Significance and prognostic value of lysosomal enzyme activities measured in surgically operated adenocarcinomas of the gastroesophageal junction and squamous cell carcinomas of the lower third of esophagus

Abstract

AIM: To establish whether there are fundamental differences in the biochemistries of adenocarcinomas of the gastroesophageal junction (GEJ) and the squamous cell carcinomas of the lower third of the esophagus (LTE).

METHODS: Between February 1, 1997 and February 1, 2000, we obtained tissue samples at the moment of resection from 54 patients for biochemical analysis. The full set of data could be comprehensively analyzed in 47 of 54 patients samples (81%). Of these, 29 were adenocarcinomas of the GEJ Siewert type I (n = 8), type II (n = 12), type III (n = 9), and 18 presented as squamous cell carcinomas of the LTE. We evaluated the mean values of 11-lysosomal enzyme and 1-cytosol protease activities of the tumorous and surrounding mucosae as well as their relative activities, measured as the ratio of activity in tumor and normal tissues from the same patient. These data were further analyzed to establish the correlation with tumor localization, TNM stage (lymph-node involvement), histological type (papillary, signet-ring cell, tubular), state of differentiation (good, moderate, poor), and survival (≤ 24 or ≥ 24 mo).

RESULTS: In adenocarcinomas, the activity of α-mannosidase (AMAN), cathepsin B (CB) and dipeptidyl-peptidase I (DPP I) increased significantly as compared to the normal gastric mucosa. In squamous cell carcinomas of the esophagus, we also found a significant difference in the activity of cathepsin L and tripeptidyl-peptidase I in addition to these three. There was a statistical correlation of AMAN, CB, and DPP I activity between the level of differentiation of adenocarcinomas of the GEJ and lymph node involvement, because tumors with no lymph node metastases histologically confirmed as well-differentiated, showed a significantly lower activity. The differences in CB and DPP I activity correlated well with the differences in survival rates, since the CB and DPP I values of those who died within 24 mo following surgical intervention were significantly higher than of those who survived for 2 years or more.

CONCLUSION: Adenocarcinomas of the GEJ form a homogenous group from a tumor-biochemical aspect, and differ from the biochemical characteristics of squamous cell carcinomas of the LTE on many points. When adenocarcinomas of the GEJs are examined at the preoperative phase, the ratio of the performed AMAN, CB, and DPP I enzymatic activity of the tissue sample from the tumor and adjacent intact mucosa within2 cm of the tumor may have a prognostic value even in the preoperative examination period, and may indicate that ranking of these patients into the neo-adjuvant treatment group should be considered.

Key Words: Prognostic value, Lysosomal enzymes, Cardiac adenocarcinomas, Siewert classification, Esophageal squamous cell carcinoma

INTRODUCTION

While the majority of carcinomas located in the upper and middle third of the esophagus are squamous cell carcinomas, the incidence of adenocarcinomas in the lower third of the esophagus (LTE) is increasing in the Western world. Although their etiology is not clear, esophageal adenocarcinomas almost always arise in areas of Barrett’s metaplasia, usually in the specialized intestinal-type mucosa in the lower esophagus, and predominantly affect elderly white men, often those with a history of heavy smoking. Barrett’s metaplasia is an accepted complication of gastroesophageal reflux disease, but gastric cancer is not linked to reflux disease. Thus, it is of interest to note that the incidence of proximal gastric adenocarcinomas is increasing. At Temple University Hospital , the percentage of cardiac carcinomas within gastric carcinomas has increased from 16% to 25% between 1976-80 and 1991-95[1]. Antonioli and Goldman[2] reported that the incidence of cardiac carcinomas has increased from 0% to 27% of all gastric carcinomas in Boston . They also showed that the number of signet-ring cell carcinomas is increased with a decrease in the male to female ratio, and an increase in the age of patients.

The classification system described by Siewert[3], which organizes adenocarcinomas of the gastroesophageal junction (GEJ) into tumors of the distal esophagus as type I (Si I), true carcinomas of the cardia as type II (Si II) and subcardial carcinomas as type III (Si III), has allowed for a comparative assessment of the data of the various sites and has facilitated the choice of surgical therapy. This classification brings light intothe “Black Box” of GEJ carcinomas. The classification itself can easily be performed by summarizing all available information from contrast radiography, endoscopy, and intraoperative findings.

Following surgical resection, there is a clear survival advantage for patients with early-stage tumors over patients with late-stage tumors, regardless of tumor location. There is, therefore, no difference in the survival rates between any of the three locations[4]. Although much effort has been made to obtain a more accurate understanding and to apply more effective therapy for carcinomas in the lower esophagus and proximal stomach, patient survival rates are still not satisfactory.

A number of molecular markers have been analyzed as possible prognostic factors in patients with esophageal cancer including expression of proliferating cell nuclear antigen, epidermoid growth factor receptor, cyclin Dl, p53, and p21. However, none of these markers are of clinical value[5-7]. Merely aberrant expression of pRB and high serum immunosuppressive acidic protein concentration seems to be useful as a prognostic factor for esophageal squamous cell carcinoma[8,9].

As the etiology of tumors may well be different, it is important to establish whether there are fundamental differences in the biochemistries of adenocarcinomas of the GEJ and the squamous cell carcinomas of the LTE. In order to answer this question, it seems evident that proteolytic enzyme activities in the tumor tissue and the histologically intact mucosa bordering, it should be measured and compared, since peptidases are primarily responsible for intracellular catabolism and turnover. The increased level of lysosomal enzymes in tumors, together with their ability to degrade extracellular matrix proteins, has led to the hypothesis that they could be involved in the process of invasion and metastasis[10-12].

MATERIALS AND METHODS

Between February 1, 1997 and February 1, 2000, we obtained tissue samples at the moment of resection from 54 patients for biochemical analysis. The average age was 57.8 years and the M/F ratio was 35/19 (65%/35%). The full set of data could be comprehensively analyzed in 47 of 54 patients samples (81%). Of these, 29 were adenocarcinomas of the GEJ Siewert type I (n = 8), type II (n = 12), type III (n = 9), and 18 presented as squamous cell carcinomas of the LTE. Tissue samples obtained during surgery were further separated into two groups: cancer tissue and normal tissue closely surrounding the cancer tissue, within 2 cm of the tumor border. The latter was checked histologically to confirm it to be tumor-free. For enzyme assays the mucosal layer was used. Reasons for exclusion included histologically confirmed synchronous multiple carcinoma (n = 3), death within early postoperative period (n = 2) and intolerance of study protocol (n = 2).

All patients in the study were so-called advanced cancer cases. Thus, the TNM stage and number of esophageal tumors were IIA:6, IIB:6, and III:6, while the tumors of the GEJ were II:10, IIIA:11, and IIIB:8.

The central-European reality is that almost 90% of malignant carcinomas in this area of the gastrointestinal tract are discovered at an advanced stage, and this fact is well reflected in this group of patients.

We evaluated the mean values of 11-lysosomal enzyme and 1-cytosol protease (as control) specific activities of the tumorous and surrounding mucosae as well as their relative activities, measured as the ratio of activity in tumor and normal tissues from the same patient. These data were further analyzed to establish the correlation with tumor localization, TNM stage (lymph-node involvement), histological type (papillary, signet-ring cell, tubular), state of differentiation (good, moderate, poor), and survival (≤ 24 or ≥ 24 mo). The term of relative activity was used for the mean ratio of enzyme activity values for tumorous and intact mucosae in individual patients expressed as a decimal fraction.

Tissue samples from the tumor and intact surrounding area were frozen on dry ice immediately after dissection and stored at -70 °C prior to use.

Samples were thawed on ice, placed in 50 volumes (w/v) of 0.15 mol/L NaCl, 0.1% Triton X-100 and homogenized with a Brinkmann Polytron homogenizer. A soluble supernatant was prepared by centrifugation at 12 000 g for 25 min at 4 °C .

Glycosidase activities were measured using 4-methylum-belliferyl (4-MU) substrates as previously described[13]. Protease assays using 7-amino-4-methylcoumarine (AMC) substrates were conducted as described by Sleat et al[14] and Sohar et al[15]. Reactions were initiated by adding 40 mL of substrate (various concentrations 20 mmol/L-1 mmol/L)-buffer (100 mmol/L) solution to 5 mL (cathepsin B) or 10 mL (other enzymes) of sample (supernatants diluted two-, four-, and eightfold in homogenization buffer in duplicate), incubated at 37 °C , and terminated by the addition of 100 mL of 0.5 mol/L glycine, pH 10.5 (4-MU substrates) or 0.1 mol/L monochloroacetic acid in 0.1 mol/L acetate, pH 4.3 (AMC substrates). Buffers consisted of 0.1 mol/L citric acid or 0.1 mol/L sodium acetate adjusted to the indicated pH using sodium hydroxide, acetic acid, or HCl respectively, contained 0.1% Triton-X-100 with 0.15 mol/L NaCl. Substrates were purchased from Sigma and prepared as stocks in dimethyl sulfoxide (remaining substrates) that were added to the reaction buffer immediately prior to assay. Samples added to substrate solutions after the addition of the termination buffer were used as blanks. Fluorescent reaction products were determined using a CytoFluor II fluorescence multiwell plate reader (PerSeptive Biosystems, Framingham , MA , USA ) with excitation at 360 nm and emission at 460 nm.

Since our variables were normally distributed, Microsoft Excel t-test with two samples, correlated t-test and single factor analysis of variance were used for the statistical analysis. Permission for the investigations was sought and obtained from the appropriate local ethical committee.

RESULTS

The analysis of 11-lysosomal enzyme and 1-cytosol protease-thimet oligopeptidase (THP)-activity levels in adenocarcinomas of the GEJ and squamous cell carcinomas of the LTE showed that α-mannosidase (AMAN), cathepsin B (CB) and dipeptidyl-peptidase I (DPP I) had a significant increase in adenocarcinoma (Table 1). AMAN, CB, DPP I, cathepsin L (CL) and tripeptidyl-peptidase I (TPP I) showed a significant increase in squamous cell carcinoma as compared to the levels of activity measured in the bordering intact mucosa (Table 2).

Table 1 Specific activities of lysosomal enzymes in 29 adenocarcinomas of the GEJ and surrounding normal mucosa (mean±SE).

Enzymes	Adenocarcinoma	Normal gastric mucosa	P
AGLU	27315 ± 3139	31210 ± 2490	NS
AMAN	89594 ± 8518	66653 ± 4066	<0.05
BGAL	318224 ± 43694	245442 ± 15421	NS
CB	324095 ± 39402	167582 ± 14842	<0.001
CL	392438 ± 44328	296166 ± 29044	NS
DPP I	2140499 ± 287604	1291759 ± 126796	<0.05
CH	491813 ± 51110	404285 ± 40939	NS
DPP II	80982 ± 9916	110728 ± 15265	NS
GCU	219615 ± 42364	178776 ± 18607	NS
HEX	213292 ± 28785	186396 ± 18779	NS
TPP I	921939 ± 328332	443130 ± 134508	NS
THP	36271 ± 14639	23382 ± 10817	NS

GEJ: gatroesophageal junction; NS: non significant; AGLU: α-glucosidase; AMAN: α-mannosidase; BGAL: β-galactosidase; CB: cathepsin B; CL: cathepsin L; DPP I: dipeptidyl-peptidase I; CH: cathepsin H; DPP II: dipeptidyl-peptidase II; GCU: β-glucuronidase; HEX: β-hexosaminidase B; TPP I: tripeptidyl-peptidase I; THP: thimet oligopeptidase (pmol /h·mg).

Table 2 Specific activities of lysosomal enzymes in 18 squamous cell carcinomas of the LTE and surrounding normal mucosa (mean±SE).

Enzymes	Squamous cell carcinoma	Normal esophageal mucosa	P
AGLU	54216 ± 16793	21596 ± 4106	NS
AMAN	86987 ± 14 625	36905 ± 2287	<0.05
BGAL	225913 ± 43791	135696 ± 24644	NS
CB	535860 ± 134809	114697 ± 13037	<0.05
CL	292996 ± 56644	60964 ± 14115	<0.001
DPP I	2328558 ± 634873	401436 ± 172042	<0.05
CH	578119 ± 50491	453922 ± 54228	NS
DPP II	41816 ± 7226	34539 ± 6833	NS
GCU	134414 ± 22391	115250 ± 19634	NS
HEX	149704 ± 30362	86965 ± 15449	NS
TPP I	191061 ± 20874	112048 ± 11472	<0.001
THP	45828 ± 7921	31459 ± 7164	NS

LTE: lower third of the esophagus; NS: non significant; AGLU: α-glucosidase; AMAN: α-mannosidase; BGAL: β-galactosidase; CB: cathepsin B; CL: cathepsin L; DPP I: dipeptidyl-peptidase I; CH: cathepsin H; DPP II: dipeptidyl-peptidase II; GCU: β-glucuronidase; HEX: β-hexosaminidase B; TPP I: tripeptidyl-peptidase I; THP: thimet oligopeptidase (pmol /h·mg).

When relative activity values were compared (Table 3) for α-glucosidase (AGLU), CB, CL, and DPP I enzymes, the activity level was twofold higher in squamous carcinomas than in adenocarcinomas.

Table 3 Relative activities of lysosomal enzymes in adenocarcinoma of the GEJ and in squamous cell carcinoma of the LTE (mean±SE).

Enzymes	GEJ	LTE	Ratio of LTE/GEJ
	Tumor/own normal mucosa	Tumor/own normal mucosa
AGLU	0.91 ± 0.09	2.55 ± 0.78	2.8
AMAN	1.39 ± 0.12	2.36 ± 0.32	1.7
BGAL	1.33 ± 0.16	2.01 ± 0.42	1.5
CB	2.15 ± 0.34	4.89 ± 1.2	2.3
CL	1.48 ± 0.17	6.21 ± 2	4.2
DPP I	2.34 ± 0.52	12.43 ± 5.3	5.3
CH	1.3 ± 0.12	1.44 ± 0.31	1.1
DPP II	0.83 ± 0.08	1.39 ± 0.23	1.7
GCU	1.36 ± 0.24	1.46 ± 0.33	1.1
HEX	1.29 ± 0.18	2.06 ± 0.45	1.6
TPP I	1.66 ± 0.33	1.77 ± 0.21	1.1
THP	2.36 ± 0.41	3.59 ± 2.3	1.5

GEJ: gatroesophageal junction; LTE: lower third of the esophagus; SE: standard error; AGLU: α-glucosidase; AMAN: α-mannosidase; BGAL: β-galactosidase; CB: cathepsin B; CL: cathepsin L; DPP I: dipeptidyl-peptidase I; CH: cathepsin H; DPP II: dipeptidyl-peptidase II; GCU: β-glucuronidase; HEX: β-hexosaminidase B; TPP I: tripeptidyl-peptidase I; THP: thimet oligopeptidase.

The values of peptidase activity in cardiac tumors with Si I, II, and III localization did not differ significantly (Table 4). There was no significant correlation between lysosomal enzyme activities and adenocarcinomas of different histological types (papillary, signet-ring cell, tubular). There was no significant difference between carcinomas of stages II, IIIA, and IIIB. However, we found a significant and relatively strong relationship between the level of differentiation of adenocarcinomas of the GEJ and the level of activity of AMAN (H: 0.78), CB (H: 0.76), and DPP I (H: 0.67) lysosomal enzymes. The lysosomal enzyme activities in well-differentiated adenocarcinomas were significantly lower than those in poorly differentiated ones. The changes of the activity of certain enzymes correlated well with lymph node involvement. Tumors with no lymph-node metastases showed a significantly lower value of activity. A moderate relation was detected between the relative activities and the presence of lymph-node metastases (H: 0.61). The change in CB and DPP I activity sensitively reflected the differences in survival rates too.

Table 4 Prognostic value of lysosomal enzyme activites in adenocarcinomas of the GEJ (mean±SE).

Prognosticcategories	Relative activity	AMAN	Specific activity	Relative activity	CB	Specific activity	Relative activity	DPP I	Specific activity
Siewert I	1.36 ± 0.19	NS	110 136 ± 20902	1.68 ± 0.37	NS	273838 ± 66573	1.37 ± 0.42	NS	1896366 ± 446446
Siewert II	1.49 ± 0.23		88413 ± 17021	2.42 ± 0.48		417978 ± 65610	2.58 ± 0.64		2673743 ± 638594
Siewert III	1.34 ± 0.19		79 101 ± 9469	2.20 ± 0.68		278994 ± 63462	2.68 ± 1.1		1861382 ± 384367
Papillary	1.54 ± 0.27	NS	92754 ± 22 466	2.79 ± 1.19	NS	368700 ± 112886	2.97 ± 1.60	NS	2156354 ± 763729
Signet-ring cell	1.42 ± 0.13		82743 ± 10 810	2.04 ± 0.27		29588 ± 58097	2.37 ± 0.59		2328450 ± 567443
Tubular	1.30 ± 0.26		90069 ± 8 722	1.80 ± 0.45		32476 ± 68549	2.10 ± 1.12		1916498 ± 504487
Grade 1	0.91 ± 0.11	P < 0.01	61931 ± 8716	0.95 ± 0.08	P < 0.01	210697 ± 54630	0.72 ± 0.06	P < 0.01	981478 ± 111507
Grade 2–3	1.63 ± 0.13		103425 ± 10283	2.75 ± 0.42		380793 ± 46390	3.14 ± 0.69		2720009 ± 332437
N0	1.00 ± 0.17	P < 0.05	67068 ± 10 377	1.03 ± 0.11	P < 0.01	201550 ± 55664	0.75 ± 0.07	P < 0.01	1053122 ± 122928
N1–N2	1.59 ± 0.13		10857 ± 10644	2.71 ± 0.43		385367 ± 44735	3.12 ± 0.69		2684187 ± 345227
Survival ≤ 2 yr	1.55 ± 0.14	NS	103466 ± 11103	2.76 ± 0.46	P < 0.01	382830 ± 49621	3.24 ± 0.74	P < 0.01	2722922 ± 373145
Survival ≥ 2 yr	1.13 ± 0.17		67052 ± 9135	1.16 ± 0.14		228650 ± 51792	0.87 ± 0.07		1194060 ± 160577
Stage II	1.19 ± 0.13	NS	80194 ± 8553	1.90 ± 0.50	NS	293720 ± 52833	1.98 ± 0.74	NS	1860244 ± 358265
Stage III	1.66 ± 0.18		102128 ± 15960	2.47 ± 0.41		36494 ± 59832	2.81 ± 0.73		2514172 ± 465458

GEJ: gastroesophageal junction; NS: non significant; N0: lymph node negative; N1–N2: lymph node positive; AMAN: α-mannosidase; CB: cathepsin B; DPP I: dipeptidyl-peptidase I.

Tables 2 and 3 show that the lysosomal enzyme activity increased significantly in the so-called advanced esophageal carcinomas compared to that in adenocarcinomas of the GEJ. However, we found no statistically measurable correlation between the rate of increased lysosomal enzyme activities and maturity of tumors, as well as the involvement of lymph nodes and TNM stage. Moreover, in contrast to adenocarcinomas of the GEJ, the significantly increased lysosomal enzyme activities in squamous cell carcinomas of the LTE did not correlate with survival time. Thus, these parameters in stages II and III had no prognostic value.

DISCUSSION

The incidence of adenocarcinomas of the esophagus (Si type I) is increasing[16,17]. These cancers are associated with Barrett’s specialized epithelium and appear to be part of the continuum of gastroesophageal reflux disease. To date, no data suggest that the prevalence of gastroesophageal reflux disease has changed over the past 20 years and thus the cause of this current “epidemic” remains unclear[18].

The epidemiology of adenocarcinomas of the “classic” gastric cardia (Si type II-III) remains even less clear. Very few studies have examined this as a distinct entity. Overall, it seems that the incidence of cancer in this location, mainly Si type II, has not changed significantly in the past 50 years[1]. Some cases of gastric cardia cancer may be associated with chronic gastritis and Helicobacter pylori infection, whereas other cases may be related to Barrett’s specialized epithelium in the proximal stomach[19-21].

A number of reports suggest that the synthesis, transport, and processing of lysosomal enzymes and their phosphorylated derivatives are altered in cancer, making these proteins interesting subjects for investigation[22]. The ability of tumor cells to invade tissues and metastasize is thought to involve an increased expression of proteinases and/or a decrease in the levels of proteinase inhibitors. Peptidases may facilitate metastasis in a number of different ways, including detachment of individual cells from the primary tumor, invasion of surrounding tissues to allow contact with vascular channels, degradation of the basement membrane during both intravasation and extravasation, and invasion of tissues during the formation of secondary tumor sites. Several classes of proteinases have been implicated in this process including matrix metalloproteinases, cysteine proteinases as CB or CL, aspartic proteinase cathepsin D, and serine proteinase plasminogen activators[23-28]. It is possible that human tumors may use combinations of these enzymes working synergistically to facilitate invasion or alternatively one specific proteinase may play a dominant role in tissue invasion for a given cancer, while some tumors exhibit an alteration in the synthesis or processing of many or all lysosomal enzymes. Defining the role of proteinases in the invasiveness of different tumors is important with regard to understanding the biology of this process and searching for potential prognostic markers and targets for therapeutic intervention.

The results of clinical investigations on cysteine cathepsins and their endogenous inhibitors in human breast, lung, brain, liver, and head and neck tumors, as well as in body fluids of ovarian, uterine, melanoma and colorectal carcinoma bearing patients, have shown that these molecules are highly predictive for the length of survival and may be used for risk assessment of relapse and death in cancer patients[29-32].

By examining the correlation of cysteine protease CB and laminin degradation, Khan et al[33] demonstrated that increased CB expression and decreased tumor-associated laminin levels may suggest a mechanism underlying the progression of colorectal adenomas to carcinomas.

As for the relationship of gastric tumors and proteolytic enzymes, b hexosaminidase and its isoenzymes[34], and plasminogen activators are in the center of interest. Cathepsins B, L and tissue type plasminogen activator activities are higher in gastric cancer tissues than in normal gastric tissues. Chung and Kawai[35] also revealed that inhibitory activities of CL, CB, urinary and tissue type plasminogen activators are stronger in normal tissue closely surrounding the gastric cancer compared to normal tissue distant from the tumor border, as a result of a defense mechanism of the host against cancer invasion.

Yamamoto et al[36] reported that the matrix metallop-roteinase matrilysin may play a key role in the progression of esophageal carcinoma and that its detection may be useful in predicting recurrence and poor prognosis, and possibly in selecting patients for anti-matrix metalloproteinase therapy. Another study on esophageal adenocarcinoma demonstrated that amplicon at 8p22, the locus of the CB gene, is identified and associated with amplification and overexpression of the CB gene in esophageal adenocarcinoma[37].

Prior to this study, no lysosomal enzyme activity measurements have been performed in adenocarcinomas of the GEJ. During our examinations we measured the activity of 11 different lysosomal enzymes and one cytosol protease THP in Si type I-III adenocarcinomas of the GEJ and in squamous cell carcinomas of the LTE. The activity of AMAN, CB, and DPP I in adenocarcinoma increased significantly as compared to that in normal gastric mucosa. In squamous cell carcinomas of the esophagus (Table 1) we also found a significant difference in the activity of CL and TPP I in addition to these three (Table 2). Moreover, when the relative activity values were compared for AGLU, CB, CL, and DPP I in adenocarcinoma of the GEJ and in squamous cell carcinoma of the LTE, marked differences in the relative activity levels by a factor of 2-5 respectively could be observed in squamous cell carcinomas as opposed to adenocarcinomas (Table 3).

From a tumor-biochemical aspect, adenocarcinomas within the Si I-III localization system did not show any differences, nor were there any significant differences in their lysosomal enzyme activities. Similarly to this there was no significant correlation between protease activities and adenocarcinomas of different histological types (papillary, signet-ring cell, tubular). However, we found a statistically measurable correlation to AMAN, CB and DPP I between the level of differentiation of adenocarcinomas of the GEJ and lymph node involvement, because histologically well-differentiated tumors with no lymph node metastases showed a significantly lower activity. The change of two proteases with cysteine catalytic site, namely CB and DPP I, activity correlated well with differences in survival rates, since the CB and DPP I values of those who died within 24 mo following surgical intervention were significantly higher than the same values for those who survived for 2 years or more (Table 4). At the same time, no statistically measurable correlation was found in relation to the proteases despite of displaying significantly higher activity levels as compared to the normal mucosa, the level of differentiation of tumors, their lymph node state or survival rates for squamous cell carcinomas of the LTE.

We conclude that adenocarcinomas of different locations in the GEJ form a homogenous group and differ in many aspects from malformations and neoformations within similar locations, though they possess a different histological structure. This conclusion is supported by data from Izutani et al[38]. The difference in radiation sensitivity could be attributed to the tissue type difference. The increased MnSOD mRNA and MnSOD proteins in adenocarcinoma are believed to indicate the activity of strong defense mechanisms protecting the active form of cysteine catalytic center at CB and DPP I against reactive oxygen species, and this activity could be a cause of resistance against radiation therapy and quinone anticancer drugs.

Our results suggest that when adenocarcinomas of the GEJ are investigated at the preoperative phase, the ratio of the actual AMAN, CB, and DPP I enzyme activity in the samples taken from the tumor and its adjacent tissues may have some prognostic value. Relative activity values within the range of 1, signaling a comparatively more favorable prognosis pattern mean that a specific kind of “static warfare” prevails between the tumor tissue and the surrounding intact mucosa. This may be due not only to the diminished invasiveness of a given tumor, but also to the more effective defensive properties of the intact bordering mucosa, or it may be a consequence of a reduction in protease inhibitor synthetic activity. It seems, therefore, that the activity levels of AMAN, CB, and DPP I in adenocarcinoma may assist us in formulating our preoperative therapeutic strategy for adenocarcinomas of the GEJ.

ACKNOWLEDGMENTS

The authors are grateful for the opportunity to measure lysosomal enzyme activities in the laboratory of Dr. Peter Lobel at CABM, UMDNJ,Piscataway , NJ , USA . We also thank Professor Charles J. Filipi, Creighton Medical University , Omaha , Nicola Pen Jackson and Laszlo Novak for critical reading of the manuscript before its submission.