Steroidogenesis in Human Adrenocortical Carcinoma:
Biochemical Activities, Immunohistochemistry, and In Situ Hybridization of Steroidogenic Enzymes and Histopathologic Study in Nine Cases
HIRONOBU SASANO, MD, TAKASHI SUZUKI, MD, HIROSHI NAGURA, MD, AND TETSUO NISHIKAWA, MD
To obtain a better understanding of steroid metabolism associated with adrenocortical malignancy we studied steroidogenesis in nine cases of adrenocortical carcinoma (six with Cushing’s syndrome, two without clinically significant adrenocortical hormonal abnor- malities, and one with primary aldosteronism) by analyzing bio- chemical enzyme activities (21-hydroxylase and 118-hydroxylase) and by immunohistochemistry and in situ hybridization of steroid- ogenic enzymes in carcinoma tissues. 21-Hydroxylase activity was markedly low but 118-hydroxylase activity was only moderately de- creased compared with normal adrenal activity. Immunohisto- chemical study of steroidogenic enzymes revealed that six of the nine cases expressed all the enzymes required for cortisol or al- dosterone biosynthesis. Immunoreactivity of these enzymes was predominantly observed in small carcinoma cells with compact and/ or clear cytoplasm and minimum morphologie nuclear atypia. In those cases with positive steroidogenic enzymes immunohisto- chemical examination of serial tissue sections revealed that a number of carcinoma cells did not express all the enzymes required for the synthesis of biologically active steroids. This may account for an increased level of precursor steroid secretion associated with ad- renocortical malignancy. In situ hybridization of cytochrome 17a- hydroxylase demonstrated that carcinoma cells with positive hy- bridization signals generally were positive for immunoreactivity, but a discrepancy between mRNA and protein expression was oc- casionally observed. Although the conclusions derived from our current study are limited by the small number of cases, ineffective corticosteroidogenesis, characteristic of steroid metabolism in hu- man adrenocortical carcinoma, was considered to be due to dis- organized expression of steroidogenic enzymes in individual car- cinoma cells. HUM PATHOL 24:397-404. Copyright @ 1993 by W.B. Saunders Company
Malignant neoplasms are associated most often with alterations of functional properties of the tissue in which they arise. Multiple pathways of steroidogenesis are
From the Department of Pathology, Tohoku University School of Medicine, Sendai, Japan; and the Department of Medicine, Yo- kohama Rosai Hospital, Yokohama, Japan. Accepted for publication July 29, 1992.
Supported in part by a grant from Ichiro Kanehara Memorial Foundation, Tokyo, Japan, and by a grant from the Ministry of Health and Welfare, Disorders of Adrenal Hormones Research Committee, Japan.
Key words: adrenal cortex, carcinoma, steroidogenesis, immuno- histochemistry, in situ hybridization.
Address correspondence and reprint requests to Hironobu Sas- ano, MD, Department of Pathology, Tohoku University School of Medicine, 2-1 Seiryou-machi, Aoba-ku, Sendai-shi, Miyagi-ken, Japan 980.
present in the normal adrenal cortex, primarily involving formation of glucocorticoids, mineralocorticoids, and weak androgens. It is known that one or more of the steroidogenic pathways are expressed in adrenocortical neoplasms, including carcinoma. Approximately 90% of the cases of adrenocortical carcinoma are associated with clinically apparent adrenocortical hormonal abnormal- ities: of those, 50% are associated with Cushing’s syn- drome, 20% with virilization, 4% with both, 12% with feminization, and 4% with hypermineralocorticoidism.1 Even the cases not associated with clinical hormonal abnormalities exhibited hypersecretion of biologically inactive precursor steroids, such as pregnenolone or 17-hydroxypregnenolone.2 Therefore, an analysis of corticosteroidogenesis in human adrenocortical carci- noma can provide a good opportunity to study how original functional features (ie, steroid pathways in nor- mal cortex) are altered through malignant transfor- mation and may provide insights into the diagnosis and treatment of adrenocortical carcinoma.
Previous studies of steroidogenesis of human ad- renocortical carcinoma usually have been performed by analyzing the plasma concentration of corticosteroids,3 measuring the steroid output of a monolayer cell culture of tumor cells with high-performance liquid chroma- tography,2,4 measuring the steroid output of a contin- uous cell line (NCI-H295; established from an invasive primary adrenocortical carcinoma) with radioimmu- noassay and mass spectrometry,5 and examining in vitro steroid production by short-term incubation of tissue slices of adrenocortical carcinoma with radioimmu- noassay.6 We have undertaken relatively direct ap- proaches to an analysis of neoplastic corticosteroido- genesis in nine cases of adrenocortical carcinoma by analyzing the steroidogenic enzyme activities (21- and 118-hydroxylase). In addition, we have performed re- cently developed immunohistochemical analyses of all the enzymes involved in corticosteroid production as well as in situ hybridization of cytochrome P-45017%, a key enzyme in cortisol production, to localize the site of steroid production in carcinoma.
MATERIALS AND METHODS Patients
Clinical and hormonal data, including prognosis of the patients, are summarized in Tables 1 and 2. Four patients died
| Patient No. | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | |
| Age at time of initial operation (yr) | 33 | 43 | 73 | 29 | 21 | 13 | 48 | 32 | 37 |
| Sex | M | F | M | F | F | F | F | F | M |
| Symptom | CS | CS | CS | CS | CS | CS | PA | NF | NF |
| 17OHCS | 126.5 | 49.0 | 18.9 | 26.5 | 41.2 | 22.3 | 8.6 | 8.7 | 3.8 |
| 17-KS | 75.0 | 45.0 | 28.6 | 57.2 | 27.9 | 14.7 | 6.0 | 16.2 | 3.5 |
| PRA | 1.52 | 6.2 | 2.29 | 0.2 | 0.08 | 0.5 | 0.28 | ||
| Preg | 3.1 | 0.612 | 0.391 | 0.3 | 6.97 | 0.051 | |||
| 17-OHpreg | - | 3.75 | - | ||||||
| Prog | 1.5 | 2.13 | 1.21 | 0.6 | |||||
| DOC | 0.372 | - | 0.266 | 0.305 | |||||
| S | 5.3 | 9.72 | 10.1 | - | - | ||||
| B | 16.6 | 5.91 | 1.81 | 0.577 | |||||
| F | 61.6 | 43.0 | 41.8 | 16.8 | 31.0 | 38.9 | 12.4 | 12.8 | 8.2 |
| Aldo | 298.8 | 158.0 | 42.1 | 88 | 41 | 1258.2 | 70 | 15.3 | |
| DHEA | 5.4 | 6.74 | 5.63 | 14 | 8.85 | 1.2 | 26 | 1.59 | |
| DHEA-S | 6606 | 6031 | 2858 | 15600 | 2586 | 740 | 3230 | 613 | |
| Testo | 3.3 | 3.37 | 5.42 | - | - | ||||
| Response to ACTH | None | None | None | None | None | Not done | Not done | Normal response | Normal response |
Abbreviations (normal values): CS, Cushing’s syndrome; PA, primary aldosteronism; NF, clinically nonfunctioning; 17OHCS, urinary 17- OHCS (3.18 to 11.5 mg/d [men], 1.33 to 5.39 mg/d [women]); 17-KS, urinary 17-KS (3.06 to 11.5 mg/d [men], 1.98 to 7.39 mg/d [women]); PRA, plasma renin activity (0.5 to 3.0 mg/mL/hr); Preg, pregnenolone (0.1 to 1.0 ng/mL [men], 0.2 to 1.5 ng/mL [women]); 17-OHpreg, 17- OH-pregnenolone (0.1 to 4.0 ng/ml); Prog, progesterone (<0.4 ng/ml [men], 0.1 to 1.5 ng/mL [women, follicular phase], 2.5 to 2.8 [women, luteal phase]); DOC, deoxycorticosterone (0.08 to 0.28 ng/mL [men], 0.03 to 0.33 ng/mL [women]); S, 11-deoxycortisol (0.04 to 1.16 ng/ml [men], 0.11 to 0.60 ng/mL [women]); B, corticosterone (0.38 to 8.42 ng/mL [men], 0.21 to 8.48 ng/ml. [women]); F, cortisol (5.22 to 20.8 ug/dL); Aldo, aldosterone (7.71 to 188 pg/mL); DHEA, dehydroepiandrosterone (1.2 to 7.5 ng/ml); DHEA-S, dehydroepiandrosterone-sulfate (400 to 1,500 ng/ml); Testo, testosterone (2.7 to 10.7 ng/ml [men], 0.06 to 0.86 ng/mL [women]). - Indicates not measured.
of carcinoma 2.5 to 27 months after the operation. Three patients were alive, but with clinically observed recurrence or metastasis 30 to 63 months after the operation. Two patients were alive and well without evidence of disease 42 and 57 months after the operation. Histopathologic examination re- vealed that all the cases met the criteria of adrenocortical ma- lignancy defined by Weiss.8 Details of Weiss’ histopathologic criteria of adrenocortical malignancy, DNA ploidy, and im- munohistochemical findings of oncogene products and inter-
mediate filaments in eight cases (cases no. 2 to 9) have been described elsewhere.9 Three patients were male and six were female; ages ranged from 13 to 73 years. Clinically, six patients exhibited Cushing’s syndrome without apparent virilization. One patient demonstrated hyperaldosteronism with sup- pressed plasma renin activity (primary aldosteronism) and two patients showed no evidence of clinical corticosteroid abnor- malities. Preoperative serum concentrations of corticosteroids measured by radioimmunoassay are summarized in Table 1.
| Case No. | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | |
| Enzymes | |||||||||
| P-450scc | + | + | + | f+ | + | + | + | f+ | |
| 38-HSD | + | + | + | + | + | + | + | f+ | |
| P-450c21 | + | + | + | + | + | + | + | ||
| P-45017a | + | + | + | + | + | + | |||
| P-450118 | + | + | f+ | + | f+ | + | - | ||
| Tumor weight (g) | 1,070 | 145 | 313 | 210 | 120 | 313 | 442 | 440 | 116 |
| Nuclear grade * | III | II | II | II | III | II | IV | II | II |
| Mitotic activityt | + | + | + | + | |||||
| Architecture# | + | - | + | + | + | + | + | ||
| Necrosis | + | + | + | + | + | + | + | + | + |
| CO | D (2.5) | D (6) | D (27) | AD (63) | AD (34) | D (16) | AD (30) | AN (42) | AN (57) |
Symbols: - , negative; +, positive; f+, focally positive.
Abbreviations: CO, clinical outcome; D, dead of disease; AD, alive with disease; AN, alive and no evidence of disease. Numbers in parentheses represent duration (months) of follow-up interval after operation.
* According to the criteria of Fuhrman et al.7
t Mitotic activity was evaluated by counting 50 random high-power fields in the area of the greatest number of mitotic figures: +, >5/50 high-power fields; - , <5/50 high-power fields.
# Architecture was evaluated by the presence (+) or absence (-) of a diffuse growth pattern.
A majority of the cases examined in this study were referred from other hospitals for histopathologic consultation. There- fore, preoperative serum concentration of all corticosteroids could not be analyzed, as shown in Table 1.
Biochemical Activities of Steroidogenic Enzymes
21-Hydroxylase (P-450(21) and 118-hydroxylase (P-450118) activities could be measured in two cases (cases no. 2 and 7) in which fresh tissue was available for examination.
Preparation of Mitochondrial and Microsomal Suspension
Mitochondrial and microsomal fractions were prepared according to the method of Barbieri et al10 with slight modi- fication. Approximately 5 g of the tissue was homogenized with 100 mL of 0.25 mol/L sucrose containing 10 mmol/L Tris-HCI, pH 7.4, by a Teflon/glass homogenizer. The super- natant was centrifuged at 8,000g for 15 minutes. This pellet was resuspended in the homogenizing buffer and centrifuged at 8,000g for 15 minutes. The final pellet was resuspended with an appropriate amount of the homogenizing buffer and used as the mitochondrial suspension. The supernatant was centrifuged at 105,000g for 60 minutes. This pellet was re- suspended in the homogenizing buffer and recentrifuged at 105,000g for 60 minutes. The final pellet was homogenized in a small volume of the homogenizing buffer and used as the microsome fraction. Protein concentration was determined by the method of Lowry et al11 with bovine serum albumin as the standard.
21-Hydroxylase
The determination of P-450c21 activity was performed ac- cording to the method of Gordon et al.12 The microsomal fraction was incubated with 150 mmol/L phosphate buffer (total volume, 200 uL) containing 11.25 umol/L [3H]17a-hy- droxyprogesterone (0.1 uCi/incubation) and saturating levels of 0.5 mmol/L NADPH. Reactions were performed at 37°℃ in air and terminated between 2 and 6 minutes later by adding 2 mL dichloromethane. The organic phase containing ex- tracted steroids was dried in N2 and reconstituted in 50 uL ethanol containing 20 ug unlabeled 17«-hydroxyprogesterone and 11-deoxycortisol acting as chromatographic carriers and aiding detection of each steroid by I2 vapor. The ethanolic extracts were subjected to thin-layer chromatography on plates precoated with silica gel using chloroform-ethyl acetate (4:1, v/v)13 as the solvent. The areas of 17@-hydroxyprogesterone and 11-deoxycortisol were scraped into scintillation vials and radioactivity was measured. The rate of conversion of 17a- hydroxyprogesterone to androstenedione by this system was less than 10% in both of the cases; this result was consistent with the findings of Gordon et al.12
118-Hydroxylase
Mitochondrial P-450118 activity was assayed as the rate of conversion of 11-deoxycortisol to cortisol by the mitochondrial fraction, which was prepared as described above according to the method of Greiner et al.14 Cortisol was measured fluoro- metrically.15
Immunohistochemistry
Immunohistochemical analysis of steroidogenic enzymes, including cholesterol side chain cleavage (P-450scc), 36-hy- droxysteroid dehydrogenase (38-HSD), P-450c21, 17a-hy-
droxylase (P-45017a), and P-450118, was performed on routinely processed, formalin-fixed, paraffin-embedded serial sections using the biotin-streptavidin amplified method and the His- tofine immunostaining system (Nichirei, Tokyo, Japan). Three to four different sections obtained from different areas of the carcinoma were examined for immunostain. For control im- munostaining, 0.01 mol/L phosphate-buffered saline and normal rabbit or mouse IgG were used instead of primary antibodies. Immunostaining procedures and properties of primary antibodies used in this study have been described pre- viously.16
In Situ Hybridization
Hybridization procedures used in this study were essen- tially according to Hayashi et al17 and Sasano et al.16 All glass- ware was washed, rinsed with distilled deionized water, and autoclaved before use. Gloves were worn when handling the glassware and slides to prevent RNase contamination of the tissue sections as much as possible. Briefly, deparaffinized sec- tions mounted on microscope slides were treated with pronase (0.25 mg/mL in 50 mmol/L Tris-HCI, pH 7.6, and 5 mmol/L EDTA) for 30 minutes at room temperature and acetylated with a freshly diluted acetic anhydride (0.25% in 0.1 mol/L triethanolamine buffer, pH 8.0) for 10 minutes. The treated sections were then processed for in situ hybridization at 45℃ for 18 hours. The cDNA probes used in this study were clone pcD17aH derived from human fetal adrenal (kindly provided by Dr J.I. Mason, University of Texas, Southwestern Medical School, Dallas, TX) and a mixture of a 433 base pair MspI- MspL fragment and a 345 base pair Xbal-MspI fragment from pcD17&H.18 Probes were labeled for 2 hours at 37℃ with 3H-dTTP by the multiprime DNA labeling method to a specific activity of 1 to 2 × 107 cpm/DNA. Hybridization mixtures contained the tritiated cDNA probe (2 ug/mL), yeast tRNA (500 µg/mL), salmon sperm DNA (80 ug/mL), 50% form- amide, 10 mmol/L Tris-HCI (pH 7.0), 0.15 mol/L NaCl, 1 mmol/L EDTA (pH 7.0), and 10% dextran sulfate. After hy- bridization and removal of the cover glass by immersing the slides in 2X standard saline-citrate buffer (SSC) (1X SSC = 0.15 mol/L NaCl, 0.015 mol/L trisodium citrate, pH 7.0) for 1 hour at room temperature, sections were washed three times in 2X SSC for 10 minutes each at 45℃. The slides were then developed in Kodak NTB-2 nuclear track emulsion and were exposed for 10 days at 4℃. The exposed slides were developed and the sections were stained with hematoxylin. Serial sections were treated with RNase A (20 ug/mL in 0.5 mol/L NaCl, 10 mmol/L Tris, pH 8.0, at 37℃ for 45 minutes) before hybrid- ization.
RESULTS
Preoperative Hormonal Data
Results are summarized in Table 1. Levels of 17- OHCS and 17-KS in 24-hour urine samples were ele- vated, except for 17-OHCS in case no. 9 and 17-KS in cases no. 7 and 9. Plasma renin activity was slightly sup- pressed in case no. 4 and markedly suppressed in case no. 7 (primary aldosteronism). In the cases not asso- ciated with apparent clinical hormonal abnormalities, one case (no. 8) demonstrated elevated levels of urinary 17-OHCS and 17-KS, pregnenolone, DHEA, and DHEA-S, but the other case (no. 9) did not show any abnormal values of adrenocortical hormones during the preoperative clinical course.
Steroidogenic Enzyme Activities
Activity of P-450c21 was 1.3 and 1.0 nmol/mg pro- tein/2 min in cases no. 2 and 7, respectively (normal range, 7.5 ± 2.5 nmol/mg protein/2 min; n= 3). Activity of P-450118 was 5.1 and 8.6 ug/mg protein/15 min in cases no. 2 and 7, respectively (normal range, 11.4 ± 3.8 ug/mg protein/15 min; n = 3). In both cases 118- hydroxylase activity was moderately low and P-450c21 activity was markedly decreased compared with normal adrenal activity.
Immunohistochemistry of Steroidogenic Enzymes
The immunohistochemistry results are summarized in Table 2. Immunoreactivity of enzymes involved in corticosteroidogenesis was observed in eight cases. In these eight cases all the enzymes involved in adrenal steroid biosynthesis were positive except for those in case no. 8, in which P-45017g and P-450118 were negative for immunostain, and case no. 9, in which P-450c21, P- 45017%, and P-450118 were negative for immunostain. In case no. 1 no immunoreactivity of any of the enzymes examined was detected. In all positive cases immuno- reactivity of the enzymes was generally prominent in small carcinoma cells with compact and/or clear cyto- plasm and with relatively low nuclear grade (ie, the ab- sence of marked nuclear atypia [Figs 1 and 2]). However, immunoreactivity was not necessarily correlated with overall nuclear grade of the cases, mitotic activity, or presence or absence of diffuse architecture and/or tu- mor necrosis. Large carcinoma cells with clear cytoplasm were negative for the enzyme except for 38-HSD, which also was observed in large tumor cells with clear abun- dant cytoplasm and minimum nuclear atypia.
Serial sections of the specimens revealed that the carcinoma cells positive for one steroidogenic enzyme did not necessarily express the other enzymes; ie, car-
cinoma cells did not produce all the enzymes required for production of biologically active steroids in most of the specimens examined. For example, in case no. 7 carcinoma cells positive for 36-HSD (Fig 3) were not necessarily positive for P-450c21 (Fig 4). In cases no. 8 and 9, which were not associated with adrenocortical hormonal abnormalities (except for elevated pregnen- olone, DHEA, and DHEA-S levels in case no. 8), P- 45017g and P-450118 were absent and the numbers of carcinoma cells positive for the enzymes were relatively smaller than those of cases no. 2 to 6, which were as- sociated with Cushing’s syndrome or primary aldoster- onism. In case no. 7, which was clinically associated with primary aldosteronism, P-45017g was observed in small compact carcinoma cells. Those tumor cells positive for P-45017a were not related to the vascular structure of the carcinoma. The immunoreactivity observed above was not present in the control sections described pre- viously.
In Situ Hybridization
In situ hybridization results are summarized in Ta- ble 3. Hybridization signals of P-45017g mRNA were detected in seven cases. In case no. 8 immunoreactivity was observed but mRNA expression could not be de- tected. In case no. 9 hybridization signals were present but immunoreactivity was not observed. Hybridization signals were observed as black dots on autoradiography (Fig 5). Heterogeneity of P-45017% hybridization signals was observed in the carcinoma cases that expressed P- 45017a MRNA (Fig 5). Carcinoma cells with positive hy- bridization signals generally were small tumor cells with minimum morphologic nuclear atypia. Examination of serial sections demonstrated that carcinoma cells that had hybridization signals were, in most cases, immu- nohistochemically positive for P-45017a, but some clus- ters of carcinoma cells demonstrated only hybridization
signals without discernable immunoreactivity and other groups of carcinoma cells exhibited only immunoreac- tivity. In the sections treated by RNase hybridization signals were not observed.
DISCUSSION
Adrenocortical carcinoma is a rare neoplasm and its incidence is approximately one case per 1,700,000 population, accounting for 0.02% of all cancer cases.19 Despite its rarity, a large number of investigators have studied this neoplasm for the following two reasons. The first is the occasional difficulty of differentiating
between adenoma and carcinoma at the time of initial surgery, even by histopathologic examination. The other is its unique feature of corticosteroidogenesis. Most of studies on steroid metabolism in adrenocortical neo- plasms have focused on the possible biochemical differ- entiation between benign and malignant adrenocortical neoplasms. For the most part, biochemical differentia- tion by hormonal studies has not been necessarily suc- cessful,20 including an attempt to correlate malignancy with increased dehydroepiandrosterone sulfate levels.21 A deficiency of 118-hydroxylation has been indicated in cases of adrenocortical carcinoma by several inves- tigators.3,22 In our present study cortisol was two to three times higher than normal, but 11-deoxycortisol was five
S
C
to 10 times higher than normal in cases no. 1 to 3, sug- gesting a relative deficiency of P-450118 activity. Enzy- matic activity of P-450118 was measured only in two cases, but in case no. 2, P-450118 activity was approximately half of the nonpathologic adrenal level. Therefore, some adrenocortical carcinoma may have, at best, a relative deficiency of P-450118, but it is unlikely that all adre- nocortical carcinoma has a defect in the 118-hydroxy- lation system.
To the best of our knowledge, no single indepen- dent marker of corticosteroidogenesis specific for ad- renocortical malignancy has been established, but it is true that steroid metabolism of adrenocortical carci- noma has a characteristic feature compared with that of both adenoma and normal adrenal cortex. This fea- ture has been summarized as low efficiency of steroid production or abnormal steroidogenesis. Neville and O’Hare and colleagues have demonstrated that carci- noma cells have at least one major difference from nor- mal cortical cells or cultured adenoma cells2,4: they in- dicated that adrenocortical tumors causing syndromes of steroid excess due to overproduction of minor prod- ucts of the normal cortex and tumors that produce no active steroids at all are likely to behave in a malignant fashion.2 Gazdar et al recently reported that multiple pathways of steroidogenesis are expressed by NCI-H295 cells described previously, including formation of glu- cocorticoids, mineralocorticoids, androgens, and estro- gens.5 In the cases examined in this study elevation of plasma levels of precursor steroids also were observed.
Adrenocortical carcinoma has been associated with morphologic diversity and even compared with cortical adenoma. Therefore, to study a characteristic steroid metabolism of adrenocortical malignancy as described above, it is extremely important to know the localization of steroidogenesis (ie, which tumor cells produce what steroids) in adrenocortical carcinoma. Biochemical studies generally cannot demonstrate the localization of
| Case No. | ||||||||
|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| - | f+ | f+ | f+ | f+ | f+ | + | - | f+ |
Symbols: +, positive hybridization signals; - , negative hybridiza- tion signals; f+, focal hybridization signals.
steroidogenesis because the tissue is processed as a mass. Classical morphologic approaches, including ultra- structural examination, can indicate the presence of steroidogenesis but cannot demonstrate which steroids are produced in the cells examined. The presence of one or more of the enzymes specifically involved in ste- roidogenesis in a cortical cell indicates the capacity of that cell to produce the specific steroid hormones cat- alyzed by a particular enzyme. Immunolocalization of steroidogenic enzymes can demonstrate expression of the enzymes directly in a formalin-fixed, paraffin- embedded specimen; thus, the presence of the enzymes can be correlated with the morphologic features of the specimen.16
In our present study of immunolocalization of steroidogenic enzymes in nine cases of adrenocortical carcinoma, immunoreactivity of all the enzymes was detected in six cases, as shown in Table 2. The absence of P-45017g and P-450118 in case no. 8 and the lack of P-450c21, P-45017a, and P-450118 in case no. 9, both nonfunctioning carcinomas, may be in accord with clinical signs. In case no. 8 raised pregnenolone levels may derive from the tumor and elevated urinary 17- OHCS and 17-KS levels may derive from the normal adrenal gland that was responsive to ACTH. In ad- dition, the degradation of pregnenolone excess to yield DHEA and DHEA-S in case no. 8 could be due
S
C
to the activity of the attached and contralateral normal adrenal gland. However, the absence of all the en- zymes in case no. 1 does not appear to make sense because the patient did have Cushing’s syndrome. The results may represent degradation of the enzyme pro- teins through tissue processing (especially prior to fixation), the presence of different forms of the en- zymes in these carcinoma cases, and/or simply that the tissue sections examined did not have tumor cells involved in corticosteroidogenesis, despite the fact that we examined sections obtained from different areas of the tumor. In addition, the absence of the enzymes in case no. 1 may be due to the fact that immunoreactivity per tumor cell may be below the detection level because the tumor in this case is the largest lesion in our current series. Further investi- gation is required to clarify this point.
In adrenocortical adenoma, especially in function- ing adenoma, the enzymes involved in steroidogenesis were expressed predominantly in small compact or clear tumor cells.28 These findings indicate that small compact or clear tumor cells play a major role in corticosteroid biosynthesis in adrenocortical adenoma. In adrenocor- tical carcinoma immunoreactivity of the enzymes also was predominantly observed in small carcinoma cells with compact and/or clear cytoplasm and minimum morphologic nuclear atypia. This finding is consistent with that of immunolocalization of adrenodoxine re- ductase, which is involved in mitochondrial steroido- genesis, in human adrenocortical carcinoma.24 However, as is shown in Figs 3 and 4, not all carcinoma cells ex- pressed the enzymes required for production of biolog- ically active steroids (ie, some carcinoma cells only ex- pressed the enzymes involved in precursor steroids, a finding that was rarely observed in adrenocortical ad- enoma).23 The precise mechanism of hormone release in adrenal cortex is disputed. However, the most widely accepted theory for steroid release is that they freely
diffuse throughout the cytosol and plasma membrane.5 Therefore, precursor steroids produced in the car- cinoma cells that do not have the enzymes further processing these steroids into biologically active hor- mones are considered to be released out of the cells. They may diffuse into the adjacent carcinoma cells that have the enzyme system producing biologically active steroids as precursors, as was shown in the cases of carcinoma associated with Cushing’s syndrome in our present study. They also may diffuse through the vascular wall and endothelium and into the circula- tion, which may result in elevated plasma levels of precursor steroids in the patients with adrenocortical carcinoma. In addition to the above immunohisto- chemical findings, although the effects of tissue prep- aration cannot be totally excluded, a discrepancy in P-45017g expression was observed between mRNA (in situ hybridization) and protein (immunohistochem- istry) levels. This discrepancy was observed rarely in normal adrenal25 and the adrenal of primary pig- mented nodular adrenocortical disease.16 This result of in situ hybridization suggests the presence of ab- normalities of transcription of cytochrome P-45017a in human adrenocortical carcinoma; nevertheless, further research, including nuclease protection assays of the respective RNAs, is required to clarify the sig- nificance of this finding.
In summary, although the data set in our current analysis may be too small to arrive at valid conclusions, results of our study suggest that ineffective or disor- ganized corticosteroidogenesis, characteristic of steroid metabolism of carcinoma, is demonstrated at the level of expression of steroidogenic enzymes in individual carcinoma cells. This “disorganized expression” of the enzymes was considered to be associated with malignant transformation of adrenocortical cells and could account for various abnormalities of corticosteroid metabolism in adrenocortical carcinoma.
REFERENCES
1. Cohn K, Gottesman L, Brennan MF: Adrenocortical carci- noma. Surgery 100:1170-1177, 1986
2. Neville AM, O’Hare MJ: The distinction between benign and malignant adrenocortical syndromes: Structural and functional criteria, in Neville AM, O’Hare MJ (eds): The Human Adrenal Cortex. Berlin, Germany, Springer-Verlag, 1982, pp 186-201
3. Doerr HG, Sippell WG, Drop SLS, et al: Evidence of 11 beta- hydroxylase deficiency in childhood adrenocortical tumors. The plasma corticosterone/11-deoxycorticosterone ratio as a possible marker for malignancy. Cancer 60:1625-1629, 1987
4. O’Hare MJ, Monagham P, Neville AM: The pathology of ad- renocortical neoplasia: A correlated structural and functional approach to the diagnosis of malignant disease. HUM PATHOL 10:137-154, 1979
5. Gazdar AF, Oie HK, Shackleton CH, et al: Establishment and characterization of a human adrenocortical carcinoma cell line that expresses multiple pathways of steroid biosynthesis. Cancer Res 50: 5488-5496, 1990
6. Sasano N, Ojima M, Masuda T: Endocrinologic pathology of functioning adrenocortical tumors, in Sommers SC, Rosen PP (eds): Pathology Annual, Part 2. Norwalk, CT, Appleton-Century-Crofts, 1980, pp 105-141
7. Fuhrman SA, Lasky LC, Limas C: Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol 6:655-663, 1982
8. Weiss LM: Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol 8: 163-169, 1984
9. Suzuki T, Sasano H, Nishikawa T, et al: Discerning malignancy in human adrenocortical neoplasms: Utility of DNA flow cytometry and immunohistochemistry. Mod Pathol 5:224-231, 1992
10. Barbieri RL, Osathanondh R, Canick JA, et al: Danazol in- hibits human adrenal 21- and 116-hydroxylation in vitro. Steroids 35: 251-263, 1980
11. Lowry OH, Rosebrough NJ, Farr AL, et al: Protein mea- surement within the folin phenol reagent. J Biol Chem 193:265-275, 1951
12. Gordon MT, Anderson DC, Mitchell R, et al: Some charac- teristics of human adrenal microsomal 21-hydroxylase activity. J Steroid Biochem 23:67-72, 1985
13. Goldman AS, Gustafsson J-Å, Stenberg Å: Neonatal imprint- ing of enzyme levels in rat adrenals. Acta Endocrinol 76:719-728, 1974
14. Greiner JW, Kramer RE, Jarrell J, et al: Mechanism of action of spironolactone on adrenocortical function in guinea pigs. J Phar- macol Exp Ther 198:709-715, 1976
15. Mejer LE, Blonchard RC: Fluorometric determination of plasma 11-hydroxycorticosteroids. II. Studies on the specificity of the method. Clin Chem 19:718-724, 1973
16. Sasano H, Miyazaki S, Sawai T, et al: Primary pigmented nodular adrenocortical disease (PPNAD): Immunohistochemical and in situ hybridization analysis of steroidogenic enzymes in eight cases. Mod Pathol 5:23-29, 1992
17. Hayashi M, Ninomiya Y, Hayashi K, et al: Secretion of col- lagen types I and II by epithelial and endothelial cells in the developing chick cornea demonstrated by in situ hybridization and immunohis- tochemistry. Development 103:27-36, 1987
18. Bradshaw KD, Watermann MR, Couch RT, et al: Character- ization of complementary DNA for human adrenocortical 17a-hy- droxylase: A probe for analysis of 17a-hydroxylase deficiency. Mol Endocrinol 1:348-357, 1987
19. Lack EE, Travis WD, Oertel JE: Adrenocortical neoplasms, in Lack EE (ed): Pathology of the Adrenal Glands. New York, NY, Churchill Livingstone, 1990, pp 115-171
20. Honour JW, Price DA, Taylor NF, et al: Steroid biochemistry of virilizing adrenal tumors in childhood. Eur J Pediatr 142:165-169, 1984
21. Gabrilove JL, Seman AT, Sabet R, et al: Virilizing adrenal venous effluent and a review of the literature. Endocr Rev 2:462-470, 1981
22. Degenhart HJ, Drop SLS, Hoogerbrugge J, et al: Virilizing adrenal adenoma with studies on the steroid content of the adrenal venous effluent and a review of literature. Horm Res 16:96-99, 1982
23. Sasano H: New approaches in human adrenocortical pa- thology: Assessment of adrenocortical function in surgical specimen of human adrenal glands. Endocr Pathol 3:4-13, 1992
24. Sasano H, Sasano N, Okamoto M, et al: Immunohistochemical demonstration of adrenodoxin reductase in bovine and human ad- renals. Pathol Res Pract 184:473-479, 1989
25. Suzuki T, Sasano H, Sawai T, et al: Immunohistochemistry and in situ hybridization of P-45017a (17a-hydroxylase/17,20-lyase). J Histochem Cytochem 40:903-908, 1992