- Email: [email protected]
Plasma Neuroendocrine Markers in Patients with Benign Prostatic Hyperplasia and Prostatic Carcinoma
oO22-5347/96/15541340$03.OO/O
Vol. 155,1340-1343,April 1996 Printed in U.S.A.
THE JOLlRNAL OF UR0LOC.Y
Copyright 0 1996 by AMERICUUROLOGICAL ASSOCLATION, INC.
PLASMA NEUROENDOCRINE MARKERS IN PATIENTS WITH BENIGN PROSTATIC HYPERPLASIA AND PROSTATIC CARCINOMA 0. CUSSENOT, J.M. VILLEITE, A. VALERI, G. CARIOU, F. DESGRANDCHAMPS, A. CORTESSE, P. MERIA, P. TEILLAC, J . FIET AND A. LE DUC From tht Departments of Urology, Hormone Biology and Pathology, Hbpital Saint Louis, Paris, France
ABSTRACT
Purpose: Approximately 50% of all malignant prostatic tumors contain neuroendocrine cells, which cannot be attributed to small cell prostatic carcinoma or carcinoid-like tumors, and which represent only 1 to 2%of all prostatic malignancies. Only limited data are available concerning the plasma levels of neurwndocrine markers in patients with prostatic tumors. Therefore, we determined the incidence of high plasma levels of neuroendocrine markers in patients with benign and malignant prostatic disease. Materials and Methods: The presence of elevated plasma neuropeptide levels was investigated in 135 patients with prostatic carcinoma and 28 with benign prostatic hyperplasia. Plasma chromogranin A, neurone-specific enolase, substance P, calcitonin, somatostatin, neurotensin and bombesin levels were analyzed by immunoassays, and were compared to clinical and pathological stages of disease. Plasma prostatic acid phosphatase and prostate specific antigen levels were also determined. All patients were followed for at least 2 years after inclusion in the study. Results: Significantly elevated levels of chromogranin A were detected in 15%of patients with prostatic carcinoma before any treatment. During hormone resistant prostate cancer progression plasma chromogranin A and neuron-specific enolase levels were elevated in 55%and 30% of the patients, respectively. In patients with stage D3 disease survival curves were generated by the Kaplan-Meier method, and log rank analysis revealed a statistically significant difference between groups positive and negative for chromogranin A. Substance P and bombesin were also occasionally elevated in prostatic tumors. Determination of neuroendocrine differentiation by neuron-specific enolase or chromogranin A immunoassays was not helpful in the prediction of progressive localized prostatic carcinoma. Conclusions: Future studies of plasma neuropeptide levels should confirm whether these parameters can be used as prognostic markers during late progression of prostatic carcinoma or for the selection of patients suitable for evaluation of new antineoplastic drugs known to be active against neuroendocrine tumors. KEY WORDS:neuroendocrine tumors, prostatic neoplasms, chromogranins, phosphopyruvate hydratase
The presence of neuroendocrine cells has been demonstrated by silver staining techniques or immunohistochemistry in the normal prostate, benign prostatic hyperplasia (BPH) and primary or metastatic prostatic adenocarcinoma before and after hormonal therapy.’ The presence of bioamines (serotonin)and peptides2v3(bombesin,4 vasoactive intestinal peptide, calcitonin5 and somatostatin9 has also been reported in prostatic fluid. Peptide receptors and autocrine secretion of various bioamines and peptides (serotonin, somatostatin, bombesin and calcitonin) have been reported in prostatic adenocarcinoma, and cell proliferation in prostatic carcinoma cell lines has been shown to be modulated by peptides and their analogues.7.s The concept of neuroendocrine differentiation in prostatic carcinoma has become more widely recognized in recent years. However, the clinical significance of this phenomenon remains unclear. Approximately 50% of all prostatic cancers contain neuroendocrine cells, which cannot be attributed to small cell prostatic carcinoma or carcinoid-like tumors, and which represent only 1 to 2% of all prostatic malignancies.9 It has been suggested that patients with tumors that show signs of neuroendocrine tissue differentiation, or with elevated plasma chromogranin A or neuron-specific enolase concentrations have a poorer prognosis than those without this Accepted for publication September 1. 1995.
type of differentiation.10-12 In contrast, expression of eutopic and/or ectopic peptide hormones in prostatic carcinoma was not associated with the “highest malignancy potential,” as evidenced by determination of tumor deoxyribonucleic acid ~ l o i d y . ’ ~The . ’ ~ presence of neuroendocrine differentiation in prostatic carcinoma suggests that new treatment strategies should be evaluated. We showed previously that stage D3 prostatic carcinomas, which express high levels of chroTheremogranin A, also express somatostatin re~eptors.1~ fore, neuroendocrine differentiation may respond particularly well to somatostatin therapy or new antagonist peptide analogues.I6 We determined in patients with BPH or prostatic adenocarcinoma the incidence of high plasma levels of neuroendocrine markers previously identified in prostatic tissue by immunohistochemistry.2 PATIENTS AND METHODS
Between September 1991 and August 1994, 163 patients were evaluated for either BPH (28) or malignant prostatic tumors (135), confirmed histologically in all patients after core biopsy, prostatectomy or transurethral resection of the prostate. When prostatectomy was performed for BPH, estimated to weigh more than 60 gm. by sonographic measurement, the weight of the tissue removed was also determined.
1340
1341
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS Elevated plasma marker ilalues Group
No.
No Pts.
Chromogranin A (normal less than 51 nglml.1 %
1 2 3 4 5 6
28 28 52 18 20 27
3.5 3.4 17 0
11 55
Mean r SD
22 r 24 2 37 2 19 2 24 2 85 %
12 13 35 8 17 78
Neuron-Specific Enolase (normal less than 10 ng./ml.) Ir
10.7 10 15 0 11 30
Mean
%
SD
7r4 8 % 11 9 r 11 6%2 622 15 r 30
PAP (normal less than 3 nghn1.J
PSA lnormal less than 3 ngJml.)
a
Mean z SD
%
Mean
25 17 66 0 11 60
6 - 11 3+2 41 f 70 2,l 3r4 165 Z 337
100
28 f 29 43 I43 423 z 945 2rl 24 r 16 1.335 f 3.070
93 I00
0 22 88
SD
I
Malignant tumors were staged according to the Interna- also indicated. Correlative analyses between clinical varitional Union Against Cancer and Whitmore classifications. ables and chromogranin A, neuron-specific enolase, PAP and After radical prostatectomy for prostatic carcinoma, tissue PSA values were performed using chi-square and Student’s t specimens were staged pathologically according to the Inter- tests. The weight of BPH specimens was compared with the national Union Against Cancer system and graded by level of each marker, correlations were determined and p Gleason score. Histological grade was determined by review c0.05 was considered significant. Survival curves were genof the original material when available (excluding core biop- erated by the Kaplan-Meier method. Statistical comparisons sies). Routine clinical studies, such as digital rectal exami- were analyzed with the log rank test. nation, ultrasonography, abdominal computerized tomography and scintigraphic bone scans, were performed for staging purposes if necessary. All patients underwent a thorough RESULTS clinical evaluation to exclude endocrine tumors. Finally, paPatient age ranged from 60 to 85 years. Clinical and pathotients were classified into 6 groups: group l-BPH, including patients with a prostatic volume estimated to be more than logical stages, and histological grade were consistent with 60 cc by digital rectal examination and sonographic determi- those of previous reports. The percentages of patients with nation justifying surgical prostatectomy, group 2-untreated high plasma levels of chromogranin A, neuron-specific enolocal (stages I1 or A2 and B) cancer, group 3-untreated lase, PSA and PAP, as well as the mean values plus or minus advanced (stages 111 and IV or C and D1 or D2) prostatic standard deviations of each marker level are shown in the carcinoma, group 4 -prostatic carcinoma in remission fol- table. No significant difference was observed between palowing radical prostatectomy, group 5-prostatic carcinoma tients with BPH and localized prostatic carcinoma for PSA, in remission with endocrine therapy and group 6-prostatic PAP, chromogranin A or neuron-specific enolase. Assay data carcinoma in progression with endocrine therapy (stage D3). measured in patients with BPH were correlated with the Performance status for patients with stage D3 disease was weight of BPH tissue. Except for PSA ( p <0.01), no other determined according to the World Health Organization markers were significantly correlated with BPH weight scale. Clinical followup starting from the date of therapy was (mean 82 2 32 gm.). designed to determine which patients had tumor progression, Patients with prostatic carcinoma were dwided into groups at defined as the presence of progression of 2 consecutive elediagnosis and according to therapeutic outcome. Significant vated plasma prostate specific antigen (PSA) levels (0.5 ng./ differences for PSA and PAP were found between localized m1.A or a positive bone scan or other imaging results consistent with metastatic disease. Tumor progression after prostatic carcinoma before treatment and after radical prostaendocrine therapy was defined as progression of symptomatic tectomy (PSA p = 0.03 and PAP p = 0.05), and between addisease (National Prostatic Cancer Project criteria, including vanced prostatic carcinoma before and after androgen deprivaurinary obstruction, bone metastases and performance sta- tion (PSA and PAP p = 0.001) but not for chromogranin A or tus but an isolated increase in PSA was not considered to neuron-specific enolase. Plasma chromogranin A and neuronIndicate progression). Plasma chromogranin-A levels were specific enolase levels were elevated in 55% and 30% of the measured by a double antibody sandwich enzyme-linked im- patients with stage D3 prostatic cancer, respectively, with a significant difference between stages D3 and D2 (p = 0.003) for munosorbent assay (normal 19 to 51 ng./ml.). Plasma neuron-specific enolase, prostatic acid phos- chromogranin A but not for neuron-specific enolase, PAP or phatase (PAP) and PSA assays were performed by immuno- PSA. Another interesting finding is that 4 patients ( 7 8 ) with radiometric techniques using monoclonal antibodies (ELSA- stage D3 prostatic carcinoma had normal plasma PSA and PAP NSE,* ELSA-PAP* and ELSA-PSA*). Blood samples were levels, with elevated plasma chromogranin A levels. Patients also obtained from 30 men 30 to 90 years old with no pros- with stage D3 and a high plasma chromogranin A level died tatic or neuroendocrine disease who served as a control earlier than those with a normal chromogranin A level. No goup. No age-related variation was observed. Normal values significant difference (Student’s t test) for World Health Orgawere less than 10 ng./ml. for neuron-specific enolase, less nization performance status was observed between groups posthan 3 ng./ml. for PSA and less than 3 ng./ml. for PAP. itive and negative for chromogranin A. Survival curves were Substance P and bornbesin were determined in plasma COI- generated by the Kaplan-Meier method and log rank analysis h e d in ice-chilled tubes containing ethylenediaminetet- showed a statistically significant difference between groups raacetic acid and aprotinin. Normal values were less than 40 positive and negative for chromogranin A (p = 0.002, see figure) Pg./ml. for substance p and less than 70 pg./ml. for bombesin. but not between those positive and negative for neuron-specific Peptides were first extracted through Sep-Pak C18t carenolase (data not shown). tndges and, after reconstitution with assay buffer, they were Other peptides, such as bombesin (3 cases with stage D2 determined by radioimmunoassay using commerically ava~lable disposable kits. Renal function was determined by se- and 2 with stage D3 disease), were occasionally markedly elevated in patients with advanced hormone resistant procreatinine assay in all arms of the study. Marker values are expressed as mean plus or minus stan- static carcinoma. NO correlation was observed between dard deviation, Percentage of values higher than normal is Gleason score (regrouped as low grade 2 to 4, medium grade 5 to 7 and high grade 8 to 10) and pathological stage (pT2 and * cis Biointemational, Saclay, France. pT3) on analysis of prostatectomy specimens. I. Waters,Milford. Massachusetts.
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS
1342
Chrom0.A- -12
Chmm0.h 11.15 1b
u 0,6
E
i0.5
.
+l
: 1
5.4 +
I
?a 0.3 7, 0.2 0.1
t
+
0
.
'
.
'
O
O
N
. '
'
4
O
N
.
'
' N
0
o
nnna
A
'
O
i
'
'
~
~
~
~
(D
m
~
w
m
O
a
,! N
(0.~1)
Kaplan-Meier plots of disease-specific survival in 27 patients with stage D3 prostatic cancer as function of plasma chromogranin A (Chrorno. A ) levels (positive more than 51 ng./ml. and negative 51 ng./ml. or less). Censored survival values represent patients still alive a t last followup. DISCUSSION
A radioimmunoassay was used to analyze plasma levels of markers of neuroendocrine differentiation in patients with BPH and prostatic carcinoma. Plasma chromogranin A was the best marker to identify neuroendocrine differentiation during progression of advanced prostatic carcinomas. Determination of neuroendocrine marker expression in the primary tumor remains controversial since, as indicated in this study, the staining is always heterogeneous and often localized. The detection of neuroendocrine markers in the blood of patients with prostatic carcinoma constitutes a more global indicator and objective quantification of neuroendotrine differentiation of tumors, since it comesponds to the entire primary tumor and its associated metastases. Elevated plasma neuroendoc,+,e cell products (chromogranin A and neuron-specific enolase) have been described in patients.17. 1s mesecell products may possibly ofiginate from a prostatic carcinoid subpopulation dispersed within common epithelial structures. positive plasma values of ectopic (neuron-specific enolase, chromogranin A, substance P and bornbesin) and eutopic (PSA and PAP) markers, therefore, seem at least to reflect different clones and in&cate the presence of a neuroendocrine prostatic cancer cell subpopulation. &dmon et a1 reported that plasma chromogranin A levels were elevated in 48% patients with stage D2 prostatic cancer.17 Similarly, bgothetis and Hoosein found that 47% of patients treated for androgen-independent growth of classical prostatic adenocacinoma had sign& cantly elevated plasma bornbesin levels, although increased bornbesin levels were rarely detected in our series (only 2 of 27 patients, 7%, with stage 3 disease).11 In the study of &dmon et a1 the frequency of high chromogranin A levels (17% of our patients with advanced prostate cancer) and the fact that the conventional PSA and PAP markers were norma1 in an unusually high percentage (33%)of patients with stage D2 prostatic cancer suggested that these patients did not demonstrate the usual progressive tumor status. None of our patients with advanced prostatic cancer (including stage D2) and 14% of those with hormone resistant cancer had a normal plasma PSA level, which is more representative of stage D2 and a hormonal resistant profile. The fact that prostatic carcinomas expressed neuropeptide markers (chromopanin A 17% and neuron-s~ecificen01as.e 15%)before any endocrine therapy shows that neuroendocrine products may be involved in prostatic carcinoma progression independently of androgen withdrawal. The incidence of elevated neuronspecific enolase levels in patients with prostatic tumors was
lower than that of elevated chromogranin A levels. Tarle and Rados found that elevated plasma neuron-specific enolase levels were more frequent in untreated subjects with localized tumors (28.6%) than in untreated subjects with disseminated disease ( 10.7c7C).ls However, at least according to COhen et a1,10 elevated neuron-specific enolase values were frequently present in nonresponders to hormonal therapy. Neuron-specific enolase levels were inversely correlated with stage in untreated prostatic carcinoma patients, which contradicts other evidence suggesting that neuroendocrine differentiation is directly correlated with grade.". lo We found that elevated neuron-specific enolase levels were more frequent during hormonal resistance (30%) but they did not have any prognostic value compared to chromogranin A expression (55%) and showed poorer concordance than immu~ ~ N nohistochemical d aOt a . 9 ~ ' ~ With a n immunohistochemical technique, Cohen et a1 detected neuroendocrine differentiation in 29% of patients with ~ this ex-~ localized prostatic carcinoma (stage 11). H pression Of neuroendocrine markers was not predictive of subsequent progression, since elevated levels of plasma markers were only detected for advanced tumors, corresponding to the primary tumor together with its metastases. The fact that 10.7% of the patients with BPH had a n abnormally high neuron-specific enolase level suggests a possible role of amine precursor uptake and decarboxylation cells during the development of BPH and requires further investigation. Cockett et al showed that small proliferating nodules of BPH contain abundant serotonin positive endocrineparacrine cells.19 CONCLUSIONS
The presence of neuroendocrine cells has been proposed by Some to be an independent prognostic variable in prostatic cancer but immunohistochemical determination of neuroendocrine differentiation in focal lesions was not predictive of clinical outcome. To our knowledge the amount of neuroendocrine cells required to produce a detectable elevation in Plasma chromogranin A has not yet been determined. Although the current limitations of the application of neuropeptides as a serological test appear to be obvious, our results will be used to delineate further the possible roles of specific neuropeptides as prostatic cancer markers in diagnostic and monitoring Protocols. Determination of Plasma chromo@anin A levels appears to be a useful addition to the analysis of neuroendocrine differentiation in prostatic adenocarcinoma, especially during hormonal resistance to this disease. The Presence of elevated Plasma levels of chromogranin A or other peptide profiles seems to predict the development of resistance to hOrmonal suppression therapy, associated with a Particularly poor prognosis. Growth and production of autocrine-paracrine factors by neuroendocrine cells or neuronal differentiation of tumor epithelial cells are probably influenced by hormones and involved in the progression of prostatiC carcinoma. It has been postulated that long-term androgen ablative therapy Produces an outgrowth of mmendocrine cells in Prostatic tumors. These findings suggest that determination O f a u W e - P m a c f i e neuropeptide production in prostatic carcinoma must be included in the development of new antiPePtider@c therapy for Prostatic carcimm, including Peptide an~ogues,8,20 and may be u s e l l to select patients during late Progression of prostatic carcinoma for treatment by drugs known be active against neuroendocrine tumors. REFERENCES
Di Sant.Agnese, p. A. and de Mesy Jensen, K, L.: Neuroendocnne differentiation in prostatic carcinoma. Hum. Path., 1 8 849, 1987. 2. Abrahamsson, P.A., Alumets, J., Wadstrom, L. B., Falkmer, S. and Grimelius, L.: Peptide hormones, serotonin, and other cell
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS differentiation markers in benign hyperplasia and in carcinoma of the prostate. Prog. Clin. Biol. Res., 243A:489, 1987. 3. Abrahamsson, P. A., Wadstrom, L. B., Alumets, J., Falkmer, S. and Grimelius, L.: Peptide-hormone- and serotonin-immunoreactive tumour cells in carcinoma of the prostate. Path. Res. Pract., 1 8 2 298, 1987. 4. Ohtake, R.: Four putative neuropeptides concentrations in the human urogenital tract. Comparison of the neuropeptides concentration between malignant and benign tissues. Nippon Naibunpi Gakkai Zasshi, 6 5 1105, 1989. 5. Di Sant’Agnese, P. A,, De Mesy Jensen, K. L. and Ackroyd, R. K.: Calcitonin, katacalcin and calcitonin gene-related peptide in the human prostate. An immunocytochemical and immunoelectron microscopic study. Arch. Path. Lab. Med., 113 790, 1989. 6. Sasaki, A. and Yoshinaga, K.: Immunoreactive somatostatin in
male reproductive system in humans. J. Clin. Endocr. Metab., 6 8 996, 1989.
7. Bologna, M., Festuccia, C., Muzi, P.. Biordi, L. and Ciomei, M.: Bombesin stimulates growth of human prostatic cancer cells in vitro. Cancer, 6 3 1714, 1989. 8. Pinski, J., Schally, A. V., Halmos, G. and Szepeshazi, K.: Effect of somatostatin analog RC-160 and bombesidgastrin releasing peptide antagonist RC-3095 on growth of PC-3 human prostate-cancer xenografts in nude mice. Int. J. Cancer, 55: 963, 1993. 9. Cohen, R. J., Glezerson, G., Haffejee, 2. and Afrika, D.: Prostatic carcinoma: histological and immunohistological factors affecting prognosis. Brit. J . Urol., 66:405, 1990. 10. Cohen, M. K., Arber, D. A., Coffield, K. S., Keegan, G. T., McClintock, J. and Speights, V. O., Jr.: Neuroendocrine differentiation in prostatic adenocarcinoma and its relationship to tumor progression. Cancer, 7 4 1899, 1994. 11. Logothetis, C. and Hoosein, N.: The inhibition of the paracrine progression of prostatic cancer a s an approach to early therapy of prostatic carcinoma. J . Cell. Biochem., suppl., 16H: 128, 1992.
1343
12. Di Sant’Agnese, P. A,: Neuroendocrine differentiation in human prostatic carcinoma. Hum. Path., 23:287, 1992. 13. Falkmer, S., Askensten, U. and Nesland, J. M.: Tumour pathol-
ogy of the neuron-paraneuron system and its evolutionary background, with special attention to neuro-endocrine differentiation in prostate and mammary carcinomas. Arch. Histol. Cytol., suppl., 52: 225, 1989. 14. Aprikian, A. G., Cordon-Cardo, C., Fair, W. R., Zhang, Z.-F., Bazinet, M., Hamdy, S. M. and Reuter, V. E.: Neuroendocrine differentiation in metastatic prostatic adenocarcinoma. J . Urol., 151: 914, 1994. 15. Cussenot, O., Duet, M., Cariou, G., Benelhadj, S., Desgrandchamps, F., Meria, P., Cortesse, A., Teillac, P., Le Duc, A. and Mundler, 0.: Apport de la scintigraphie a l’Octreoscan@ dans la detection des localisations metastatiques des adenocarcinomes prostatiques. Prog. Urol., 4 B34, 1994.
16. Schally, A. V. and Redding, T. W.: Somatostatin analogs as adjuncts to agonists of luteinizing hormone-releasing hormone in the treatment of experimental prostatic cancer. Proc. Natl. Acad. Sci., 84: 7275, 1987. 17. Kadmon, D., Thompson, T. C., Lynch, G. R. and Scardino, P. T.: Elevated plasma chromogranin-A concentrations in prostatic carcinoma. J. Urol., 1 4 6 358, 1991. 18. Tarle, M. and Rados, N.: Investigation on serum neurone-specific enolase in prostatic cancer diagnosis and monitoring: comparative study of a multiple tumor marker assay. Prostate, 19 23, 1991. 19. Cockett, A. T. K., Di Sant’Agnese, P. A,, Gopinath, P., Schoen, S.
and Abrahamsson, P. A.: Relationship of neuroendocrine cells of prostate and serotonin to benign prostatic hyperplasia. Urology, 4 2 512, 1993. 20. Di Sant’Agnese, P. A. and Cockett, A. T. K.: The prostatic endocrine-paracrine (neuroendocrine) regulatory system and neuroendocrine differentiation in prostatic carcinoma: a review and future directions in basic research. J. Urol., part 2, 1 5 2 1927, 1994.
Vol. 155,1340-1343,April 1996 Printed in U.S.A.
THE JOLlRNAL OF UR0LOC.Y
Copyright 0 1996 by AMERICUUROLOGICAL ASSOCLATION, INC.
PLASMA NEUROENDOCRINE MARKERS IN PATIENTS WITH BENIGN PROSTATIC HYPERPLASIA AND PROSTATIC CARCINOMA 0. CUSSENOT, J.M. VILLEITE, A. VALERI, G. CARIOU, F. DESGRANDCHAMPS, A. CORTESSE, P. MERIA, P. TEILLAC, J . FIET AND A. LE DUC From tht Departments of Urology, Hormone Biology and Pathology, Hbpital Saint Louis, Paris, France
ABSTRACT
Purpose: Approximately 50% of all malignant prostatic tumors contain neuroendocrine cells, which cannot be attributed to small cell prostatic carcinoma or carcinoid-like tumors, and which represent only 1 to 2%of all prostatic malignancies. Only limited data are available concerning the plasma levels of neurwndocrine markers in patients with prostatic tumors. Therefore, we determined the incidence of high plasma levels of neuroendocrine markers in patients with benign and malignant prostatic disease. Materials and Methods: The presence of elevated plasma neuropeptide levels was investigated in 135 patients with prostatic carcinoma and 28 with benign prostatic hyperplasia. Plasma chromogranin A, neurone-specific enolase, substance P, calcitonin, somatostatin, neurotensin and bombesin levels were analyzed by immunoassays, and were compared to clinical and pathological stages of disease. Plasma prostatic acid phosphatase and prostate specific antigen levels were also determined. All patients were followed for at least 2 years after inclusion in the study. Results: Significantly elevated levels of chromogranin A were detected in 15%of patients with prostatic carcinoma before any treatment. During hormone resistant prostate cancer progression plasma chromogranin A and neuron-specific enolase levels were elevated in 55%and 30% of the patients, respectively. In patients with stage D3 disease survival curves were generated by the Kaplan-Meier method, and log rank analysis revealed a statistically significant difference between groups positive and negative for chromogranin A. Substance P and bombesin were also occasionally elevated in prostatic tumors. Determination of neuroendocrine differentiation by neuron-specific enolase or chromogranin A immunoassays was not helpful in the prediction of progressive localized prostatic carcinoma. Conclusions: Future studies of plasma neuropeptide levels should confirm whether these parameters can be used as prognostic markers during late progression of prostatic carcinoma or for the selection of patients suitable for evaluation of new antineoplastic drugs known to be active against neuroendocrine tumors. KEY WORDS:neuroendocrine tumors, prostatic neoplasms, chromogranins, phosphopyruvate hydratase
The presence of neuroendocrine cells has been demonstrated by silver staining techniques or immunohistochemistry in the normal prostate, benign prostatic hyperplasia (BPH) and primary or metastatic prostatic adenocarcinoma before and after hormonal therapy.’ The presence of bioamines (serotonin)and peptides2v3(bombesin,4 vasoactive intestinal peptide, calcitonin5 and somatostatin9 has also been reported in prostatic fluid. Peptide receptors and autocrine secretion of various bioamines and peptides (serotonin, somatostatin, bombesin and calcitonin) have been reported in prostatic adenocarcinoma, and cell proliferation in prostatic carcinoma cell lines has been shown to be modulated by peptides and their analogues.7.s The concept of neuroendocrine differentiation in prostatic carcinoma has become more widely recognized in recent years. However, the clinical significance of this phenomenon remains unclear. Approximately 50% of all prostatic cancers contain neuroendocrine cells, which cannot be attributed to small cell prostatic carcinoma or carcinoid-like tumors, and which represent only 1 to 2% of all prostatic malignancies.9 It has been suggested that patients with tumors that show signs of neuroendocrine tissue differentiation, or with elevated plasma chromogranin A or neuron-specific enolase concentrations have a poorer prognosis than those without this Accepted for publication September 1. 1995.
type of differentiation.10-12 In contrast, expression of eutopic and/or ectopic peptide hormones in prostatic carcinoma was not associated with the “highest malignancy potential,” as evidenced by determination of tumor deoxyribonucleic acid ~ l o i d y . ’ ~The . ’ ~ presence of neuroendocrine differentiation in prostatic carcinoma suggests that new treatment strategies should be evaluated. We showed previously that stage D3 prostatic carcinomas, which express high levels of chroTheremogranin A, also express somatostatin re~eptors.1~ fore, neuroendocrine differentiation may respond particularly well to somatostatin therapy or new antagonist peptide analogues.I6 We determined in patients with BPH or prostatic adenocarcinoma the incidence of high plasma levels of neuroendocrine markers previously identified in prostatic tissue by immunohistochemistry.2 PATIENTS AND METHODS
Between September 1991 and August 1994, 163 patients were evaluated for either BPH (28) or malignant prostatic tumors (135), confirmed histologically in all patients after core biopsy, prostatectomy or transurethral resection of the prostate. When prostatectomy was performed for BPH, estimated to weigh more than 60 gm. by sonographic measurement, the weight of the tissue removed was also determined.
1340
1341
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS Elevated plasma marker ilalues Group
No.
No Pts.
Chromogranin A (normal less than 51 nglml.1 %
1 2 3 4 5 6
28 28 52 18 20 27
3.5 3.4 17 0
11 55
Mean r SD
22 r 24 2 37 2 19 2 24 2 85 %
12 13 35 8 17 78
Neuron-Specific Enolase (normal less than 10 ng./ml.) Ir
10.7 10 15 0 11 30
Mean
%
SD
7r4 8 % 11 9 r 11 6%2 622 15 r 30
PAP (normal less than 3 nghn1.J
PSA lnormal less than 3 ngJml.)
a
Mean z SD
%
Mean
25 17 66 0 11 60
6 - 11 3+2 41 f 70 2,l 3r4 165 Z 337
100
28 f 29 43 I43 423 z 945 2rl 24 r 16 1.335 f 3.070
93 I00
0 22 88
SD
I
Malignant tumors were staged according to the Interna- also indicated. Correlative analyses between clinical varitional Union Against Cancer and Whitmore classifications. ables and chromogranin A, neuron-specific enolase, PAP and After radical prostatectomy for prostatic carcinoma, tissue PSA values were performed using chi-square and Student’s t specimens were staged pathologically according to the Inter- tests. The weight of BPH specimens was compared with the national Union Against Cancer system and graded by level of each marker, correlations were determined and p Gleason score. Histological grade was determined by review c0.05 was considered significant. Survival curves were genof the original material when available (excluding core biop- erated by the Kaplan-Meier method. Statistical comparisons sies). Routine clinical studies, such as digital rectal exami- were analyzed with the log rank test. nation, ultrasonography, abdominal computerized tomography and scintigraphic bone scans, were performed for staging purposes if necessary. All patients underwent a thorough RESULTS clinical evaluation to exclude endocrine tumors. Finally, paPatient age ranged from 60 to 85 years. Clinical and pathotients were classified into 6 groups: group l-BPH, including patients with a prostatic volume estimated to be more than logical stages, and histological grade were consistent with 60 cc by digital rectal examination and sonographic determi- those of previous reports. The percentages of patients with nation justifying surgical prostatectomy, group 2-untreated high plasma levels of chromogranin A, neuron-specific enolocal (stages I1 or A2 and B) cancer, group 3-untreated lase, PSA and PAP, as well as the mean values plus or minus advanced (stages 111 and IV or C and D1 or D2) prostatic standard deviations of each marker level are shown in the carcinoma, group 4 -prostatic carcinoma in remission fol- table. No significant difference was observed between palowing radical prostatectomy, group 5-prostatic carcinoma tients with BPH and localized prostatic carcinoma for PSA, in remission with endocrine therapy and group 6-prostatic PAP, chromogranin A or neuron-specific enolase. Assay data carcinoma in progression with endocrine therapy (stage D3). measured in patients with BPH were correlated with the Performance status for patients with stage D3 disease was weight of BPH tissue. Except for PSA ( p <0.01), no other determined according to the World Health Organization markers were significantly correlated with BPH weight scale. Clinical followup starting from the date of therapy was (mean 82 2 32 gm.). designed to determine which patients had tumor progression, Patients with prostatic carcinoma were dwided into groups at defined as the presence of progression of 2 consecutive elediagnosis and according to therapeutic outcome. Significant vated plasma prostate specific antigen (PSA) levels (0.5 ng./ differences for PSA and PAP were found between localized m1.A or a positive bone scan or other imaging results consistent with metastatic disease. Tumor progression after prostatic carcinoma before treatment and after radical prostaendocrine therapy was defined as progression of symptomatic tectomy (PSA p = 0.03 and PAP p = 0.05), and between addisease (National Prostatic Cancer Project criteria, including vanced prostatic carcinoma before and after androgen deprivaurinary obstruction, bone metastases and performance sta- tion (PSA and PAP p = 0.001) but not for chromogranin A or tus but an isolated increase in PSA was not considered to neuron-specific enolase. Plasma chromogranin A and neuronIndicate progression). Plasma chromogranin-A levels were specific enolase levels were elevated in 55% and 30% of the measured by a double antibody sandwich enzyme-linked im- patients with stage D3 prostatic cancer, respectively, with a significant difference between stages D3 and D2 (p = 0.003) for munosorbent assay (normal 19 to 51 ng./ml.). Plasma neuron-specific enolase, prostatic acid phos- chromogranin A but not for neuron-specific enolase, PAP or phatase (PAP) and PSA assays were performed by immuno- PSA. Another interesting finding is that 4 patients ( 7 8 ) with radiometric techniques using monoclonal antibodies (ELSA- stage D3 prostatic carcinoma had normal plasma PSA and PAP NSE,* ELSA-PAP* and ELSA-PSA*). Blood samples were levels, with elevated plasma chromogranin A levels. Patients also obtained from 30 men 30 to 90 years old with no pros- with stage D3 and a high plasma chromogranin A level died tatic or neuroendocrine disease who served as a control earlier than those with a normal chromogranin A level. No goup. No age-related variation was observed. Normal values significant difference (Student’s t test) for World Health Orgawere less than 10 ng./ml. for neuron-specific enolase, less nization performance status was observed between groups posthan 3 ng./ml. for PSA and less than 3 ng./ml. for PAP. itive and negative for chromogranin A. Survival curves were Substance P and bornbesin were determined in plasma COI- generated by the Kaplan-Meier method and log rank analysis h e d in ice-chilled tubes containing ethylenediaminetet- showed a statistically significant difference between groups raacetic acid and aprotinin. Normal values were less than 40 positive and negative for chromogranin A (p = 0.002, see figure) Pg./ml. for substance p and less than 70 pg./ml. for bombesin. but not between those positive and negative for neuron-specific Peptides were first extracted through Sep-Pak C18t carenolase (data not shown). tndges and, after reconstitution with assay buffer, they were Other peptides, such as bombesin (3 cases with stage D2 determined by radioimmunoassay using commerically ava~lable disposable kits. Renal function was determined by se- and 2 with stage D3 disease), were occasionally markedly elevated in patients with advanced hormone resistant procreatinine assay in all arms of the study. Marker values are expressed as mean plus or minus stan- static carcinoma. NO correlation was observed between dard deviation, Percentage of values higher than normal is Gleason score (regrouped as low grade 2 to 4, medium grade 5 to 7 and high grade 8 to 10) and pathological stage (pT2 and * cis Biointemational, Saclay, France. pT3) on analysis of prostatectomy specimens. I. Waters,Milford. Massachusetts.
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS
1342
Chrom0.A- -12
Chmm0.h 11.15 1b
u 0,6
E
i0.5
.
+l
: 1
5.4 +
I
?a 0.3 7, 0.2 0.1
t
+
0
.
'
.
'
O
O
N
. '
'
4
O
N
.
'
' N
0
o
nnna
A
'
O
i
'
'
~
~
~
~
(D
m
~
w
m
O
a
,! N
(0.~1)
Kaplan-Meier plots of disease-specific survival in 27 patients with stage D3 prostatic cancer as function of plasma chromogranin A (Chrorno. A ) levels (positive more than 51 ng./ml. and negative 51 ng./ml. or less). Censored survival values represent patients still alive a t last followup. DISCUSSION
A radioimmunoassay was used to analyze plasma levels of markers of neuroendocrine differentiation in patients with BPH and prostatic carcinoma. Plasma chromogranin A was the best marker to identify neuroendocrine differentiation during progression of advanced prostatic carcinomas. Determination of neuroendocrine marker expression in the primary tumor remains controversial since, as indicated in this study, the staining is always heterogeneous and often localized. The detection of neuroendocrine markers in the blood of patients with prostatic carcinoma constitutes a more global indicator and objective quantification of neuroendotrine differentiation of tumors, since it comesponds to the entire primary tumor and its associated metastases. Elevated plasma neuroendoc,+,e cell products (chromogranin A and neuron-specific enolase) have been described in patients.17. 1s mesecell products may possibly ofiginate from a prostatic carcinoid subpopulation dispersed within common epithelial structures. positive plasma values of ectopic (neuron-specific enolase, chromogranin A, substance P and bornbesin) and eutopic (PSA and PAP) markers, therefore, seem at least to reflect different clones and in&cate the presence of a neuroendocrine prostatic cancer cell subpopulation. &dmon et a1 reported that plasma chromogranin A levels were elevated in 48% patients with stage D2 prostatic cancer.17 Similarly, bgothetis and Hoosein found that 47% of patients treated for androgen-independent growth of classical prostatic adenocacinoma had sign& cantly elevated plasma bornbesin levels, although increased bornbesin levels were rarely detected in our series (only 2 of 27 patients, 7%, with stage 3 disease).11 In the study of &dmon et a1 the frequency of high chromogranin A levels (17% of our patients with advanced prostate cancer) and the fact that the conventional PSA and PAP markers were norma1 in an unusually high percentage (33%)of patients with stage D2 prostatic cancer suggested that these patients did not demonstrate the usual progressive tumor status. None of our patients with advanced prostatic cancer (including stage D2) and 14% of those with hormone resistant cancer had a normal plasma PSA level, which is more representative of stage D2 and a hormonal resistant profile. The fact that prostatic carcinomas expressed neuropeptide markers (chromopanin A 17% and neuron-s~ecificen01as.e 15%)before any endocrine therapy shows that neuroendocrine products may be involved in prostatic carcinoma progression independently of androgen withdrawal. The incidence of elevated neuronspecific enolase levels in patients with prostatic tumors was
lower than that of elevated chromogranin A levels. Tarle and Rados found that elevated plasma neuron-specific enolase levels were more frequent in untreated subjects with localized tumors (28.6%) than in untreated subjects with disseminated disease ( 10.7c7C).ls However, at least according to COhen et a1,10 elevated neuron-specific enolase values were frequently present in nonresponders to hormonal therapy. Neuron-specific enolase levels were inversely correlated with stage in untreated prostatic carcinoma patients, which contradicts other evidence suggesting that neuroendocrine differentiation is directly correlated with grade.". lo We found that elevated neuron-specific enolase levels were more frequent during hormonal resistance (30%) but they did not have any prognostic value compared to chromogranin A expression (55%) and showed poorer concordance than immu~ ~ N nohistochemical d aOt a . 9 ~ ' ~ With a n immunohistochemical technique, Cohen et a1 detected neuroendocrine differentiation in 29% of patients with ~ this ex-~ localized prostatic carcinoma (stage 11). H pression Of neuroendocrine markers was not predictive of subsequent progression, since elevated levels of plasma markers were only detected for advanced tumors, corresponding to the primary tumor together with its metastases. The fact that 10.7% of the patients with BPH had a n abnormally high neuron-specific enolase level suggests a possible role of amine precursor uptake and decarboxylation cells during the development of BPH and requires further investigation. Cockett et al showed that small proliferating nodules of BPH contain abundant serotonin positive endocrineparacrine cells.19 CONCLUSIONS
The presence of neuroendocrine cells has been proposed by Some to be an independent prognostic variable in prostatic cancer but immunohistochemical determination of neuroendocrine differentiation in focal lesions was not predictive of clinical outcome. To our knowledge the amount of neuroendocrine cells required to produce a detectable elevation in Plasma chromogranin A has not yet been determined. Although the current limitations of the application of neuropeptides as a serological test appear to be obvious, our results will be used to delineate further the possible roles of specific neuropeptides as prostatic cancer markers in diagnostic and monitoring Protocols. Determination of Plasma chromo@anin A levels appears to be a useful addition to the analysis of neuroendocrine differentiation in prostatic adenocarcinoma, especially during hormonal resistance to this disease. The Presence of elevated Plasma levels of chromogranin A or other peptide profiles seems to predict the development of resistance to hOrmonal suppression therapy, associated with a Particularly poor prognosis. Growth and production of autocrine-paracrine factors by neuroendocrine cells or neuronal differentiation of tumor epithelial cells are probably influenced by hormones and involved in the progression of prostatiC carcinoma. It has been postulated that long-term androgen ablative therapy Produces an outgrowth of mmendocrine cells in Prostatic tumors. These findings suggest that determination O f a u W e - P m a c f i e neuropeptide production in prostatic carcinoma must be included in the development of new antiPePtider@c therapy for Prostatic carcimm, including Peptide an~ogues,8,20 and may be u s e l l to select patients during late Progression of prostatic carcinoma for treatment by drugs known be active against neuroendocrine tumors. REFERENCES
Di Sant.Agnese, p. A. and de Mesy Jensen, K, L.: Neuroendocnne differentiation in prostatic carcinoma. Hum. Path., 1 8 849, 1987. 2. Abrahamsson, P.A., Alumets, J., Wadstrom, L. B., Falkmer, S. and Grimelius, L.: Peptide hormones, serotonin, and other cell
PLASMA NEUROENDOCRINE MARKERS IN PROSTATIC TUMORS differentiation markers in benign hyperplasia and in carcinoma of the prostate. Prog. Clin. Biol. Res., 243A:489, 1987. 3. Abrahamsson, P. A., Wadstrom, L. B., Alumets, J., Falkmer, S. and Grimelius, L.: Peptide-hormone- and serotonin-immunoreactive tumour cells in carcinoma of the prostate. Path. Res. Pract., 1 8 2 298, 1987. 4. Ohtake, R.: Four putative neuropeptides concentrations in the human urogenital tract. Comparison of the neuropeptides concentration between malignant and benign tissues. Nippon Naibunpi Gakkai Zasshi, 6 5 1105, 1989. 5. Di Sant’Agnese, P. A,, De Mesy Jensen, K. L. and Ackroyd, R. K.: Calcitonin, katacalcin and calcitonin gene-related peptide in the human prostate. An immunocytochemical and immunoelectron microscopic study. Arch. Path. Lab. Med., 113 790, 1989. 6. Sasaki, A. and Yoshinaga, K.: Immunoreactive somatostatin in
male reproductive system in humans. J. Clin. Endocr. Metab., 6 8 996, 1989.
7. Bologna, M., Festuccia, C., Muzi, P.. Biordi, L. and Ciomei, M.: Bombesin stimulates growth of human prostatic cancer cells in vitro. Cancer, 6 3 1714, 1989. 8. Pinski, J., Schally, A. V., Halmos, G. and Szepeshazi, K.: Effect of somatostatin analog RC-160 and bombesidgastrin releasing peptide antagonist RC-3095 on growth of PC-3 human prostate-cancer xenografts in nude mice. Int. J. Cancer, 55: 963, 1993. 9. Cohen, R. J., Glezerson, G., Haffejee, 2. and Afrika, D.: Prostatic carcinoma: histological and immunohistological factors affecting prognosis. Brit. J . Urol., 66:405, 1990. 10. Cohen, M. K., Arber, D. A., Coffield, K. S., Keegan, G. T., McClintock, J. and Speights, V. O., Jr.: Neuroendocrine differentiation in prostatic adenocarcinoma and its relationship to tumor progression. Cancer, 7 4 1899, 1994. 11. Logothetis, C. and Hoosein, N.: The inhibition of the paracrine progression of prostatic cancer a s an approach to early therapy of prostatic carcinoma. J . Cell. Biochem., suppl., 16H: 128, 1992.
1343
12. Di Sant’Agnese, P. A,: Neuroendocrine differentiation in human prostatic carcinoma. Hum. Path., 23:287, 1992. 13. Falkmer, S., Askensten, U. and Nesland, J. M.: Tumour pathol-
ogy of the neuron-paraneuron system and its evolutionary background, with special attention to neuro-endocrine differentiation in prostate and mammary carcinomas. Arch. Histol. Cytol., suppl., 52: 225, 1989. 14. Aprikian, A. G., Cordon-Cardo, C., Fair, W. R., Zhang, Z.-F., Bazinet, M., Hamdy, S. M. and Reuter, V. E.: Neuroendocrine differentiation in metastatic prostatic adenocarcinoma. J . Urol., 151: 914, 1994. 15. Cussenot, O., Duet, M., Cariou, G., Benelhadj, S., Desgrandchamps, F., Meria, P., Cortesse, A., Teillac, P., Le Duc, A. and Mundler, 0.: Apport de la scintigraphie a l’Octreoscan@ dans la detection des localisations metastatiques des adenocarcinomes prostatiques. Prog. Urol., 4 B34, 1994.
16. Schally, A. V. and Redding, T. W.: Somatostatin analogs as adjuncts to agonists of luteinizing hormone-releasing hormone in the treatment of experimental prostatic cancer. Proc. Natl. Acad. Sci., 84: 7275, 1987. 17. Kadmon, D., Thompson, T. C., Lynch, G. R. and Scardino, P. T.: Elevated plasma chromogranin-A concentrations in prostatic carcinoma. J. Urol., 1 4 6 358, 1991. 18. Tarle, M. and Rados, N.: Investigation on serum neurone-specific enolase in prostatic cancer diagnosis and monitoring: comparative study of a multiple tumor marker assay. Prostate, 19 23, 1991. 19. Cockett, A. T. K., Di Sant’Agnese, P. A,, Gopinath, P., Schoen, S.
and Abrahamsson, P. A.: Relationship of neuroendocrine cells of prostate and serotonin to benign prostatic hyperplasia. Urology, 4 2 512, 1993. 20. Di Sant’Agnese, P. A. and Cockett, A. T. K.: The prostatic endocrine-paracrine (neuroendocrine) regulatory system and neuroendocrine differentiation in prostatic carcinoma: a review and future directions in basic research. J. Urol., part 2, 1 5 2 1927, 1994.