Cytokeratin Expression in Adrenocortical Neoplasia:
An Immunohistochemical and Biochemical Study With Implications for the Differential Diagnosis of Adrenocortical, Hepatocellular, and Renal Cell Carcinoma
MICHAEL J. GAFFEY, MD, S. THOMAS TRAWEEK, MD, STACEY E. MILLS, MD, WILLIAM D. TRAVIS, MD, ERNEST E. LACK, MD, L. JEFFREY MEDEIROS, MD, AND LAWRENCE M. WEISS, MD
The immunostaining patterns of adrenocortical tumors are not clearly defined, primarily due to their inconsistent expression of cytokeratins (CK). To address this issue and to investigate whether adrenocortical tumors can be immunohistochemically differentiated from histologically similar tumors arising from the kidney and liver, we studied four normal adrenal glands, two adrenocortical ade- nomas (ACAs), 31 adrenocortical carcinomas (ACCs), 37 renal cell carcinomas (RCCs), and 33 hepatocellular carcinomas (HCCs) with anti-CK antibodies AE1, CAM 5.2, UCD/PR10.11, 35BH11, PKK1, and Ks19.1, as well as antibodies to vimentin (VIM), epithelial membrane antigen (EMA), and HMFG-2. Normal adrenal cortical cells showed variable staining with all anti-CK antibodies on fixed and frozen sections. In contrast, only one of two fixed ACAs stained with a single anti-CK, although both neoplasms reacted with multiple anti-CK antibodies on frozen sections. Similarly, 20 of 31 fixed ACCs contained VIM, but only one tumor stained for CK; frozen sections of this and another, previously negative tumor, however, stained with most of the anti-CK antibodies tested. One-dimensional West- ern immunoblot analysis confirmed the presence of CKs 18 and 19 in two examples of normal adrenal cortex, one ACA, and the ACC immunohistochemically positive on fixed and frozen sections, with CK 19 identified in the ACC that was positive on frozen section alone. All fixed HCCs and most RCCs stained with multiple anti- CK antibodies (33 and 34 cases, respectively), with a proportion of tumors positive for VIM (six and 22 cases, respectively), EMA (seven and 30 cases, respectively), and HMFG-2 (15 and 28 cases, respec- tively). The results suggest that CK expression is diminished in most adrenocortical tumors to levels too low to be recognized fol- lowing the deleterious effects of fixation. While the immunohis- tochemical absence of CK, EMA, and HMFG-2 in fixed sections in the majority of ACCs is distinctive, sufficient phenotypic overlap exists such that differentiation between RCC and HCC may not be possible in an individual case. HUM PATHOL 23:144-153. Copyright C) 1992 by W.B. Saunders Company
From the Departments of Pathology, Divisions of Surgical Pa- thology, the University of Virginia Health Sciences Center, Char- lottesville, VA; City of Hope National Medical Center, Duarte CA; the National Cancer Institute, National Institutes of Health, Bethesda, MD; and the Georgetown University Hospital, Washington DC. Ac- cepted for publication April 18, 1991.
Key words: adrenocortical tumors, cytokeratin, renal cell carci- noma, hepatocellular carcinoma, immunohistochemistry.
Address correspondence and reprint requests to Michael J. Gaffey, MD, Department of Pathology, Box 214, University of Virginia Health Sciences Center, Charlottesville. VA 22908.
Adrenocortical carcinomas (ACCs) are uncommon neoplasms that may be confused with tumors of similar histologic appearance arising in adjacent sites. For ex- ample, differentiation of a primary ACC from an invasive renal cell carcinoma (RCC) can be problematic. Renal cell carcinoma may also present as a solitary adrenal metastasis,’ and the simultaneous presence of hepatic lesions brings hepatocellular carcinoma (HCC) into the differential diagnosis as well. Accordingly, a reliable method for distinguishing these tumors is desirable.
Intermediate filament (IF) typing is useful in tumor identification, as most neoplasms express the IF types characteristic of their parental tissue.2 Epithelial cells and their corresponding neoplasms characteristically contain cytokeratins (CKs), a multigene-coded, complex IF family of 19 polypeptides distinguished on the basis of their distinctive molecular weights and isoelectric points.2,3 The recent development of monoclonal antibodies directed against one or more of the 19 CK polypeptides has permitted CK subtyping by immuno- histochemical methods. To this end, the immunohisto- chemical expression of CK in RCC and HCC has been well described.4-16 The IF patterns in ACC, however, are poorly defined, with the relatively few published studies reporting a predominant absence of CK expres- sion.17-20 Most of these studies, however, either failed to use extended periods of enzymatic digestion of fixed sections to uncover masked antigenic expression, or used antibody panels with limited recognition of those low molecular weight CKs characteristically found in simple epithelia.
In an effort to clarify the expression of CK in ad- renocortical neoplasia, we have applied an extended panel of anti-CK and other antibodies to fixed, paraffin- embedded sections of 31 ACCs, two adrenocortical ad- enomas (ACAs), and four normal adrenal glands. Snap- frozen tissue for immunohistochemical and Western blot analysis was available for two of the ACCs, both of the ACAs, and all four of the normal adrenal glands. For purposes of differential diagnosis, 37 RCCs and 33 HCCs were examined with an identical antibody panel applied to fixed, paraffin-embedded sections.
CYTOKERATIN IN ADRENOCORTICAL NEOPLASIA (Gaffey et al)
| Reagents | Specificity | Source | Dilution | Reference | |
|---|---|---|---|---|---|
| PFN | FZN | ||||
| Anticvtokeratins: | |||||
| AE-1*+ | CK 10, 14, 15, 16. 19 | Hybritech (San Diego, CA) | 1:50 | (1:25) | 30 |
| CAM 5.2* + | CK 8. 18. 19 | Becton Dickinson (Mountain View, CA) | 1:5,000 | (1:2,500) | 28 |
| ['CD/PR10.11%+ | CK 8. 18 | R. Cardiff. UC at Irvine | 1:1.000 | (1:500) | 26 |
| 35BHI1* + | CK 8 | Enzo Biochem (New York, NY) | 1:9,000 | (1:4,500) | 27 |
| PKK1* | CK 8. 18, 19 | Labsystems (Raleigh, NC) | 1:400 | (1:100) | 29 |
| Ks 19.1* | CK 19 | Progen (Heidelberg, Germany) | 1:200 | (1:100) | 42 |
| Amivimentin* | DAKO (Carpenteria, CA) | 1:100 | (1:50) | ||
| Antiepithelial membrane antigen* | DAKO | 1:40 | (1:20) | 33 | |
| Anti-HMFG-2* | OXOID (Columbia, MO) | 1:5,000 | (1:2,500) | 43 | |
* Murine monoclonal preparations; PFN denotes dilutions used in avidin-biotin immunohistochemical procedures on fixed, paraffin-embedded tissues, whereas FZN denotes those dilutions used in cryostat sections.
+ Constituents of the KC-2 anticytokeratin cocktail used at working dilutions of 1:100 and 1:50 for deparaffinized and crvostat sections. respectively: CK denotes cytokeratin according to the nomenclature of Moll et al.”
| Feature | No. of Cases | Percentage |
|---|---|---|
| Nuclear grade | ||
| 1 | 1/31 | 3 |
| 11 | 18/31 | 58 |
| III | 7/31 | 23 |
| IV | 5/31 | 16 |
| Mitotic rate (per 50 HPF) | ||
| 0-5 | 4/31 | 13 |
| 6-20 | 11/31 | 35 |
| 21-50 | 9/31 | 29 |
| Greater than 50 | 7/31 | 23 |
| Atypical mitoses | ||
| Present | 21/31 | 68 |
| Absent | 10/31 | 32 |
| Cytoplasm | ||
| 26% to 100% clear | 5/31 | 16 |
| 0% to 25% clear | 26/31 | 84 |
| Tumoral architecture | ||
| Nondiffuse | 9/31 | 29 |
| Diffuse | 22/31 | 71 |
| Necrosis | ||
| Present | 23/31 | 74 |
| Absent | 8/31 | 26 |
| Invasion of venous structures | ||
| Present | 12/31 | 39 |
| Absent | 19/31 | 61 |
| Invasion of sinusoidal structures | ||
| Present | 20/31 | 65 |
| Absent | 11/31 | 35 |
| Invasion of capsule | ||
| Present | 11/31 | 35 |
| Absent | 20/31 | 65 |
| Weight (g) | ||
| 0-99 | 0/17 | 0 |
| 100-500 | 6/17 | 35 |
| 501-1000 | 3/17 | 18 |
| >1,000 | 8/17 | 47 |
| Size (cm) | ||
| 0-5 | 0/16 | () |
| 6-10 | 2/16 | 12 |
| 11-20 | 10/16 | 63 |
| >20 | 4/16 | 25 |
Abbreviation: HPF, high-power field.
MATERIALS AND METHODS
Case Selection
Thirty-one ACCs were retrieved from the surgical pa- thology files of the University of Virginia Health Sciences Center, the City of Hope National Medical Center, and the Georgetown University Hospital. Twenty ACCs were primary lesions localized to the adrenal gland at the time of diagnosis. The remaining 11 tumors were metastatic ACCs surgically re- sected from patients on treatment protocol. Two ACAs and four normal adrenal glands were also examined. Formalin- fixed, paraffin-embedded material was available in 28 ACCs, both ACAs, and all four of the normal adrenal glands, with B5-fixed material available in the remaining three ACCs. Fresh tissue obtained at the time of surgical resection, snap-frozen in isopentane, and stored at -70℃ was available for two of the ACCs, both ACAs, and all four of the normal adrenal glands.
All primary tumors were classified as either ACC or ACA by previously defined criteria.21,22 Briefly, adrenocortical le- sions were evaluated for each of the following nine criteria of malignancy: (1) nuclear grade of III or IV according to the criteria of Fuhrman et al,23 (2) mitotic rate greater than 5/50 high-power fields, (3) atypical mitotic figures, (4) clear cells constituting ≤25% of the tumor, (5) diffuse growth pattern, (6) necrosis, (7) invasion of venous structures, (8) invasion of sinusoidal structures, and (9) capsular penetration. Details and illustrations of these criteria are presented elsewhere.21,22 A tumor with two or fewer criteria was categorized as an ade- noma; tumors with three or more criteria were considered to be carcinomas. Tumor size and weight were also recorded when available. All tumors were simultaneously reviewed by three of us (M.J.G., T.T., L.M.W.) and the diagnoses were confirmed. For comparative purposes, 33 and 37 cases of clinically and pathologically clear-cut HCCs and RCCs, respectively, were also culled from the surgical pathology files of the four con- tributing institutions.
Immunohistochemistry
Tissue sections were mounted on slides coated with 1% polylysine solution and stained by the avidin-biotin complex method as detailed elsewhere.24,25 The antibodies used and their respective specificities, dilutions, and sources are listed in Table 1. All adrenocortical CK preparations were subjected to digestion with 0.1% trypsin (ICN Nutritional, Cleveland,
| Tissue Type | Anticytokeratins* | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| KC2 | AE1 | CAM | UCD | 35BH | PKK1 | Ks19.1 | VIM | EMA | HMFG2 | |
| Adrenal | ||||||||||
| Normal cortex | 4/4 (4/4) | 4/4 (3/4) | 4/4 (4/4) | 4/4 (4/4) | 2/4+ (4/4) | 0/4 (4/4) | 2/4 (4/4) | 0/4 (0/4) | 0/4 (0/4) | 0/4 (0/4) |
| Cortical adenomas | 0/2 (2/2) | 0/2 (2/2) | 0/2 (2/2) | 1/2+ (2/2) | 0/2 (2/2) | 0/2 (1/2) | 0/2 (1/2) | 0/2 (0/2) | 0/2 (0/2) | 0/2 (0/2) |
| Cortical carcinomas | 1/31 (2/2) | 1/31 (1/2) | 1/31 (2/2) | 1/31 (2/2) | 1/31 (2/2) | 1/31 (2/2) | 1/31 (1/2) | 20/31 (0/2) | 0/31 (0/2) | 0/31 (0/2) |
| Liver | ||||||||||
| Normal hepatocytes | 19/19 | 0/19 | 13/19 | 19/19 | 13/19 | 1/19 | 0/19 | 0/19 | 0/19 | 0/19 |
| Hepatocellular carcinoma | 33/33 | 16/33 | 13/33 | 33/33 | 6/33 | 27/33 | 15/33 | 6/33 | 7/33 | 15/33 |
| Kidney | ||||||||||
| Normal tubular epithelium | 12/12 | 12/12 | 10/12 | 12/12 | 12/12 | 12/12 | 12/12 | 0/12 | 12/12 | 12/12 |
| Renal cell carcinoma | 34/37 | 24/37 | 10/37 | 31/37 | 8/37 | 16/37 | 20/37 | 22/37 | 30/37 | 28/37 |
Note. The presence of vimentin in stromal and endothelial cells is not included in the table. For the relative extent of positive staining, see text. Values without parentheses refer to cases studied with fixed, paraffin-embedded material, whereas the values in parentheses refer to cases studied by frozen section as well. Values are expressed as cases positive/cases studied.
Abbreviations: KC2, cytokeratin cocktail; CAM, CAM 5.2; UCD, UCD/PR10.11: 35BH, 35BH11; VIM, vimentin; EMA, epithelial membrane antigen; HMFG. human milk fat globule 2.
* Adrenocortical sections subjected to 30-, 60-, 90-, and 180-minute enzymatic digestion; positive cases recorded after 30-minute preliminary trypsinization unless specified otherwise. All liver and kidney sections subjected to 30-minute enzymatic digestion alone.
+ Positive after 60-minute trypsinization.
OH) in phosphate-buffered saline for 30, 60, 90, and 180 minutes at room temperature; all liver and kidney sections for CK staining were treated with only preliminary trypsinization for 30 minutes. Enzymatic digestion was not used for the re- maining antibodies. Antibody localization was detected by in- cubation with 3,3’ diaminobenzidine tetrahydrochloride. In addition to the individual anti-CK preparations, a monoclonal anti-CK cocktail (KC-2) was used, the combined constituents of which recognize CK filaments with molecular weights of 40, 45, 50, 52, 54, and 56.5 kd.26-30 Snap-frozen tissue sections were immunohistochemically stained by the avidin-biotin complex method with chromogen substrate 3-amino-9-ethyl- carbaxole (AEC; Polyscience Inc, Warrington, PA).
All stained tissue sections were reviewed simultaneously by three of us (M.J.G., T.T., L.M.W.) using a multi-headed microscope to determine both the presence and extent of im- munohistochemical staining. Appropriate positive and negative controls were stained and studied concurrently.
Intermediate Filament Extraction
Intermediate filaments were extracted from control MCF- 7 cells and frozen tissue from two examples of normal adrenal cortex, one ACA, and two ACCs. There was insufficient frozen tissue from the remaining ACA for analysis. Briefly, the frozen tissues and cells were minced and homogenized at 4℃ in ap- proximately 100 volumes of Triton X-100 buffer (0.5 M Tris- HCL, 50 mmol/L EDTA, 10% Triton X-100, and 1.4 M NaCl) and centrifuged at 10,000 rpm for 15 minutes at 4°℃. The insoluble protein pellets were subsequently homogenized in high-salt buffer (1.5 mol/L KCI, 0.5 mol/L Tris-HCL, 50 mmol/L EDTA, and 10% Triton X-100) at 4℃ and centrifuged at 12,000 rpm for 20 minutes at 4℃. The pellets were then homogenized in low-salt buffer (0.5 mol/L Tris-HCI, 50 mmol/ L EDTA, 1.4 mol/L NaCl, and 10% Triton X-100) and cen- trifuged at 15,000 rpm for 30 minutes at 4℃. The resulting IF pellets were stored at -20℃ until use.
One-Dimensional Sodium Dodecyl Sulfate- Polyacrylamide Gel Electrophoresis
The IF pellets were solubilized by heating at 80℃ for 15 minutes in 0.5 ml. to 1.0 mL of sodium dodecyl sulfate buffer
(50 mmol/L. Tris-HCL, 1% sodium dodecyl sulfate, 2% b-mer- captoethanol, 1 mmol/L MgCl2) and were clarified at 2,500g for 15 minutes. Cytoskeletal proteins were then electropho- resed on 10 X 12 cm slab gels in the presence of 0.1% sodium dodecyl sulfate by previously published methods.34
Immunoblotting
The separated proteins were transferred electrophoreti- cally to Immunobiolon-P membranes (Millipore Corp, Bed- ford, MA) using a Bio-Rad Transblot apparatus (Bio-Rad I.ab- oratories, Palo Alto, CA) and stained with anti-CK antibodies AEI (dilution 1:30) and CAM 5.2 (dilution 1:10) using an indirect immunoperoxidase technique and 3’3 diaminoben- zidine tetrahydrochloride as chromogen.
RESULTS
Tumor Characterization
Adrenocortical tumors. The gross features of the pri- mary ACCs and ACAs studied were characteristic of their respective diagnoses.21.22 Tumor size and weight were recorded in 16 and 17 ACCs, respectively, with an average diameter of 16.8 cm (range, 9 to 30 cm; median, 17 cm) and a mean weight of 1,323 g (range, 100 to 5,720 g; median, 1,150 g). The two ACAs studied were small, measuring 3.5 and 4.3 cm and weighing 30 and 34 g, respectively.
Microscopically, all 31 ACCs exhibited three or more of the histologic criteria for adrenocortical malig- nancy as summarized in Table 2, with five or more cri- teria observed in 61% of cases. By comparison, of the two ACAs studied, one contained greater than 26% clear cells while the other failed to satisfy any of the micro- scopic criteria of malignancy.
Other tumors. The 33 HCCs and 37 RCCs examined exhibited gross and microscopic appearances typical for their respective diagnoses and will not be described in detail. Of the 33 HCCs studied, 31 tumors were eosin-
ophilic and trabecular with one predominantly clear cell HCC and a single fibrolamellar variant. Of the 37 RCCs studied, 28 were clear cell, six were eosinophilic, and the remaining three were papillary.
Immunohistochemistry
Normal adrenal cortex. As detailed in Table 3, fixed- section immunohistochemical examination showed the cortical cells of all four normal adrenal glands positive with KC-2, CAM 5.2, and UCD/PR10.11 (Fig 1). The cortical cells of two cases stained with 35BH11; one of these two was positive only after 60 minutes of enzymatic digestion. Immunoreactivity for the latter preparations was cytoplasmic and distributed in random transcortical trabeculae without apparent relationship to any func-
tional zone of the adrenal cortex. In contrast, staining with AE1 (four cases) and Ks19.1 (two cases) was focal and restricted to subcapsular cortical cells. The cortical cells of all cases were uniformly negative with PKK1 and antibodies to vimentin (VIM), epithelial membrane an- tigen (EMA), and HMFG-2. Frozen section immuno- histochemistry, however, revealed all four cases to be diffusely reactive with KC-2, CAM 5.2, UCD/PR10.11, 35BH11, and PKK1. Four and three cases showed focal, subcapsular positivity for Ks19.1 and AE1, respectively (Fig 1). Staining with antibodies to VIM, EMA, and HMFG-2 was uniformly negative.
Adrenocortical adenoma. Immunohistochemical ex- amination of two ACAs in formalin-fixed sections showed one tumor to be focally positive with UCD/ PR10.11 alone after 60 minutes of digestion (Table 3).
In contrast, frozen section immunostaining showed scattered cortical cell positivity with KC-2, AE-1, CAM 5.2, UCD/PR10.11, and 35BH11 in both cases, with PKK1 and Ks19.1 reactivity noted in one case each (Fig 2). Neither tumor stained for VIM, EMA, or HMFG-2 in fixed or frozen sections.
Adrenocortical carcinoma. Of the 31 ACCs studied in fixed sections, 30 were negative for all of the anti-CK antibodies tested following 30, 60, 90, and 180 minutes of digestion (Table 3). A single tumor, however, stained diffusely with KC-2, CAM 5.2, and UCD/PR10.11 and focally with AE1, 35BH11, PKK1, and Ks19.1 after 30 minutes of enzymatic digestion (Fig 3). Vimentin im- munopositivity was noted in the cortical cells of 20 ACCs, with all tumors negative for EMA and HMFG-2. In contrast, frozen section immunostaining of two ACCs, including the single CK-positive tumor and an- other immunohistochemically negative case, demon- strated both tumors to be strongly positive for all of the anti-CK preparations tested (Fig 3). Frozen section staining with antibodies to VIM, EMA, or HMFG-2 was negative in both cases.
Uninvolved liver. In 19 cases, hepatic parenchyma uninvolved by carcinoma was present on the stained sections. In all 19 cases, the hepatocytes were diffusely positive with KC-2 and UCD/PR10.11, with focal reac- tivity observed for CAM 5.2 and 35BH11 in 13 cases
each (Table 3). Reactivity for PKKI was present in a single case, with staining for AE1 and Ks19.1 uniformly negative. In contrast, the bile ducts consistently stained with all the anti-CK preparations tested. No staining for VIM was seen in either the hepatocytes or bile ducts, while immunopositivity for EMA and HMFG-2 was noted in occasional bile ducts in a luminally accentuated, cytoplasmic pattern. Although rare, nonneoplastic he- patocytes also stained for EMA in a coarse, granular distribution; the latter pattern was considered nonspe- cific in view of its dissimilarity from the finely diffuse cytoplasmic staining pattern characteristic of these an- tibodies.3 31,35
Hepatocellular carcinoma. Immunostaining of 33 paraffin-embedded HCCs after 30 minutes of enzymatic digestion revealed all tumors to be diffusely positive with KC-2 and UCD/PR10.11. In addition, 27 HCCs were positive with PKK1, 13 with CAM 5.2, and six with 35BH11 (Table 3). Sixteen HCCs reacted with AEl and 15 were labeled with Ks19.1 (Fig 4). Six HCCs were reactive for VIM, with seven and 15 tumors diffusely positive for EMA and HMFG-2, respectively.
Uninvolved kidney. Nonneoplastic kidney adjacent to tumor was present in 12 cases. The tubules and col- lecting ducts of all 12 cases were diffusely positive with KC-2 and UCD/PR10.11 and focally reactive with AE1, 35BH11, PKK1, Ks19.1, EMA, and HMFG-2 (Table 3).
Ten cases stained with CAM 5.2. with the tubules and ducts of all 12 kidneys negative for VIM.
Renal cell carcinoma. Of the 37 RCCs studied by fixed section immunostaining, 34 were strongly positive with KC-2, 31 were immunoreactive with UCD/ PR10.11, 24 reacted with AE1, and 20 stained with Ks19.1 (Fig 5). Diffuse staining was also observed with PKKI in 16 RCCs, with 10 and eight tumors positive with CAM 5.2 and 35BH11, respectively. Three RCCs were CK- and VIM-negative; all were of the clear cell type. The remaining clear cell RCCs also tended to react with fewer anti-CK antibodies than the granular tumors, but too few of the latter variants were studied for mean- ingful comparison. Twenty-two tumors coexpressed VIM, and 28 RCCs stained for both HMFG-2 and EMA, including the three CK-negative RCCs (Fig 5). Two ad- ditional RCCs stained for EMA alone.
Immunoblots
Immunoblots were performed on two examples of normal adrenal cortex, one ACA, and two ACCs. Stain- ing with AE1 and CAM 5.2 clearly demonstrated bands of 43 kd and 40 kd, corresponding to CKs 18 and 19 of Moll’s catalog,2 respectively, in control MCF-7 cells, both examples of normal adrenal cortex, one ACA, and the ACC immunohistochemically positive on fixed and frozen sections. Cytokeratin 19 was identified in the other ACC positive on frozen section alone (Fig 6). Cy- tokeratin 8 (53 kd) was not detected in any sample.
DISCUSSION
In the current study, the cortical cells of all four normal adrenal glands studied were immunoreactive for CK, as compared with one of two ACAs and only one of 31 ACCs in fixed tissue sections. This suggests that a loss of CK expression might occur on adrenocortical neoplastic transformation. The observed CK immuno- reactivity in two adrenocortical tumors on frozen but not fixed sections, however, argues that CK is present in these tumors, but in amounts too low to be detected following the deleterious effects of fixation. Inconsistent CK expression in adrenocortical neoplasia has been re- ported elsewhere, but prolonged periods of enzymatic digestion and extended monoclonal antibody panels specific for CK polypeptides of low molecular weight were, for the most part, not used in these studies. 19-20) Accordingly, whether CK expression could be demon- strated in most adrenocortical tumors by methodologic manipulation had not been completely explored. Our results indicate that despite the biochemically docu- mented presence of CK in these tumors, the majority of ACCs will not label with anti-CK antibodies in fixed sections, a significant consideration in the differential diagnosis of these tumors.
Our studies also demonstrate that normal and neoplastic adrenal cortical cells contain low molecular weight CK polypeptides characteristically present in simple epithelia. The immunoreactivity of both normal
and neoplastic cortical cells for CAM 5.2, UCD/ PR10.11, 35BH11, and PKK1, monoclonal antibodies directed against CKs 8 and 18, is strong evidence for the presence of these polypeptides. The observed im- munoreactivity for AEI and Ks19.1, both of which are reactive with CK 19, also indicates the additional presence of at least CK 19 in the cases studied. The presence of CKs 18 and 19 was subsequently con- firmed by one-dimensional gel electrophoresis in two examples of normal adrenal cortex, one ACA, and one ACC, with CK 19 alone demonstrated in the re- maining ACC. The absence of CK 18 in one ACC, as well as the identification of CK 18 in the apparent absence of its paired member, CK 8, is inconsistent with established patterns of CK expression in epithe- lial cells.2 The inability to demonstrate by immunoblot
all CK subtypes predicted to be present immunohis- tochemically, however, is not uncommon, and is thought to reflect differential epitope accessibility in- herent in denaturizing sodium dodecyl sulfate-poly- acrylamide gels.36
While immunopositivity for those antibodies di- rected against CKs 8 and 18 was unrelated to any func- tional zone of the adrenal cortex, the subcapsular stain- ing noted with AE1 and Ks19.1 suggests that the expression of CK 19 is localized to cortical cells within the zona glomerulosa. As CK 19 is the simplest, lowest molecular weight CK and the first to appear during em- bryogenesis, those cortical cells reactive with AE1/ Ks19.1 may be less differentiated than their nonreactive counterparts.37 The latter observation is of interest in view of previous proposals for a stem cell population
within the zona glomerulosa responsible for cortical re- generation.38
The observed lack of cortical VIM immunoreactivity in the four normal adrenal glands and two ACAs studied is in agreement with the results of Miettinen et al,18 but at variance with those of Wick et al,19 Henzen-Logmans et al,17 and Cote et al,20 who reported uniform VIM positivity in most examples of nonneoplastic adrenal cortex and in 18 of 20 ACAs. That fixation-induced antigenic masking is not responsible for these discrep- ancies is evident because the current cases, as well as those of Miettinen et al18 and Henzen-Logmans et al, 17 were studied on frozen section, whereas Wick et al19 and Cote et al20 observed VIM positivity on formalin- fixed material. The presence of VIM in 20 of 31 ACCs, however, is in accord with the latter studies, with VIM
reactivity reported in 10 of 25 ACCs studied by Miet- tinen et al18 and in the combined 27 ACCs examined by Wick et al.19 Henzen-Logmans et al,17 and Cote et al.2” In addition, the observed absence of EMA and HMFG-2 staining in normal adrenal cortex, ACA, and ACC studied is consonant with the results of both Wick et al19 and Sloane et al.31.35
Both normal renal tubular epithelium and RCC have been shown by two-dimensional gel electrophoresis to contain CKs 7, 8, 18, and 19, consistent with the variable reactivity of all 12 examples of nonneoplastic renal parenchyma and with that of 34 of 37 RCCs for most of the anti-CK preparations tested.4.13 The obser- vation by us and others, however, that a small proportion of RCCs may lack CK expression suggests that a loss or diminution of CK expression may also occur in renal
3
40-
43 -
40-
123
1 2 3 4 5
epithelium on neoplastic transformation.12 The staining of most RCCs with antibodies to EMA and HMFG-2, including those tumors negative for CK, reflects the reactivity of normal renal tubules and ducts for the latter preparations, 12,19,31,39 whereas the absence of VIM in normal renal epithelium contrasts with the presence of this IF in 22 of 37 RCCs.
Two-dimensional gel electrophoretic studies have shown that normal hepatocytes express CKs 8 and 18, whereas bile duct epithelium expresses both these and CKs 7 and 192,7; this accounts for the observed immu- noreactivity of both cell types for antibodies directed against CKs 8 and 18, whereas anti-CK 19 preparations stained the bile ducts alone. Previous reports as to whether the normal pattern of hepatocyte CK expres- sion is altered on neoplastic transformation are con- flicting, with biochemical and immunohistochemical studies reporting either the presence8,14,15 or ab-
sence2,7,10,11,40 of aberrant CK expression in HCC. In the current study, 16 of 33 HCCs stained with antibodies directed against CK subtypes not normally expressed by normal hepatocytes,2.7 supporting the concept of aberrant CK expression in these tumors. Additional ev- idence of IF infidelity is evident in the absence of de- tectable VIM in normal hepatocytes, as compared with the presence of this IF type in a small proportion of HCCs. The reactivity of seven HCCs for EMA is at vari- ance with previous studies reporting the uniform ab- sence of EMA in normal hepatocytes and their tu- mors.19,31,35 Results similar to ours were reported by Christensen et al, however, who found EMA positivity in 25 of 62 tumors.5
The current results indicate that ACC, RCC, and HCC may not be immunohistochemically separable in the individual case, and other histologic, biochemical, and clinical criteria must be utilized for their separation. The absence of CK and VIM reactivity in fixed sections of ACC appears to be distinctive, but an occasional RCC
may also manifest a similar phenotype. In addition, while the uniform absence of EMA and HMFG-2 reactivity in ACC contrasts with the presence of these markers in the current RCC negative for CK, Medeiros et al re- ported five of 55 RCCs negative for all markers of ep- ithelial differentiation.12 Similarly, the EMA and HMFG- 2 reactivity of an occasional HCC delineates these lesions from ACC but not from RCC. The variable presence of an aberrant CK profile and VIM in those HCCs negative for EMA and HMFG-2, however, resembles the phenotype of some RCCs and CK-positive ACCs. He- patocellular carcinomas may express other, more char- acteristic antigens, such as alpha-1 antitrypsin and alpha- fetoprotein, but the immunohistochemical detection of these markers is often technically difficult and lacks the necessary sensitivity for diagnostic use.41.42 Finally, the ACAs and ACCs studied were immunohistochemically indistinguishable on both fixed and frozen sections, in- dicating that other clinicopathologic criteria must be used for their distinction, a finding in agreement with the results of Wick et al.19
The diminished expression of CK by adrenocortical tumors, the expression of aberrant CK by HCC, and the de novo appearance of VIM in both ACC and RCC indicates that the IF expression of a given cell type is not fixed and may be subject to modulation on neo- plastic transformation. Accordingly, differential IF expression should not be used as evidence of histogen- esis or as a sole criterion in the differential diagnosis of these tumors.
REFERENCES
1. Foucar E, Dehner LP: Renal cell carcinoma occurring with contralateral adrenal metastasis. A clinical and pathological trap. Arch Surg 114:959-963, 1979
2. Moll R, Franke WW, Schiller DL, et al: The catalog of human cytokeratins: Patterns of expression in normal epithelia, tumors, and cultured cells. Cell 31:11-24, 1982
3. Sun TT, Eichner R, Schermer A, et al: Classification, expression and possible mechanisms of evolution of mammalian epithelial keratins: A unifying model. Cancer cells 1. The transformed phenotype. Cold Spring Harb Lab 169-176, 1984
4. Achstatter T, Moll R, Moore B, et al: Cytokeratin polypeptide patterns of different epithelia of the human male urogenital tract. Immunofluorescence and gel electrophoretic studies. J Histochem Cxtochem 33:415-426. 1985
5. Christensen WN, Boitnott JK, Kuhajda FP: Immunoperoxidase staining as a diagnostic aid for hepatocellular carcinoma. Mod Pathol 2:8-12. 1989
6. Gown AM, Vogel AM: Monoclonal antibodies to intermediate filament proteins of human cells: Unique and cross-reacting antibodies. J Cell Biol 95:414-424, 1982
7. Denk. H, Krepler R. Lackinger F. et al: Biochemical and im- munocytochemical analysis of the intermediate filament cytoskeleton in human hepatocellular carcinomas and in hepatic neoplastic nodules of mice. Lab Invest 46:584-596, 1982
8. Fischer HP, Altmannsberger M, Weber K, et al: Keratin poly- peptides in malignant epithelial liver tumors. Differential diagnostic and histogenetic aspects. Am J Pathol 127:530-537, 1987
9. Holthofer H, Miettinen A, Paasivuo R, et al: Cellular origin and differentiation of renal carcinomas. A fluorescence microscopic study with kidney-specific antibodies, antiintermediate filament anti- bodies, and lectins. Lab Invest 49:317-326, 1983
10. Johnson DE, Herndier BG, Medeiros LJ, et al: The diagnostic utility of the keratin profiles of hepatocellular carcinoma and cholan- giocarcinoma. Am J Surg Pathol 12:187-197. 1988
11. Lai Y-S, Thung SN, Gerber MA, et al: Expression of cyto- keratins in normal and diseased livers and in primary liver carcinomas. Arch Pathol Lab Med 113:134-138, 1989
12. Medeiros LJ. Michie SA. Johnson DE, et al: An immuno- peroxidase study of renal cell carcinomas: Correlation with nuclear grade, cell type, and histologic pattern. HUM PATHOL. 19:980-987, 1988
13. Pitz S, Moll R, Storkel S, et al: Expression of intermediate filament proteins in subtypes of renal cell carcinomas and in renal oncocytomas. Lab Invest 56:642-653. 1987
14. Van Eyken P, Sciot R, Paterson A, et al: Cytokeratin expres- sion in hepatocellular carcinoma: An immunohistochemical study. HUM PATHOL 19:562-568, 1988
15. Van Eyken P, Sciot R, Brock P, et al: Abundant expression of cytokeratin 7 in fibrolamellar carcinoma of the liver. Histopathology 17:101-107, 1990
16. Van Eyken P, Sciot R, Desmet VJ: A cytokeratin immuno- histochemical study of cholestatic liver disease: Evidence that hepa- tocytes can express “bile duct-type” cytokeratins. Histopathology 15: 125-135, 1989
17. Henzen-Logmans SC, Stel HV, Van Muijen GNP, et al: Expression of intermediate filament proteins in adrenal cortex and related tumors. Histopathology 12:359-372, 1988
18. Miettinen M. Lehto V-P, Virtanen I: Immunofluorescence microscopic evaluation of the intermediate filament expression of the adrenal cortex and medulla and their tumors. Am J Pathol 118:360- 366. 1985
19. Wick MR. Cherwitz DI., McGlennen RC, et al: Adrenocortical carcinoma. An immunohistochemical comparison with renal cell car- cinoma. Am J Pathol 122:343-352, 1986
20. Cote RJ, Cordon-Cardo C. Reuter VE, et al: Immunopa- thology of adrenal and renal cortical tumors. Coordinated change in antigen expression is associated with neoplastic conversion in the ad- renal cortex. Am J Pathol 136:1077-1084, 1990
21. Weiss LM: Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol 8: 163-169, 1984
22. Weiss I.M, Medeiros LJ. Vickery AL .: Pathologic teatures of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol 13:202-206, 1989
23. Fuhrman SA, Lasky I.C, Limas C: Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol 6:655-663, 1982
24. Battifora H, Kopinski M: The influence of protease digestion and duration of fixation on the immunostaining of keratins. A com- parison of formalin and ethanol fixation. J Histochem Cytochem 34: 1095-1100. 1986
25. Sheibani K. Tubbs RR: Enzyme immunohistochemistry: Technical aspects. Semin Diagn Pathol 1:235-250, 1984
26. Chan R, Edwards BF, Hu R, et al: Characterization of two monoclonal antibodies in an immunohistochemical study of keratin 8 and 18 expression. Am J Clin Pathol 89:472-480, 1988
27. Gown AM. Vogel AM: Monoclonal antibodies to human in- termediate filament proteins II. Distribution of filament proteins in normal human tissues. Am J Pathol 114:309-321, 1984
28. Makin CA, Bobrow LG, Bodmer WF: Monoclonal antibody to cytokeratin for use in routine histopathology. J Clin Pathol 37:975- 983, 1984
29. Miettinen M, Virtanen I, Talerman A: Intermediate filament proteins in human testis and testicular germ-cell tumors. Am J Pathol 120:402-410, 1985
30. Woodcock-Mitchell J, Eichner R. Nelson WG, et al: Immu- nolocalization of keratin polypeptides in human epidermis using monoclonal antibodies. J Cell Biol 95:580-588, 1982
31. Sloane JP. Ormerod MG: Distribution of epithelial membrane antigen in normal and neoplastic tissues and its value in diagnostic tumor pathology. Cancer 47:1786-1795. 1981
32. Karsten U, Papsdorf G, Roloff G. et al: Monoclonal anti- cytokeratin antibody from a hybridoma clone generated by electro- fusion. Eur J Cancer Clin Oncol 21:733-740, 1985
33. Burchell J. Durbin H. Taylor-Papadimitriou J: Complexity of antigenic determinants, recognized by monoclonal antibodies HMFG-1 and HMFG-2, in normal and malignant human mammary epithelial cells. J Immunol 131:508-513, 1983
34. Laemmli UK: Cleavage of structural proteins during the as- sembly of the head of bacteriophage T4. Nature 227:680-685, 1970
35. Sloane JP. Hughes F, Ormerod MG: An assessment of the value of epithelial membrane antigen and other epithelial markers in solving diagnostic problems in tumour histopathology. Histochemical J 15:645-654, 1983
36. Gown AM, Boyd HC, Chang Y, et al: Smooth muscle cells can express cytokeratins of “simple” epithelium. Immunocytochemical and biochemical studies in vitro and in vivo. Am J Pathol 132:223- 232, 1988
37. Banks-Schlegel SP: Keratin alterations during embryonic epidermal differentiation: A presage of adult epidermal maturation. J Cell Biol 93:551-559, 1982
38. Zajicek G, Ariel I, Arber N: The streaming adrenal cortex: Direct evidence of centripetal migration of adrenocytes by estimation of cell turnover rate. J Endocrinol 111:477-482, 1986
39. Pinkus GS, Kurtin PJ: Epithelial membrane antigen. A di- agnostic discriminant in surgical pathology. Immunohistochemical profile in epithelial, mesenchymal, and hematopoietic neoplasms using paraffin sections and monoclonal antibodies. HUM PATHOL 16:929- 940, 1985
40. Osborn M, Van Lessen G, Weber K, et al: Methods in lab- oratory investigation. Differential diagnosis of gastrointestinal carci- nomas by using monoclonal antibodies specific for individual keratin polypeptides. Lab Invest 55:497-504, 1986
41. Roncalli M, Borzio M, DeBiagi G, et al: Liver cell dysplasia and hepatocellular carcinoma: A histological and immunohistochemical study. Histopathology 9:209-221, 1985
42. Thung SN, Gerber MA, Sarno E, et al: Distribution of five antigens in hepatocellular carcinoma. Lab Invest 41:101-105, 1979