relationship between colorectal cancer glutathione levels and patient survival

Relationship Between Colorectal Cancer Glutathione Levels and Patient Survival Early Results Sam C. Barranco, Ph.D., *a Roger R. Perry, M.D.,-~ Mary E. Durm, B.S.,* Mohammed Quraishi, M.D.,~: Alice L. Werner, M.D.,t Sharon G. Gregorcyk, M.D.,t Paul Kolm, Ph.D.t

b>om the *Quest Biomedical Research Center, Virginia Beach, Virginia, tBiostatistics, the Departments of Surgery and Pathology, Eastern Virginia Medical School, Norfolk, Virginia, and }Pasadena General Hospital, Pasadena, Texas

PURPOSE: Elevated glutathionc is a cause of resistance to anticancer agents and x-rays. The purpose of tb_is study was to determine the frequency and clinical significance of glutathione elevan'on in human colorectal cancer. METHODS: Gluta- thione levels were measured in 41 colon cancers, 24 rectal cancers, and corresponding normal tissues. The patients were then followed up prospectively for tumor recurrence and survival. Survival was analyzed by the Kaplan-Meir method and Cox proportional hazards regression. RESULTS: Glutathione levels in pmnary colorectal cancers were significantly higher than in the corresponding normal tissues. Elevated glutathione levels had a significant negative effect on survival in patients with colorectal cancer, whether based on the mean (P = 0.02) or median (P = 0.04) normal tissue levels. A negative effect of glutathione levels on survival was apparent in patients with colorectal cancer, whether or not they were treated with postoperative therapy. The larger the ratio of tumor glutathione to normal tissue glutathione, the poorer the prognosis. When adjusted for other covariates, glutatbione was still a significant predictor of survival. CONCLUSIONS: An elevated tumor glutathione level at the time of diagnosis appears to confer a poor prognosis in patients with cotorectal cancer. Longer-term study using a larger number of patients will be required to confirm these findings. Knowledge of tumor glutathione content may help identifT patients requiring more intensive therapy, lKey words: Glutathione; Colorectal cancer; Survival; Prognostic factors]

Barranco SC, Perry RR, Duma ME, Quraishi M, Wemer AL, Gregorcyk SG, Kolm P. Relationship between colorectal cancer glutathione levels and patient survival: early results. Dis Colon Rectum 2000;43:1133-1140.

F ai lure o f cer tain tumors to r e s p o n d to drugs and

radia t ion m a y b e b e c a u s e o f i nhe ren t res is tance

or res is tance i n d u c e d b y p rev ious t reatments . :-3 Sev-

eral causes o f res is tance have b e e n ident i f ied, 2-6 one

o f w h i c h is e l eva ted g lu ta th ione (GSH) content . GSI t

inf luences the eff icacy of a n u m b e r of an t i cancer

agents , specif ical ly the n i t rosoureas , adr iamycin ,

b l eomyc in , me lpha l an , mi tomyc in C, a n d x-rays. 7

These agents kill cells at least in par t t h rough the

p r o d u c t i o n of h igh ly react ive fi'ee radicals; GSH acts

as a free radical scavenger , thus r educ ing their effec-

t iveness. The h igher the GSH content , the m o r e resis-

tant are the cancer ceils to t reatment . 3, 7, 8 Consider-

a t ion of intracel lular res is tance factors is impor t an t

whi l e p l ann ing cance r t rea tment reg imens , b e c a u s e

reversa l of res is tance m a y i m p r o v e the efficacy of

c h e m o t h e r a p y . 9

Data is emerging on the potential clinical relevance of

GSH and related enzymes in hematologic mal ignan-

cies. :° However , there are only limited data available on

the clinical significance of GSH elevations in h u m a n

solid tumors. Prior w o r k by our group in a series of

breast cancer patients showed that GSH levels were

significantly e levated in breast tumors and lymph node

metastases c o m p a r e d with nom:a l t issues.:: In the study

repor ted here w e analyzed the GSH content in tumor

tissues and in normal tissues of a series of patients with

colorectal cancer to determine whe the r a relationship

exists be tw e e n tumor GSH content and survival.

PATIENTS AND METHODS

'~Deceased. Supported by National Cancer Institute grant CA 15397, by Small Business Innovative Research/National Cancer Institute grant R43 CA57062, American Cancer Societ 3, Career Development Award 93-283, and the Sentara Endowment Fund. Address reprint requests ~o Dr. Perw: Department of Surgery, East- ern Virginia Medical School, 825 Fairfax Avenue, Norfolk, Virginia 23507.

Tissue Acquisition and Handling

Fresh rumor a n d c o r r e s p o n d i n g n o r m a l t issue (0.5

grams) w e r e o b t a i n e d at surgery and f rozen immed i -

ately. The samples w e r e ma in t a ine d at - 8 0 ° C until

ana lyzed . All assays w e r e p e r f o r m e d o n samples ob-

1133

1134 BARRANCO E T A L Dis Colon Rectum, August 2000

rained from multiple sites in the tumor to overcome the extreme variability known to exist for prognostic

factors in heterogeneous tumors, unlike normal tissue where little variability is noted. ~2 Based on our previ-

ous work, ~2 three to seven sites were sampled in each

tumor, depending on the tumor size, and the results

were averaged to obtain a mean GSH level for the tumor. Both the periphery and central area of the

tumor were sampled whenever feasible. The investi- gators who performed the assays were blinded to the relevant clinical and pathologic data.

GSH Determination

GSH assay was performed as previously de-

scribed. 11 Tissue samples were weighted and homogenized in ice-cold 3 percent sulfosalicytic acid using a Tissuemizer TM SDT 1810 (Tekmar Co., Cincinnati,

OH). Typically, 50 mg of homogenized tissue from

each normal or tumor sample were used for each assay; determinations were made in duplicate. Protein was determined on the crude homogenate using the

Bradford methodJ 3 The homogenate was centrifuged

at 14,000g and the supematant assayed for total GSH using the Tietze cyclic reduction assay . 14 The absor-

bance was monitored at 412 nm using a UV160U spectrophotometer (Shimadzu, Kyoto, Japan). Results are expressed as nmol GSH/mg protein.

D N A I n d e x a n d P r o l i f e r a t i o n P r o p e r t i e s

Samples of tissues from normal or tumor sites were

dissociated into single nuclei by exposure to 0.5 percent pepsin, 0.1 percent RNase, fixed with 70 percent ethanol, stained with 1:1 mixture of ethidium bro-

mide-mithramycin for 20 minutes, 12 and processed on a Coulter PROFILE ® flow cytometer (Coulter Electron-

ics, Hialeah, FL). The DNA index and cell proliferation properties were determined using MULTICYCLE TM

(Phoenix Flow Systems, Inc., San Diego, CA).

Clinical Data and Statistical Analysis

Forty-one patients with colon cancer and 24 patients with rectal cancer were followed up prospec-

tively. Patient follow-up ranged up to 33 months with a median follow-up of 10 months. All patients under- went surgical resection of their tumors and also received standard treatment based on stage. Such treatment included chemotherapy or radiation therapy or both where appropriate.

The effect of tumor GSH content on patient survival was analyzed in two different ways. First, an upper

limit of a 95 percent confidence interval for the mean normal tissue GSH level was obtained and used as the

cutoff for categorizing patients with elevated (High

GSH) vs. nonelevated (Low GSH) groups. Because

the distribution of GSH levels tended to be positively

skewed, an upper limit of a 95 percent confidence interval for the median normal tissue GSH level was

also obtained. The 95 percent confidence interval for the median was obtained by percentile bootstrap re-

sampling with 200 resamples. Both the mean and median were used in the analysis to see whether differences in the cutoff for the definition of normal

affected the results. Survival and disease-free survival (DFS) differences between the High and Low groups

(defined by the mean and median) were compared

using the log-rank test, and survival and DFS func- tions were estimated by the Kaplan-Meier method. 15

Differences between the High and Low GSH groups with respect to age, gender, pathologic stage, tumor

size, DNA content, s-phase, G2M and carcinoembryonic antigen were analyzed by independent groups

t-tests or Wilcoxon's rank-sum test for continuous variables and Fisher's exact probability test for cate- gorical variables.

For the second analysis a ratio of tumor GSH to normal tissue GSH was obtained to standardize tumor

GSH levels individually. In this case a ratio of i would mean that tumor GSH and normal tissue GSH were

equal, a ratio tess than 1 that tumor GSH was less than normal tissue GSH, and a ratio greater than 1 that tumor GSH was greater than normal tissue GSH. Cox proportional hazards reo~ession was used to assess the influence of GSH and other patient and tumor characteristics on survival and DFS. 16

RESULTS

The clinical and tumor characteristics of patients with colon cancer did not differ significantly from patients with rectal cancer except for age and normal tissue GSH levels. Patients with colon cancer were, on average, six years older than patients with rectal cancer. GSH levels in tumors were significantly higher than in normal tissue for both colon and rectal cancers (P < 0.05 for both; Table 1). Normal tissue GSH levels in patients with rectal cancer were more elevated than in patients with colon cancer. Because of this, cutoff values for High and Low GSH groups were calculated for each cancer group.

Vol. 43, No. 8 COLORECTAL CANCER GLUTATHIONE 1135

Based on mean normal tissue GSH, there were 33

patients in the High group, 11 (33 percent) of whom

died, and 32 in the Low group, 2 (6 percent) of whom died. Based on median normal tissue GSH, there were 45 patients in the High group and 20 in the Low

group. M1 13 deaths occurred in the High group. Figure 1 plots the Kaplan-Meier survival curves for

High and Low groups based on mean normal tissue GSH (A) and median normal tissue GSH (B). Whether

based on the mean or the median, the High group had

a significantly poorer prognosis for survival than the

Low group. At 24 months the estimated survival rate

for the High group was 54 percent compared with 89

Table 1. GSH Values in Colorectal Cancer

Tissue Normal Tumor

Colon (n = 41) 8.48 +__ 0.9 15.33 +_ 1.4" Rectum (n = 24) 13.04 _+ 2.1 16.56 _+ 2.3* Entire group (N = 65) 10.24 +_ 1 15.92 -+ 1.4*

GSH = glutathione. Figures are nmol GSH/mg protein + standard error. *P < 0.05.

percent for the low group (Fig, 1A). Tumor GSH

group was not a significant predictor of DFS (P =

0.287). Table 2 presents a comparison of High and Low GSH

groups (based on upper limit of the mean 95 percent confidence interval) with respect to clinical and demo-

graphic variables and tumor characteristics. The two

groups did not differ with respect to any of the variables

studied, whether the mean (Table 2) or median (data

not shown) was used to define the groups,

Cox proportional hazards regression was used to

assess the effect of different variables on survival.

Table 3 presents the hazard ratios, 95 percent confidence intervals, and P values for each variable ana-

lyzed. The mean GSH tumor to normal tissue ratio

was 1.89 + 1.48. Tumor to normal tissue ratio was a

significant predictor of survival, with larger ratios as-

sociated with poorer survival. Figure 2 plots the Cox

regression survival estimates for tumor to normal ratios of 1, 2, 3, 4 and 5. For a ratio of I ( tumor GSH and

normal tissue GSH equal), the estimated survival rate

at 24 months was 72 percent, but for a ratio of 5

(tumor GSH 5 times that of normal tissue level), the

o~

'E :3 tO g o 2 n

A

0.9 -~q- 4~4 --~ ~

0.7 , , ,

. . . . . . L

0.e 11 it

0.5

0.4" p = .0213

0.3

> Mean ~H 0.2" < Mean GSH

0.1"

0.0" t I

0 6 12

It I I I/)

o

ft.

1.0 [ : -+IH~-PI--I--f---H--+-I

O.

0.7

0.8

0.#

02r

0.2

O.t

0.0

0

> Median GSH < Median GSH

p = . 0 ~ 0

18 24 30 36 6 12 18 24 30

Time (months) Time (months)

36

Figure 1. Kaplan-Meier survival curves of the influence of glutathione (GSH) levels on survival of patients with colorectal cancer. High and low tumor GSH levels based on (A) mean normal tissue GSH levels, and (B) median normal tissue GSH levels.

1136 BARRANCO ETAL Dis Colon Rectum, August 2000

estimated survival rate at 24 months was only 38 percent. Pathologic stage was the only other significant predictor of survival. Higher stage was associated with poorer survival.

Table 2. High GSH and Low GSH Patient Characteristics

Variable High GSH Low GSH P

Age (yr) 66.2 _+ 13.1 62.3 + 13,7 0.278 % males 42 53 0.46 Tumor Size (cm) 4.77 - 2.4 5.10 -+ 2.91 0.633 Stage (%)

I 0 3 II 48 41 III 21 25 IV 30 31 0.871

DNA Index 1,517 -+ 0.438 1.521 -+ 0.387 0.97 S-phase (median) 9.32 9.58 0.578 G2M (median) 3,97 4.21 0,94 CEA (median) 2.8 2.4 0.479

GSH = glutathione; CEA = carcinoembryonic antigen.

Table 3. Cox Proportional Hazards Regression,

Univariate Results

Hazard 95% Confidence Variable P

Ratio Interval

Age 1.02 0.97-1.06 0.504 Tumor size 1,01 0.84-1.22 0.907 CEA 1 0.51-1.99 0.99 DNA Index 1.73 0.45-6.65 0.422 Stage 5.15 2.07-12.83 <0.001 GSH ratio 1.32 1.03-1.69 0.03

CEA = carcinoembryonic antigen; GSH = glutathione.

GSH tumor to normal tissue ratio was then paired in a multivariate Cox regression with the other variables. Table 4 presents these results. For all analyses, GSH tumor to normal tissue ratio was a significant predictor of survival even when adjusted for the other variable. A test of difference in log-likelihoods indicated that the model including pathologic stage and GSH was a significantly better predictor of survival than the model of pathologic stage alone (P < 0.001). As an illustration of the joint effect of stage and GSH, Figure 3 plots estimated survival for Stages 1 through 4 and GSH tumor to normal tissue ratios of 2 and 3. These GSH ratios were commonly observed in our patients. Similar plots can be constructed for higher GSH tumor to normal tissue ratios (data not shown).

Patients were stratified with respect to whether they received postoperative therapy. Thirty patients received some type of postoperative therapy-: 13 received chemotherapy only, 5 received postoperative radiation only, and 12 received both. Stratified by postoperative or no postoperative therapy, GSH tumor to normal tissue ratio was still a significant predictor of survival (P = 0.050). Figure 4 plots estimated survival for GSH tumor to normal tissue ratios of 1 to 5 for patients who received postoperative therapy (A) and those who did not (B).

DISCUSSION

This study demonstrates that GSH is elevated in colorectal cancer compared with corresponding nor-

==

0 0

o = 1::: O

9 E L

1.0"

0,9"

0.8

0.7

0.(~

0.5-

0.4-

0.3"

0.2-

0.1-

0.0"

odNormal Ratio

1

2

3

4

5

I" l I I I I I

0 6 12 18 24 30 36

Time(months)

Figure 2. Cox regression survival estimates based on glutathione tumor to normal tissue ratios,

VoL 43, No. 8 COLORECTAL CANCER GLUTATHIONE 1137

mal tissue. The results reported here are particularly

important because, for the first time, they suggest a

significant association between colorectal cancer GSH levels at the time of diagnosis and survival. Significant survival differences are apparent even though the

Table 4. Cox Proportional Hazards Regression,

Multivariate Results

Hazard 95% Confidence Variable P Ratio Interval

Age 1.02 0,97-1,08 0.342 GSH ratio 1.4 1.05-1.86 0.02

Tumor size 1.04 0,86-1.28 0.67 GSH ratio 1.33 1.03-1.71 0.028

CEA 0.95 0.44-2.05 0.899 GSH ratio 1.33 1.01-1,73 0.04

DNA index 1.36 0.32-5.85 0.415 GSH ratio 1.3 1,01-1.68 0,045

Stage 5 . 9 1 2.24-15.56 <0.001 GSH ratio 1.67 1,15-2.41 0.007

GSH = glutathione; CEA = carcinoembryonic antigen.

follow-up is relatively short. Another interesting find-

ing is the higher GSH levels noted in normal rectum

compared with normal colon. The reason for this is

unknown, but may somehow be related to differences

in stool volume or composition or to basic biochem-

ical differences between the colon and rectum.

In-vitro studies have shown that elevated GSH lev-

els contribute to the resistance of tumor cells to anti-

cancer drugs such as melphalan, bleomycin, galacti-

tol, adriamycin, actinomycin D, mitomycin C, the

nitrosoureas, and x-rays. ? Others have noted a similar

relationship in patients with ovarian cancer who de-

veloped resistance to alkylating agents. 17 Almost one-

half of patients with colorectal cancer in our study

received regimens containing 5-fluorouracil (5-FU)

alone or with radiation therapy, and GSH remained a

significant predictor of outcome in this subgroup.

5-FU is an antimetabolite, and GSH is generally not

thought to affect the sensitivity- of tumors to 5-FU. We

have previously shown that multiple drug resistance

mechanisms can occur simultaneously within a cell. 18

It may be that colorectal cancers with elevated levels

of GSH also possess other drug resistance mecha-

nisms which affect 5-FU sensitivity. Because of tissue

1D

0.9"

0.8-

0.7"

L~ O.e~

¢q

o .u

o

o.4-

0.3

0.2

0.1-

Pathological Stage and GSH tumor/normal ratio = 2

0.O

Stage I

Pathoiogical Stage and GSH tumor/normal ratio = 3

==

07

O r~

O

1.0"

0.9.

0.8

0.7

0.e

0.9

0.4

0.3

0,2

0.1-

0.0

Stage 1

4

0 6 12 18 24 30 36 0 6 12 18 24 30 36

Time(months) Time(months)

Figure 3. Cox regression survival estimates based on glutathione tumor to normal tissue ratio and pathologic stage.


A

E 03 o¢ o 2 n

1 . 0 -

0.9-

0.8-

0,7-

0.6-

0.5-

0.4-

0,3-

0.2-

0,1-

0.0-

~',..,:~...

t~ x i

~ \ \ " .~ ' , ",,,,,

• \ ' , , ".,. \ , \ ",, . '...

' \ N

\ \ \ ,

x

Tumor/Normal G$H

1

.......... " ...... 2

. . . . . . . . . . 3

. . . . . . . . . . . . . . 4

. . . . . 5

1.0-

0.9"

0.8-

0.7-

.~ o . e -

~_ 0 .5 -

2 0.4- CL

0.3-

0.2"

0.1"

0.0-

B

v~< :::..

\ . \

Tumor/Normal GSH

1

................... 2

. . . . . . . . . . 3

. . . . . . . . . . . . . . . 4

. . . . . 5

I I I I I I I l I I I ! I

0 6 12 18 24 30 36 0 6 12 18 24 30 36

Time(months) Time(months)

Figure 4. Cox regression survival estimates based on glutathione (GSH) tumor to normal tissue ratios, and postoperative therapy. A. Patients who received postoperative chemotherapy or radiation therapy or both. B. Patients who did not receive postoperative therapy.

limitations, it was not possible to measure multiple drug resistance mechanisms in this study. Elevated GSH levels also had a negative effect on survival in

patients who did not receive postoperative chemotherapy or radiation therapy. The reason for this is

uncertain. Elevated levels of GSH may be a marker for tumors which are more biologically aggressive.

Reports suggest that tumors with elevated GSH levels may be made more responsive to therapy when the

GSH levels are reduced by exposure to buthionine sulfoximine (BSO) shortly before treatment with anticancer agents. A small clinical study by O'Dwyer e t al. 9 showed

that marked depletion of cellular GSH may be achieved

by BSO, without untoward nonhematologic toxicity. An- other small trial showed that BSO can be given safely together with melphalan, although with some increase in hematologic toxicity. 19

We have previously shown that GSH is elevated in breast cancer. 2° Elevated GSH levels have also been demonstrated in hematologic, ovarian, head and neck, and lung malignancies, i° In the present study we show that GSH is elevated in colorectal cancer, as has been noted by others, zl However, we believe this is the first study which attempts to determine whether

GSH levels correlate with survival in cotorectal can-

cer. There was no significant correlation between GSH levels in the primary tumor and a variet T of

parameters, including tumor size and pathologic stage. A similar finding has been reported by Hanigan e t al . 22 for the cell surface enzyme gamma-glutamyl

transpeptidase, which is involved in amino acid trans- port and GSH synthesis. Overall, these results suggest that the average GSH content of the primary tumor

does not increase as tumors progress toward a more advanced stage.

Tumor heterogeneity is believed to be responsible for certain treatment failures. 23 In addition, tumor het-

erogeneity makes the study- of prognostic factors difficult, because parameters may differ in different lo- cations of the tumor. We agree with the conclusions of Cook e t al . 24 and Lee et al . 25 that large variations in

GSH levels can occur depending on the location of the biopsy, possibly because of different growth sta- tus and cell cycle kinetics in different regions of the tumor. In addition, tumor GSH levels can be influ- enced by the presence of other cells, including leu- kocytes, red blood cells, and fibroblasts. 26 Thus, the enzymatic method of measuring GSH levels, because

Vol. 43, No. 8 COLORECTAL CANCER GLUTATHIONE 1139

it represents a tissue average, may underestimate the

variability present and may underestimate the maxi-

mal level present. However, the use of multiple samples from each tumor, as was done in this study,

should increase the accuracy. Our previous work sug-

gests that sampling at three to seven sites per tumor is

required to be 90 percent certain of identifying eleva-

tions in GSH or abnormalities in other prognostic

factors. 12 Other methods, including flow cytometry,

have been used to measure GSH levels; however, the

results are qualitative only and are highly dependent

on the cell type and assay conditions. 26

A major focus of cancer research in recent years has

been the development of methods for predicting sur-

vival and the likely effectiveness of different treatment

alternatives. These prognostic methods are based on

the results of tests performed on solid tumor samples.

For example, factors that have been shown in some

studies to effect outcome in patients with colorectal

cancer include pathologic stage, tumor grade, tumor fixation, and carcinoembryonic antigen level. 27 How-

ever, the ability to predict survival or response to

therapy remains imprecise, and the most effective

treatment cannot always be selected. The uniqueness

of each individual tumor makes it difficult to prescribe

the most effective treatment based solely on past

experience with other patients.

The results of this study suggest that measuring

tumor GSH levels at the time of diagnosis may be

useful in predicting survival in patients with colorectal

cancer. Longer-term follow-up and the addition of

larger numbers of patients are warranted to confirm

these findings. Measurement of tumor GSH levels may

also provide an important means of identifying pa-

tients with a poorer prognosis who might benefit from

more intensive therapy.

A C K N O W L E D G M E N T S

The authors thank K. MacLean, M. Rogerson, P.

Roberts, M. Levin, and J.A. Mazetta for excellent tech-

nical assistance, and B.R. Barranco for preparation of the artwork.

REFERENCES

1. Barranco SC, Townsend CM Jr, Jenkins VK, Koester SK, Ho BY, Reumont KJ. Treatment-induced changes in sensitivity in a multiclonal human tumor mixture model in vitro. Cancer Res 1988;48:2749-55.

2. Barranco SC, Townsend CM Jr, Weintraub B, el al.

Changes in glutathione content and resistance to anti-

cancer agents in human stomach cancer cells induced by treatments with melphalan in vitro. Cancer Res 1990; 50:3614-8.

3. Meister A. Glutathione, ascorbate, and cellular protec- tion. Cancer Res 1994;54(Suppl 7):1969s-75s.

4. Barranco SC, Ho DH, Drewinko B, Rhomsdahl MM, Humphrey RM. Differential sensitivities of human mel- anoma cells grown in vitro to arabinosylcytosine. Can- cer Res 1972;32:2733-6.

5. Biedler JL, Riehm H, Peterson RH, Spengler BA. Mem- brane-mediated drug resistance and phenotypic rever- sion to normal growth behavior of Chinese hamster ovary cells. J Natl Cancer Inst 1975;55:67140.

6. Rothenberg M, Ling V. Multidrug resistance: molecular biology and clinical relevance. J Natl Cancer Inst 1989; 81:907-10.

7. Barranco SC, Weimraub B, MacLean KK, Beasley EG, Jenkins VK, Townsend CM Jr. Relationship between glutathione levels and drug or radiation sensitivities in human gastric cell lines in vitro. Invest New Drugs 1991;9:29-36.

8. Ozols RF, Hamilton TC, Masuda H, Young RC. Manip- ulation of cellular thiols to influence drug resistance. In: Woolley PV, Tew KD, eds. Mechanisms of drug resistance in neoplastic cells. Bristol Myers Symposium No. 9. New York: Academic Press, 1988:289-305.

9. O'Dwyer PJ, Hamilton TC, Young RC, et al. Depletion of glutathione in normal and malignant human cells in

vivo by buthionine sutfoximine: clinical and biochemical results. J Natl Cancer Inst 1992;84:264-7.

10. Hall AG, Autzen P, Cattan AR, Malcolm AJ, Cole M, Kemahan J, Reid MM. Expression of mu class glutathione S-transferase correlates with event-free survival in childhood acute lymphoblastic leukemia. Cancer Res 1994;54:5251-4.

11. Perry RR, Mazetta J, Levin M, Barranco SC. Glutathione levels and variability in breast tumors and normal tissue. Cancer 1993;72:783-7.

12. Barranco SC, Perry RR, Durm ME, et al. Intratumor variability in prognostic indicators may be the cause of conflicting estimates of patient survival and response to therapy. Cancer Res 1994;54:5351-6.

13. Bradford MlVI. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248-54.

14. Tietze F. Enzymatic method for quantitative determination of nanogram amounts of total and oxidized glutathione: applications to mammalian blood and other tissues. Anal Biochem 1969;27:502-22.

15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat 1958;53:457-81.

16. Mehta C, Patel N. StatXact: stati~cal software for exact nonparametric inference. Cambridge: Cytel Software, 1991.


17. Britten BA, Green JA, Warerfius tLM. Cellular glutat~ione (GSH) and GSH-s-transferase (GST) activity- in human

ovarian tumor biopsies following exposures to alkylating

agents, IntJ Radiat Oncol Biol Phys 1992;24:527-31. 18. Perry RR, Kang Y, Greaves B. Biochemical characteriza-

tion of a Mitomycin C resistant colon cancer cell line

variant. Biochem Pharmacol 1993;46:1999-2005. 19. Bailey HH, Mulcahy RT, Tutsch KD, et al. Phase I clinical

trial of intravenous L-buthionine, sulfoximine and melphalan: an attempt at modulation of glutathione. J Clin Oncol

1994;12:194-205. 20. Perry RR, Mazetta J, Levin M, Barranco SC. Glutathione

levels and variability in breast tumors and normal tissue.

Cancer 1993;72:784-7. 21. Mekhail-Ishak K, Hudson N, Tsao M, Batist G. Implica-

tions fbr therapy of drug-metabolizing enzymes in human colon cancer. Cancer Res 1989;49:4866-9.

22. Hanigan MH, Frierson HF Jr, Brown JE, Lovell MA, Taylor PT. Human ovarian tumors express gamma-

g lu t amyl t r a n s p e p t i d a s e . Cancer Res 1994;54: 286-90.

23. Barranco SC, Townsend CM Jr, Quraishi MA, et al.

Heterogeneous responses of an in vitro model of human stomach cancer to anticancer drugs. Invest New

Drugs 1983;1:117-27. 24. Cook JA, Pass HI, Iype SN, et al. Cellular glutathione

and thiol measurements from surgically resected human

tung tumor and normal lung tissue. Cancer Res 1991; 51:4287-94.

25. Lee FY, Vessey A, Rofstad E, Siemann DW, Southerland RM. Heterogeneity of glutathione content in human

ovarian cancer. Cancer Res 1989;49:5244-8. 26. Cook JA, Mitchell JB. Measurement of thiols in cell

populations from tumor and normal tissue. Methods

Enzymol 1995;17:203-12. 27. Wiggers T, Arends JSV, Volovics A. Regression analysis

of prognostic factors in colorectal cancer after curative resections. Dis Colon Rectum 1988;31:33-41.

relationship between colorectal cancer glutathione levels and patient survival

Documents