cells in culture

(14). Russo et a!. (12) showed that OTZ administration can

increase the GSH content in a normal lung fibroblast cell line

but not in a lung cancer cell line, resulting in a selective

potentiation of the cytotoxicity of several anticancer drugs in the

cancer cells. We have recently shown that administration of

OTZ to tumor-bearing rats increases the GSH level in normal

tissue (including bone marrow) while paradoxically decreasing

the GSH bevel in the tumor (15), and administration of OTZ

together with the anticancer drug melphalan significantly en-

hances the in vivo tumor response to melphalan ( 16). Theseobservations stimulated our investigation of 5-OPase in exper-

imental models and clinical specimens. We have purified the

enzyme from rat kidney and generated a high-affinity antibody

against it. Here we present our results on the tissue distribution

of 5-OPase as well as the preliminary enzymatic and immuno-

examination of 5-OPase in rat and clinical human specimens.

MATERIALS AND METHODS

Reagents and Tissues

S-Oxo-L-[’4C]-proline was prepared by cyclization of

L-[’4C]-glutamate (DuPont New England Nuclear, Mississauga,

Canada) as described previously (17).

The animal model used in this study was the MatB tumor-

bearing female Fisher 344 rat. Five X i05 MatB cells (rat

mammary carcinoma cells) were s.c. injected into 10-12-week-

old rats, and a solid tumor was palpable in approximately 10

days. At that time, the rats were sacrificed, and tissue samples of

kidney, liver, lung, and bone marrow from limb long bones and

tumors were excised, assayed immediately for 5-OPase activity,

and processed routinely for histology. The bone marrow cells

were collected by centrifugation through Ficoll-Paque (Pharma-

cia Biotech, Uppsala, Sweden).

Human tissue samples were obtained immediately after

surgical excision. The tumor and normal tissues were dissected

by the pathologist (L. C. A.) and processed separately for

5-OPase activity and for histological analysis within 30 mm.

Samples of normal kidney included both cortex and medulla.

Samples of normal colon and stomach consisted of mucosa and

scant amounts of attached submucosa. Samples of normal lung

contained only subsegmental bronchi and alveolar parenchyma.

Human peripheral blood mononuclear cells were collected

from 13 normal donors in the fasting state before 10 a.m. and

purified by centrifugation through Ficoll-Paque.

Enzymatic Assay of 5-OPase

Tissues or cells were homogenized in ice-cold buffer A [50

mM Tris (pH7.3), 0.1 msi EDTA, 2 msi DTf, and 5 nmi

S-oxo-L-probine]. Purification of 5-OPase followed the liquid

chromatography method described previously (18). The activity

of 5-OPase was determined by measuring the conversion rate of

5-oxo-L-[’4C1-proline to L-[’4C]-gbutamate (18). The stability of

5-OPase in tissues was tested by assaying the 5-OPase activity

in rat kidney after the kidney had been kept at 4#{176}Cfor different

time intervals up to 30 h after sacrificing the rat, and after the

dialyzed rat kidney solution was frozen at -80#{176}Cfor up to 1

week. Results showed no decline of 5-OPase activity in the

delayed assay samples as compared to the fresh tissue samples.

The enzyme activity was therefore stable under the conditions

tested. The standard curve of this assay method was tested with

purified rat kidney 5-OPase and proven to be linear and sensi-

tive even within the low range of 10-100 milbiunits (data not

shown).

Western Blot

An ECL Western blotting protocol (Amersham, Bucking-

hamshire, United Kingdom) was followed. The primary anti-

body at a 1:50,000 dilution was incubated for 3 h, and the

secondary antibody at a 1:2000 dilution was incubated for 1 h at

room temperature.

Immunohistochemistry

Histological sections of formabin-fixed paraffin-embedded

tissue samples were examined for immunodetectable 5-OPase

using standard techniques. The protein blocking agent, the sec-

ondary antibody solution, and the streptavidin peroxidase rea-

gent were from Lipshaw Immunon (Pittsburgh, PA). The avidin/

biotin blocking kit was from Vector Laboratories, Inc.

(Burlingame, CA). The primary antibody, diluted to 1:500 and

1: 1000 in 5% normal goat serum, was applied and incubated

overnight at 4#{176}C.Preimmune rabbit serum diluted to 1:750 was

used as control. Incubation with the biotinylated secondary

antibody and then with the streptavidin-peroxidase conjugate

was 30 mm each.

Antibody Production and Validation

5-OPase Purification and Immunization. Rat kidney

5-OPase was partially purified from frozen kidneys using a

previously described method (18). The 5-OPase in the semipu-

rifled pool was separated by preparative SDS-PAGE. The

5-OPase band, identified by its molecular weight (140,000; Ref.

19) and by the fact that it was the most dominant band on the

gel, was then cut from the gel, minced, and emulsified with an

equal volume of incomplete Freunds adjuvant (WA; Sigma

Chemical Co., St. Louis, MO) and injected into multiple i.m. or

s.c. sites (0.5-1 mI/site) of a New Zealand strain White rabbit.

The rabbit was bled 12 days after the priming immunization.

The same procedure was followed for additional boost immu-

nizations (once in 4-8 weeks or when the antibody titer was

found to be declining).

ELISA to Detect Antibody. 5-OPase was coated onto a

96-well microtiter plate (Dynatech Laboratories, Chantilly, VA)

at 0.2 p.g/well overnight under vacuum. ELISA followed pro-

tocobs described previously (20). A goat antirabbit IgG (H+L)

horseradish peroxidase conjugate (Bio-Rad, Richmond, CA)

was used as secondary antibody at a 1:4,000 dilution. The

primary antibody exposure was 2 h, and that of secondary

antibody was 1 h. The developing solution was 0.03% tetram-

ethylbenzidine [stock solution was 0.125% (v/v) tetramethyb-

benzidine in methanol] and 0.02% H202 in citrate phosphate

buffer [0.1 M Na2HPO4 (pH 5.0) and 0.05 M citric acid mono-

hydrate]. Development time was 30 mm. The plate was read on

a Dynatech ELISA reader (Dynatech; Fisher) at 405 nm wave-



0 20000 40000 60000

60

50

40

30

20

0 100 200 300 400 500

�‘ c� � eqt� Co ,-� C�

Clinical Cancer Research 133

A

0)C

(8a,

Cl)Ca)

�0

(I)

-Jw

B0)

E

E

>

0

(8

a,U)(8a.0Cl,

CCa(8

Ea)0.

C’)

C

D

antiserum dilution (1 :)

Antiserum dilution (1:)

1 2 3

5-OPase.

activity (mU/mg)-#{248}

$4

‘0- 5-OPase

� $4‘O �‘

length. All experiments were done in triplicate. All steps were

carried out at room temperature. The result (Fig. lA) showed

that the preimmune and the priming immune serum did not

generate any antibody/antigen binding signal. After the first

booster immunization, even at a 1 :50,000 dilution of the serum,

a 25-fold stronger signal in the immune serum than in the

preimmune serum was detected. Therefore, this antibody was

produced in a significant titer in the rabbit.

Immunoprecipitation to Determine the Affinity of An-

tibody. Twenty-five p.g of 5-OPase (47 milbiunits) and the

primary antibody at various dilutions were mixed, and the

mixture was rotated for S h at 4#{176}C.This antigen/antibody

mixture was then mixed with protein A-Sepharose CL-4B

(Pharmacia LKB), rotated for 2 h at 4#{176}C,and then centrifuged.

The supernatant and the Sepharose precipitate were used for the

enzymatic activity assay and Western blot experiment, respec-

lively. All samples were in duplicates. The result showed that

the preimmune serum had no specific effect on 5-OPase activity

in the supernatant (the small decline in the activity observed is

more likely due to the dilution of the reaction mixture), whereas

the amount of 5-OPase activity left in the supernatant after

precipitation by the immune serum was proportional to the

dilution of the serum. A I : I 1 .5 dilution of the immune serum

precipitated over 90% of the 5-OPase activity compared to that

of the control (Fig. 1B). The Sepharose resin of the 1 : 1 1.5

antiserum dilution sample was washed with SDS sample buffer

and subjected to Western blot analysis together with the super-

natant and 47 milliunits of purified 5-OPase as control. The

S-OPase band was not seen in the supernatant but appeared in

the ebution of the Sepharose (data not shown). These results

demonstrated that this polyclonal antibody recognizes and binds

to 5-OPase with very high affinity.

Western Blot to Test the Specificity of the Antibody.

An ECL Western blot was performed with the semipurified

5-OPase preparation. As shown in Fig. I C, at the position of the

5-OPase protein (Mr 140,000) on an ECL Western blot film, no

band was detected in the lane blotted with either the preimmune

serum (Lane 1) or the primary immune serum (lime 2), whereas

there was a strong band in the lane blotted with the first booster

serum (Lane 3). To test if the band density corresponds to the

amount of 5-OPase activity in tissue samples, both an ECL

Western blot and a 5-OPase activity assay were performed with

protein extracts from rat kidney, liver, lung, and MatB tumor.

The same amount of protein (20 p.g) was loaded on the gel. As

shown in Fig. 1D, a specific band with the correct molecular

weight was detected, and its density correlated with the S-OPase

activity measured in these samples.

Immunohistochemical Method to Test Specificity of the

Antibody. Immunohistological examination of 5-OPase in rat

and human kidney samples confirmed that the polyclonal anti-

Fig. I A, ELISA assay of the rabbit antirat antibody of 5-OPase. B.immunoprecipitation of rat kidney 5-OPase with rabbit antirat 5-OPase

antibody. In A and B: #{149}.immune serum; 0, preimmune serum. C.Western blot of rat kidney 5-OPase probed with rabbit antirat 5-OPaseserum. Lane I was blotted with preimmune serum, Lane 2 was blotted

with primary immune serum, and Lime 3 was blotted with first booster

serum. 1), Western blot band density of 5-OPase in comparison with

5-OPase activity in rat tissues.



I . .4

I - r� � f

Ep a�

.j. �.,

�

4�

�. � ‘

I � ;�:�

41�

4-� C


I..

4

-,. . S - -

-� ‘

#.s, ‘� �;‘

,p%_ �.

.1

‘S

. � ��‘a”L” ‘4 C’s.-j: � � * .�. *4� ,,�t � �#{149}�,

Fig. 2 Immunohistochemical localization of 5-OPase in the kidney.

Normal human kidney reacted with preimmune rabbit serum served as

a negative control (A, X200). 5-OPase is localized in the normal kidney

of a human (B, X260) with kidney carcinoma and a rat (C. X200)bearing MatB tumor most strongly to cytoplasm of distal tubules (wide

arrows) and less strongly to proximal tubules (thin arrows). In theglomerulus, 5-OPase is localized to the podocytes (open arrow) lining

Bowman’s capsule.

body directed against rat kidney 5-OPase can specifically detect

5-OPase distribution in rat tissues and cross-reacts specifically

with human 5-OPase. As indicated in Fig. 2, the distribution of

5-OPase was very similar between the normal kidneys of hu-

mans and rats: 5-OPase was localized to the cytoplasm of cells

of the proximal and distal tubules and collecting ducts, being

slightly more immunointense in the distal than proximal tubules

and somewhat less intensively localized to collecting ducts in

. . , Table 1 5-OPase activity in the normal tissues of female Fisher ratsbearing MatB tumors, and patients with the corresponding cancer

, The experimental error was within 10% of the values. See “Mate-

. riabs and Methods” for experimental procedures.

5-OPase activity (milbiunits/mg)

Tissue Rat Human

Kidney 187.6 ± 23.5 (n = 6) 8.3 ± 2.3 (n = 6)Liver 26.1 ± 2.9 (n = 5) 19.6 (n 2)”

Lung 3.4 ± 1.6 (n = 5) 3.3 ± 2.4 (n = 4)

Colon 9.6±2.7(n= 12)Stomach 6.37 (n = 1)Breast 11.2 ± 7.2(n = 3)

Uterine 4.4 (n = 1)

BMb 2.1 (n = 2)

PBMC5C 11.3 ± 5.8(n 13)

Tumor 3.4 ± 2.6 (n

a Samples from two patients were assayed, and the 5-OPase activ-

ity was 30.0 and 9. 1 milliunits/mg, respectively.b BM, bone marrow.

‘. PBMCs, peripheral blood mononuclear cells.d Data from MatB mammary tumors grown in female Fisher rats.

the medulla of both human and rat kidneys. Immunopositivity

for 5-OPase was also found in the cells lining Bowman’s cap-

sule and in the podocytes of the glomeruli in both species.

5-OPase immunoreactivity was most pronounced in renal ade-

nocarcinoma cells with a granular cell phenotype.

These results strongly indicated that this polycbonal anti-

body had great affinity and specificity for 5-OPase and could be

used for additional studies on the bevel and distribution of

5-OPase in tissue samples of both rats and humans.

RESULTS

Enzymatic Characterization of 5-OPase in Rat and Hu-man Tissues. The 5-OPase activity in rat tissues presented a

very specific distribution pattern (Table 1). Rat kidney had the

highest activity [187.6 ± 23.5 milliunits/mg (n = 6)], followed

by liver [26. 1 ± 2.9 milliunits/mg (n = 5)], lung [3.4 ± 1.6

milliunits/mg (n = 5)], and bone marrow [2.1 milliunits/mg

(n 2)]. 5-OPase activity in MatB tumors [3.4 ± 2.6 milli-

units/mg (n = 5)] was relatively low. As a comparison, the

mean values of 5-OPase activity in normal tissues from patients

were: liver, 19.6 mibliunits/mg (two specimens, 30 and 9.11

milliunits/mg, respectively); peripheral blood mononuclear

cells, 1 1 .3 ± 5.8 milliunits/mg; breast tissue, 1 1 .2 ± 7.2 milli-

units/mg (n = 3); colon, 9.6 ± 2.7 milliunits/mg (n = 12);

kidney, 8.3 ± 2.3 milliunits/mg (n = 6); uterine smooth muscle,

4.4 milliunits/mg (n = 1); and lung, 3.3 ± 2.4 milliunits/mg

(n 4). Therefore, humans do not have a tissue distribution

pattern of 5-OPase activity similar to that of rat. It is noted that:

(a) whereas rat kidney had strikingly high 5-OPase activity [up

to 187.6 ± 23.5 milhiunits/mg (n 6)], human kidney only had

an average 5-OPase activity of 8.3 ± 2.3 milliunits/mg (n 6),

a 30-fold difference between species; and (b) 5-OPase activity

in humans presented a significant individual variation; in the

peripheral mononuclear cells, 5-OPase activity ranged from

5-25 milliunits/mg (5-fold intenndividual variation), and in the

normal colon mucosa, it ranged from 3.6-12.8 milliunits/mg

(3-fold interindividual variation).




-�i

E�!.

:�

�tCa0C,)(8a-0Lb

..‘,III

I n

ioIuij�i

�t#{149}.,ri I.#{149},.iII lulJ1J111�I III IIII I�i I�II III i.iI I.. III�0 ‘�.‘�-

U NormalDTumor

.L.�

IIII

�Colon Lung

Fig. 3 The 5-OPase activity in the paired normal (U) and tumortissues (EJ) of cancer patients. The experimental error was within 10%

of the values. See “Materials and Methods” for experimental proce-dures.

To investigate whether 5-OPase activity changes along

with tumorigenesis in humans, paired specimens, i.e. , tumor and

the adjacent normal tissue from the same patient, were assayed

for 5-OPase activity. Among the eight colonic tumor and normal

tissue pairs studied, 5-OPase activity was higher in normal

cobonic mucosa than in the matched colonic carcinoma in three

pairs, lower in three pairs, and equal in two pairs. Therefore, no

consistent pattern of 5-OPase activity in normal cobonic mucosa

versus colonic carcinoma could be documented. 5-OPase activ-

ity level in these cobonic carcinomas did not seem to be related

to the grade of tumor differentiation, nor to the age or gender of

the patient. Similar results were found in four pairs of lung

samples but were less confident due to the limited number of

samples (Fig. 3).

Immunocharacterization of 5-OPase. Because the

amount of tissue required for the activity assay of 5-OPase is

rarely available, and the heterogeneity of cells within tissue

samples may complicate interpretation of the 5-OPase activity,

we characterized 5-OPase in greater detail using Western blot

and immunohistochemistry methods. Paired normal and tumor

tissues from patients with stomach, lung, and colon cancer were

examined, because our interest in GSH has focused on 5-OPase

in the context of chemical carcinogenesis, and in these tumors,

dietary and environmental xenobiotics have been implicated.

Seven pairs of normal and neoplastic lung specimens were

examined by Western blot. Cases 1 and 2 and 4-6 showed a

13.0-, 47.8-, 4.1-, 10.3-, and 4.0-fold higher 5-OPase bevel in

normal lung specimens than in the corresponding tumors, re-

spectively. In case 7, 5-OPase was only detected in the normal

lung tissue and not in the tumor. Case 2 was classified as a large

cell undifferentiated carcinoma, and cases 1 and 4-7 were

classified as adenosquamous carcinomas. The tumor in case 3

was a bronchioboalveolar cell carcinoma, and here 5-OPase

bevels in normal and tumor lung were equivalent (Fig. 4A).

Another four separate pairs of tumor and normal lung tissues

were examined by immunohistochemistry; the columnar bron-

chial epithelial cells and chondrocytes in normal lung tissues

were strongly 5-OPase positive, as were alveolar macrophages.

Weak-to-moderate 5-OPase immunopositivity was found in

cells of tumors (bronchioloalveobar tumors or adenosquamous

tumors; data not shown).

Four pairs of normal stomach tissue and gastric tumor from

patients were examined by Western blot. 5-OPase levels in the

normal stomachs were shown to be 4.1-, 1.8-, 2.0-, and 10. 1-fold

(cases 1-4, respectively) higher than in the corresponding tu-

mors (Fig. 4A). By immunohistochemistry, the four normal

stomach specimens showed a consistent pattern in which

5-OPase immunopositivity was localized strongly to the cyto-

plasm of normal parietal cells and localized weakly to glands in

the normal gastric antral mucosa. In the areas where intestinal

metaplasia was observed, 5-OPase was also localized to the

cytoplasm of enterocytes in the superficial mucosa, being

mostly found in cytoplasm of columnar absorptive cells. How-

ever, immunohistochemistry showed a rather different pattern of

5-OPase immunoreactivity among the four stomach tumor spec-

imens, ranging from completely negative in the first case to

moderately positive in some other cases. This might be due to

the pathological differences among these tumors. The tumor in

case 1 was classified as a mucinous adenocarcinoma, in which

tumor cells were dispersed in abundant mucus, and virtually no

5-OPase immunopositivity was found by immunohistochemis-

try (Fig. 4B). Tumors from the three other cases (cases 2-4)

were solid and histologically classified as intestinal-type gastric

adenocarcinomas. In these tumors, tumor cell cytoplasm was

variably but convincingly immunoreactive for 5-OPase. On

average, tumors from cases 2 and 3 showed moderate immu-

nopositivity, whereas case 4 showed weak and only focally

moderate immunoreactivity (data not shown). These results

from immunohistochemistry correlated with those obtained

from Western blot experiments (Fig. 4A), in which we observed

a greater difference in 5-OPase level between normal and tumor

tissues in case 4 than in cases 2 and 3. Case 1 provided enough

tissue to perform 5-OPase activity assay and Western blotting as

well as immunohistochemistry. The activity assay revealed a

4.9-fold higher 5-OPase activity in the normal stomach tissue

(6.4 mibliunits/mg) compared to that in the tumor ( I .3 mibliunits/

mg). Western blot showed a 4.1-fold higher 5-OPase band

density in the normal stomach tissue than in the tumor (Fig. 4A).

Immunohistochemistry detected strong 5-OPase positivity in the

parietal cells in the normal tissue, but the tumor was 5-OPase

negative (Fig. 4B). These three approaches showed essentially

the same result and serve to validate the observation as well as

the utility of these different assay techniques.

In three paired human colon specimens, Western blot de-

tected no obvious or consistent difference of 5-OPase level

between the normal and neoplastic tissues, which is compatible

with the enzyme activity measured in eight other pairs of colon

samples. By immunohistochemistry, 5-OPase was localized to

colonic enterocyte cytoplasm in the superficial mucosa in two of

three cases. Tumor cell cytoplasm was immunopositive, varying

from nonreactive to focally moderately reactive (data not

shown).

The localization of 5-OPase in human breast, liver, and

kidney was also investigated by immunohistochemistry. In the

breast, 5-OPase was shown to be localized to cells of the

terminal ducts and ductules in normal breasts and also to the

cytoplasm of invasive ductal cancer cells (data not shown). In

normal liver, 5-OPase was detected in both hepatocytes and bile

duct epithelium. Hepatocyte cytoplasm was diffusely, mildly to

moderately immunoreactive, and no regional pattern was ob-



ALung

__i_____a__-3----4 5 6 7NTNT NTN T NTNTNT

5-OPase-

Band density 130:1

ratio (N vs. T)

Stomach

B

__478:1 1:1 4.1:1 10.3:1 4.0:1

1 2 3 4NT NTNTNT

-�, “��“w

%i� ‘

. -a


. � � � - -

4.1:1 1.8:1 2.0:1 10.1:1

Fig. 4 A. Western blot of 5-OPase in seven pairs of normal and tumor lung tissues and in four pairs of normal and tumor stomach tissues. The densityof bands was measured using NIH Image 1.60 software. N, normal tissue; T, tumor tissue. B, immunohistochemical localization of 5-OPase in normal

human stomach (b, X 130) and stomach cancer (c, X260) in the tissues of stomach case 1, as in A. Negative control (a, X 130) was reacted with

preimmune serum. 5-OPase is localized to cytoplasm of parietal cells (arrow) in benign body mucosa (b) and is weakly reactive in thismucin-producing gastric adenocarcinoma (c).

served. Immunointensity of S-OPase was slightly more marked

in bile duct epithelium, where positive staining was mostly

supernuclear (data not shown). In the kidney, as shown in

“Materials and Methods,” 5-OPase was localized to the cyto-

plasm of cells of the proximal and distal tubules and collecting

ducts, being slightly more immunointense in the distal than

proximal tubules and somewhat less intensively localized to

collecting ducts in the medulla. Immunopositivity for 5-OPase

was also found in the cells lining Bowman’s capsule and in the

podocytes of the gbomeruli (Fig. 2). Renal adenocarcinoma cells

were weakly to moderately 5-OPase positive as compared to the

normal renal cells (data not shown); 5-OPase immunoreactivity

was most pronounced in cells with a granular cell phenotype.

DISCUSSION

This is the first report on 5-OPase activity in rat tissues by

measuring the conversion rate of L-[’4C1-5-oxOproline to

L-[’4C]-glutamate, which directly reflects the activity of

5-OPase. Our results (Table 1) reflected a pattern of 5-OPase

distribution in rat tissues consistent with a previous study (2)

that showed that the relative rates of oxidation of L�[i4C]�5�

oxoprobine to 14CO2 in rat kidney (9.7 milliunits/mg dry tissue

weight), spleen (2.7 milliunits/mg dry tissue weight), liver (1.6

milliunits/mg dry tissue weight), intestine (0.26 milliunit/mg dry

tissue weight), heart muscle (0.25 milliunit/mg dry tissue

weight), and brain (0. 14 milliunit/mg dry tissue weight) are

100:29:16:2.7:2.7:1.4.

Previous studies of 5-OPase focused on its catalytic mech-

anism and its activity in some animal tissues (2, 17). This is also

therefore the first report on 5-OPase activity and cellular distri-

bution in human tissues. The present study shows that: (a)

compared to rats, human 5-OPase activity is generally lower,

and, unlike rats, whose 5-OPase is mainly localized to the

kidney and liver (2), human 5-OPase shows no obvious prefer-

entiab organ distribution pattern; and (b) in humans, there is

significant individual variation of 5-OPase, as indicated by the




enzymatic assay of 5-OPase in the normal colon (n = 12; 3-fold

individual variation) and peripheral mononuclear cells (ii = 13;

5-fold individual variation). Despite the limited number of sam-

ples, this individual variation in 5-OPase activity may suggest

heterogeneity of cells within the tissue samples, different expo-

sures to inducers of 5-OPase, or a genetically based polymor-

phism yet to be defined.

We have raised a polyclonal antibody against rat kidney

5-OPase and proved that it has high affinity and specificity and

cross-reacts with human 5-OPase in various tissues. Using this

antibody, we showed a similar distribution pattern of 5-OPase in

the normal kidney of both rats and humans: 5-OPase is localized

to the renal tubular cells. This result is consistent with those of

a previous study that measured the 5-OPase activity in micro-

disected renal segments and showed that rat 5-OPase activity

was highest in the renal tubules and low in the gbomeruli (21).

Because the physiological role of renal tubules, especially the

proximal tubules, is to reabsorb the glomerular filtrate including

massive amount of amino acids (22), the -y-glutamyl cycle is an

active inward amino acid transporter system (22), and rat kidney

is the richest source for 5-OPase in rats (2), our results on

5-OPase distribution in the kidney are consistent with these

physiological aspects and suggest that 5-OPase has an important

function in the -‘y-glutamyb cycle at this site. This deduction is

further substantiated by our observation of 5-OPase distribution

in colon tissue: 5-OPase is strongly, if not exclusively, localized

to the enterocytes in the superficial mucosa, whose function is to

reabsorb several types of substances, including amino acids

(23).

5-OPase was also clearly localized to cells involved in

secretory functions such as the parietab cells that secrete hydro-

chloric acid in gastric mucosa and the cells of the ducts,

ductules, and lobules that secrete milk in breasts as well as the

cells of bile ducts, which also have secretory functions (23). The

functional relevance of this distribution might be that the ab-

sorptive function of cells may be up-regulated when its secretion

is elevated as compared to other cells.

The focus of this report is to compare the bevels of 5-OPase

in tumors and their adjacent normal tissues in humans. Because

of the implication of chemical carcinogens in their genesis and

the robe of GSH in their detoxification mechanisms, we focused

on lung, colon, and stomach cancers. Two of four pairs of lung

specimens studied by enzymatic assay, six of seven separate

pairs by Western blot, and four of six separate pairs by immu-

nohistochemistry showed lower levels of 5-OPase in tumors as

compared to the paired normal tissues. Four pairs of stomach

specimens were examined by both Western blot and immuno-

histochemistry, all of which showed a higher 5-OPase level in

the normal tissues than in the paired tumors. These results are

consistent with the previous finding that in culture, tumor cells

usually have lower 5-OPase levels than normal cells (14). In all

cases studied, the heterogeneity of tumor cells is significant, and

their 5-OPase positivity ranged from completely negative in

some cases to moderately positive in some other cases. Because

the stomach and lung contain several cell types, the cellular

origins of gastric and lung neoplasms have yet to be conclu-

sively elucidated. The absence of 5-OPase in foveolar cells (that

typically secrete mucin) and in mucinous adenocarcinoma may

thus be significant. The presence of 5-OPase in intestinal meta-

plastic columnar absorptive cells and in intestinal-type cancers

may also reflect derivation of this histological subset from this

cell type. In comparison with paired stomach and lung tissues,

paired colon specimens showed a rather inconsistent relation-

ship between the 5-OPase levels in normal and tumor tissues.

The mechanism by which 5-OPase expression is decreased

in particular tumors is of interest to future studies. We have

previously demonstrated decreased expression of another GSH-

related enzyme, OSH S-transferase a and p., in human breast

tumor tissues compared to adjacent normal tissues (24). The

significance and mechanism of this phenomenon will be more

easily studied once the 5-OPase gene and its regulatory elements

are isolated and characterized.

Our findings in this report also have potential implication

for cancer chemotherapy. 017, a 5-oxo-L-proline analogue and

a substrate for 5-OPase, has been shown to differentially in-

crease the GSH level in normal tissues while decreasing it in

tumor tissues, thus accentuating the relative susceptibility of

tumor cells to the anticancer drug melphalan (13, 14). An ideal

model in which to explore this selectivity of OTZ should have

significantly lower 5-OPase levels in the tumor than in the

paired normal tissue. Although the physiological significance of

differences in 5-OPase level in normal versus tumor tissue is not

yet certain in both stomach and lung cancers, a lower 5-OPase

level was found in the tumor tissues as compared to normal

tissues, suggesting that a decrease in this particular enzyme

occurs at some point in the process of tumorigenesis. Because

OTZ is metabolically activated by 5-OPase, these data suggest

a potential use of OTZ as a biochemical modulator in the

treatment of these cancers. Although we have not assessed

human bone marrow 5-OPase activity, previous reports demon-

strated that OTZ is activated in this tissue and may result in a

protective effect (25).

REFERENCES

1. Van Der Werf, P., and Meister, A. The metabolic formation andutilization of 5-oxo-L-proline (L-pyroglutamate. L-pyrolidone carboxy-

late). Adv. Enzymol., 43: 5 19-556, 1975.

2. Van Der Verf, P., Orlowski, M., and Meister, A. Enzymatic conver-

sion of 5-oxo-L-proline (L-pyrolidone carboxylate) to L-glutamate cou-

pled with cleavage of adenosine tnphosphate to adenosine diphosphate.a reaction in the y-glutamyl cycle. Proc. NatI. Acad. Sci. USA, 68:

2982-2985, 1971.

3. Van Der Werf, P., Stephani, R. A., Orlowski, M., and Meister, A.

Inhibition of 5-oxoprolinase by 2-imidazolidone-4-carboxylic acid.

Proc. Natl. Acad. Sci. USA. 70: 759-761, 1973.

4. Griffith, 0. W., and Friedman, H. S. Inhibition of metabolic drug

inactivation: modulation of drug activity and toxicity by perturbation of

glutathione metabolism. in: T. C. Chou and D. C. Rideout (eds.),

Synergism and Antagonism in Chemotherapy. pp. 245-284. New York:

Academic Press, 1991.

5. O’Dwyer, P. J., Hamilton, T. C., Young, R. C.. LaCreta, F. P.. Carp,N.. Tew. K. D., Padavic, K., Comis, R. L., and Ozols, R. F. Depletionof glutathione in normal and malignant human cells in viva by buthi-

onine sulfoximine: clinical and biochemical results. J. Nail. Cancer Inst..84: 264-267, 1992.

6. Mulcahy, R. 1., Bailey, H. H., and Gipp, J. J. Transfection ofcomplementary DNAs for the heavy and light subunits of human -y-glu-

tamylcysteine synthetase results in an elevation of intracellular gluta-thione and resistance to melphalan. Cancer Res., 55: 4771-4775. 1995.

7. Moore, W. R.. Anderson, M. E., Meister, A., Murata, K., and

Kimura, A. Increased capacity for glutathione synthesis enhances re-




sistance to radiation in Escherichia coli: a possible model for mamma-

han cell protection. Proc. NatI. Acad. Sci. USA, 86: 1461-1464, 1989.

8. Hanigan, M. H., Frierson, H. F., Jr., Brown, J. E., Lovell, M. A., andTaylor, P. T. Human ovarian tumors express y-glutamyl transpeptidase.

Cancer Res., 54: 286-290, 1994.

9. Meister, A. 5-Oxoprolinuria (pyroglutamic aciduria) and other dis-orders of the y-glutamyl cycle. in: J. B. Stanbury, J. B. Wyngaarden,D. S. Fredrickson, J. L. Goldstein, and M. S. Brown (eds.), The Meta-bolic Basis of Inherited Diseases, fifth edition, pp. 348-359. New York:

McGraw-Hill, Inc., 1983.

10. Henderson, M. J., Larsson, A., Carisson, B., and Dear, P. R. F.5-Oxoprolinuria associated with 5-oxoprobinase deficiency: further cv-

idence that this is a benign disorder. J. Inherited Metab. Dis., 16:

1051-1052, 1993.

I I . Williamson, J. M., Boettcher, B., and Meister, A. Intracellular

cysteine delivery system that protects against toxicity by promoting

glutathione synthesis. Proc. NatI. Acad. Sci. USA, 79: 6246-6249,

1982.

12. Russo, A., DeGraff, W., Friedman, N., and Mitchell, J. B. Selective

modulation of glutathione levels in human normal versus tumor cellsand subsequent differential response to chemotherapy drugs. Cancer

Res., 46: 2845-2848, 1986.

13. Williamson, J. M., and Meister, A. Stimulation of hepatic glutathi-one formation by administration of L-2-oxothiazolidine-4-carboxylate, a

5-oxo-L-prolinase substrate. Proc. NatI. Acad. Sci. USA, 78: 936-939,1981.

14. Meister, A. Selective modification of glutathione metabolism. Sci-

ence (Washington DC), 220: 472-477, 1983.

15. Baruchel, S., Wang. T., Farah, R., Jamali, M.. and Batist. G. in viva

selective modulation of tissue glutathione in a rat mammary carcinomamodel. Biochem. Pharmacol., 50: 1505-1508, 1995.

16. Wang, T., Chen, X., Schecter, R. L., Baruchel, S., Alaoui-Jamali,M.. Melnychuk. D., and Batist, G. Modulation of glutathione by a

cysteine pro-drug enhances in vivo tumor response. J. Pharmacol. Exp.Ther., 276: 1 169-1 173, 1996.

17. Van Der Werf, P., and Meister, A. 5-Oxoprolinase (L-pyroglutamate

hydrolysis): purification and catalytic properties. J. Biol. Chem., 250:

6686-6692, 1975.

18. Meister, A., Griffith, 0. W., and Williamson, J. M. 5-Oxo-L-prolinase from rat kidney. Methods Enzymol., 113: 445-451, 1985.

19. Williamson, J. M., and Meister, A. Effect of subthydryl groupmodification on the activities of 5-oxo-L-prolinase. J. Biol. Chem., 257:

9161-9172, 1982.

20. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D. Seidman,

J. G., Smith, J. A., and Struhl, K. (eds). Current Protocols in MolecularBiology. Volume 2, Chapter 11, pp. 11.2.2-11.2.5. New York: John

Wiley and Sons, Inc., 1987.

21. Heinle, H., and Wendel, A. The activities of the key enzymes of the

y-glutamyl cycle in microdissected segments of the rat nephron. FEBSLett., 73: 220-224, 1977.

22. Orlowski, M., and Meister, A. The -y-glutamyl cycle: a possibletransport system for amino acids. Proc. Nail. Acad. Sci. USA, 67:

1248-1255, 1970.

23. Ganong, W. F. Gastrointestinal Function. in: Review of MedicalPhysiology, fifth edition, chapter 5, pp. 354-374. Los Altos, CA: Lange

Medical Publications, 1971.

24. Alpert, L. C., Schecter, R. L., Berry, D. A., Melnychuk, D., Peterr,

W., Caruso, J., Townsend, A. J., and Batist, G. Relation of glutathioneS-transferase a and p. isoforms to response to therapy in human breast

cancer. Clin. Cancer Res., 3: 661-667, 1997.

25. Wilson, D. M., White, R. D., Webb, L. E., Bender, J. G., Pippin,

L. L., and Goldberg, D. I. Amelioration of AZT-induced bone marrowhypoplasia in mice cotreated with the glutathione (GSH) prodrug, Pro-

cysteine. Abstract Book, IXth International Conference on AIDS. Ber-

lin, Germany: Wellcome Foundation Ltd., 1993.



1998;4:131-138. Clin Cancer Res X Chen, R L Schecter, O W Griffith, et al. biochemical modulation of glutathione.tissues of humans and rats: a potential new target for Characterization of 5-oxo-L-prolinase in normal and tumor

Updated version

http://clincancerres.aacrjournals.org/content/4/1/131

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/4/1/131To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Characterization of 5-Oxo-L-prolinase in Normal and Tumor ... · enhanced in vivo tumor response to...

Documents

Transcript of Characterization of 5-Oxo-L-prolinase in Normal and Tumor ... · enhanced in vivo tumor response to...