Differential Expression of Melanocytic Markers in Myoid, Lipomatous, and Vascular Components of Renal Angiomyolipomas Open Access

The majority of the renal tumors in adults are renal cell carcinomas. The incidence of benign renal tumors is low, and most cases are detected incidentally or at autopsy.1 Renal angiomyolipoma is one of these benign tumors, classically a triphasic tumor with varying amounts of myoid spindle cells, adipose cells, and dysmorphic vessels.1–3 Although angiomyolipoma was originally considered to be a hamartoma, several lines of evidence now strongly support its neoplastic nature.4,5 Angiomyolipomas are now included in a family of tumors called PEComas (tumors showing perivascular epithelioid cell differentiation) that also includes lymphangiomyomatosis, clear cell “sugar” tumor of the lung, and other rare, morphologically and immunophenotypically similar lesions arising at a variety of visceral and soft tissue sites. The PEComa concept is still controversial and not universally accepted, as perivascular epithelioid cells are not detected in normal tissues.6 

The distinction of angiomyolipoma from renal cell carcinoma is critical in terms of clinical management. As the majority of the renal angiomyolipomas behave in a benign fashion and malignant outcome is rare, most cases are managed conservatively with watchful waiting for small tumors and surgical removal for larger tumors with a high risk of spontaneous rupture and hemorrhage.7 The majority of the cases are readily diagnosable if they exhibit the typical triphasic morphologic features. However, angiomyolipomas that are composed predominantly of 1 component may pose significant diagnostic challenge. For example, epithelioid angiomyolipoma may be mistaken for a renal cell carcinoma with “granular cytoplasm.” A smooth muscle–predominant or fat-predominant angiomyolipoma may be confused with a smooth muscle or a lipomatous tumor. Fortunately, renal angiomyolipoma, like other tumors included in the so-called PEComas, expresses a variety of melanocytic markers, muscle-specific actin, and vascular markers, a feature very helpful for establishing a correct diagnosis.3,8–12 HMB-45 and Melan-A are the 2 best-characterized melanocytic markers in angiomyolipomas. Recently, several other melanocytic markers, including tyrosinase and micropthalmia transcription factor, have been described to have excellent diagnostic sensitivity in cutaneous melanocytic lesions,13 although they have rarely been studied in renal angiomyolipomas.3,14–16 Kit (CD117), a transmembrane growth factor receptor expressed in cells of melanocytic lineage and a variety of other cell lineages, has been reported in a single study in renal angiomyolipomas.15 NK1-C3, another melanocytic marker with excellent diagnostic sensitivity in cutaneous melanocytic lesions,13 has also been reported in a single study in renal angiomyolipoma.16 

Furthermore, no studies so far have compared the expression of different melanocytic markers in the 3 different components of renal angiomyolipomas. Such information would be important when applying melanocytic markers on renal angiomyolipomas, especially those composed predominantly of 1 component. This study was therefore designed to evaluate the expression patterns of 5 melanocytic markers in the 3 different components of renal angiomyolipoma. In addition, we also compared the diagnostic sensitivities of these 5 melanocytic markers used alone or in combination and proposed a panel of markers with maximum diagnostic sensitivity for the diagnosis of renal angiomyolipomas.

Twenty classic cases of renal angiomyolipomas were retrieved from the surgical pathology files. Histologic materials were reviewed in all cases, and the original diagnoses were confirmed. One or more tissue blocks containing all 3 components (smooth muscle, fat, and blood vessels) were selected.

A tissue microarray (TMA) of the 20 renal angiomyolipomas was constructed. For each case, three 1.5-mm cores, each containing fat, blood vessel, and smooth muscle components, were taken to construct the TMA. Normal renal parenchyma was also included as a control.

The TMA was then stained for 5 different melanocytic markers, including HMB-45 (1:40 dilution, Dako, Carpinteria, Calif), Melan-A (clone A103, dilution 1:40, Dako), tyrosinase (clone T311, dilution 1:60, Novocastra, Newcastle upon Tyne, United Kingdom), NK1-C3 (dilution 1:50, BioGenex, San Ramon, Calif), and CD117 (dilution 1:400, Santa Cruz Biotechnology, Santa Cruz, Calif). Briefly, the 5-μm tissue sections were antigen-retrieved in 0.1M citrate buffer, pH 6.0, in a pressure steamer for 15 minutes. The slides were then incubated sequentially with primary antibody, biotinylated secondary antibody, avidin-peroxidase complex (Ventana, Tucson, Ariz), and chromogenic substrate diaminobenzidine. The immunostain was performed on a Ventana Benchmark automatic stainer (Ventana).

Using normal renal parenchyma as a control, each core was graded as no staining (score 0), weak (similar to the background staining in normal renal tissue, score 1), moderate (score 2), or strong (score 3) staining. Only the cores with >5% of cells exhibiting moderate-to-strong staining were considered positive. A component (smooth muscle, fat, or blood vessel) was graded positive for a marker if 1 or more of the 3 cores from that component was positive. A case was graded positive for a marker if 1 or more of the 9 cores from the 3 components of that case was positive.

Twenty consecutive patients with classic angiomyolipoma were included in the study (Figure 1, A through C). They consisted of 17 women and 3 men, with a mean age of 55.6 years (range, 33–79 years). All tumors were solitary with a mean size of 4.6 cm (range, 1.2–15 cm). The surgical modality included radical nephrectomy in 4 patients and partial nephrectomy in the remaining 16 patients.

Of these 20 cases, HMB-45 was positive in 95% (19/20), Melan-A in 85% (17/20), NK1-C3 in 70% (14/20), tyrosinase in 50% (10/20), and CD117 in 40% (8/20) of the cases. HMB-45 and Melan-A combined (case positive for either HMB-45 or Melan-A) were positive in 100% (20/ 20) of the cases (Table 1).

The 5 markers had different sensitivities in the fat, blood vessel, or smooth muscle components (Table 1, Figures 2 through 4). In the fat component, 90% (18/20) of the cases were positive with HMB-45, 70% (14/20) with Melan-A, 55% (11/20) with NK1-C3, 30% (6/20) with tyrosinase, and 20% (4/20) with CD117. In the smooth muscle component, 85% (17/20) of the cases were positive with HMB-45, 60% (12/20) with Melan-A, 55% (11/20) with NK1-C3, and 40% (8/20) with tyrosinase and CD117. In the blood vessel component, 80% (16/20) of the cases were positive with HMB-45, 45% (9/20) with NK1-C3, 40% (8/20) with Melan-A, 10% (2/20) with tyrosinase, and 10% (2/20) with CD117. Of samples of the fat or smooth muscle component, 95% (19/20) were positive for any 1 of the 5 markers, as were 85% (17/20) of samples of the blood vessel component.

Either HMB-45 or Melan-A was positive in 90% (18/20) of samples of the smooth muscle component, 95% (19/20) of samples of the fat component, and 80% (16/20) of samples of the blood vessel component. Inclusion of NK1-C3 improved the positive staining only in the smooth muscle component from 90% to 95% when only HMB-45 or Melan-A was used, but not in the fat or blood vessel components.

In the blood vessel component, only the adventitia and a few scattered cells in the tunica media and the cells around the blood vessels were positive for these melanocytic markers. No endothelial cells were found to be positive. In the smooth muscle component, the majority of the cells were positive, although there was some intratumoral heterogeneity in the staining intensity. In the fat component, some adipose cells showed partial membranous staining. On the other hand, most intensively positive cells were situated between the adipose cells and had an epithelioid, eosinophilic, and granular cytoplasm. A few of these cells showed nuclear atypia or multinucleation.

In this study we compared the sensitivities of 5 melanocytic markers in the diagnosis of renal angiomyolipoma. Our results are similar to those of previously published studies.3,14,16 HMB-45 and Melan-A are the most sensitive markers, positive in 95% and 85%, respectively, of renal angiomyolipomas that have classical triphasic components. HMB-45 and Melan-A stains complement each other. One case was negative for HMB-45 but was positive for Melan-A. The other 2 cases were negative for Melan-A, yet positive for HMB-45. All (20/20) cases were positive for either HMB-45 or Melan-A; therefore, a panel consisting of HMB-45 and Melan-A would detect all the classical renal angiomyolipomas. Other melanocytic markers have a relatively lower sensitivity: 70% for NK1-C3, 50% for tyrosinase, and 40% for CD117. Inclusion of any of these 3 markers did not improve the diagnostic sensitivity conferred by HMB-45 and Melan-A. Therefore, we recommend performing HMB-45 and Melan-A staining in the workup of potential renal angiomyolipoma, while the use of tyrosinase, NK1-C3, or CD117 is not necessary.

We also compared the staining patterns of these 5 melanocytic markers in the 3 components of angiomyolipomas (fat, blood vessels, and smooth muscle). Such information is not available in the literature, yet it would be important when ordering and interpreting melanocytic markers on angiomyolipomas, especially those with 1 predominant component, or biopsy specimens that include predominantly 1 component. These 5 markers have different sensitivities in the fat, blood vessel, and smooth muscle components. However, in all 3 components, HMB-45 and Melan-A are still the best markers, with the former positive in 90%, 85%, and 80% of fat, smooth muscle, and blood vessel components, respectively, and the latter positive in 70%, 60%, and 40% of these 3 components, respectively. In addition, HMB-45 and Melan-A combined were positive in 90% of samples of the smooth muscle component, 95% of samples of the fat component, and 80% of samples of the blood vessel component. Inclusion of other markers did not improve the sensitivity already conferred by HMB-45 and Melan-A, except for NK1-C3, which slightly improved the staining sensitivity in the smooth muscle component from 90% to 95%, compared with HMB-45 and Melan-A combined. On the other hand, CD117 and tyrosinase have much lower sensitivity in these 3 components, with sensitivity in the range of 10% to 20% for CD117 and 10 to 40% for tyrosinase. Therefore, a panel of HMB-45 and Melan-A is adequate to work up renal angiomyolipomas, even for those cases with 1 predominant component. NK1-C3 may be included in cases where the differential diagnosis includes smooth muscle– predominant angiomyolipoma.

The fat and smooth muscle components are more likely than the blood vessel component to exhibit positive staining of any of the 5 melanocytic markers. The implication is that when choosing a tumor block to perform immunostaining for melanocytic markers, one should avoid a block in which the blood vessel component predominates; we would suggest using a block with predominantly smooth muscle or fat components, particularly one that includes cells with epithelioid, eosinophilic, and granular cytoplasm situated between the adipose cells.

Cells that are positive for these melanocytic markers include those with smooth muscle and fat differentiation, cells in vascular adventitia, and occasional cells in tunica media, as well as cells with “PEC” morphology among smooth muscle, fat, and blood vessels. Similarly, Stone et al16 detected HMB-45 and NK1-C3 immunoreactivity predominantly in epithelioid and spindle cells and premelanosome–like structures in epithelioid cells in 9 cases by electron microscopy. Kaiserling et al17 and Liwnicz et al18 also detected premelanosomes in angiomyolipomas.

In summary, HMB-45 and Melan-A combined are positive in 100% of renal angiomyolipomas. We recommend the use of these 2 markers in the workup of these neoplasms, including those with predominantly one component. Other markers, including NK1-C3, tyrosinase, and CD117, are of limited use. Since different melanocytic markers in general exhibit higher sensitivity in the fat and smooth muscle components than the blood vessel component, a tissue block containing the former 2 components is preferred when performing melanocytic marker staining to confirm the diagnosis of renal angiomyolipoma.